Your SlideShare is downloading. ×

3645754 content

739

Published on

The International Institute for Science, Technology and Education (IISTE) , International Journals Call for papaers: http://www.iiste.org/Journals

The International Institute for Science, Technology and Education (IISTE) , International Journals Call for papaers: http://www.iiste.org/Journals

Published in: Business, Technology
0 Comments
0 Likes
Statistics
Notes
  • Be the first to comment

  • Be the first to like this

No Downloads
Views
Total Views
739
On Slideshare
0
From Embeds
0
Number of Embeds
0
Actions
Shares
0
Downloads
3
Comments
0
Likes
0
Embeds 0
No embeds

Report content
Flagged as inappropriate Flag as inappropriate
Flag as inappropriate

Select your reason for flagging this presentation as inappropriate.

Cancel
No notes for slide

Transcript

  • 1. Journal of Economics and Sustainable Development www.iiste.orgISSN 2222-1700 (Paper) ISSN 2222-2855 (Online)Vol.2, No.6, 2011Supply Response of Rice in Ghana: A Co-integration Analysis John K.M. Kuwornu (Corresponding author) Department of Agricultural Economics and Agribusiness, P. O. Box LG 68, University of Ghana, Legon-Accra, Ghana Tel: +233 245 131 807 E-mail: jkuwornu@ug.edu.gh / jkuwornu@gmail.com Maanikuu P. M. Izideen Department of Agricultural Economics and Agribusiness, P. O. Box LG 68, University of Ghana, Legon-Accra, Ghana Tel: +233 241 913 985 E-mail: maanikuu@yahoo.com Yaw B. Osei-Asare Department of Agricultural Economics and Agribusiness, P. O. Box LG 68, University of Ghana, Legon-Accra, Ghana Tel: +233 209 543 766 E-mail: yosei@ug.edu.ghReceived: October 1st, 2011Accepted: October 11th, 2011Published: October 30th, 2011AbstractThis study presents an analysis of the responsiveness of rice production in Ghana over the period1970-2008. Annual time series data of aggregate output, total land area cultivated, yield, real prices of riceand maize, and rainfall were used for the analysis. The Augmented-Dickey Fuller test was used to test thestationarity of the individual series, and Johansen maximum likelihood criterion was used to estimate theshort-run and long-run elasticities. The land area cultivated of rice was significantly dependent on output,rainfall, real price of maize and real price of rice. The elasticity of lagged output (12.8) in the short run wassignificant at 1%, but the long run elasticity (4.6) was not significant. Rainfall had an elasticity of 0.004 andsignificant at 10%. Real price of maize had negative coefficient of -0.011 and significant at 10%significance level. This is consistent with theory since a rise in maize price will pull resources away fromrice production into maize production. The real price of rice had an elasticity of 2.01 and significant at 5%in the short run and an elasticity of 3.11 in the long run. The error correction term had the expected negativecoefficient of -0.434 which is significant at 1%. It was found that in the long run only real prices of maizeand rice were significant with elasticities of -0.46 and 3.11 respectively. The empirical results also revealedthat the aggregate output of rice in the short run was found to be dependent on the acreage cultivated, thereal prices of rice, rainfall and previous output with elasticities of 0.018, 0.01, 0.003 and 0.52 respectively.Real price of rice and area cultivated are significant 10% level of significance while rainfall and laggedoutput are significant 5%. In the long run aggregate output was found to be dependent on acreage cultivated,real price of rice, and real price maize with elasticities of 0.218, 0.242 and -0.01 respectively at the 1%significance level. The analysis showed that short-run responses in rice production are lower than long-runresponse as indicated by the higher long-run elasticities. These results have Agricultural policy implicationsfor Ghana.Key Words: Supply response, Rice, Error Correction Model, Co-integration Analysis, Ghana1. IntroductionOne of the most influential policy prescriptions for low-income countries ever given by developmenteconomists has been to foster industrialization by withdrawing resources from agriculture (e.g., Lewis,1954). There is robust evidence that the majority of policy makers followed this prescription at least untilthe mid-1980s. The results of a comprehensive World Bank study (Krueger et al., 1992), for example, showfor the period 1960 –1985 that in most countries examined, agriculture was taxed both directly via 1
  • 2. Journal of Economics and Sustainable Development www.iiste.orgISSN 2222-1700 (Paper) ISSN 2222-2855 (Online)Vol.2, No.6, 2011interventions in agricultural markets and indirectly via overvalued exchange rates and import substitutionpolicies. It is not obvious whether the disincentives for agricultural production have continued to exist sincethe mid-1980s. On the one hand, most developing countries have adopted structural adjustment programswhich explicitly aim at a removal of the direct and indirect discrimination against agriculture. But, on theother hand, it is known that many of these programs were not fully implemented, especially in Sub-SaharanAfrica (Kherallah et al., 2000; Thiele and Wiebelt 2000; World Bank, 1997). Hence it can be concluded thata certain degree of discrimination still prevails.One of the most important issues in agricultural development economics is supply response of crops. Thisis because the responsiveness of farmers to economic incentives determines agricultures’ contribution to theeconomy especially where the sector is the largest employer of the labour force. This is often the case inthird world low income countries. Agricultural pricing policy plays a key role in increasing farm production.Supply response is fundamental to an understanding of this price mechanism (Nerlove and Bachman,1960).There is a notion that farmers in less developed countries respond slowly to economic incentives such asprice and income. Reasons cited for poor response vary from factors such as constraints on irrigation andinfrastructure to a lack of complementary agricultural policies. The importance of non-price factors drewadequate attention in the literature: rainfall, irrigation, market access for both inputs and output, and literacy.The reason cited for a low response to prices in less developed economies is the limited access to input andproduct markets or high transaction costs associated with their use. Limited market access may be eitherdue to physical constraints such as absence of proper road links or the distances involved between the roadsand the markets, or institutional constraints like the presence of intermediaries (Mythili, 2008).In Ghana, the goals of agricultural price policy are among others fair incomes for farmers, low food pricesfor urban consumers, cheap raw materials for manufacturing (Ministry of Food and Agriculture (MOFA),2005). Price support was one policy that was used by government in targeting these goals. Price support hasbeen used in many countries across the world. The prices of major commodities have been set below worldprices using subsidies and trade barriers, guaranteed prices (act as floors) and domestic market forcesdetermining actual prices (Food and Agriculture Organization (FAO), 1996). The effect of liberalization onthe growth of agriculture crucially depends on how the farmers respond to various price incentives.For many low-income countries, the impact of structural reforms on economic growth and povertyalleviation crucially depends on the response of aggregate agricultural supply to changing incentives.Despite its policy relevance, the size of this parameter is still largely unknown (Thiele, 2000).Rice is a very important crop especially for those areas where it is produced. Rice is gradually taking theposition as the main staple food for the majority of families in Ghana especially in the urban centres. Thegrowth in consumption of rice has been not been matched with a corresponding growth in local productionof the crop. Per capita consumption of rice has steadily increased over the years since the 1980s from about12.4kg/person /year in 1984 to about 20kg/person /year (Statistical Research and Information Division(SRID), 2005). It continued to rise to 38kg/person/year in 2008 (National Rice Development Policy(NRDP), 2009). Over the last decade rice per capita consumption has increased by more than 35%. This isattributed mainly to the rapid urbanization and changes in food consumption patterns. It is estimated basedon population and demand growth rates that per capita consumption will reach 41.1kg by 2010 and 63.0kgby 2015 giving an aggregate demand of 1,680,000tons/year (Statistical Research and Information Division(SRID), 2005). Changes in population dynamics and the taste or preference for foreign products iscontributing to this trend. These changes are significant considering the rate of population growth. Ghana’spopulation has grown by about 70% from 12.3 million in 1984 to about 21 million in 2004. The populationis now 24 million (Ghana Statistical Service (GSS), 2011).The growing quantum of rice imports into the country has also been triggered partly by the fall in worldmarket price. This is largely due to increased output levels in India and South Asia. Massive subsidies inrice exporting countries have also contributed to this phenomenon. Despite all these developments thereseem to be a lack of coherent and comprehensive national policy for rice in Ghana. From the period ofeconomic recovery and structural adjustment, the country has had to embark on trade liberalization and theremoval of all forms of subsidies on agriculture. These policies no doubt have adversely affected riceproduction in the country. The smallholder rural farmer is faced with unfair competition from abroad. The 2
  • 3. Journal of Economics and Sustainable Development www.iiste.orgISSN 2222-1700 (Paper) ISSN 2222-2855 (Online)Vol.2, No.6, 2011role of incentives to farmers has generally been sidelined or ignored in most developing countries due toconditions of trade liberalization and global trade integration. However, a number of empirical studies inother developing economies addressed the question of farmers’ response to economic incentives andefficient allocation of resources (e.g. Chinyere, 2009). The agricultural sector in Ghana has undergonevarious policy regimes which has affected both the factor and product market resulting in changes in thestructure of market incentives (prices) faced by farmers. Most of these policies have, however, not beencrop specific and therefore has wide variations in the quantum of changes in the incentives. This studytherefore follows the supply response framework of analysis to examine the dynamics of the supply of ricein Ghana. Effort in this direction will have to be preceded by a thorough analysis of the factors that affectsthe supply of rice. These teething problems lead to the following research questions;How responsive is rice production to price and non-price factors? What are the long run and short runelasticities of rice production? What are the trends in area cultivated, output and real prices of output?Therefore, objective of the study is threefold: (a) Examine the acreage and output response of riceproduction in Ghana. (b) Estimate and compare the long run and short run elasticities of rice production.(c) Analyze the trends in output, area cultivated and real prices.The rest of the paper is structured as follows. Section 2 outlines the methodology. In section 3 we presentthe empirical applications and the results. Section 4 provides the conclusions.2.0 MethodologyThis section presents the methodology of the study.2.1 Linear RegressionLinear regression was conducted to determine the growth rates of the variables over the study period. Timein years was regressed on acreage cultivated, aggregate output, real price of rice separately.2.2 Time series analysisIn empirical analysis using time series data it is important that the presence or absence of unit root isestablished. This is because contemporary econometrics has indicated that regression analysis using timeseries data with unit root produce spurious or invalid regression results (e.g. Townsend, 2001). Most timeseries are trended over time and regressions between trended series may produce significant parameterswith high R2s, but may be spurious or meaningless (Granger and Newbold, 1974). When using the classicalstatistical inference to analyze time series data, the results are only stationary when the series are stationary.The solution to this problem was initially provided by Box and Jenkins (1976), by formulating regressionsin which the variables were expressed in first difference. Their approach simply assumed thatnon-stationary data can be made stationary by repeated differencing until stationarity is achieved and thento perform the regression using these differenced variables. However, according to Davidson et al., (1978),this process of repeated differencing even though leads to stationarity of the series; it is achieved at theexpense of losing valuable long run information. This posed a new challenge to time series econometrics.The concept of cointegration was introduced to solve these problems (Granger, 1981; Engle and Granger,1987). By using the method of cointegration an equation can be specified in which all terms are stationaryand so allow the use of classical statistical inference. It also retains information about the long runrelationship between the levels of variables, which is captured in the stationary co-integrating vector. Thisvector will comprise the parameters of the long run equilibrium and corresponds to the parameters of theerror correction term in the second stage regression (Mohammed, 2005). The cointegration approach takesinto consideration the long-run information such that spurious results are avoided.2.3 Stationarity TestsA data series is said to be stationary if it has a constant mean and variance. That is the series fluctuates aroundits mean value within a finite range and does not show any distinct trend over time. In a stationary seriesdisplacement over time does not alter the characteristics of a series in the sense that the probabilitydistribution remains constant over time. A stationary series is thus a series in which the mean, variance andcovariance remain constant over time or in other words do not change or fluctuate over time. In a stationary 3
  • 4. Journal of Economics and Sustainable Development www.iiste.orgISSN 2222-1700 (Paper) ISSN 2222-2855 (Online)Vol.2, No.6, 2011series the mean always has the tendency to return to its mean value and to fluctuate around it in a more or lessconstant range, while a non-stationary series has a changing mean at different points in time and its variancechange with the sample size (Mohammed, 2005). The conditions of stationarity can be illustrated by thefollowing:Yt = ѳYt-1 + µ t t=1………T (1)Where µ t is a random walk with mean zero and constant variance. If ѳ < 1, the series Yt stationary and ifѳ = 1 then the series Yt is non-stationary and is known as random walk. In other words the mean, varianceand covariance of the series Yt changes with time or have an infinite range. However Yt can be madestationary by differencing. Differencing can be done multiple times on a series depending on the number ofunit roots a series has. If a series becomes stationary after differencing d times, then the series contains dunit roots and hence integrated of order d denoted as I (d). Thus, in equation (1) where ѳ = 1, Yt has a unitroot. A stationary series could also exhibit other properties such as when there are different kinds of timetrends in the variable.The DF (Dickey-Fuller)-statistic used in testing for unit root is based on the assumption that µ t is whitenoise. If this assumption does not hold, it leads to autocorrelation in the residuals of the OLS regressionsand this can make invalid the use of the DF-statistic for testing unit root. There are two approaches to solvethis problem (Towsend, 2001). In the first instance the equations to be tested can be generalized. Secondlythe DF-statistics can be adjusted. The most commonly used is the first approach which is the AugmentedDickey-Fuller (ADF) test. µ t is made white noise by adding lagged values of the dependent variable to theequations being tested, thus:∆Yt = (ѳ1 – 1) Yt-1 + IYt-I + µ t (2)∆Yt = α2 + (ѳ2 – 1) Yt-1 + I ∆Yt-I + µ t (3)∆Yt= α3+β3t(ѳ3–1)Yt-1+ I ∆Yt-I + µ t (4)The ADF test uses the same critical values with DF. The results of the ADF test for unit roots for each ofthe data series used in this study are presented in in the next section using equation (4) where Yt is theseries under investigation, t is the time trend, α3 is the constant term and µ t are white noise residuals.Eviews was used in the analysis, all the data series was tested for stationarity and the results are presentedin section three.2.4 CointegrationCointegration is founded on the principle of identifying equilibrium or long run relationships betweenvariables. If two data series have a long run equilibrium relationship it implies their divergence from theequilibrium are bounded, that is they move together and are cointegrated. Generally for two or more seriesto be co-integrated two conditions have to be met. One is that the series must all be integrated to the sameorder and secondly a linear combination of the variables exist which is integrated to an order lower thanthat of the individual series. If in a regression equation the variables become stationary after firstdifferencing, that is I (1), then the error term from the cointegration regression is stationary, I (0)(Hansen and Juselius, 1995). If the cointegration regression is presented as:Yt = α + βXt + µ t (5)where Yt and Xt are both I (1) and the error term is I (0), then the series are co-integrated of order I (1,0 )and β measures the equilibrium relationship between the series Yt and Xt and µ t is the deviation from thelong-run equilibrium path. An equilibrium relationship between the variables implies that even though Ytand Xt series may have trends, or cyclical seasonal variations, the movement in one are matched bymovements in the other. The concept of cointegration has implications for economists. The economicinterpretation that is accepted is that if in the long-run two or more series Yt and Xt themselves arenon-stationary, they will move together closely over time and the difference between them is constant(stationary) (Mohammed 2005).2.4.1 Testing for CointegrationThere are two most commonly used methods for testing cointegration. The Augmented Dickey-Fullerresidual based test by Engle and Granger (1987), and the Johansen Full Information Maximum Likelihood(FIML) test (Johansen and Juselius, 1990). For the purpose of this study the Johansen Full InformationMaximum Likelihood test is used due to its advantages. The major disadvantage of the residual based test is 4
  • 5. Journal of Economics and Sustainable Development www.iiste.orgISSN 2222-1700 (Paper) ISSN 2222-2855 (Online)Vol.2, No.6, 2011that it assumes a single co-integrating vector. But if the regression has more than one co-integrating vectorthis method becomes inappropriate (Johansen and Juselius, 1990). The Johansen method allows for allpossible co-integrating relationships and allows the number of co-integrating vectors to be determinedempirically.2.4.2 Johansen Full Information Maximum Likelihood ApproachThe Johansen approach is based on the following Vector AutoregressionZt = AtZt-1 + … + AkZt-k + µ t (6)Where Zt is an (n×1) vector of I(1) variables (containing both endogenous and exogenous variables), At is(n×n) matrix of parameters and µ t is (n×1) vector of white noise errors. Zt is assumed to be nonstationaryhence equation (6) can be rewritten in first difference or error correction form as;∆Zt = Γ1∆Zt-1 + … + Γk-1∆Zt-k+1 + πZt-k + µ t (7)where Γ1 = - ( 1- A1 - A2 - … -Ai), (i = 1, …, k-1) and π = - (1- A1-A2- …-Ak).Γ1 gives the short run estimates while π gives the long run estimates. Information on the number ofco-integrating relationships among variables in Zt is given by the rank of the matrix π. If the rank of πmatrix r, is 0 < r > n, there are r linear combinations of the variables in Zt that are stationary. Thus π can bedecomposed into two matrices α and β where α is the error correction term and measures the speed ofadjustment in ∆Zt and β contains r co-integrating vectors, that is the cointegration relationship betweennon-stationary variables. If there are variables which are I(0) and are significant in the long runco-integrating space but affect the short run model then equation (7) can be rewritten as:∆Zt = Γ1∆Zt-1 + πZt-k + vDt + µ t (8)where Dt represents the I(0) variables.To test for co-integrating vector two likelihood ratio (LR) tests are used. The first is the trace test statistic;Λtrace = -2lnQ = -T i) (9)Which test the null hypothesis of r co-integrating vectors against the alternative that it is greater than r. Thesecond test is known as the maximal-eigen value test:Λmax = -2 ln(Q: r 1 r + 1 = -T ln(1-λr+1 ) (10)which test the null hypothesis of r co-integrating vectors against the alternative of r+1 co-integratingvectors. The trace test shows more robustness to both skewness and excess kurtosis in the residuals than themaximal eigen value test (Harris, 1995). The error correction formulation in (7) includes both the differenceand level of the series hence there is no loss of long run relationship between variables which is acharacteristic feature of error correction modeling.It should be noted that in using this method, the endogenous variables included in the VectorAutoregression (VAR) are all I(1), also the additional exogenous variables which explain the short runeffect are I(0). The choice of lag length is also important and the Akaike Information Criterion (AIC), theScharz Bayesian Criterion (SBC) and the Hannan-Quin Information Criterion (HQ) are used for theselection.According to Hall (1991) since the process might be sensitive to lag length, different lag orders should beused starting from an arbitrary high order. The correct order is where a restriction on the lag length isrejected and the results are consistent with theory.2.5 Error Correction Models (ECMs)The idea behind the mechanism of error correction is that a proportion of disequilibrium from one period iscorrected in the next period in an economic system (Engle and Granger, 1987). The process of making adata series stationary is either done by differencing or inclusion of a trend. A series that is made stationaryby including a trend is trend stationary and a series that is made stationary by differencing is differencestationary. The process of transforming a data series into stationary series leads to loss of valuable long runinformation (Engle and Granger, 1987). Error correction models helps to solve this problem.The Granger representation theorem is the basis for the error correction model which indicates that if thevariables are cointegrated, there is a long-run relationship between them and can be described by the errorcorrection model. The following equation shows an ECM of agricultural supply response involving thevariables Y and X in its simplest form: 5
  • 6. Journal of Economics and Sustainable Development www.iiste.orgISSN 2222-1700 (Paper) ISSN 2222-2855 (Online)Vol.2, No.6, 2011∆Yt = α∆Xt – ѳ(Yt-1 – γXt-1)+µ t (11)Where µ t is the disturbance term with zero mean, constant variance and zero covariance. Parameter α takesinto account the short run effect on Y of the changes in X, while γ measures the long-run equilibriumrelationship between Y and X that is:Yt = γXt + µ t (12)Where Yt-1 – γYt-1 + µ t-1 measures the divergence (errors) from long-run equilibrium. Also ѳ measures theextent of error correction by adjustment in Y and its negative sign indicates that the adjustment is in thedirection which restores the long-run relationship (Hallam and Zanoli, 1993). In order to estimate equation(5), Engle and Granger (1987) proposed a two stage process. Firstly the static long run cointegrationregression (6) is estimated to test cointegration between the two variables. If cointegration exists the laggedresiduals from equation (5) are used as error correction term in the Error Correction Model (in equation 6)to estimate the short run equilibrium relationship between the variables in the second stage. The validity ofthe Error Correction Models (ECMs) depends upon the existence of a long-run or equilibrium relationshipamong the variables (Mohammed, 2005).The Error Correction Model (ECM) has several advantages. It contains a well-behaved error term andavoids the problem of autocorrelation. It allows consistent estimation of the parameters by incorporatingboth short-run and long-run effects. Most importantly all terms in the ECM are stationary. It ensures that noinformation on the levels of the variables is lost or ignored by the inclusion of the disequilibrium terms(Mohammed, 2005). ECM solves the problems of spurious correlation because ECMs are formulated interms of first difference which eliminates trends from the variables (Ganger and Newbold, 1974). It avoidsthe unrealistic assumption of fixed supply based on stationary expectations in the partial adjustment model.2.6 Supply Response ModelsThis study estimated the total area cultivated and aggregate output of rice in Ghana using doublelogarithmic regression models.Area cultivated of rice (lgarea) is a function of own price (Lrp), rainfall (lgrain), aggregate output (lgoutput)and price of maize (lmp). The equation used for the regression is:Lgarea = f1 (Lgoutput, Lrp, Lgrain, Lmp) (13)Aggregate output of rice (lgoutput) is a function of the area cultivated (lgarea), the price of rice (Lrp), priceof maize (Lmp), rainfall (lgrain). The estimating equation is:Lgoutput = f2 (lgrain, Lrp, Lmp, Lgarea)1. (14)3. Empirical Application and Results3.1 Results for linear regressionTable 1: The results of the regression analysis for the trend of the variables against timeVariable Coefficient Std error t-statistic R2 F-statistic ProbLgarea 1.334274 1.149167 1.61079 0.035154 0.253046 0.2530Lgoutput -13.86625 68.09380 -0.20364 0.005889 0.844432 0.8444lgyield -25.99348 35.51112 -0.73198 0.071100 0.487957 0.4880Lrp 2.588941 5.221601 4.956136 0.399186 24.58311 0.0000As can be seen from table 1 the results indicated that the regression for area, output and yield turned outinsignificant. Only the trend of real price of rice was significant at 1% significance level. Real rice priceyielded a positive coefficient of 2.589 which implies that for each year the real price of rice grew by 2.589units.3.2 Unit Root Test ResultsAs a requirement for cointegration analysis the data was tested for series stationarity and to determine theorder of integration of the individual variables. For cointegration analysis to be valid all series must beintegrated of the same order usually of order one (Towsend, 2001). Eviews was used to perform these tests.1 Price of maize is included because we assume that the same resources (land type, fertilizer etc.) can be used toproduce both maize and rice. Hence, a rise in the price of maize will pull resources away from rice production to maizeproduction. 6
  • 7. Journal of Economics and Sustainable Development www.iiste.orgISSN 2222-1700 (Paper) ISSN 2222-2855 (Online)Vol.2, No.6, 2011The data series on annual acreage cultivated (Lgarea), aggregate output (Lgoutput), aggregate yield(Lgyield), real price of rice (Lrp), real price of maize (Lmp), rainfall (Lgrain) was tested for unit root forthe study period 1970 - 2008. The Augmented Dickey-Fuller test was used for this test. The results arepresented below.Table 2: Results of unit root test at levelsSeries ADF test Mackinnon Lag-length Prob Conclusion statistic critical valueLgarea 2.221125 3.615588 2 0.2024 Non-stationaryLgoutput 1.519278 5.119808 2 0.4582 Non-stationaryLgyield 0.103318 4.580648 2 0.9419 Non-stationaryLmp 0.378290 3.621023 2 0.9026 Non-stationaryLrp 1.429037 3.615588 2 0.5579 Non-stationaryLgrain 3.006033 2.963972 7 0.0457 StationaryThe results of the unit root test after first differencing are presented in table 3.Table 3: Results of unit root test at first differencesSeries ADF test Mackinnon Lag-length Prob Conclusion statistic critical valueLgarea 7.517047 3.621023 2 0.0000 I(1)Lgoutput 9.577098 3.621023 2 0.0000 I(1)Lgyield 7.517047 3.621023 2 0.0000 I(1)Lmp 10.02189 3.621023 2 0.0000 I(1)Lrp 6.346392 3.626784 2 0.0000 I(1)Note; All variables are in log form. The ADF method test the hypothesis that H0 : X ~ I(1), that is, has unitroot (non-stationary) against H1 : X ~ I(0), that is, no unit root (stationary). The Mackinnon critical valuesfor the rejection of the null hypothesis of unit root are all significant at 1%. Lgarea denotes log of areacultivated, Lgoutput denotes log total output, Lgyield denotes log of yield, Lmp denotes log of maize price,Lrp denotes log of rice price and Lgrain denotes log of rainfall.The results of the unit root tests showed that all the series are non-stationary at levels except for rainfallwhich is stationary at levels as shown in table 2 above. However as expected all the non-stationary seriesbecame stationary after first differencing. From table 2 the null hypothesis of unit root could not be rejectedat levels since none except rainfall of the ADF test statistics was greater than the relevant MackinnonCritical values. Hence the null of the presence of unit root is accepted.However the hypothesis of unit root in all series was rejected at 1% level of significance for all series afterfirst difference since the ADF test statistics are greater than the respective Mackinnon critical values asshown in table 3 above.3.3 Cointegration ResultsWhen the order of integration of the data series have been established, the next step in the process ofanalysis is to determine the existence or otherwise of cointegration in the series. This is to establish theexistence of valid long-run relationships between variables. Basically there are two most commonly usedmethods to test for cointegration. These were suggested by Engle and Granger (1987) and Johansen andJuselius (1990). This study applies the Johansen approach which provides likelihood ratio tests for thepresence of number of co-integrating vectors among the series and produces long-run elasticities. The Errorcorrection model was then used to estimate short-run elasticities. 7
  • 8. Journal of Economics and Sustainable Development www.iiste.orgISSN 2222-1700 (Paper) ISSN 2222-2855 (Online)Vol.2, No.6, 20113.3.1 Supply Response of Rice OutputFirstly, the Vector Error Correction Modeling (VECM) procedure using the Johansen method involvesdefining an unrestricted Vector Autoregression (VAR) using the following equation.Zt=AtZt-1+…+AkZt-1+µ t (15)Likelihood Ratio (LR) tests were conducted with maximum of three lags due to the short time series. Theresults are shown in table 4.Table 4: Results for lag selection output model VAR Lag Order Selection Criteria Endogenous variables: OUTPUT AREA MP RP Exogenous variables: C RAIN Lag LogL LR FPE AIC SC HQ0 -1208.968 NA 6.64e+32 86.92626 87.30689* 87.042621 -1184.928 37.77692 3.84e+32 86.35198 87.49387 86.701072 -1174.563 13.32579 6.37e+32 86.75452 88.65766 87.336333 -1138.068 36.49514* 1.92e+32* 85.29058* 87.95499 86.10511* * indicates lag order selected by the criterion LR: sequential modified LR test statistic (each test at 5% level) FPE: Final prediction error AIC: Akaike information criterion SC: Schwarz information criterion HQ: Hannan-Quinn information criterion LogL: Log-likelihoodThe results indicate that the Akaike Information Criterion (AIC) and the Hannan-Quin information criterion(HQ) selected lag order three while the Scharz Bayesian Criterion (SBC) selected the lag order zero. Thus,this study selects the lag order three as the order for the VAR models. The next step in the Johansen methodis to test for the number of co-integrating vectors among the series in the model.The results for the cointegration test imply that both the trace test and the maximum eigen value test selectsthe presence of one co-integrating vector, thus it can be concluded that the variables in the model areco-integrated. The Johansen model is a form of Error Correction Model. When only one co-integratingvector is established its parameters can be interpreted as estimates of long run co-integrating relationshipbetween the variables (Hallam and Zanoli, 1993). This implies that the estimated parameter values fromthis equation when normalized on output are the long run elasticities for the model. Eviews automaticallyproduces the normalized estimates. These coefficients represent estimates of long-run elasticities withrespect to area cultivated, real price of rice and real price of maize. The normalized cointegration equationfor rice output is given below;Output=0.216748area+0.241522rp-0.009975mp-12841.69 (16)Since cointegration has been established among the variables, then the dynamic ECM can be used forsupply response analysis since it provides information about the speed of adjustment to long runequilibrium and avoids the spurious regression problem between the variables (Engle and Granger, 1987).The ECM for rice output is presented as;∆output = α0+ 1i∆outputt-i + 2i∆areat-i + 3i∆rpt-I + 4i∆mpt-I + 5rain – θECt-I (17)Where θECt-I = α (β1outputt-I – β2areat-I – β3rpt-I – β4mpt-I) 8
  • 9. Journal of Economics and Sustainable Development www.iiste.orgISSN 2222-1700 (Paper) ISSN 2222-2855 (Online)Vol.2, No.6, 2011In the ECM model above, the α’s explain the short run effect of changes in the explanatory variables on thedependent variable whereas β’s represent the long run equilibrium effect. θECt-I is the error correction termand correspond to the residuals of the long run cointegration relationship, that is the normalized equation(16). The negative sign on the error correction term indicates that adjustments are made towards restoringlong run equilibrium. This representation ensures that short run adjustments are guided by and consistentwith the long run equilibrium relationship. This method provides estimates for the short run elasticities, thatis the coefficients of the difference terms and, whereas the parameters from the Johansen cointegrationregression are estimates of the long run elasticities (Townsend and Thirtle, 1994). In selecting the best ECMestimates, models with lag lengths are estimated and those with insignificant parameters are eliminated.Note that the estimates presented above represent the short run effect of the explanatory variables on thedependent variable. The long run effect is captured by the estimates of the normalised Johansen regressionresults presented in equation (16). The diagnostic tests are the t-ratio test of the coefficients, LM test forautocorrelation and the Jarque Berra test for normality of residuals.Table 5: ECM results for rice outputVariable Coefficient t-statistic Prob∆lgarea 0.017611 0.668907 0.0510∆lgoutput(-1) 0.523393 2.283959 0.0319lgRain 0.003740 0.997918 0.0328∆lgmp -0.001107 -0.298174 0.7683∆lgrp 0.009922 0.307758 0.0761Residual -0.174253 -1.321005 0.0043R2 (0.754242) F-statistic (2.523411) Prob(F-stat) 0.058261It indicates that aggregate rice output is dependent on area cultivated, previous year’s output, and previousyear’s price of rice. The coefficient of maize price was not significant.An R2 of 0.754 indicates that 75% of the variation in the dependent variable can be accounted for byvariation in the explanatory variables. The results indicate that area cultivated was significant at 10%, withelasticity of 0.0176, which implies that a one percent increase in area cultivated of rice will lead to a0.018 % increase in output in the short run. This is rather on the low side. This can be attributed to the factthat an increase in the land cultivated without a necessary increase in the resources of the farmer willincrease adverse effect on the minimal resources available to the farmer hence resulting in this marginalincrease in output. This result also indicates that the productivity of the farmer reduces with increase infarm size. The long run elasticity of area cultivated is 0.2167 (in equation (16)) is higher than the 0.0176 forthe short run. This indicates that over the long run farmers adjust their farm sizes more to output than in theshort run. That is, a 1% increase in area cultivated will lead to 0.217% increase in output in the long run.Lagged output has elasticity of 0.523393 in the short run and is significant at the 5% significance levelimplying that a one percent increase in output in a year will lead to a 0.52% increase in output thesubsequent year. This simply means that an increase in output will make capital available for the farmer toinvest in the subsequent year’s rice production activities. This is only true if the additional resource isinvested into the following year’s rice production activities.Rainfall is significant at 5%. An elasticity of 0.003740 for rainfall indicates that a 1% increase in rainfallwill result in a 0.004% rise in output. This low impact of rainfall can be explained by the fact that a largeproportion of total rice output is produced under the irrigation projects across the country. This makes theresponse of output to rainfall highly inelastic.A coefficient of 0.009922 which is significant at 10% for real price of rice indicates that a 1% rise in real 9
  • 10. Journal of Economics and Sustainable Development www.iiste.orgISSN 2222-1700 (Paper) ISSN 2222-2855 (Online)Vol.2, No.6, 2011prices will lead to 0.01% increase in output the subsequent year in the short run. This makes prices alsoinelastic. This low rice price elasticity suggests that farmers do not always necessarily benefit fromincreasing prices due to the structure of the marketing system. Middlemen and other marketing channelmembers purchase the rice from farmers at the farm gate and thereafter transport it to market centres to besold. Hence when there is a rise in prices a little fraction of it is transmitted to the famer. The low priceelasticity could also be attributed to the fact that farmers are hindered by an array of constraints such asland tenure issues, rainfall variability, lack of capital resources and credit facilities which limits theircapacity to respond to price incentives. The long run coefficient of 0.2415 indicates that 1% percentincrease in real prices will lead to a 0.24% increase in output. The long run elasticity far exceeds the shortrun. This is because over the long run when prices show a continual rise, farmers are able to accumulatecapital enough to enhance production.The residual which is the error correction term is significant at 1% and has the expected negative sign. Itmeasures the adjustment to equilibrium. Its coefficient of -0.174 indicates that the 17.4% deviation of riceoutput from long run equilibrium is corrected for in the current period. This slow adjustment can beattributed to the fact that famers in the short run are constrained by technical factors as mentioned earlierwhich limits their ability to adjust immediately to price incentives.In the long run maize price is significant at 1%. With a negative coefficient of 0.009975 it implies that a 1%increase in the price of maize will lead to 0.01% reduction in the output of rice. This is to say that resourceswill be diverted to maize production relative to rice production leading to the fall in rice output. However,Maize price was not significant in the short run. The results of the LM test of serial correlation for up tofifth order show that there is no serial correlation in the data set.3.3.2 Supply Response of Area CultivatedTesting for the selection of lag length yielded the same results for both the acreage and output models.Hence the same lag order three is therefore used for the Vector Error Correction model (VEC).The trace test selects one co-integrating vector while the eigen value test selects two co-integrating vectors.But since the trace test is the more powerful test (Mohammed, 2005), the result from the trace test is usedhere. Thus the acreage model has one co-integrating vector. The normalised cointegration equation for areacultivated is given as;lgarea =4.613649lgoutput+3.11429lrp–0.46020lmp–59247 (18)The ECM for area cultivated is presented as;∆lgarea = α0 + 1i ∆lgoutputt-i + 2i ∆lgareat-i + 3i ∆lgrpt-I + 4i ∆lgmpt-I + 5l lgrain –θECt-I (19)Where θECt-I = α (β1areat-I – β2outputt-I – β3rpt-I – β4mpt-I)The result of the estimated ECM is given in table 6 below.Table 6: ECM results for area cultivatedVariable Coefficient t-statistic Prob∆lgarea(-1) 0.169747 1.343021 0.1192∆lgoutput(-1) 12.80810 1.816372 0.0824lgRain 0.003984 0.138622 0.0891∆lmp(-1) -0.011350 -0.401819 0.0691∆lrp(-1) 2.016747 1.440683 0.0265Residual -0.435411 -1.043641 0.0047R2 (0.770669) F-statistic Prob(F-stat) 0.017301 (1.707148)An R2 of 0.77067 indicates that 77% 0f the variation in the dependent variable is accounted for by variationin the explanatory variables. Output is significant at 10% with an elasticity of 12.80 indicating that a 1%increase in output this year will increase land cultivated in the subsequent year by 12.8% in the short run.Thus acreage cultivated is highly elastic with respect to output. An increase output level gives the farmersan opportunity to acquire necessary resources and equipment to put more land under cultivation. Output 10
  • 11. Journal of Economics and Sustainable Development www.iiste.orgISSN 2222-1700 (Paper) ISSN 2222-2855 (Online)Vol.2, No.6, 2011elasticity is even higher than that for real prices. This can be attributed to the fact that a large proportion ofoutput is for household consumption and thus is independent of prices. Increased output alone is enoughmotivation to increase acreage under cultivation. In the long run output is insignificant. This because in thelong run land will get exhausted and output will be dependent on productivity (and not the area of landcultivated).Rainfall is significant at 10% with a coefficient of 0.003984; this implies a 1% increase in rainfall leads to0.004% increase in area cultivated. This low response to rain can be attributed to the fact that large areasunder rice cultivation are in irrigated fields which depends less on rainfall for cultivation. Rainfall is anexogenous variable (I(0)), hence it measures short run effect.Maize price is significant at 10% with a negative coefficient of -0.011350 in the short run and -0.460 in thelong run. This means over the long run if maize prices are continually increasing farmers will commit moreresources into maize farming relative to rice farming. Thus as the price of maize rises area under ricecultivation reduces both in the short and long run. In the short run a 1% increase in maize price reduces areacultivated of rice by 0.0114% and by 0.46% in the long run.Real price of rice was significant at 5% with elasticities of 2.017 and 3.11 in the short run and long runrespectively, thus price is elastic. This implies a 1% percent increase in real price of rice will result 2.017%and 3.11% increase in acreage cultivated in the short run and long run respectively. This is expected sincein the long run farmers are able to adjust to overcome some the major challenges of production and henceare able to adjust more to price incentives.The error correction term has the expected negative sign and with a coefficient of 0.4354 indicating that43.54% of the deviation from long run equilibrium is corrected for in the current period.The results of the test for serial correlation show that is no autocorrelation in the series. The Jarque-Berratest statistic of 10.18 with probability value of 0.25 implies that the residuals are normally distributed.4. ConclusionsEconomic theory in the past had been based on the assumption that time series data is stationary and hencestandard statistical techniques (Ordinary Least Squares regression (OLS)) designed for stationary series wasused. However, it is now known that many time series data are non-stationary and therefore usingtraditional OLS methods will lead to invalid or spurious results. To address this problem, the method ofdifferencing was introduced. Differencing according to Granger however leads to loss of valuable long-runinformation. Granger and Newbold (1974) introduced the technique of cointegration which takes intoaccount long-run information therefore avoiding spurious results while maintaining long-run information.This study presents an analysis of the responsiveness of rice production in Ghana over the period1970-2008. Annual time series data of aggregate output, total land area cultivated, yield, real prices of riceand maize, and rainfall were used for the analysis.The Augmented-Dickey Fuller test was used to test the stationarity of the individual series. The Johansenmaximum likelihood criterion is used to estimate the short-run and long-run elasticities.The trend analysis for rice output, rice price, acreage cultivated, and yield revealed that only rice price issignificant at 1% significance level. The results imply that for each year, the price of rice will increase by2.589 units.All the time series data that was used were tested for unit root. They were found to be non-stationary atlevels but stationary after first differencing at the one percent significance level except for rainfall whichwas stationary at levels at the 5% significance level. The Likelihood Ratio tests selection of lag orderselected order three by the AIC and HQ criteria for both models. The Johansen cointegration test selectedone cointegration vector (in both rice output and rice price models) indicating that the variables areco-integrated. The diagnostic tests of serial correlation and normality test was done using the LagrangeMultiplier test for autocorrelation and the Jarque-Berra test for normality of residuals. The results indicateno serial correlation and normally distributed residuals.The land area cultivated of rice was significantly dependent on output, rainfall, real price of maize and realprice of rice. The elasticity of lagged output was 12.8 in the short run and was significant at 1%. However,this elasticity was not significant in the long run.Rainfall had an elasticity of 0.004 and significant at 10%. Also real price of maize had negative coefficientof -0.011 which was significant at 10%. This is consistent with theory since a rise in maize price will pull 11
  • 12. Journal of Economics and Sustainable Development www.iiste.orgISSN 2222-1700 (Paper) ISSN 2222-2855 (Online)Vol.2, No.6, 2011resources away from rice production into maize production. The coefficient of the real price of maizeestimated by Chinere (2009) was -0.066 for rice farming in Nigeria. This is higher than the -0.011 estimatedfor Ghana (in this study).The real price of rice had an elasticity of 2.01 and significant at 5% in the short run and an elasticity of 3.11in the long run. The error correction term had the expected negative coefficient of -0.434 which issignificant at 1%. It was found that in the long run only real prices of maize and rice were significant withelasticities of -0.46 and 3.11 respectively. The error correction estimated by Chinere (2009) was -0.575.This indicates that adjustment to long run equilibrium is faster in Nigeria. This could be attributed to betteragricultural infrastructure in Nigeria compared to Ghana. The error correction for Indian rice in the ricezone as estimated by Mohammed (2005) was -0.415. This could be attributed to the fact that Indianagriculture was highly constrained by land problems hence leaving room for little adjustments in terms ofincreasing acreage cultivated.The aggregate output of rice in the short run was found to be dependent on the acreage cultivated, the realprices of rice, rainfall and previous output with elasticities of 0.018, 0.01, 0.004 and 0.52 respectively. Realprice of rice and area cultivated are significant 10% level of significance while rainfall and lagged outputare significant 5%. In the long run aggregate output was found to be dependent on acreage cultivated andthe real price of rice and real price maize with elasticities of 0.218, 0.242 and -0.01 respectively at the 1%significance level.Mythili (2008) used panel data to estimate the supply response of Indian farmers. His findings also supportthe view that farmers’ response to price is low in the short-run and their adjustment to reaching desiredlevels is low for food grains.The analysis showed that short-run response in rice production is lower than long-run response as indicatedby the higher long-run elasticities. This is because in the short run the farmers are constrained by the lack ofresources needed to respond appropriately to incentives. In the short-run inputs such as land, labour, andcapital are fixed. To address these concerns government should devise policies to make land available tofarmers so that prospective farmers could increase acreage cultivated. More irrigation facilities should beconstructed to put more land under cultivation.It was also observed that the acreage model had higher elasticities than the output model. Thus farmers tendto increase acreage cultivated in response to incentives. This implies that farmers have more control overland than the other factors that influence output.Efforts should be put in place to make the acquisition of inputs such as tractors and fertilizer moreaccessible and affordable to farmers, and to improve the road network linking farming communities and theurban centres.Price control policy should be introduced and enforced to address the problem of frequent price fluctuationwhich is the main reason for the low response to prices. Since farmers are aware of these price fluctuationsthey are reluctant to immediately respond positively to price rises.Though there is a market for rice in Ghana, recent developments have shown that consumers prefer foreignpolished rice; therefore, government should put in place the needed infrastructure to process the locallyproduced rice to ensure the sustainability of local rice production.ReferencesBox G. E. P. and Jenkins G. M. (1976). ‘Time Series Analysis: Forecasting and Control’, 2nd ed, SanFrancisco, Holden-day. Cambridge University Press.Chinyere G. O. (2009). ‘Rice output supply response to changes in real prices in Nigeria; An AutoregressiveDistributed Lag Model approach’. Journal of Sustainable Development in Africa, 11(4), 83 -100. ClarionUniversity of Pennsylvania, Clarion, Pennsylvania.Davidson, J. E. H., Handry, D. F., Srba F., and Yeo S. (1978). ‘Econometric modeling of aggregate timeseries relationships between consumer’s expenditures and income in the UK’, Economic Journal. 88,661-692. 12
  • 13. Journal of Economics and Sustainable Development www.iiste.orgISSN 2222-1700 (Paper) ISSN 2222-2855 (Online)Vol.2, No.6, 2011Dickey D. A. and Fuller W. A, (1981). ‘Likelihood ratio statistics for autoregressive time series with a unitroot’. Econometrica, 49, 1057-1072.Engle R. F, and Granger C. W. J. (1987). ‘Cointegration and error correction: Representation, estimationand testing’, Econometrica, 55, 251 -276.FAO (1996). ‘Report of expert consultation on the technological evolution and impact for sustainable riceproduction in Asia and Pacific’. FAO-RAP Publication No. 1997/23, 206 pp. Regional Office for the Asiaand the Pacific.Granger C. W. J. (1981). ‘Some properties of time series data and their use in econometric modelspecification.’ Journal of Econometrics, 16 (1), 121-130.Granger C. W. J. and Newbold P. (1974). ‘Spurious Regression in Econometrics’, Journal of Econometrics2, 111-120.Ghana Statistical Service (2011). ‘Ghana Statistical Service release’.Hall S. G. (1991). ‘The effect of varying length VAR on the maximum likelihood estimates ofco-integrating vectors.’ Scottish Journal of Political Economy, 38, 317- 323.Hallam D. and Zanoli R. (1993). ‘Error Correction Models and Agricultural Supply Response’. EuropeanReview of Agricultural Economics, 20, 151-166.Hansen S. and Juselius K. (1995). ‘Cats in rats: Cointegration analysis of time series, Evanston, Illinois.Harris R, (1995). Using cointgration analysis in econometric modeling. Prentice Hall, HarvesterWheatsheaf, England.Johansen S. and Juselius K. (1990). ‘Maximum likelihood estimation and inference on cointegration- withapplication to the demand for money’. Oxford bulletin of Economics and statistics, 52, 170-209.Kherallah M C., Delgado E., Gabri-Madhin N. M., and Johnson M., (2000). ‘The Road Half Travelled:Agricultural Market Reform in Sub-Saharan Africa’. International Food Policy Research Institute, FoodPolicy Report 10. Washington D.C. Kiel Working Paper No. 1016.Krueger A. O., Schiff M, and Valdés A. (1992). ‘The Political Economy of Agricultural Pricing Policies’. AWorld Bank Comparative Study, Baltimore.Lewis W. A, (1954). ‘Economic Development with Unlimited Supplies of Labour’. Manchester School 22:139–191.Ministry of Food and Agriculture, Ghana (2005), Ministry of Food and Agriculture report.Mohammed S. (2005). ‘Supply response analysis of major crops in different agro-ecological zones inPunjab.’ Pakistan Journal of Agricultural Research, 2007, 124 (54).Mythili G. (2008). ‘Acreage and yield response for major crops in the pre – and post reform periods in India:A dynamic panel data approach’, Report prepared for IGIDR-ERS/USDA project: Agricultural markets andpolicy, Indira Gandhi Institute of Development Research, Mumbai. 13
  • 14. Journal of Economics and Sustainable Development www.iiste.orgISSN 2222-1700 (Paper) ISSN 2222-2855 (Online)Vol.2, No.6, 2011Nerlove M. and Bachman K L., (1960). ‘Analysis of the changes in agricultural supply, Problems andApproaches’. Journal of farm economics. 3(XLII), 531-554.National Rice Policy Development (2009). National Rice Policy Development Report.Statistical Research and Information Division (SRID). (2005). Statistics Research and InformationDepartment Release.Thiele R., (2000). ‘Estimating the aggregate supply response, a survey of techniques and results fordeveloping countries.’ Kiel Institute of world Economics Duuesternbrooker Weg 120 241015 Kiel, Germany.Kiel working paper No 1016 Rainer Thiele.Thiele R, and Wiebelt M, (2000). ‘Adjustment lending in Sub Saharan Africa Kiel Institute of WorldEconomics.’ Kiel Discussion Papers 357, Kiel. Tobacco In Zimbabwe Contributed Paper Presented at the41st Annual.Townsend R and Thirtle C. (1994). ‘Dynamic acreage response. An error correction model for maize andtobacco in Zimbabwe’. Occasional paper number 7 of the International Association AgriculturalEconomics.Townsend T. P. (2001). "World Cotton Market Conditions". Beltwide Cotton Conferences, Proceedings,Cotton Economics and Marketing Conference, National Cotton Council, Memphis, TN, pp. 401-405.World Bank. (1997). "Cotton and Textile Industries: Reforming to Compete." Vol. I, II, Rural DevelopmentSector Unit; South Asian Region, World Bank. Report No. 16347- IN, Washington, D.C, 1997. 14
  • 15. Journal of Economics and Sustainable Development www.iiste.orgISSN 2222-1700 (Paper) ISSN 2222-2855 (Online)Vol.2, No.6, 2011Sustainable Use of Water Resources in the Form of Pisciculture to Generate Income in West Bengal -A study report. Bairagya Ramsundar Department of Economics, SambhuNath College, Labpur, Birbhum, West Bengal, India, Pin-731303 Mobile No. +919474308362 Fax: +913463 266255, E-mail: ramsundarbairagya@gmail.comReceived: October 1st, 2011Accepted: October 11th, 2011Published: October 30th, 2011AbstractThough the two-third of earth’s volume is comprised of water the world is facing the problem of scarcity offresh- water. Because most of the water is sea water which is salt in nature. It is the modern day tragedy thatdue to the scarcity of these valuable lives saving natural resources life will exist no more. So the sustainableuse of this resource is very much important today. India was blessed with vast inland natural waterresources. But Indian Economy faces the problem of proper utilization of these huge water resourcesspread over its vast stretches of land. A proper policy for utilization of these resources would become agoverning direction of economic growth. The role of fisheries in the country’s economic development isamply evident. It generates employment, reduces poverty, generates income, increases food supply andmaintains ecological balance between flora and fauna. Scientists have shown that 3 bighas forest areas areequivalent to 1 bigha plank origin in water bodies which create more O2. This paper intends to develop ascientific plan use of water with a view to sustainable management of water resources to generate incomefrom fishery in the field of pisciculture by using the scarce water resources in a sustainable eco-friendlymanner.Keywords: Ecology, Growth, Perishable, Pisciculture, Pollution, Project, Sustainable development,Composite fish culture1. Introduction According to 2001 Census India is the second (next to China) largest populous country in the worldabout 17.5% of world populations live in India which covers only 2.4% (Dutt R. and Sundharam K.P.M.2009) of geographical land area. The density of population is 324 persons per sq. km. If the current growthrate of population (i.e. 2.11% per annum) continues within 2030 India will be the most populous country inthe world creating a food scarcity and a huge amount of unemployment in a massive scale. About 68% oftotal population spends their livelihood from agriculture and the per capita income is very low. Waterscarcity is too much related to food scarcity. Hence it is very crucial to develop a scientific plan of wateruse with a view to sustainable management of water resource. Management of shortage of water andmanagement of water pollution are complex task. These issues have drawn the attention of the developedand developing countries as well as the various national and international organizations including theJohannesburg Summit 2002, Year 2003 has been declared as year of Fresh -Water. But in India there arehuge prospect to utilize the unused fresh water resources for pisciculture which can play an important rolein this aspect. Realizing its importance during the 5th - five year plan the Government of India introduced 15
  • 16. Journal of Economics and Sustainable Development www.iiste.orgISSN 2222-1700 (Paper) ISSN 2222-2855 (Online)Vol.2, No.6, 2011beneficiary-oriented programme in the form of a pilot project entitled ‘Fish Farmers Development Agency’(FFDA) to provide self employment, financial, technical and extension support to fish farming in ruralareas. In 1974-75 this Programme was further extended under World Bank-assisted, Inland fisheries projectto cover about 200 districts of various states in India. 2. Sustainable development and Pisciculture Sustainable development is a pattern of resource use that aims to meet human needs while preservingthe environment so that these needs can be met not only in the present, but also for future generations. Theterm was used by the Brundtland Commission which coined what has become the most often-quoteddefinition of sustainable development as development that “meets the needs of the present withoutcompromising the ability of future generations to meet their own needs”. It is usually noted that thisrequires the reconciliation of environmental, social and economic demands - the “three pillars” ofsustainability. This view has been expressed as an illustration using three overlapping ellipses indicatingthat the three pillars of sustainability are not mutually exclusive and can be mutuallyreinforcing. Sustainable development ties together concern for the carrying capacity of natural systems withthe social challenges facing humanity. As early as the 1970s “sustainability” was employed to describe aneconomy in equilibrium with basic ecological support systems [Wikipedia]. A primary goal of sustainabledevelopment is to achieve a reasonable and equitable distributed level of economic well being that can beperpetuated continually for next generation. Thus the field of sustainable development can be broken intothree constituent parts i.e. environmental, economic and social sustainability. It is proved that socio-economic sustainability is depended on environmental sustainability because the socio- economic aspects,like agriculture, transport, settlement, and other demographic factors are born and raised up in theenvironmental system. All the environmental set up is depended on a piece of land where it exists. Water is a renewable natural resource and is a free gift of nature. In the early days the supplyof water was plenty in relation to its demand and the price of water was very low or even zero but in courseof time the scenario has totally changed (Bairagya R. and Bairagya H. 2011). Water is a prime need forhuman survival and also an essential input for the development of the nation. For sustainable developmentmanagement of water resources is very important today. Due to rapid growth of population, expansion ofindustries, rapid urbanization etc. the demand of water rises many-fold which is unbearable in relation to itsresources in the earth. All the environmental set up is depended on a piece of land where it exists. So, to getsustainable environmental management, sustainable land management is necessary. As a result the use ofwater resources in the form pisciculture in a sustainable manner may gain priority. 4. Importance of the study Social scientists have always identified the rural areas for investigation. In the case of India to a largenumber of studies have been carried out in rural situations including panchayats and co-operative societies.Though many research works have been done in the biological and marine sciences, the economicinvestigations of pisciculture have not yet been done so far. In this respect the present study has a cleareconomic importance for the upliftment of the rural economy at the grass root level. Thus pisciculture has apositive effect on ecology and hence to maintain ecological balance a meaningful use of unused waterresources has an important role. Scarcity of water is a modern day tragedy human society will exist no more 16
  • 17. Journal of Economics and Sustainable Development www.iiste.orgISSN 2222-1700 (Paper) ISSN 2222-2855 (Online)Vol.2, No.6, 2011due to these scarce resources. So a meaningful use of this resources in the form of pisciculture to generateincome of the rural poor gain topmost priority.The state of West Bengal plays an important role for the implementation of the programme. Though the twodistricts Burdwan and Birbhum of W.B. are primarily agricultural districts, there is huge scope forpisciculture. A few studies have been undertaken by several experts, notably, D. Prasad (1968), R. Charan(1981), A.V. Natarajan (1985), K. M. B. Rahim (1992,93), A Chakravorty (1996), I Guha and R. Neogy(1996), P.K. Ghosh (1998) and others on the economic evaluation of pisciculture. Though these are usefulguides to researchers, yet there is ample scope for further works relating to pisciculture in the rural areas ofW.B. Besides, there is the necessity of developing studies concerning the impact of these programmes onthe rural economy. The present study is a modest attempt in remedying this inadequacy.4.1 ObjectivesThis pilot study was conducted on the following objectives i) To generate employment in rural India. ii) Torise in food supply and reduce mal-nutrition. iii) For proper utilisation of unused water resources. iv) Tomaintain ecological balance in a sustainable manner. V) Identification of fish farmers suitable and willingto develop fish farming in ponds.vi) Arranging training in organized manner and ensuring extensionservices to fish farmers.4.2 Methodology of the study4.2.1 Selection of study areaThe present study was confined to survey the rural areas of Burdwan and Birbhum districts of W. B. inrespect of implementation of the programme of Fish Farmers Development Agency (Mishra S. 1987). Inthe selection of districts following considerations weighed most: i) Both the districts are covered with waterareas constituting half of the total inland water resources. ii) There is a heavy concentration of tanks andponds in both the districts. iii) Both the districts are considered as having one homogeneous agro-climaticzone in view of the broad similarities of soils, climate and other features. Since they are also neighboringdistricts a suitable comparison can easily be made. iv) In both the districts, the FFDA programmes are beingimplemented in full fledged form by the Government authority. v) Data from both the districts can beobtained because of my personal knowledge about the two districts.4.2.2 Selection of sampleKeeping in view the time factor, limited fund and limited ability it was not possible to collect data from allthe recorded fish farmers of the two districts. At first 5 blocks from each district have been purposefullyselected. Then 10 recorded fish farmers from each block have been selected purposefully. Thus 50 fish 17
  • 18. Journal of Economics and Sustainable Development www.iiste.orgISSN 2222-1700 (Paper) ISSN 2222-2855 (Online)Vol.2, No.6, 2011farmers from each district have been selected for interview. For comparative analysis of data each farmerwas taken as the unit. For this study data have been collected on different aspects of the programme such asfarmers’ income, water area, finance, total production, product price, cost of production, profit, duration oftraining period etc. The data have been collected by personal interview method through a questionnaire.Thus the collected data are entirely primary in nature.4.2.3 Location map of the study areaThe state of West Bengal lies between latitude 21038/ to 27010/ North and longitude 85038/ to 89050/ East.The district Burdwan lies between 22056/ and 25053/ North latitude and 86048/ and 88025/ East longitude.The district Birbhum lies between 23032/30// and 24035/00// North Latitudes and 88001/40// and 87005/25//East Longitudes shown in figure-1.5. Fisheries in IndiaIndian fisheries can broadly be divided into two categories: i) Inland fisheries & ii) Marine fisheries.Further Inland fisheries can be classified into two types: i) Capture & ii) Culture. Capture fisheries consistof rivers, lakes canals etc. where farmers do not cultivate fishes. Natural breeding process is the commonphenomenon there. On the other hand, culture fisheries consist of ponds, tanks, swamps, marshes etc. Inthis case, farmers have to sow fish seeds, nurse it and send it to proper size before harvesting. India is thethird largest producer of fish in the world and the second in inland fish production. Indian fishing resourcescomprising of 2 million sq.km.of EEZ for deep sea fishing, 7,520 km. of coast line, 29,000 km. of Rivers,1.7 million Ha. of reservoirs, 1 million Ha. of brackish water and .8 million Ha. of ponds, lakes and tanksfor inland and marine fish production (Giriappa S. (ed) 1994) . About 14 million fishermen draw theirlivelihood from fishery. During the period 1981 to 2002 the contribution of fishery to GDP has increasedfrom Rs.1230 crores to Rs.32060 crores. The fish production increased from 0.7 million tons in1951 to 6.8million tons in 2006.5.1 Fisheries in West BengalWest Bengal has a vast water resource potentiality. By utilizing these water resources there is a hugeprospect of pisciculture (Chakraborty S. K.1991). These resources can be divided into two categories: i)Inland water resources and ii) Marine water resources. Inland resources constitute ponds, rivers, marshylands, canals, reservoirs etc. Water Resources of West Bengal shown in figure-2.It should be noted that tanks/ponds occupy the major share i.e. 46.70% of total inland water resources. Butout of 2, 76,202 Ha. area under ponds and tanks only 2, 20,000 Ha. i.e. 79.65% are presently used forpisciculture which means 20.35% remains unused. Moreover, out of 5, 91,476.71 Ha. total inland waterresource only 2.87000 Ha. water area is brought under pisciculture i.e. 48.56% are presently used and51.44% remains unused. These unused water resources can be brought under pisciculture through properutilization. 18
  • 19. Journal of Economics and Sustainable Development www.iiste.orgISSN 2222-1700 (Paper) ISSN 2222-2855 (Online)Vol.2, No.6, 2011In West Bengal marine fishery has a substantial share amounting to a coast line of 158 km. inshore area upto 10 fathoms depth is 770 sq. km. (Mamoria C.B. 1979), offshore area (10-40) fathoms depth is 1813 sq.km. and a continental shelf up to 100 Fathom is 17,049 sq.km. Out of 19 districts of West Bengal only twodistricts East- Midnapur and South 24-Parganas are coastal. West Bengal is on the top of the list in fishproduction in the country. With the passage of time, more and more people are getting themselves involvedin fishery. As fish constitutes the staple food of the people efforts are being made to augment fishproduction.From the period 1986 to 2000, the total fish production increased from 424000 tons to 1045,000 tons (i.e.2.46 times). At the same period the inland fish production increased 2.25 times and marine fish productionincreased 4.50 times. In the year 1986, the share of inland and marine fisheries to total fish production were91% and 9% respectively. But in the year 2000, the share of inland and marine fisheries to total fishproduction are 83% and 17% respectively. Thus we see that during the period 1986 to 2000, the share ofinland fish to total production declined (from 91% in 1986 to 83% in 2000) and the share of marinefisheries increased (from 9% in 1986 to 17% in 2000).Not only in fish production but also in the demand for fish West Bengal is the highest in the country. Thedomestic demand for fish in West Bengal is high because almost all the people of West Bengal arefish-eating. But the state has a higher demand for fish than its production of fish i.e. this state has a deficitin fish supply. To meet this gap the state West Bengal has to import fish from other states like AndhraPradesh, Tamil Nadu etc. At the same time various efforts have been made to augment fish production tobridge this gap.5.2 Fisheries in Burdwan districtThe district of Burdwan is mainly an agricultural district though it is also well advanced in industrialproduction. It is called the ‘granary’ of West Bengal. The district is filled with well fertile productive land.Not only that, the district is also well endowed with a large number of water bodies in the form of rivers,ditches, canals, marshy lands, ponds, tanks etc. The principal rivers are Ajoy, Damodar and Kunur etc.There is a good prospect of pisciculture by utilizing these water resources. The distribution of total waterresources in Burdwan district (i.e. 66480.82 Ha.) is shown in figure-3.Though the district has huge water resources, during the period 1982-1998 only 50% of this water area wasbrought under scientific pisciculture. In the year 1999, the district’s total fish production was 55,500 metrictons while demand in that year was 65,000 metric tons. Thus there was a deficit of 9,500 metric tons. Tomeet this demand the district had to import fish from neighboring states. The district will be self-sufficientin fish production if the unused 50% water resources are brought under pisciculture.5.3 Fisheries in Birbhum district 19
  • 20. Journal of Economics and Sustainable Development www.iiste.orgISSN 2222-1700 (Paper) ISSN 2222-2855 (Online)Vol.2, No.6, 2011The district of Birbhum is a part of Rarh area. The district is well-drained by a number of rivers and rivulets.The principal rivers of this district are: Ajoy, Brahmani, Dwarka, Kopai, Mayurakshi etc. But most of therivers remain dried up during a greater part of the year. Due to this adverse natural factor pisciculture has notmade any significant progress in this district. Agriculture is the main occupation of the common people ofthis district. The main water areas for pisciculture of this district are rivers, tanks, ponds, khals, Bills, baors,reservoirs etc. The distribution of total water resources in Birbhum district (i.e. 45215.81 Ha.) is shown infigure-4.The figure-4 shows that this district has a large number of water resources. But during the period 1980-1999,only 46% of this water area was brought under pisciculture through FFDA. So there is a huge prospect ofpisciculture by utilizing these unused water resources. In the year 1999 total number of fish farmers in thisdistrict was 2, 00747 out of which 56.5% were males and 43.5% were females. The total number of fishermanfamily in this district was 4,975. The average fish production of this district was 21,000 metric tons and theannual demand was 24,000 metric tons. Thus the district had a deficit (3000 mt) in fish production. To meetits own demand the district has also to import fish from the neighbouring states.6. Findings6.1 Income distribution of sample fish farmers before and after assistance from FFDAIt was observed that 26 farmers in Burdwan district & 32 farmers in Birbhum district have pisciculture asthe main occupation while others have pisciculture as a subsidiary occupation. Their basic occupations areagriculture, business and service and have various types of incomes from those occupations. But as incomegeneration scheme only income from fishery was taken into account i.e. income means income earned fromselling fishes. From sample survey we have collected two sets of income data (one showing income beforeand other showing income after the assistance of FFDA) for each fish farmer. Thus the amount of incomereveals the income of the fish farmer before the receipt of assistance of FFDA and the income after theassistance of FFDA. In this regard the income comparison has done in two areas (namely within thedistrict, between the district and the overall change of income) and‘t’ (Gujarati D.N. 2009) test was used forstatistical analysis. Now the various types of incomes distribution of fish farmers of the two districts areshown in terms of four tables.From table- 1 it is clear in both the districts the average income of most of the fish farmers (i.e. 89 out of100 fish farmers of the two districts taken together) has monthly income before assistance below Rs. 4000i.e. they are basically come from poor family income group i) From the field survey we get that the per capita availability of fish in Burdwan district is 24 kg. per headper year while in Birbhum district it is 23.5 kg. per head per year. Now the world’s per capita availability offish is 11.5 kg. per head per year and for developed countries it is 25 kg. per head per year. But India has alow per capita availability of 3.5 kg. per head per year only. Thus we see that the fishermen have a standardlevel of per capita fish consumption. The point is very important to reduce mal-nutrition (which is acommon phenomenon for the country) by means of intensive pisciculture. But India has a poor per capita 20
  • 21. Journal of Economics and Sustainable Development www.iiste.orgISSN 2222-1700 (Paper) ISSN 2222-2855 (Online)Vol.2, No.6, 2011availability of fish of only 3.5 kg per head per year. Thus we see that the per capita availability of fish forthese 100 farmers’ families of these two districts are much higher than that of India and world’s averageand even equal to the standard of the developed countries.ii) It was found that the average monthly income of fish farmers before the assistance of FFDA in Burdwandistrict was Rs. 2680 and in Birbhum district it was Rs. 2045. But after the assistance of FFDA the averagemonthly incomes of the fish farmers of the two districts increased to Rs.3935 and Rs.3721 respectively.Thus we see that the average monthly income of fish farmers of Burdwan district has increased 1.47 timesand in case of Birbhum district it has increased 1.8 times. The result indicates that the average income offish farmers of Burdwan district was higher than the average income of those of Birbhum district in respectof both before and after the assistance of FFDA. However, the rate of increment was higher in Birbhumthan that of Burdwan district. Thus we can say that the incomes of the fish farmers have improved due topisciculture. iii) Statistical methods were adopted is to judge whether the change of income of the sample fish farmersbefore and after the assistance of FFDA was significant or not. This has been done in two areas: - a) Changeof income within the district and b) Overall change of income of the two districts.6.2 Change of income within the districtBurdwan district- It is seen that the value of‘t’ = 3.31 is significant at .01 level, meaning thereby that thechange income of sample fish farmers of Burdwan district is significant. The gain is in favor of theassistance of FFDA. It may be concluded from the result that there is a positive impact of assistance on fishfarming (i.e. the production of fish). It may also be concluded that production has increased significantly inBurdwan district.Birbhum district-The result indicates that “t” is significant at .01 level. Hence it may be said that inBirbhum district the change of income of fish farmers is significant. 6.3 Overall change of income taking the two districts togetherIt is seen from the above table that the value of‘t’ is significant at .01 level which indicates that the incomeof fish farmers before and after assistance differ significantly. The result reveals that the mean income after assistance is significantly greater than that before assistance. It may be concluded from the overall results considering both the districts that the impact of assistance on fish production is significant.The finding of these results was that the assistance of FFDA programme has raised the fish farmer’s incomeundoubtedly in both the districts. Not only that the change of income was statistically significant but alsothe result revealed that there was a positive impact of FFDA assistance on fish production (i.e. fish 21
  • 22. Journal of Economics and Sustainable Development www.iiste.orgISSN 2222-1700 (Paper) ISSN 2222-2855 (Online)Vol.2, No.6, 2011production has significantly increased). Moreover, it was also found that the gain of income of fish farmersof Birbhum district was higher than that of Burdwan district. Thus we can conclude that in case of theimplementation of the programme of FFDA the district Burdwan has more successful achievements thanthe district of Birbhum.7 Suggestions7.1 General suggestionsi) In pisciculture, fishermen are not only directly employed in fishing but also some other alternativeoccupations like net making, marketing of fish seed and fishery product , transport, boat making etc. manyrural people earn income. Since fish is a perishable commodity proper marketing channels should beestablished. Hence to reduce pressure from agriculture pisciculture may be alternative occupations forgenerating income and employment for a large number of poor people. ii) In Burdwan district there aresome open caste pits (OCP) in Ranigang coal belt and in Birbhum district also such pits are available atKhoirasole block, calamines at Md.Bazar block. Fish production may increase by utilizing these waterresources for pisciculture purposes.iii) The urban waste (i.e. garbage) may be recycled as fish feed (to thoseponds and water areas lying near the towns) to raise fish production and to prevent the environmentalpollution in those areas. v) To make financial support for the poor fishermen Bank should grant loan on along-term basis and at a low rate of interest in proper amount and in proper time.vi) the selection of actualbeneficiary is very much essential and it should be made on the basis of need and neutrally not in politically.vii) Since fish is a perishable commodity storage facilities should be provided such that the fishermen arenot forced to sell their product at a lower price. Prices of organic manures and fish seeds should be kept aslow as possible or the Government should give more subsidies in the case of chemical fertilizers so that thepoor fish farmers can buy them. viii) For welfare measures of the rural poor fishermen some Group andPersonal Insurance Schemes, Old- Age Pension Schemes should be taken and the Fishery Dept. shouldissue “Identity Card” to each fisherman. ix) Since both the districts are agricultural based we shouldinterlink agriculture with pisciculture shown in figure-5. Along with pisciculture in ponds other alliedculture can be inter-linked in composite farming. The concept of composite farming e.g. in the pond- thereare pisciculture and on one side of the pond mulberry trees can be cultivated for the development ofsericulture industry-from there silk industry can be grown. Thus the final products (i.e. silk yarn and silkcloth) come to the market and their waste materials are drained off to the pond and used as fish feed. In thesame pattern on another side of the pond animal husbandry can be practiced (e.g. poultry, duccary andpiggery). Their waste can be used as a valuable manure for fish feed and the residual can be utilized foragricultural production. The excreta of the animal husbandry can also be used in the bio-gas plant for fueland light. The products of animal husbandry e.g. milk, meat and egg come to the market directly. Onanother side of the pond some fruit plants such as papine, guava, mango etc. can be cultivated by using theexcess manure of animal husbandry and the products can be sold in the market.x) Composite fish culture: Stocking of various species should be in a certain proportion such that varioustypes of fishes live in various layers and eat the entire food organism (so called Polyculture or CompositeFish Culture). Stocking is an important factor to raise fish production. On an average the survibility ofstocking is 80% (Jhingran V.G. 1991). This means that 20% of stocking is lost for various reasons such asimproper handling, netting and also eaten by preratory animals like snakes, frogs etc. The main speciescultured in India are Rahu, Catla, Mrigel, Silver Carp (S.C.), Grass Carp (G.C.), Cyprinus Carpio (Cy, Ca.) 22
  • 23. Journal of Economics and Sustainable Development www.iiste.orgISSN 2222-1700 (Paper) ISSN 2222-2855 (Online)Vol.2, No.6, 2011and Bata etc. In polyculture (where various fishes are cultured simultaneously) (Agarwal S.C. 1990) systemall these species are cultured in certain ratios. It is found that all of these species do not generally live in thesame layer. In mixed (or polyculture or composite culture) culture all these fishes are cultured in such a waythat all layer’s feed are used. If we divide the whole layer into 3 parts i.e. upper, middle and lower layers,then Cattla and S.C live in the upper layer, Rahu and G.C. live in the middle layer and Mrigel and Cy. Ca.live in the lower layer. It is also to be noted that these two species in the same layer are not competitors ofeach other. In their feed habit one eats waste weeds and the other eats green weeds. Hence in polyculture allthe pond’s feed are used scientifically and economically. The production of Big head and Big head Africangiant hybrid magur are totally banned because their food habits are too much and hamper the foods of othercommon species. Suppose 1000 fish seed is cultivated in the pond. Then the stocking ratios for variousfishes are Rahu: Catla: Mrigel: S.C.: G.C.: Cy.Car = 3 : 1 : 1.5 : 1.5 : 1 : 2 Thus the number of seeds are,Rahu = 300, Catla = 100, Mrigel = 150, S.C = 150, G.C. = 100, Cy. Ca. = 200. Now the survibility rate is80%.Thus the number of produced matured fishes are Rahu = 240, Catla = 80, Mrigel = 120, S.C. = 120,G.C. = 80, Cy.Car. = 160. Fishes are cultivated for 9-12 months. The culture period may extend up to 18months with partial harvesting between. Moreover, the general annual growth of these fishes are, Rahu =1kg, Catla 1.5 kg. Mrigel 0.75kg, SC. = 0.8 kg, G.C. = 1.5 kg, Cy. Car. = 0.8 kg. The fish farmers collectfish seeds from private traders or from Government fish farm.xi) Since training is a necessary ingredient to raise fish production the Government has initiated somespecial training programmes for fish farmers. Moreover to motivate the poor fish farmers some amount ofremuneration or stipend is paid to the participant during training. The amount of remuneration dependsupon the kind of training, place of training organization and duration of the training programmes conductedby the FFDA. To attend the district level training programmes for the farmers of distant villages, travelingallowances are also paid.xii) It is very essential to use mahua cake because it has dual roles in pisciculture. It firstly acts as apesticide but later it acts as organic manure and produces a desired quantity of fish food organism for thebaby fishes. The presence of ‘Saponin’ in mahua cake kills the pre-ratory animals / fishes of the pond. Thispoisonous effect continues about 10-15 days and after that it acts as manure. On the other hand, the use ofpesticides has a negative long-term effect to ecological balance. So the pesticides are not generally usedthough they are cheaper. The mahua cake should be applied at the rate of 333.33kg/ Yr. / bigha.7.2 Limitations of the studyThe following are the limitations of the study: i) Regarding the change in the level of income before andafter the assistance of FFDA, the statements of the fish farmers have been taken into consideration on goodfaith. ii) Due to limited time, ability and resource constraints, data have been collected from a small numberof fish farmers of the two districts and results are assumed to be the representative for the district as awhole. Iii) Due to difficulties in getting responses from the sample fish farmers sometimes we had to relyon the different FEOs’ opinions and also on different official records on good faith. iv) The observationsmade on the basis of collected data are obviously particularistic in nature in so far as these data relate tomicro-study like the present one (i.e. only 100 fish farmers from the two districts were taken intoconsideration). Micro-studies do not attempt to build general theories but the utility of this type of study isthat large number micro-studies may, in course of time, be helpful in constructing meaningfulgeneralizations. Moreover, it is an explorative study which seeks to explore the conditions of fish farmers 23
  • 24. Journal of Economics and Sustainable Development www.iiste.orgISSN 2222-1700 (Paper) ISSN 2222-2855 (Online)Vol.2, No.6, 2011before and after the assistance of FFDA programme. A much larger study may be undertaken to vindicatethe results obtained from this explorative study.8. ConclusionThe study indicates that the introduction of the programme of FFDA had a clear positive impact on the ruraleconomy through employment and income generation and also raising the standard of living andsocio-economic performances of the rural community of the two districts. It is environmentally viable,sustainable and eco-friendly in nature. So for sustainable use of the scarce water resources it is veryessential in the present context. Unemployment is a curse for the society today and pressure of populationon agricultural sector is rising day by day. Thus fish farming may be an alternative occupation for thoseunemployed people and undoubtedly generate income in a viable method. Therefore it is recommended thatthe present programme should be further spread in the rural areas by means of proper planning, adequatesupervision, effective implementation and better monitoring in a sustainable manner.References:Agarwal S.C. (1990), Fishery Management, Ashis Publishing House, New-Delhi. Bairagya R. and Bairagya H. (2011), “Water Scarcity a Global Problem- An Economic Analysis”,Indian Journal of Landscape Systems and Ecological Studies, Vol.-34, Institute of Landscape, Ecology &Ekistics, Kolkata, 127-132 Bairagya R. (2004), “A Comparative Study of the Functioning of Fish Farmers DevelopmentAgency (FFDA) in the Districts of Burdwan and Birbhum (1985-1995)”, PhD Thesis, Department ofCommerce, the University of Burdwan. Bureau of Applied Economics, Burdwan (1998), Key Statistics of Burdwan, Govt. of W.B.Bureau of Applied Economics, Birbhum (1998), Key Statistics of Birbhum, Govt. of W.BChakraborty S. K. (1991), A Study in Bhery Fisheries, Agro-Economic Reaserch Centre, Visva- Bharati.Dutt R. and Sundharam K.P.M. (2009), Indian Economy, S. Chand and Co. New-Delhi, 101-103Jhingran V.G. (1991), Fish and Fisheries in India, Hindustan Publishing Cor.-Delhi.Giriappa S. (ed) (1994), Role of Fisheries in Rural Development, Daya Publishing House, Delhi. Gujarati D.N. (2009), Basic Econometrics, McGraw-Hill., 98-133Mamoria C.B. (1979), Economic & Commercial Geography of India, Shiblal Agarwala Company, Agra-3Mishra S. (1987), Fisheries in India, Ashis Publishing House, New-DelhiAcknowledgement: In preparation of this paper I deeply acknowledge my respected sir Prof. JaydebSarkhel, Department of Commerce, Burdwan University, West Bengal, India, e-mail: 24
  • 25. Journal of Economics and Sustainable Development www.iiste.orgISSN 2222-1700 (Paper) ISSN 2222-2855 (Online)Vol.2, No.6, 2011jaydebsarkhel@gmail.com Tables and Figures: Figure-1 25
  • 26. Journal of Economics and Sustainable Development www.iiste.orgISSN 2222-1700 (Paper) ISSN 2222-2855 (Online)Vol.2, No.6, 2011 26
  • 27. Journal of Economics and Sustainable Development www.iiste.orgISSN 2222-1700 (Paper) ISSN 2222-2855 (Online)Vol.2, No.6, 2011 Figure-2Figure-3 Figure-4Table -1 Change of income distribution of Burdwan. 27
  • 28. Journal of Economics and Sustainable Development www.iiste.orgISSN 2222-1700 (Paper) ISSN 2222-2855 (Online)Vol.2, No.6, 2011 Table -2 Change of income distribution of Birbhum districtTable--3 Classification of sample fish farmers according to their monthly incomes after the assistance 28
  • 29. Journal of Economics and Sustainable Development www.iiste.orgISSN 2222-1700 (Paper) ISSN 2222-2855 (Online)Vol.2, No.6, 2011 Table- 4 Classification of sample fish farmers according to their monthly incomes after the assistance Source: All tables data sources are computed from field survey. 29
  • 30. Journal of Economics and Sustainable Development www.iiste.orgISSN 2222-1700 (Paper) ISSN 2222-2855 (Online)Vol.2, No.6, 2011 MARKET Light, Fuel Silk yarn & Gobar gas Silk cloth Milk, Meat & Egg Silk industry Cow, Piggary, waste materials Duccary, Poultry w as Sericulture te m at er ia Agricultural production ls Produced Fish (Fruit, papaine, guava, Mulburyculture mango etc.) Side POND Side Circular flow of pisciculture and other allied culture Figure: 5 30
  • 31. Journal of Economics and Sustainable Development www.iiste.orgISSN 2222-1700 (Paper) ISSN 2222-2855 (Online)Vol.2, No.6, 2011 The Nexus of Private Savings and Economic Growth in Emerging Economy: A Case of Nigeria Adelakun, Ojo Johnson Department of Economics, Joseph Ayo Babalola University P.M.B. 5006, Ilesa, Osun State, Nigeria, email joadelakun@yahoo.co.ukReceived: October 11st, 2011Accepted: October 19th, 2011Published: October 30th, 2011AbstractThis study discusses the trend in Nigerian saving behaviour and reviews policy options to increasedomestic saving. It also examines the determinants of private saving in Nigeria during the period covering1970 – 2007. It makes an important contribution to the literature by evaluating the magnitude and directionof the effects of the following key policy and non-policy variables on private saving: Income growth,interest rate, fiscal policy, and financial development. The framework for analysis involves the estimationof a saving rate function derived from the Life Cycle Hypothesis while taking into cognizance the structuralcharacteristics of a developing economy. The study employs the Error-Correction modelling procedurewhich minimizes the possibility of estimating spurious relations, while at the same time retaining long-runinformation. The results of the analysis show that the saving rate rises with both the growth rate ofdisposable income and the real interest rate on bank deposits. Public saving seems not to crowd out privatesaving; suggesting that government policies aimed at improving the fiscal balance has the potential ofbringing about a substantial increase in the national saving rate. Finally, the degree of financial depth has anegative but insignificant impact on saving behaviour in Nigeria.Keywords: Private Saving, Saving Rate, Macroeconomic Policy, Interest Rate, Economic Growth.IntroductionResearchers and policy makers are known to be having growing concern among researchers and policymakers over the declining trend in saving rates and its substantial divergence among countries. This is dueto the critical importance of saving for the maintenance of strong and sustainable growth in the worldeconomy. Over the past three decades, saving rates have doubled in East Asia and stagnated in Sub-SaharanAfrica, Latin America and the Caribbean (Loayza, Schmidt-Hebbel and Serven, 2000). The personal savingrate has been drifting downward for the last two decades. According to the latest statistics, personal savingdeclined from about 10% of disposable income in the early 1980s to 1.8% in 2004. The decline hasreceived particular attention recently because saving was negative in 2005 for the first time since the GreatDepression. Although saving declined in other developed countries during this period, the U.S decline wasmore pronounced than in most of the three countries. Development economists have been concerned for decades about the crucial role of domesticsaving mobilization in the sustenance and reinforcement of the saving-investment-growth chain indeveloping economies. For instance, Aghevli et al (1990) found that the saving rate and investment inhuman capital are indeed closely linked to economic growth. The relationship among saving, investmentand growth has historically been very close; hence, the unsatisfactory growth performance of several 31
  • 32. Journal of Economics and Sustainable Development www.iiste.orgISSN 2222-1700 (Paper) ISSN 2222-2855 (Online)Vol.2, No.6, 2011developing countries has been attributed to poor saving and investment. This poor growth performance has generally led to a dramatic decline in investment. Domesticsaving rates have not fared better, thus worsening the already precarious balance of payments position(Chete, 1999). In the same vein, attempts to correct external imbalances by reducing aggregate demandhave led to a further decline in investment expenditure, thus aggravating the problem of sluggish growthand declining saving and investment rates (Khan and Villanueva, 1991). In addition, low personal savinghas created short-run concerns that a sudden increase in the saving rate could reduce growth of consumerspending, and output and employment.Statement of the Problem The strong positive correlation which exists between saving, investment and growth is wellestablished in the literature. The dismal growth record in most African countries, relative to other regions ofthe world has been of concern to economists. This is because the growth rate registered in most Africancountries is often not commensurate with the level of investment. In Nigeria for instance, the economywitnessed tremendous growth in the 1970s and early 1980s as a result of the oil boom and this led to theinvestment boom especially in the public sector. However, with the collapse of the oil market in the 1980s,investment fell, thereby resulting in a fall in economic growth. For instance, during the investment boom,gross investment as a percentage of Gross Domestic Product (GDP) was 16.8 and 31.4 percent in 1974 and1976 respectively, whereas it declined to 9.5 and 8.9 percent, respectively in 1984 and 1985(CBN 2008). One question begging for an answer is: What is the impact of saving and investment on growth? Ithas been argued that saving affects investment, which in turn influences growth in output. The transformationof initial growth into sustained output expansion requires the accumulation of capital and its correspondingfinancing. An output expansion in turn sets in motion a self reinforcing process by which the anticipatedgrowth encourages investment, which supports growth, as well as financial development. It is certain thatwithout a significant increase in the level of investment (public and private), no meaningful growth in outputwould be achieved. Indeed if private investment remains at the current low level, it will slow down potentialgrowth and reduce long run level of per capita consumption and income, thereby leading to low savings andinvestment.Objective of the Study • This study has the following objectives: • To know the impact of private saving in economic growth in Nigeria. • To also carry out an analysis of the sources and trend of saving in Nigeria. • To also know the motivations of saving and how savings are measured. • To also know how saving affects the economic performance in the country. • To also evaluate the impact of the main determinants of saving identified in the literature on private saving in Nigeria.Method of the Study (Methodology) The framework for this analysis is derived from the life-cycle model which has withstood the test oftime in explaining the changes in private saving over time. It is appropriately modified to accommodate thepeculiarities of a developing country and builds on the existing cross-country literature on saving whichquantifies the effects of a variety of policy and non-policy variables on private saving. Its attractiveness liesin its elegant formulation of the effects of interest rate and growth on saving. In addition, its flexibility makesit possible for other relevant theoretical considerations to be incorporated, thus forming an integratedanalytical framework, without altering its fundamental structure. This framework makes a new contributionto the literature by employing time series data in evaluating the determinants of private saving in Nigeria 32
  • 33. Journal of Economics and Sustainable Development www.iiste.orgISSN 2222-1700 (Paper) ISSN 2222-2855 (Online)Vol.2, No.6, 2011between 1970 and 2009. It does this while explicitly addressing some of the econometric problems arisingfrom the use of time-series data. Literature Review, Theoretical Framework and Empirical EvidenceIntroduction Keynes (1936) defined savings as the excess of income over expenditure on consumption.Meaning that savings is that part of the disposable income of the period which has not passed intoconsumption (Umoh, 2003 and Uremadu, 2005). Given that income is equal to the value of current output;and that current investment (i.e. Gross capital formation) is equal to the value of that part of current output,which is not consumed; savings is equal to the excess of income over consumption. Hence, the equality ofsavings and investment necessarily follow thus:Income = Value of output = Consumption + InvestmentSavings = Income – ConsumptionSavings = Investment ex-post. There abound numerous theoretical evidences concerning the functional relationships betweensavings and a wide range of causal variables. For instance, Juster and Taylor (1975) report that savings is anincreasing function of income. Moreover, Modigliani (1970), Madison (1992), Bosworth (1993), Caroll andWeil (1993), Schmidt-Hebbel, Sarven and Solimano (1994), Modigliani (1992), Jappeli and Pugano (1994),Edwards (1995), Collins (1991) and Uremadu (2000) maintain that there exists a positive relationshipbetween savings and income growth rates. Aghevli (1990) in Ozigbo (1999) reported that there is consensusthat the level of savings is largely determined by the level of income. In Nigeria and other developing economies, there are other evidences that interest rate has significanteffect on financial savings especially time and savings deposits while the structure of deposits wasdetermined by differentials in deposit rates as has been demonstrated in Ndekwu, (1991). He also showedusing monthly data that interest rates deregulation in Nigeria have a positive impact on financial savingsduring the period, 1984-1988.Literature Survey Franco Modigliani in his Life Cycle model determined that over the typical individual’s lifetimehis level of income will fluctuate from low levels in his younger years, to high levels in his middle-agedworking years, back to low levels in his retirement years. However, this individual prefers to maintain arelatively stable level of consumption. In order to maintain this steady consumption, the individual will beforced to borrow during his younger years, save during his middle-aged years and then spend down hissavings in his retirement years. From estimating his model, Modigliani concludes that individuals have amarginal propensity to consume (MPC) out of income of approximately 0.70 and an MPC out of net worthif approximately 0.07 to 0.08. (Ando and Modigliani, 1962) Many researchers have studied the possible determinants of private savings behavior. InAmaoteng’s survey article, he shows that saving has been found to be positively correlated with income,wealth, education, age, a high level of risk tolerance, and a favorable perception of one’s own financialstatus; and negatively correlated with a larger family size. (Amaoteng 2002) Modigliani’s life cycle modelillustrates that age structure can have a strong impact on the level of savings in an economy. Sinceindividuals in the middle-aged working years (which we will define as ages 25-55) tend to save more thanindividuals in the younger (ages 0-24) or retirement years (ages 56+), a population with a higherconcentration of individuals in the middle-aged range will have higher savings rates. (Amaoteng 2002)Trend of Saving in Nigeria In mobilizing funds from the surplus units of the economy, banks incur some costs mainly in interestpayments on deposit accounts. In order to recover the cost of deposit mobilization and other operatingoverheads, banks lend at higher interest rates. The difference between the two types of rates is referred to as 33
  • 34. Journal of Economics and Sustainable Development www.iiste.orgISSN 2222-1700 (Paper) ISSN 2222-2855 (Online)Vol.2, No.6, 2011the interest rate spread or the intermediation spread. The spread measures the efficiency of the intermediationprocess in the market, such that, a high intermediation spread implies that there is inefficiency in the market,especially as it discourages potential savers and borrowers, thus, hampering investment and growth. Prior to the deregulation of the banking sector, interest rates were administratively determined bythe Central Bank. Both the deposit and lending rates were fixed by the CBN on the basis of policy decisions.At that time, the major goals were socially optimum resource allocation, promotion of orderly growth of thefinancial market, as well as reduction of both inflation and the internal debt service burden on thegovernment. During the period 1970 to 1985, the rates were unable to keep pace with prevailing inflationrate, resulting in negative real interest rates. Moreover, the performance of the preferred sectors of theeconomy was below expectation, thus, leading to the deregulation of the interest rate in August 1987 to amarket-based system. This enabled banks to determine their deposit and lending rates according to the marketconditions through negotiations with their customers. However, the minimum rediscount rate (MRR) which is the central bank’s nominal anchorcontinued to be determined by the CBN. The lack of responsiveness of the structure of deposit and lendingrates to market fundamentals makes the interest rate inefficient. The wide divergence between the deposit andlending rates (interest rate spread) is inimical to economic growth and development of the Nigerian economy.Between 1980 and 1984, interest rate differentials averaged 3.9 per cent. Even though this was reasonablewithin the accepted limit, the spread widened between 1985 and 1989, averaging 4.3 per cent per annum. Thisimpacted negatively on the amount of loanable funds available to the private sector for investment. The interest differential further widened to an average of 7.9 per cent between 1990 and 1994.Thereafter, the yearly interest rate spread maintained an upward trend, rising from 8.2 per cent in 1995 to 24.6per cent in 2002, before declining to 15.7 per cent in 2005 (see Figure 1). The widening gap between thedeposit and lending rates reflects the prevailing inefficiencies in the Nigerian banking sector and has deterredpotential investors from borrowing, and thus lowered the level of investment in the economy. Interest Rate Spread (in Percentage)Source: Central Bank of Nigeria i) Statistical Bulletin, 2006 and ii) Annual Report and Statement of Accounts, various years. The use of interest rate spread has however been criticized given that higher levels of interest ratesare usually associated with higher inflation rates, and therefore a higher cost of holding money. In addition,higher inflation rates tend to be associated with higher country premia. As a result of these disadvantages ofinterest rate spread as an indicator of efficiency, net interest margin has been proposed as a better alternative.Net interest margin is equal to total interest revenues minus total interest expenditure divided by the value ofassets. Higher values of net interest margin indicate a higher spread on deposit and lending rates and thereforelower efficiency. Figure 2 shows the interest rate figures in Nigeria between 1970 and 2009. A cursory look revealsthat the nominal interest rate was institutionally determined by the monetary authorities throughout the 1970sand the first half of the 1980s. However, with the advent of the structural adjustment programme in the mid1980s which brought with it a rash of financial sector reforms, Nigeria abandoned its fixed interest rateregime that saw nominal interest rates rising from 9.3 percent in 1985 to 26.8 percent in 1989, and reaching apeak of 29.8 percent in 1992. The figure has since hovered between 13.5 percent and 24.4 percent. It stood at16.5 percent in 2009.Real Interest Rate (in Percentage)Source: Central Bank of Nigeria i) Statistical Bulletin, 2006 andii) Annual Report and Statement of Accounts, various years. 34
  • 35. Journal of Economics and Sustainable Development www.iiste.orgISSN 2222-1700 (Paper) ISSN 2222-2855 (Online)Vol.2, No.6, 2011 The real interest rate figures present an interesting picture. Between 1970 and 2009, the figure wasnegative 20 times, attaining positive figures on 18 occasions. The fixed interest rate regime of the 1970s andearly 1980s no doubt contributed to this negative trend by fixing the interest rate at artificially low levels. Forinstance, in the first two decades (1970 to 1989) when the fixed regime dominated, real interest rate wasnegative 14 times and positive only 6 times. However, in the last two decades (1990 to 2009), when marketforces took over, the real interest rate was negative on only 6 occasions. The inflation rate also played a veryimportant role in making the real interest rate negative for most of the period. A cursory glance at figure 2shows that the years when the realinterest rate was negative usuallycoincided with those of double-digitinflation rates. Table 1 shows the componentsof saving in Nigeria including savingsand time deposits with deposit moneybanks, the national provident fund,federal savings bank, federal mortgagebank, life insurance funds and otherdeposit institutions. Saving and timedeposits in banks is by far the single mostimportant component of saving inNigeria and has witnessed a continuousgrowth over the years. Beginning with asum of N337 million in 1970, it rose toN5.2 billion in 1980. By 1990, the figurehad climbed to N23.1 billion, risingfurther by 2000 to N343.2 billion. As at2005, the figure stood at N1.3 trillion. Its contribution to total savinghas however been mixed. In 1970,savings in banks consisted of 98.8 percentof total saving, with this figure reducinggradually to 89.5 percent in 1980, andfurther declining to 78 percent in 1990.From then the percentage of savings inbanks in total saving has witnessed anupward trend, rising to 89.1 percent in2000. Since 2003, this percentage hasbeen 100 percent showing that it hasbecome the only component of saving. The National Provident Fundand the Federal Mortgage bank were bothestablished in 1974. Beginning withN130 million, the National ProvidentFund rose to N724 million in 1990,reaching a peak of N1.37 billion in 1998.The fund maintained this figure till 2002when it was scrapped by the government.The Federal Mortgage Bank on the otherhand experienced a more rapid growth,rising from a paltry N7.3 million at its inception in 1974 to N305 million in 1990. By 2002 when it ceased to 35
  • 36. Journal of Economics and Sustainable Development www.iiste.orgISSN 2222-1700 (Paper) ISSN 2222-2855 (Online)Vol.2, No.6, 2011exist, it had mobilized N22.3 billion. The figures for the Federal Savings Bank have been mixed. It stood atN4.9 million in 1970, increasing to N8.1 million in 1978. It thereafter declined to N4.0 billion in 1982, afterwhich the figure climbed steadily till it reached N37.5 billion in 1989 when it was discontinued. Lifeinsurance funds were established in the same year 1989 with the sum of N1.1 billion. The figure rose sharplyto N19.4 billion in 1994 thereafter witnessing a rapid decline. The amount mobilized stood at N8.5 billion in2002 when the federal government scrapped it.Savings, Growth and Fiscal Deficit (in percent)Notes: i) Savings is the ratio of private saving to Gross National Disposable Income (GDNI); ii) Growth is the growth rate of real per capita GNDI; iii) Fiscal Balance is the surplus or deficit of the entire federation as a percentage of GDP.Source: Central Bank of Nigeria i) Statistical Bulletin, 2006 andii) Annual Report and Statement of Accounts, various years. Figure 3 shows the other macroeconomic variables of interest, including private saving rate, growthand fiscal balance. The Nigerian economy has witnessed several fluctuations in its chequered history, witheconomic growth fluctuating between 45 percent and -31 percent in the period between 1970 and 2009. In the27 year period between 1974 and 2001, the economy experienced negative growth 14 times, while making apositive showing only 13 times. However, growth has been positive since 2002. Fiscal balance was evenmore troubling given that Nigeria experienced a budget surplus only six times out of the 38 year periodbetween 1970 and 2009. The State governments have been as culpable as the government at the centre, witheach level seemingly competing to outspend the other. Private saving witnessed much less volatility, with the variable recording a negative value only oncein the 38 year period. The saving rate fluctuated between 20 percent and 41 percent between 1970 and 1979.This figures changed to 14 percent and 36 percent in the next decade. Between 1990 and 1999, the saving ratehovered between -0.6 percent and 39 percent, reaching an impressive range of between 20 percent and 65percent in the period 2000 to 2009. The private saving rate stood at 58 percent in 2009.Theoretical Framework The life-cycle hypothesis was formulated by Modigliani (1970) and is the principal theoreticalunderpinning that has guided the study of savings behaviour over the years. A critical analysis of this theoryhowever shows that it seems to mirror what happens in developed economies with little or no regard to thepeculiarities of developing countries like Nigeria. There are a number of reasons that make it imperative forsaving behaviour in developing countries to be modelled separately from that in developed economies. First,at the microeconomic level, developing-country households tend to be large and poor. They have a differentdemographic structure, more of them are likely to be engaged in agriculture, and their income prospects aremuch more uncertain. The problem of allocating income over time thus looks rather different in the twocontexts, and the same basic models have different implications for behaviour and policy. Second, at the macroeconomic level, both developing and developed countries are concerned withsaving and growth, with the possible distortion of aggregate saving, and with saving as a measure ofeconomic performance. However, few developing countries possess the sort of fiscal system that permitsdeliberate manipulation of personal disposable income to help stabilize output and employment. Third, muchof the literature in the last five decades expresses the belief that saving is too low, and that development andgrowth are impeded by the shortfall. Sometimes the problem is blamed on the lack of government policy,other times on misguided policy. Lastly, saving is even more difficult to measure in developing than inadvanced economies, whether at the household level or as a macroeconomic aggregate. The resulting datainadequacies are pervasive and have seriously hampered progress in answering basic questions. 36
  • 37. Journal of Economics and Sustainable Development www.iiste.orgISSN 2222-1700 (Paper) ISSN 2222-2855 (Online)Vol.2, No.6, 2011 Given the above, and following Deaton (1989), this paper appropriately modifies the life-cycletheory by developing a model of households which cannot borrow but which accumulate assets as a bufferstock to protect consumption when incomes are low. Such households dissave as often as they save, do notaccumulate assets over the long term, and have on average very small asset holdings. However, theirconsumption is markedly smoother than their income. Following McKinnon (1973) and Shaw (1973), we argue that for the typical developing country, thenet impact of a change in real interest rate on saving is likely to be positive. This is because, in the typicaldeveloping economy where there is no robust market for stocks and bonds, cash balances and quasi-monetaryassets usually account for a greater proportion of household saving compared to that in developed countries.In addition, in an environment where self-financing and bank loans constitute the major source of investmentfunds, accumulation of financial saving is driven mainly by the decision to invest and not by the desire to liveon interest income. Given the peculiarities of saving behaviour, in addition to the fact that the bulk of savingcomes from small savers, the substitution effect is usually larger than the income effect of an interest ratechange.Empirical Evidence There is an abundance of empirical studies that deal with the impact of the different variables ofinterest on savings mobilization. Some authors have found a strong positive relationship between real percapita growth and saving rates (see for example, Modigliani, 1970; Bosworth, 1993; and Carrol and Weil,1994). However, its structural interpretation is controversial, since it is viewed both as evidence that growthdrives saving (Modigliani, 1970; and Carrol and Weil, 1994) and that saving drives growth through thesaving-investment link (Levine and Renelt, 1992; and Mankiw, Romer and Weil, 1992). Given the importance of controlling for the joint endogeneity of saving and income growth, a panelinstrumental-variable approach to estimate the effect of income growth on saving was carried out by Loayza,Schmidt-Hebbel, and Serven (2000). They found that a one percentage point rise in growth rate increases theprivate saving rate by a similar amount, although this effect may be partly transitory. In their study, theyutilized the world saving database, whose broad coverage makes it the largest and most systematic collectionof annual time series on country saving rates and saving-related variables, spanning 35 years (1960 – 1994)and 134 countries (112 developing and 12 industrial). Obadan and Odusola (2001) employed both graphicalanalysis as well as Granger Causality tests to determine the impact of growth on saving. Their resultsrevealed that growth of income does not Granger-cause saving, suggesting that saving is not income-inducedin Nigeria. Evidence on the reverse causation argument also shows that saving does not Granger-causegrowth. The findings therefore do not show any direct relationship between saving and income growth. Analytically, the effect of financial liberalization on private saving rates works through theexpansion of the supply of credit to previously credit-constrained private agents. This allows households andsmall firms to use collateral more widely, and reduces down payments on loans for consumer durables andhousing. Quantitative evidence strongly supports the theoretical prediction that the expansion of credit shouldreduce private saving as individuals are able to finance higher consumption at their current income level.Loayza, Schmidt-Hebbel, and Serven (2000), find that a 1 percentage point increase in the ratio of privatecredit flows to income reduces the long-term private saving rate by 0.75 percentage point. Bandiera andothers (2000), on carrying out a deeper analysis of eight episodes of financial liberalization, failed to find asystematic direct effect on saving rate: it was positive in some cases (Ghana and Turkey), clearly negative inothers (Mexico and Korea), and negligible in the rest. These studies however, have a number of shortcomings. To begin with, each of them focuses ononly one of the determinants of saving. They therefore do not identify the determinants of saving and analyzetheir impact on the saving rate. In addition, the conclusion of Essien and Onwioduokit (1998) should be takenwith a measure of caution. This is because the time span of their study is relatively short (1987-1993). It istherefore difficult to separate the effect of financial development from the effect of recovery and increasedcapital inflow to the economy, all of which were taking place concurrently. Our study will try to overcomethis problem of simultaneity by using a longer time frame dating from 1970-2009. 37
  • 38. Journal of Economics and Sustainable Development www.iiste.orgISSN 2222-1700 (Paper) ISSN 2222-2855 (Online)Vol.2, No.6, 2011Research Methodology The methodology used in this study is the Cointegration and Error-Correction Methodology (ECM).The ECM is made up of models in both levels and differences of variables and is compatible with long-runequilibrium behaviour.Model Specification Drawing from the analysis above on the life cycle framework, the following model was specified: PSR= β0 + β1GRCY +β2RIR + β3FB + β4DFD + εWhere: β1 β2 and β4 ˃0, while β3 ˂ 0 andPSR = private saving rateGRCY = growth rate of real per capita GNDIRIR = real interest rateFB = fiscal balanceDFD = degree of financial depth The saving equation was estimated using annual data for the period 1970-2009. The estimationperiod was determined largely by the availability of adequate data on all variables.Descriptive Statistics. The characteristics of the distribution of the variables are presented in Table 1 below. Jarque-Bera is atest statistic for testing whether the series is normally distributed. The test statistic measures the difference ofthe skewness and the kurtosis of the series with those from the normal distribution. Evidently, theJarque-Bera statistic rejects the null hypothesis of normal distribution for the real interest rate. On thecontrary, the null hypothesis of normal distribution is accepted for degree of financial depth, fiscal balance,income growth and private saving. In Nigeria, as in most developing countries, due to the absence of detailed statistical coverage ofsectoral financial activity, most of the data on saving are obtained from the national accounts statistics asthe difference between measurable aggregates. This residual or indirect approach to the calculation ofsaving has some drawbacks. First, the saving of one group of economic units used by another forconsumption is not captured. Second, capital gains and losses induced by price changes are not treatedadequately. Third, consumer durables and certain elements of government expenditure are also notadequately treated (see Shafer, Elmeskov, and Tease, 1992). For these reasons, the results obtained shouldbe interpreted with caution. Table 2. Summary of the Descriptive Statistics of the Variables DFD FB GRCY PSR RIR Mean 24.24 -3.46 2.02 28.69 -5.31 Median 24.00 -3.50 3.00 26.00 -0.60 Maximum 35.00 9.80 45.00 65.00 18.00 Minimum 12.00 -11.10 -31.00 -0.60 -52.60 Std. Dev. 6.39 4.29 17.84 12.79 16.01 Skewness -0.07 0.52 0.48 0.56 -1.05 Kurtosis 2.009 4.01 3.33 4.05 3.74 Jarque-Bera 1.54 3.24 1.61 3.65 7.61 Probability 0.46 0.20 0.45 0.16 0.02 38
  • 39. Journal of Economics and Sustainable Development www.iiste.orgISSN 2222-1700 (Paper) ISSN 2222-2855 (Online)Vol.2, No.6, 2011 Sum 897.00 -127.99 74.70 1061.40 -196.40 Sum Sq. Dev. 1472.81 661.12 11459.88 5886.52 9229.21 Observations 37 37 37 37 37 Source: National accounts statisticsResults of Stationarity Tests Testing for the existence of unit roots is a principal concern in the study of time series models andcointegration. The presence of a unit root implies that the time series under investigation is non-stationary;while the absence of a unit roots shows that the stochastic process is stationary (see Iyoha and Ekanem,2002). The time series behaviour of each of the series using the Augmented Dickey-Fuller andPhillips-Perron tests are presented in Tables 3 and 4, respectively. The results show that while the privatesaving rate (PSR), growth rate of real per capita GNDI (GRCY) and fiscal balance (FB) are I(0) variables(stationary before differencing), real interest rate (RIR) and the degree of financial depth (DFD) are I(1)variables (stationary after first differencing). This is deduced from the fact that the absolute values of both theADF and PP test statistics of RIR, GRCY and FB before differencing are greater than the absolute value ofthe critical values at the 1 percent significance level. For the other variables, this is the case only afterdifferencing once. Table 3. Results of Augmented Dickey Fuller (ADF) Unit Root Test Variable ADF Value ADF Value Critical Value Level of before After Integration Differencing Differencing PSR -3.657* n.a 3.621 I(0) GRCY -5.068* n.a 3.627 I(0) RIR -3.204 -6.275* 3.621 I(1) FB -4.450* n.a 3.621 I(0) DFD -1.979 -5.784* 3.621 I(1)Notes: * denotes significant at 1 percent; the null hypothesis is that there is a unit root. n.a = not applicable Table 4. Results of Phillips-Perron (PP) Unit Root Test Variable PP Value PP Value After Critical Value Level of Before Differencing Integration Differencing PSR -3.683* n.a 3.621 I(0) GRCY -5.019* n.a 3.627 I(0) RIR -3.045 -13.017* 3.621 I(1) FB -4.405* n.a 3.621 I(0) DFD -2.047 -5.784* 3.621 I(1)Notes: * denotes significant at 1 percent; the null hypothesis is that there is a unit root. n.a = not applicableCointegrated Models 39
  • 40. Journal of Economics and Sustainable Development www.iiste.orgISSN 2222-1700 (Paper) ISSN 2222-2855 (Online)Vol.2, No.6, 2011 In this study, the method established by Johansen (see Johansen, 1991) was employed in carryingout the cointegration test. This is a powerful cointegration test, particularly when a multivariate model isused. Moreover, it is robust to various departures from normality in that it allows any of the five variables inthe model to be used as the dependent variable while maintaining the same cointegration results. Accordingly, Johansen’s test was carried out to check if the saving equation is cointegrated. Table 5shows that both the Trace and Maximum Eigen statistics rejected the null of no cointegration at the 5 percentlevel; while Trace test indicated that there are two cointegrating equations at the 5 percent level; MaximumEigen test indicated only one cointegrating equation at the 5 percent level. The implication is that a linearcombination of all the five series was found to be stationary and thus, are said to be cointegrated. In otherwords, there is a stable long-run relationship between them and so we can avoid both the spurious andinconsistent regression problems which otherwise would occur with regression of non-stationary data series. Table 5 Johansen’s Cointegration Test Results Maximum Eigenvalue Test Trace Test Null Alternative Eigen-value Critical Alternative LR Ratio Critical Hypothesis Hypothesis Value Hypothesis Value 95% 99% 95% 99% r=0 r=1 39.79* 37.52 r≥1 108.69** 87.31 96.58 42.36 r≤1 r=2 31.30 31.46 36.65 r≥2 68.90* 62.99 70.05 r≤2 r=3 18.02 25.54 30.34 r≥3 37.60 42.44 48.45 r≤3 r=4 16.09 18.96 23.65 r≥4 19.58 19.58 30.45 r≤4 r=5 3.49 12.25 16.26 r≥5 3.49 12.25 16.26 Notes: * denotes significant at the 5% level ** denotes significant at the 1% levelLong run Model We now present the results for the long run relationship.PSR = +0.4013 +0.5016GRCY +0.0028RIR -0.0190FB -0.1226DFD (3.346)** (2.233)* (3.769)** (0.459) As postulated by our modified version of the lifecycle hypothesis, the income growth variable(GRCY) is an important determinant of the private saving rate. The coefficient of GRCY is both positivelysigned and statistically significant at the 1 percent level. An increase in the growth rate by one percent leadsto a long-run increase in the saving rate by 0.5 percent. These results are consistent with those obtained byModigliani (1970), Maddison (1992), Bosworth (1993) and Carroll and Weil (1994). Thus, as the incomes ofprivate agents grow faster, their saving rate increases. This is consistent with the existence of consumptionhabits and our modified version of the Lifecycle model. The implication is that any policy that encouragesincome growth in the long run will have a strong impact on private saving rate. Given the historical close linkbetween saving and investment rate, a rise in growth rate will lead to a virtuous cycle of higher income andsaving rates. The result for the real interest rate variable suggests that the real rate of return on bank deposits hasa statistically significant positive effect on saving behaviour in Nigeria. A one percent increase in RIR isassociated with a 0.003 percentage point increase in the private saving rate. This finding is consistent with theMcKinnon-Shaw proposition which states that, in an economy where the saving behaviour is highly intensivein money and near-money assets, the direct incentive effect of high real interest rates on saving behaviour(i.e. the income effect) generally overwhelms the substitution of other assets for financial assets in response 40
  • 41. Journal of Economics and Sustainable Development www.iiste.orgISSN 2222-1700 (Paper) ISSN 2222-2855 (Online)Vol.2, No.6, 2011when faced with such interest rate changes (i.e. the substitution effect). The implication is that governmentshould find an effective mechanism for increasing the abysmally low interest rate on bank deposits if thepresent crusade to increase the private saving rate is to achieve any measure of success. The result for fiscal balance points to a significant substitutability between public and private savingin the Nigerian context. However, there is no support for full Ricardian equivalence, which predicts fullcounterbalancing of public saving by private dis-saving. Specifically, an improvement in the fiscal balanceby one percent is associated with 0.019 percentage point reduction in the private saving rate. The rather weakprivate saving offset to changes in the fiscal balance behaviour may be explained by substantial uncertainty inthe economy, widespread liquidity (or wealth) constraints, tax-induced distortions and limits in households’attempts to smooth consumption over time. Thus in the Nigerian context, policies geared to improvement infiscal balance has the potential of bringing about a substantial net increase in total domestic saving. Thisfinding is consistent with cross-country results of Corbo and Schmidt-Hebbel (1991) and those of Athukoralaand Sen (2004) for India. The degree of financial depth failed to attain statistical significance in the saving function. Thus,there is no empirical support for the view that the development of the financial sector has contributed to thegrowth in private saving. The implication is that financial deepening may not bring about an automaticimprovement in the saving rate. For this, one requires a deeper analytical understanding of the various factorsat work here.Empirical Results Dynamic Error-Correction Model Having identified the cointegrating vector using Johansen, we proceed to investigate the dynamicsof the saving process. Table 6 reports the final parsimonious estimated equation together with a set ofcommonly used diagnostic statistics. The estimated saving function performs well by the relevant diagnostictests. In terms of the Chow test for parameter stability conducted by splitting the total sample period into1970-1986 and 1987-2009 there is no evidence of parameter instability.The results show that the coefficient of the error-correction term for the estimated saving equation is bothstatistically significant and negative. Thus, it will rightly act to correct any deviations from long-runequilibrium. Specifically, if actual equilibrium value is too high, the error correction term will reduce it,while if it is too low, the error correction term will raise it. The coefficient of -0.4415 denotes that 44percent of any past deviation will be corrected in the current period. Thus, it will take more than two yearsfor any disequilibrium to be corrected. The Keynesian absolute income hypothesis is found to hold for saving behavior in Nigeria. Thecoefficient for real per capita GNDI (GRCY) is positive and statistically significant at the 1 percent level.Thus the Nigerian experience provides support for the argument that, for countries in the initial stages ofdevelopment, the level of income is an important determinant of the capacity to save. In this respect, ourresults are consistent with the cross-country results of Modigliani (1993), Hussein and Thirlwall (1999),Loayza et al (2000) and the results for India of Athukorala and Sen (2004). This implies that the highunemployment rate which results in low disposable income is a strong impediment in raising the saving ratein Nigeria. Contrary to the postulation of the Life-Cycle Model, the income growth variable (GRCY) wasfound to have a significant negative impact on the private saving rate. This result is interesting given that itdoes not conform to those obtained from earlier studies (see Modigliani, 1970; Madison, 1992; Bosworth,1993 and Carroll and Weil, 1994). Our Nigerian experience seems to provide support for the simplepermanent income theory which predicts that higher growth (i.e. higher future income) could reduce currentsaving. In other words, at sufficiently high rates of economic growth, the aggregate saving rate maydecrease if the lifetime wealth of the young is high enough relative to that of their elders (see Athukoralaand Sen, 2004). There are two plausible explanations for this finding. The first is the penchant of Nigeriansto indulge in conspicuous consumption. As a result, growth in per capita income could actually lead to a 41
  • 42. Journal of Economics and Sustainable Development www.iiste.orgISSN 2222-1700 (Paper) ISSN 2222-2855 (Online)Vol.2, No.6, 2011decrease in saving. The second is that income growth was actually negative in roughly half of the periodunder observation.Table 6. Estimated Short Run Regression Results for the Private Saving Model Dependent Variable:DPSRIncluded observations: 35 after adjusting endpoints. Variable Coefficient T-Statistic Probability C 0.1137 2.9728 0.0063 DPSR(-1) 0.0303 0.1952 0.8467 DGRCY 0.3047 3.5435 0.0015 DRIR(-1) -0.0016 -1.6013 0.1214 DFB -0.0054 -1.2194 0.2337 DDFD 0.8020 1.6733 0.1063 ECM(-1) -0.4415 -3.3118 0.0027 Adjusted R-squared 0.3356 S.D Dependent Var. 0.1064 S.E of regression 0.0867 F-Statistic 3.6936 Durbin-Watson stat 2.2200 Prob. (F-statistic) 0.0087JBN – χ2 (1) = 0.33 LM – χ2 (1) = 1.92Probability (JBN) = 0.85 Probability (LM) = 0.18ARCH – χ2 (1) = 1.0 CHOW – χ2 (1) = 1.6Probability (ARCH) = 0.32 Probability (CHOW) = 0.20 Furthermore, it is only the income growth variable that is statistically significant at the 1 percentlevel, indicating that in the short run, it is only growth in income that has a relationship with the private savingrate. The implication is that short run changes in private saving rate that correct for past deviations emanateprincipally from changes in income growth. The coefficient estimate shows that a unit change in incomegrowth will bring about a 0.3 percent change in private saving. The other four explanatory variables (PSR(-1), RIR, FB and DFD) do not have any short run impact on the private saving rate. This result is in keepingwith the long run relationship where over 50 percent of changes in private saving are explained by changes inincome growth.Conclusion This paper has investigated the determinants of private saving in Nigeria for the period 1970-2009.In the first place, it attempts to shed more light on the problems associated with the conventional models ofdeterminants of saving. Drawing on econometric analysis, it goes on to propose the alternative of anError-Correction Model of the determinants of saving function. The estimation results for the long runmodel point to the growth in income and the real interest rate as having statistically significant positiveinfluences on domestic saving. There is also a clear role for fiscal policy in increasing total saving in theeconomy, with the private sector considering public saving as an imperfect substitute for its own saving.The Ricardian equivalence was thus, found not to hold in Nigeria contrary to what obtains in industrializedand semi-industrialized economies. Finally, financial development seems not to have any impact on thesaving rate. We began this study by asking what the relevant policies for raising the Nigerian saving rate are.Our results help to understand the effectiveness of policy variables in raising the saving rate in terms oftheir magnitude and direction.Policy Implications 42
  • 43. Journal of Economics and Sustainable Development www.iiste.orgISSN 2222-1700 (Paper) ISSN 2222-2855 (Online)Vol.2, No.6, 2011 A stronger policy framework is imperative in bringing about improved macroeconomicperformance. The government should sustain its National Economic Empowerment and DevelopmentStrategy (NEEDS) programme which is partly responsible for the increasing diversification emerging in theeconomy. The growing contribution of non-oil sectors in GDP growth in recent years is a positivedevelopment and should be encouraged. Agriculture has grown strongly in recent years and was the largestindustry contribution to GDP in 2009. With about 70 per cent of the working population employed in theagricultural sector, the strong agricultural contribution to GDP bodes well for employment. Moreimportantly, government’s efforts to diversify the economy appear to be yielding results and should besustained.Recommendations Some major recommendations for policy can be drawn from the analysis. First, the focus ofdevelopment policy in Nigeria should be to increase the productive base of the economy in order to promotereal income growth and reduce unemployment. For this to be achieved, a diversification of the country’sresource base is indispensable. This policy thrust should include a return to agriculture; the adoption of acomprehensive energy policy, with stable electricity as a critical factor; the establishment of a viable iron andsteel industry; the promotion of small and medium scale enterprises, as well as a serious effort at improvinginformation technology. Second, contrary to popular belief, income growth has a negative influence on private saving inNigeria. Policy makers should thus take explicit account of this result in the formulation of economic policy.For instance past experience has shown that rapid increases in wages of urban sector workers did not result inany appreciable increase in private saving. Rather, the extra income was used in the purchase of mainlyimported consumer goods, thus increasing our dependence on imports. Third, public saving has been shown to be a complement rather than a substitute for private saving inNigeria. Government should therefore sustain its oil- price-based fiscal rule (OPFR) which is designed to linkgovernment spending to notional long run oil price, thereby de-linking government spending from current oilrevenues. This mechanism will drastically reduce the short term impact of fluctuations in the oil price ongovernment’s fiscal programmes. State governments should also desist from spending their share of excesscrude oil revenue indiscriminately. This is because this practice can severely test the absorptive capacity ofthe economy in addition to risking the fuelling of inflation. The challenge is for state governments to saveexcess revenue or spend it directly on imported capital goods in order to sustain Nigeria’s hard-wonmacroeconomic stability. Fourth, monetary policy should focus on ways of increasing the abysmally low real interest rate onbank deposits. It should also devise means of substantially reducing the interest rate spread. Lastly, it ispertinent to note that even though this paper has concentrated on Nigeria, its results can be applied to otherAfrican countries not previously studied. They contain some valuable lessons for informing policy measuresin the current thrust towards greater mobilization of private saving in the African continent.ReferencesAghevli, Bijan, James Boughton, Peter Montiel, Delano Villanueva, and Geoffrey Woglom. 1990. “The Roleof National Saving in the World Economy: Recent Trends and Prospects.” IMF Occasional Paper No. 67.Ajakaiye,Obadan, and Ayodele Odusola 1995. “Real Deposit Rates and Financial Savings Mobilization inNigeria: An Empirical Test”. Journal of Economic Management. Vol.2. No.2 October 1995.Amoateng, Kofi A. 2002. Do Stock and Home Ownership Influence U.S. Personal Savings? ManagerialFinance 28 no.4: 1-11. www.emeraldinsight.com (accessed September 5, 2006)Ando, Albert and Franco Modigliani. 1962. The “Life Cycle” Hypothesis of Saving: Aggregate Implicationsand Tests. The American Economic Review 52 no.5 (Dec):55-82Athukorala, Prema-Chandra, and Kunal Sen. 2004 “The Determinants of Private Saving in India”, WorldDevelopment, Vol. 32, No. 3, pp. 491-503. 43
  • 44. Journal of Economics and Sustainable Development www.iiste.orgISSN 2222-1700 (Paper) ISSN 2222-2855 (Online)Vol.2, No.6, 2011Bandiera, Oriana, Gerard Caprio, Patrick Honohan, and Fabio Schiantarelli. 2000. “Does Financial ReformRaise or Reduce Private Saving?, Review of Economics and Statistics, 82(2): 239-263.Barro, Robert. 1974. “Are Government Bonds Net Wealth?”, Journal of Political Economy, 82(6):1095-1117.Bosworth, B.P. 1993. “Saving and Investment in a Global Economy. Washington, D.C.: BrookingsInstitution.Carroll, Christopher, and David Weil. 1994. “Saving and Growth: A Reinterpretation.” Carnegie-RochesterConference Series on Public Policy 40:133-192.Central Bank of Nigeria. 2003 & 2006. “CBN Statistical Bulletin.” Abuja, NigeriaChete, Louis. 1999. “Macroeconomic Determinants of Private Savings in Nigeria.” Monograph Series No. 7.Nigerian Institute of Social and Economic Research (NISER).Collins, S.M. 1991. “Savings Behaviour in Ten Developing Countries”, in B.D.Corbo, Vittorio, and Klaus Schmidt-Hebbel. 1991. “Public Policies and Saving in Developing Countries”,Journal of Development Economics, 36(1) :89-115.Craigwell, R. 1991. “Demand for money in Jamaica: A Cointegration Approach.” Money Affairs, Vol. 4, pp19-41Dickey, D. and W. Fuller. 1981. “Likelihood ratio statistics for autoregressive time series with a unit root.”Econometrica, 50, 1057-1072Douglas and J.B. Shoven, (eds): National Savings and Economic Performance. National Bureau of economicresearch, University of Chicago PressDownes, A., C. Holder, and H. Leon. 1991. “A Cointegration Approach to Modelling Inflation in a SmallOpen Economy”, Journal of Economic Development, Vol. 16, No. 1, pp. 57-67.Edwards, S. 1995. “Why are Savings Rates so Different across Countries?An International Comparative Analysis”, NBER Working Paper No. 5097Essien, E, and Onwioduokit. 1989. “Recent Developments in Econometrics: An Application to FinancialLiberalization and Saving in Nigeria”. NDIC Quarterly, Vol.8, No. 112.Gersovitz, M. 1988. “Saving and Development”, in H. Chenery and T.N. Srinivasan, eds., Handbook ofDevelopment Economics, Vol. 1, Amsterdam: Elsevier.Hussein, K. and A. Thirlwall. 1999. “Explaining differences in the domestic savings ratio across countries:Apanel data study.”Journal of Development Studies,36,31-52.International Monetary Fund (IMF). 1999, 2006. “International Financial Statistics Yearbook.” Washington,D.C.Iyoha, Milton, and O.T. Ekanem. 2002. “Introduction to Econometrics.” Mareh Publishers, Benin City,Nigeria.Johansen, S. 1991. “Estimation and hypothesis testing of cointegration vectors, in Gaussian vectorautoregressive models.” Econometrica, 59:1551-1580.Juster, F.T. and L.D. Taylor 1975. “Towards a Theory of Saving Behaviour”,American Economic Review 65 (May) pp.203-204Keynes, J.M. 1936. The General Theory of Employment, Interest and Money,London: MacMillan and Company LtdKhan, M and D Villanueva. 1991. “Macroeconomic Policies and Long-term Growth.” AERC Special Paper,13, May, p.39.Kimbal, M. 1990. “Precautionary Saving in the Small and the Large”, Econometrica, 58, 53-73. 44
  • 45. Journal of Economics and Sustainable Development www.iiste.orgISSN 2222-1700 (Paper) ISSN 2222-2855 (Online)Vol.2, No.6, 2011Levine, Ross, and David Renelt. 1992. A Sensitivity Analysis of Cross-Country Growth Regressions.”American Economic Review 82(4): 942-963Lewis, Arthur. 1955. “Theory of Economic Growth”, London: Allen and Unwin.Loayza, Norman, Klaus Schmidt-Hebbel, and Luis Serven. 2000. “What Drives Private saving Across theWorld?” Review of Economics and Statistics 82(2):165-181.Maddison, Angus. 1992. “A Long-Run Perspective on saving.” Scandinavian Journal of Economics 94(2):181-196.Mankiw, N. Gregory, David Romer, and David N. Weil. 1992. “A Contribution to the Empirics of EconomicGrowth.” Quarterly Journal of Economics, 107(2): 407-437.McKinnon, R. 1973. “Money and Capital in economic development.” Washington, D.C: The BrookingInstitute.Modigliani, Franco. 1970. “The Life-Cycle Hypothesis of Saving and Inter-country Differences in the savingRatio.” In W.A Eltis, M.F.G. Scott, and J.N. Wolfe, eds., Induction, Trade, and Growth: Essays in Honour ofSir Roy Harrod. Oxford: Clarendon Press.Modigliani, F. 1992. “Savings in Developing Countries: Income, Growth andOther Factors” Pacific Basin Capital Markets Research, Vol. 3, pp.23-35Mwega, Francis. 1997. “Saving in Sub-Saharan Africa: A Comparative Analysis.” Journal of AfricanEconomies. 6(3) (Supplement) 199-228.Nwankwo, Green O. (1994). The Nigerian Financial system. London: Macmillan Publishers Ltd.Schmidt-Hebbel, Klaus, Steven B. Webb, and Giancarlo Corsettii. 1992. “Household Saving in DevelopingCountries: First cross-country Evidence.” The World Bank Economic Review 6(3): 529-547.Shafik, N. 1992. “Modelling Private Investment in Egypt.” Journal of Development Economics, Vol. 39, pp.263-277.Shaw, E. 1973. “Financial deepening in economic development.” Oxford: Oxford University Press.Summers, L. H., and C. Carroll. 1989. “The Growth-Saving Nexus”, Paper presented to National Bureau ofEconomic Research Conference of Saving, Maui, Hawaii, January.Umoh,O.J. 2003. “An Empirical Investigation of the Determinants of Aggregate National Savings inNigeria”, Journal of Monetary and Economic Integration, Vol. 3, No. 2 (December 2003) pp.113-132Uremadu, S.O. 2006a. “The Impact of Real Interest Rate on Savings Mobilisation in Nigeria: An ErrorCorrection Approach”, An Unpublished Project Proposal to CBN’s Application for Diaspora CollaborativeResearch Program (DCRP)/Visiting Research Scholars Program (VRSP), 2006----- 2006b. “The Impact of Real Interest Rate on Savings Mobilisation in Nigeria”, An unpublished Ph.D.Thesis Proposal submitted to the Department of Banking and Finance, UNN, Enugu Campus----- 2002. Introduction to Finance. Benin: Mindex Publishing Company Ltd. pp, 1-7 45
  • 46. Journal of Economics and Sustainable Development www.iiste.orgISSN 2222-1700 (Paper) ISSN 2222-2855 (Online)Vol.2, No.6, 2011 Monetary Policy and Output-inflation Volatility Interaction in Nigeria: Evidence from Bivariate GARCH-M Model Onyukwu E. Onyukwu (corresponding author) Department of Economics, University of Nigeria, Nsukka, Enugu State, Nigeria Tel: +234-8034741642 E-mail: oonyukwu@yahoo.com Emmanuel O. Nwosu Department of Economics, University of Nigeria, Nsukka Enugu State, Nigeria E-mail: emmanuel.nwosu@unn.edu.ng Diejomaoh Ito African Institute for Applied Economics 54 Nza Street, Independence Layout, Enugu, Nigeria E-mail: idiejomaoh@aiaenigeria.orgReceived: October 11st, 2011Accepted: October 19th, 2011Published: October 30th, 2011AbstractThis article reports on a recent study that applies bivariate GARCH methodology to investigate theexistence of a tradeoff between output growth and inflation variability in Nigeria and to ascertain theimpact of monetary policy regime changes (from direct control regime to indirect or market based regime)on the nature of the volatility tradeoffs. Investigations reveal the existence of a short run tradeoffrelationship between output growth and inflation within and across both regimes. However, no strongevidence of long run volatility relationship could be established. Our results further reveal that regimechanges affected the magnitude of policy effects on output and inflation. Monetary policy had a strongereffect on output growth than on price stability during the period of direct control while it has a much largerimpact on inflation during the current period of market-based regime. Also volatility of output and inflationbecame more persistent during the period of indirect control.Keywords: Monetary Policy, Output-Inflation Volatility, Bivariate GARCH-M Model 46
  • 47. Journal of Economics and Sustainable Development www.iiste.orgISSN 2222-1700 (Paper) ISSN 2222-2855 (Online)Vol.2, No.6, 20111. Introduction and Background to the StudyOver the past four decades, economists and policy makers have shown considerable interest inunderstanding causes of macroeconomic volatility and how to reduce it. Conceptually, macroeconomicinstability refers to conditions in which the domestic macroeconomic environment is less predictable.It is of concern because unpredictability hampers resource allocation decisions, investment, andgrowth. This article focuses on the volatility interaction of the growth rate of output and the levels ofinflation rates. Changes in the behavior of these endogenous variables usually reflect changes in themacroeconomic policy environment as well as external shocks.Both Okonjo-Iweala & Phillip (2006) and Baltini (2004) have described the Nigerian macroeconomicenvironment as one of the most volatile among emerging markets. Among emerging marketeconomies, Nigeria exhibits the highest inflation and exchange rate variability, the lowest outputvolatility, and an interest rate volatility that is slightly smaller than that of South Africa and muchsmaller than that of Brazil, but slightly larger than that of Chile (Baltini 2004).Nigeria offers unique opportunity to the study of output-inflation volatility interaction and itsrelationship with monetary policy. This is due to the fact that monetary policy conduct in Nigeria haswitnessed two alternative regimes since the mid-1970s, the direct control regime and indirect ormarket-based regime with different relative weights attached to output growth and price stabilityobjectives.1.1 Statement of the ProblemThis study evaluates the effect of monetary policy regime changes, by estimating a bivariateGARCH-M model of output and, thus examines the nature of output-inflation variability trade-off aswell as volatility persistence which are of major interest in macroeconomic policy debates. Our studyfurther ascertains the efficacy of monetary policy regime change from direct to indirect approaches(Note 1) in reducing macroeconomic volatility, particularly in the light of the Taylor curve tradeoffhypothesis. Thus, our basic research objectives are:• To ascertain if there is evidence of output-inflation volatility trade-off in Nigeria;• To investigate if a change in monetary policy regime affects the nature of output growth-inflation volatility tradeoff.• To ascertain how monetary policy shocks affect inflation and output growth variability dynamics.What is remaining of this article has been organized into three sections. Section two is devoted to anoverview of relevant theoretical framework and methodology of analysis. Section three presents anddiscusses the results of our analyses. Section four summarizes the study findings and makes somepolicy recommendations. 47
  • 48. Journal of Economics and Sustainable Development www.iiste.orgISSN 2222-1700 (Paper) ISSN 2222-2855 (Online)Vol.2, No.6, 2011The period covered by the study is 1981 to 2007, essentially due to dataavailability. The study period cuts across two monetary policy regimes,thus affording us the opportunity to make sub-sample comparisons. Thestudy uses quarterly data in the analysis in order to capture substantialvariability in the variables of interest.2. Methodology of StudyOn the relationships between monetary policy, output volatility, and inflation volatility, one theoryholds that the volatility of output and inflation will be smaller the stronger monetary policy reacts toinflation than output gap (Gaspar & Smets 2002) (Note 2). In this case it is customary to assume thatthe economy’s social loss function is the sum of the variance of inflation and output. If the economy ishit by demand shocks, the central bank will never face a trade-off between output stability andinflation stability, that is, a monetary policy action that reduces output volatility is consistent withstable inflation.However, in a more general case when the economy is hit by demand shocks as well as supply shocks,the central bank faces an inescapable tradeoff between output stability and inflation stability if itchooses, (as is commonly the case), to minimize the social loss function. Thus, monetary policy actionwhich reduces the variance of inflation will increase the variance of output, and vice versa. Thesehypotheses have been jointly studied using bivariate GARCH-M class of models (Fountas et al 2002;Grier et al 2004; Lee 2002). According to Lee (2004), the GARCH approach has two majoradvantages over the conventional measure of volatility, such as moving standard deviations andsquared residual terms in vector autoregression (VAR) models. The first advantage is that conditionalvolatility, as compared to unconditional volatility, better represents perceived uncertainty which is ofparticular interest to policy makers. The second advantage is that the GARCH model offers insightsinto the hypothesized volatility relationship in both the short run and the long run. Whereastime-varying conditional variances reveal volatility dynamics in the short run, the model alsogenerates a long run measure of the output-inflation covariance that will be helpful in evaluatingmonetary policy tradeoffs. These inform the use of bivariate GARCH model in this study.2.1 Our Empirical ModelsFollowing Fountas et al. (2002), Grier et al. (2004) and Lee (2002) we use a bivariate GARCH modelto simultaneously estimate the conditional variances and covariance of inflation and output growth inorder to address our first and second research questions. We employ the following bivariate VAR (p)model for estimating the conditional means of output and inflation: 48
  • 49. Journal of Economics and Sustainable Development www.iiste.orgISSN 2222-1700 (Paper) ISSN 2222-2855 (Online)Vol.2, No.6, 2011 p pΧ t = Φ 0 + ∑ Φ1i Χ t −i + Φ 2j ∑ Yt − j +ε t .......... .........1 .1 i =1 j=1where ε t /Ω t ~ N(0, H t ) and Ω is information set available up to time t - 1such that H t = (h yt , h πt ) ′ [ ]Φ 0 = Φ yo , Φ π0 ′ is a 2x1 vector of constantsΧ t = [y t , π t ]′ is a 2 x 1 vector of real output growth y t and inflation rate π t .Υ t is a vector of additional explanatory variables such as inflation uncertainty, etcΦ 1i , Φ 2j are vectors of 2x2 matrices of parameters to be estimated. [ ]ε t = ε yt , ε πt ′ is a 2x1 vector of output and inflation innovations.The vector of conditional variances of output and inflation are specified as follows: H t = C0C0 + A′H t − 1 A + B ′ε t − 1ε t − 1B + D′Ft − 1D.............1.2 ′where H t = (h yt , h πt )′ is a 2x1 vectorof the conditiona variancesof output and inflation, lC, A, B, and D are 2x2 upper tria ngular matricesof parameters and F is a vectorof explanator variable ; y s,that is, F = (∆MPR, ∆Oilprices,etc).More explicitlythe matriceof parameterscan be specifiedin upper tria ngular form as : c c  a a  b b  δ δ C 0 =  11 12 ; A =  11 12 ; B =  11 12 ; and D =  11 12 . 0 c 22     21 22  a a  b 21 b 22    δ 21 δ 22   2.2 Economic Meaning of Coefficients and Apriori ExpectationsThe diagonal elements in matrix Co represent the means of conditional variances of output growth andinflation, while the off diagonal element represents their covariance. The Parameters in matrix Adepict the extents to which the current levels of conditional variances are correlated with their pastlevels. In specific terms, the diagonal elements (a11 and a22) reflect the levels of persistence in theconditional variances; a12 captures the extent to which the conditional variance of output is correlatedwith the lagged conditional variance of inflation. For the existence of output-inflation volatilitytrade-off, the variable is expected to have negative sign and be statistically significant. The parametersin matrix B reveal the extents to which the conditional variances of inflation and output are correlatedwith past squared innovations; b12 depicts how the conditional variance of output is correlated with thepast innovation of inflation. This measures the existence of cross-effect from an output shock toinflation volatility. In order to estimate the impact of monetary policy on the conditional variances, weinclude one period lagged change in the Central Bank’s monetary policy rate (MPR) (or VAR-basedgenerated monetary surprises) in the vector F. The resulting coefficients in matrix D measure theeffects of these variables on inflation and output volatility. For monetary policy to have a trade-off onthe conditional variances, the diagonal elements have to alternate in signs.In order to address the third research question we augment the VAR model specified in (1) by adding 49
  • 50. Journal of Economics and Sustainable Development www.iiste.orgISSN 2222-1700 (Paper) ISSN 2222-2855 (Online)Vol.2, No.6, 2011monetary policy variable in the vector X and then compute the generalized impulse responses andgeneralized variance decompositions to analyse the short run dynamic response of output growth andinflation monetary policy shocks/innovations. The generalized variance decomposition and impulseresponse functions are unique solution and invariant to the ordering of the variables in the VAR(Pesaran & Shin 1998). Also, it has been argued, however, that in the short run unrestricted VARsperform better than a cointegrating VAR. For example, Naka & Tufte (1997) studied the performanceof VECMs and unrestricted VARs for impulse response analysis over the short-run and found that theperformance of the two methods is nearly identical. We adopt unrestricted VARs in attempting toanswer our third research question because of the short-term nature of the variance decomposition andimpulse response analysis. AIC and SBC will be used for lag order selection.2.3 Method of Estimation and Data SourcesThe models are estimated by the method of Broyden, Fletcher, Goldferb and Shanno (BFGS) simplexalgorithm. We employed the RATS software in the estimation of the system. This is due to the fact thatRATS has an inbuilt bivariate GARCH system that supports simultaneous estimation and thus is ableto implement different restrictions that might be assumed on the variance-covariance structure of thesystem.The data used for the estimations are quarterly data from the Central Bank of Nigeria StatisticalBulletin of various years, the Annual Report and Statement of Account of various years.Monetary Policy Measure: According to Nnanna (2001) the Minimum Rediscount Rate (MRR) is thenominal anchor, which influences the level and direction of other interest rates in the domestic moneymarket. Its movements are generally intended to signal to market operators the monetary policy stanceof the CBN. Similarly, Agu (2007) observes that “the major policy instrument for monetary policy inNigeria is the minimum rediscount rate (MRR) of the Central Bank and notes that while both interestand inflation rates are high, a worrisome problem in the observed response to these macroeconomicimbalances is the lack of policy consistency and coherence. This could be on account of inadequateinformation on the nature and size of impact of the MRR on key macroeconomic aggregates.” Thistype of inconsistency in the conduct of monetary policy is likely to increase rather than stabilizemacroeconomic volatility. Hence, we shall use MRR (MPR) as a measure of monetary policy stance.3. Presentation of Results and DiscussionTable 1 shows the summary statistics of the variables used in the study. Inflation and GDP growthrates are calculated as annualized quarterly growth rates of consumer price index (CPI) and real GDPrespectively. As the table indicates, average nominal GDP was almost two times greater in 1995-2007period compared to 1981-1994 period. But, the consumer price index for the period, 1995-2007, wasalmost twenty fold of that in the period, 1981-1994. For this reason average real GDP for the period1995 to 2007 was lower than that of 1981-1994.However, the average growth rate of real GDP was higher in the second period compared to the first;while average inflation rate for the second period was lower compared to the first period. This shows 50
  • 51. Journal of Economics and Sustainable Development www.iiste.orgISSN 2222-1700 (Paper) ISSN 2222-2855 (Online)Vol.2, No.6, 2011that the economy performed better on the average under market-based monetary regime compared tothe controlled regime. The standard deviation of inflation is lower in the second period implying lowerunconditional volatility and there is no significant change in the unconditional volatility of real GDPgrowth which appears to suggest that real quarterly GDP fluctuations are modest over the two periodsbut slightly higher in the second period. The minimum rediscount rate, a measure of monetary policystance of the Central Bank of Nigeria (CBN) was on the average lower during the controlled regimeperiod compared to the indirect regime period. This suggests that monetary policy became tighterduring the period of indirect or market-based regime aimed specifically to control inflation byreducing the rate of money growth.Oil prices being used to control for exogenous shocks in the model had a low average during theperiod of controlled regime. At that time lower oil prices affected GDP growth adversely as theeconomy entered into a recession that led to the introduction the Structural Adjustment Programme(SAP), which further depressed the economy. The adverse supply effect resulting from SAP and loweroil revenue together with tighter controls on interest rate helped to put a strain on the economy. Duringthe period of indirect monetary approach, oil prices began to increase rapidly in the internationalmarket and this resulted in positive output growth for most of the period and quick recovery of theeconomy from the adverse effects of SAP. However, the Central Bank has been very cautious withrising oil prices and as a result has been setting the minimum rediscount rate in order to accommodatethe adverse effects of the rise in oil prices on inflation.Table 2 shows the tests for serial correlation, autoregressive conditional heteroskedasticity effect, andnormality. A series of Ljung-Box (1978) tests for serial correlation suggests that there is a significantamount of serial dependence in the data. Output growth is negatively skewed, and inflation ispositively skewed and output growth failed to satisfy the Jarque-Bera tests for normality (Jarque &Bera 1980). The ARCH tests also reveal the presence of first order serial dependence in theconditional variances of output growth and inflation suggesting that our application of GARCH (1, 1)model is appropriate to the data.Valid inference from GARCH model requires that the variables be stationary, at least in theirconditional means (Lee 2002). As a result, unit root tests were conducted on the variables using theDickey & Fuller (1981) methodology. The analysis, however, was eventually based on the augmentedDickey-Fuller unit root tests. The results are presented in table 3 below.The results show that only inflation rate is level stationary while other variables are stationary intheir first differences.Table 4 shows the results from GARCH estimations of the two sub-samples and the overall sample.The results would help to address our concerns in the first and second research questions, which are:whether or not the change in the approach to monetary policy in Nigeria from direct to indirect led to achange in volatility interactions, transmissions and tradeoff between output growth and inflation. Theresults are in two parts. The first part shows the estimations of the conditonal mean equation while thesecond part shows the time-varying conditional variance equation which is of particular interest to us.The first part of the conditional mean and conditonal variance eqautions represents output growthwhile the second vector denotes inflation rate. The conditional mean for GDP growth shows that 51
  • 52. Journal of Economics and Sustainable Development www.iiste.orgISSN 2222-1700 (Paper) ISSN 2222-2855 (Online)Vol.2, No.6, 2011inflation volatility does affect output growth negatively and this is statistically significant across thetwo regimes but was highly significant during the period of indirect regime. But the effect of inflationvolatility on inflation was not certain as the cofficient was neither stable nor significant across the twoperiods. Past levels of inflation are found to have positive effects on currrent inflation levels and this issignificant across the two samples while the coefficient was almost the same across the two periods.Oil price shocks do not have any meaningful effect on inflation but do have positive and significanteffects on output growth especially in the second period. This may be due to the fact that oil priceincreases which characterized most of the sample period from 1995 to 2007 may have contributedpositively to the output growth in Nigeria as a largely oil dependent economy. This result should notbe surprising as most studies have found positive output effect of oil price shocks in majoroil-exporting countries.We are especially interested in the conditional variance equations. The estimates show that the longrun or unconditional volatilities of inflation and output growth were higher in the first period than inthe second period. The estimates for the mean covariance terms are negative and but significant onlyin the first sample period. The estimates also indicate that over the sample periods, the meanunconditional variance of output is significant across the two sample periods.The parameters in matrix A show the extents to which the current levels of conditional variance arecorrelated with their past levels. The higher estimates in the second period seem to suggest that acurrent shock will have relatively long lasting effects on the future levels of the conditional variancesof output growth and inflation than it had in the first sample period. The estimates also reveal thatfollowing change in monetary policy regime, inflation volatility and output volatility have relativelybecome more persistent in Nigeria.The off-diagonal elements in A that is A (1, 2), on the other hand, reveal the extent to which theconditional variance of one variable is correlated with the lagged conditional variance of anothervariable. The estimate for A (1, 2) appears with the expected negative sign and is statistically differentfrom zero for all sub-samples and the overall sample. This confirms the existence of Taylor-curvevolatility tradeoff in Nigeria. Interestingly the estimate for the sample period 1981 to 1994 is relativelylarger than the estimates for the sample period 1995 to 2007 suggesting that low inflation variability isnow associated with higher output gap variability. This seems to suggest that CBN’s efforts to stabilizeprices or specifically to target low inflation must come at a heavy cost of output fluctuations. Thisresult is therefore, consistent with the finding by Castelnuovo (2006) that the tighter the monetarypolicy, the higher is the inflation-output gap volatility.The parameters in B matrix reveal the extents to which the conditional variances of inflation andoutput are correlated with past squared innovations (deviations from their conditional means). Ofparticular interest is the off-diagonal elements B (1, 2) and B (2, 1) which depict how the conditionalvariance of inflation is correlated with the past squared innovations of output. In the first sampleperiod there is a positive and significant volatility cross-effect from inflation to output. While in thesecond period there is positive and significant volatility cross effect from output growth to inflationvariability.In order to address the third research question we computed the impulse responses and variance 52
  • 53. Journal of Economics and Sustainable Development www.iiste.orgISSN 2222-1700 (Paper) ISSN 2222-2855 (Online)Vol.2, No.6, 2011decompositions from the VAR specification augmented by including monetary policy variable as oneof the endogenous variables in equation (1.1) and then using oil prices as exogenous variable. Wecomputed separate impulse responses and forecast error variance decompositions for each variable forthe two regime periods in order to understand how output growth and inflation respond to innovationsin monetary policy over the two regimes (see Figure 1 and Table 5 in the appendix). The impulseresponse functions are interpreted in conjunction with the variance decompositions. For example, inthe period of direct control regime, inflation responded negatively to innovations in monetary policybut the variance decompositions show that this response was not significant because monetary policyonly account for a small part of the forecast error variance of inflation and this seems to remainconstant in the long run. During the period of indirect regime inflation also responded negatively toinnovations to monetary policy but monetary shocks accounted for larger part of its forecast errorvariance, which is almost twice that of the direct control period.Real GDP growth rate responded positively to innovations in monetary policy during the directapproach. This may be due to the positive effects of low interest rates pursued during most of thatperiod. The variance decomposition of output growth shows that monetary policy accounted for alarger part of the forecast error variance of output growth during the period of direct control regimethan it accounts for inflation. This result is expected because the primary objective of monetary policythen was to achieve rapid output growth. However, inflation innovations had larger effect on outputgrowth in both periods of monetary regimes showing that inflation volatility is very crucial indetermining movements in output variance.During the two regimes output growth responded negatively to shocks on inflation. This finding isconsistent with the results in the GARCH estimations. During the period of the indirect regime outputgrowth responded negatively to innovations to monetary policy and this shows there is existence ofpolicy tradeoff between inflation and output growth. The pursuance of low inflation objective duringthe period of indirect regime does trigger off negative output reactions. The variance decompositionshows that this reaction is not significant since monetary policy shocks account for an insignificantpart of forecast error variance of output growth even in the long run.4. Summary, Policy Recommendations and Conclusion4.1 Summary of FindingsThe study on which report this article focuses investigated the existence of tradeoff relationshipbetween output growth and inflation in Nigeria and the impact of alternative monetary policy regimeson inflation and output growth. The study findings show evidence of short-run tradeoff relationshipbetween the variability of output growth and inflation but no evidence strong long run volatilityrelationship was found. The study also found that monetary policy accounted for a larger part of theforecast error variance of output growth during the period of direct control monetary policy than in theperiod of indirect control monetary policy. This result was expected given that the objective ofmonetary policy during the direct control regime was to achieve rapid and stable output growth. On 53
  • 54. Journal of Economics and Sustainable Development www.iiste.orgISSN 2222-1700 (Paper) ISSN 2222-2855 (Online)Vol.2, No.6, 2011the other hand, the response of inflation to monetary policy changes in the period of indirect ormarket-based regime was larger compared to its response during the period of direct control. Again,this result is not surprising because the major focus of monetary policy in Nigeria during the period ofindirect control was to achieve low inflation.The results of the study further reveal that the volatility of output growth and inflation during theperiod of market-based policy regime are more persistent compared to the period of direct controls.Evidence of volatility cross-effect from output to inflation and vice versa was present but notsignificant across the two sample periods. The results fail to establish clearly any evidence to suggestthat monetary policy tradeoff is a long run phenomenon. From these findings, the following policyrecommendations could be made.4.2 Policy RecommendationsMarket-based or indirect control approach to monetary policy conduct in Nigeria should be carefullyexamined regularly in order to ascertain its desirability and workability. This would help to determinewhen changes in monetary policy stance actually affect the variability of output and inflation and inwhat direction. Policy makers should be careful not to believe too fervently that the market works inNigeria. Policy changes could trigger off more volatility than demand or supply shocks. This reflectsthe fact that market imperfection is very typical of developing countries with underdeveloped financialmarkets. In Nigeria, it is hard to believe that inflation is caused by excessive money growth or thegrowth of credit. Instead, inflation has been driven largely by high cost of doing business, rising costof energy prices and depreciating exchange rate that has made the cost of imported raw materialsexceedingly high. This has intensified the effect of adverse supply shocks on inflation. Again, the sizeof the informal and non-monetized sector of the economy is quite substantial making it possible formonetary policy to have a big impact. In such a macroeconomic environment, tightening of monetarypolicy in response to high inflation would exacerbate an already heated environment by increasing thecost of credit to firms that depend on borrowing as the major source of finance.Our study reveals that volatility tradeoff is higher during the period of indirect monetary regime thanduring the period of direct control and that output is responding negatively to monetary shocks. Thisimplies that the Central Bank should be very cautious of the objective of targeting low inflation assuch a policy could trigger off not only low output growth but also high output growth volatility.Finally, we suggest that monetary policy instruments be supported by other fiscal and physicalmeasures such as ensuring that energy cost, the cost of imported raw materials and possibly the cost ofhousing are reduced through other improved supply-side processes. While, attention is being focus onlow inflation it is also pertinent to realize that high and stable output growth objective is equallyimportant to Nigeria as a developing economy. There should be a balance between output growthobjective and low inflation. The study reveals that monetary policy objective that targets low inflationwould be likely achieved at a heavy cost in terms of adverse output growth effect. 54
  • 55. Journal of Economics and Sustainable Development www.iiste.orgISSN 2222-1700 (Paper) ISSN 2222-2855 (Online)Vol.2, No.6, 20114.3 ConclusionThis study has shown that there is very little empirical evidence to suggest that monetary policyregime change necessarily alters existing inflation-output growth variability tradeoff. It could not findstrong evidence of long run tradeoff between output growth and inflation, which is required in order toascertain the effectiveness of monetary policy regime changes from that perspective. This is notaltogether surprising as half of the studies undertaken on this same issue in other jurisdictions havethus far found no evidence of long run policy tradeoff (see Lee 2004 for example). However, moststudies did find that volatility tradeoff changed when monetary policy regime changed, this studyseems to corroborate the same findings. It is perhaps important to observe here that the availability ofgood quality macroeconomic data at short-time intervals like monthly or quarterly series remains amajor challenge to policy-relevant research in Nigeria. However, the situation is not very muchdifferent in most other developing countries. Using extrapolated quarterly GDP data in empiricalstudies of this nature may influence the research outcomes since such data were econometricallygenerated under certain assumptions. Further research is therefore recommended in this issue in thefuture, particularly as high frequency and good quality data begin to be available. It would indeed bevery informative to policy makers in Nigeria who are currently experimenting with the adoption ofinflation targeting monetary policy regime to read this research output. It will perhaps assist them inappreciating the cost of such regime in terms of output growth volatility.ReferencesAgu, C. (2007). What Does the Central Bank of Nigeria Target: An Analysis of Monetary PolicyReaction Function in Nigeria. African Economic Research Consortium (AERC), Nairobi, Kenya.Baltini, N. (2004). Achieving and Maintaining Price Stability in Nigeria. IMF Working Paper,WP/04/97.Castelnuovo, E. (2006). “Monetary Policy Switch, the Taylor Curve and Great Moderation”, [Online]Available: http://ssrn.com/abstract=880061].Dickey, D. A., & Fuller, W. A. (1981). Likelihood Ratio Statistics for Autoregressive Time Series witha Unit Root. Econometrica, 49, 1057–1072.Fountas, S., Menelaos, K. & J. Kim (2002). Inflation and Output Growth Uncertainty and theirRelationship with Inflation and Output Growth. Economics Letters, 75, 293-301.Fountas, S. & Menelaos, K. (2007). Inflation, Output Growth, and Nominal and Real Uncertainty:Empirical Evidence for the G7. Journal of International Money and Finance, 26, 229-250.Fuhrer, J. (1997). Inflation/Output Variance Trade-offs and Optimal Monetary Policy. Journal ofMoney, Credit and Banking, 29(2), 214-34.Gaspar, V. & Smets, F. (2002). Monetary Policy, Price Stability and Output Gap Stabilization.International Finance, 5(2), 193-211.Grier, K. B., Henry, O. T., Olekalns, N. & Shields, K. (2004). The Asymmetric Effects of Uncertainty 55
  • 56. Journal of Economics and Sustainable Development www.iiste.orgISSN 2222-1700 (Paper) ISSN 2222-2855 (Online)Vol.2, No.6, 2011on Inflation and Output Growth. Journal of Applied econometrics, 19, 551–565.Jarque, C. M. & Bera, A. K. (1980). Efficient tests for normality, homoscedasticity and serialindependence of regression residuals. Economics Letters, 6, 255–259.Lee, J. (2002). The Inflation-Output Variability Tradeoff and Monetary Policy: Evidence from aGARCH Model. Southern Economic Journal, 69(1), 175-188.Lee, J. (2004). The Inflation-Output Variability Tradeoff: OECD Evidence. Journal of ContemporaryEconomic Policy, 22(3), 344-356.Ljung, G. M. & Box, G. E. P. (1978). On a Measure of Lack of Fit in Time Series Models. Boimetrika,66: 66–72.Naka, A. & Tufte, D. (1997), Examining the Impulse Response Functions in Cointegrated Systems;Applied Economics, 29, 1593-1603.Nnanna, O. (2001). Monetary Policy Framework in Africa: the Nigerian Experience. Central Bank ofNigeria, Working Paper.Okonjo-Iweala, N. & Philip Osafo-Kwaako. (2006). Nigeria’s Economic Reforms: Progress andChallenges. Brookings Global Economy and Development, Working Paper, 6, 1-30.Pesaran, M. H. and Y. Shin (1998). Generalised impulse response analysis in linear multivariatemodels. Economic Letters, 58 , 17-29.Roberts, J. M. (2006). Monetary Policy and Inflation Dynamics. International Journal of CentralBanking, September, 2006.Taylor, J. (1979). Estimation and control of Macroeconomic Model with Rational Expectations.Econometrica, 47(5), 1267-86.Taylor, J. (1994). The Inflation/Output Variability Trade-off Revisited, in Goals, Guidelines andConstraints Facing Monetary Policymakers. FRB of Boston, Conference Series 38, 21-38.NotesNote 1. Other studies failed to distinguish the two periods in their analyses and thus evaluate theeffectiveness of the CBN’s monetary policy even over the period when monetary policy in Nigeria hadoverbearing political interference.Note 2. Gaspar, Victor and Frank Smets (2002) “Monetary Policy, Price Stability and Output GapStabilization.” International Finance, 5:2, 193-211FiguresImpulse Response Functions 56
  • 57. Journal of Economics and Sustainable Development www.iiste.orgISSN 2222-1700 (Paper) ISSN 2222-2855 (Online)Vol.2, No.6, 2011 Plot of responses of inflation rate 1981-1994 0.08 0.06 0.04 0.02 0.00 -0.02 -0.04 -0.06 -0.08 0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 Real GDP Growth Rate 1981-1994 MRR 1981-1994 inflation rate 1981-1994 Figure 1. Plot of responses of inflation rate 1981-1994 Plot of responses of MRR 1981-1994 10 8 6 4 2 0 -2 -4 0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 Real GDP Growth Rate 1981-1994 MRR 1981-1994 inflation rate 1981-1994 Figure 2. Plot of responses of MRR 1981-1994 Plot of responses of Real GDP Growth Rate 1981-1994 2.5 2.0 1.5 1.0 0.5 0.0 -0.5 0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 Real GDP Growth Rate 1981-1994 MRR 1981-1994 inflation rate 1981-1994 Figure 3. Plot of responses of Real GDP Growth Rate 1981-1994 57
  • 58. Journal of Economics and Sustainable Development www.iiste.orgISSN 2222-1700 (Paper) ISSN 2222-2855 (Online)Vol.2, No.6, 2011 Plot of responses of inflation rate 1995-2007 0.050 0.025 -0.000 -0.025 -0.050 -0.075 -0.100 -0.125 0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 Real GDP Growth Rate 1995-2007 MRR 1995-2007 inflation rate 1995-2007 Figure 4. Plot of responses of inflation rate 1995-2007 Plot of responses of MRR 1995-2007 14 12 10 8 6 4 2 0 -2 -4 0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 Real GDP Growth Rate 1995-2007 MRR 1995-2007 inflation rate 1995-2007 Figure 5. Plot of responses of MRR 1995-2007 58
  • 59. Journal of Economics and Sustainable Development www.iiste.orgISSN 2222-1700 (Paper) ISSN 2222-2855 (Online)Vol.2, No.6, 20112.52.01.51.00.50.0-0.5-1.0-1.5 0 5 10 15 20 Real GDP Growth Rate 1995-2007 inflation rate 1995-2007 MRR 1995-2007 Figure 6. Plot of responses of Real GDP Growth Rate 1995-2007TablesTable 1. Summary Statistics of Study VariablesSeries Sample obs mean Std Dev Minimum MaximumGDP 1981:1 2007:1 108 409541.01 220926.82 205045 1317391.3 1981:1 1994:4 56 283126.25 44004.84 205045 347271.1 1995:1 2007:4 52 545679.97 252724.21 348746.7 1317391.3CPI 1981:1 2007:1 108 2459.17 2697.68 47.9 8722.6 1981:1 1994:4 56 291.44 320.28 47.9 1458.4 1995:1 2007:4 52 4793.65 2107.46 1669.9 8722.6INFLAcbn 1981:1 2007:1 108 23.66 19.91 1.3 77.9 1981:1 1994:4 56 27.77 19.89 3.0 66.7 1995:1 2007:4 52 19.23 19.14 1.3 77.9Oilprices 1981:1 2007:1 108 28.07 15.23 11.4 90.7 1981:1 1994:4 56 23.56 6.94 13.6 38 1995:1 2007:4 52 32.94 19.71 11.4 90.7Mrr 1981:1 2007:1 108 14.07 4.41 6 26 1981:1 1994:4 56 13.28 5.24 6 26 1995:1 2007:4 52 14.93 3.12 8 20LOGRGDP 1981:1 2007:1 108 10.7 1.43 9.0 14.1 1981:1 1994:4 56 11.94 0.81 10.1 13.1 1995:1 2007:4 52 9.35 0.20 9.0 10.0GDPGRT 1981:1 2007:1 107 0.54 1.1 -1.1 6.8 1981:1 1994:4 55 0.31 1.02 -1.1 6.8 59
  • 60. Journal of Economics and Sustainable Development www.iiste.orgISSN 2222-1700 (Paper) ISSN 2222-2855 (Online)Vol.2, No.6, 2011 1995:1 2007:4 52 0.77 1.1 -0.6 4.6Table 2. Tests for Serial Correlation, ARCH and NormalityTests for Serial Correlation, ARCH and NormalitySeries Q(8) Q(16) Q(24) ARCH(1) ARCH(2) JB-STATGDPGRWT 21.778[0.0006] 54.765[0.000] 82.40[0.000] 16.636[0.000] 5.846[0.00537] 2.228[0.3283]INFLACBN 65.281[0.0000] 76.992[0.000] 82.846[0.00] 11.787[0.006] 40.51[0.0000] 21.037[0.07]Table 3. Unit Root Tests of Variables of the ModelVariable Level First Difference LagsLogrgdp -3.139162* -3.755954** 1Inflacbn -2.450268** -3.935362** 1Oilprices 1.576064 -4.293117** 1Logoilprices -0.001255 -4.789354** 1MRR -2.890303 -5.668382** 1GDPGRT -2.385600 -7.695887** 3Table 4. Bivariate GARCH Estimations of Output Growth and InflationBIVARIATE GARCH ESTIMATIONS OF OUTPUT GROWTH AND INFLATIONCONDTIONAL MEAN EQUATIONS 1981:1 2007:4 1981:1 1994:4 1995:1 2007:4Variables Mod1 Mod11 Mod21CONSTANT -7.51810** -5.546* -3.43389Trend 0.07571** -0.011 0.021079GRGDP{1} 0.05173 0.073 -0.1475INFLA_VOL -0.09100** -0.0897* -0.1000**OILP_SHOCK 3.98553** 6.7809 6.3037**CONSTANT 7.65128** 10.3289** -0.1001Trend -0.07907** -0.10556 0.009257INFLACBN{1} 0.70064** 0.7497** 0.70817**INFLA_VOL 0.2668** 0.28209 -0.10738OILP_SHOCK 0.05470 -18.9548** 0.10766CONDITIONAL VARIANCE EQUATIONSC(1,1) 2.55319** 4.4927** 3.7733**C(1,2) -0.51521 -4.57815** -0.4757C(2,2) 1.9726** 0.7345 0.000055A(1,1) 0.9095** 0.1876 0.87385**A(1,2) -0.41810** -1.0158** -0.37536** 60
  • 61. Journal of Economics and Sustainable Development www.iiste.orgISSN 2222-1700 (Paper) ISSN 2222-2855 (Online)Vol.2, No.6, 2011A(2,2) 0.73918** 0.15856 0.74536**B(1,1) 0.63512** 0.7660** 0.24098B(1,2) -0.1518 -0.10767 0.2372**B(2,1) 0.3460** 0.5393** -0.23189B(2,2) -0.6345** 0.05555 -0.23189Convergence 72 Iters 54 Iters 55 ItersTable 5. Variance Decompositions Decomposition of Variance for Series MRR 1981-1994Step Std Error MRR INFLACBN LOGRGDP1 2.2635274 100.000 0.000 0.0002 3.0798643 99.840 0.113 0.0473 3.5912226 99.454 0.503 0.0434 3.9572332 98.817 1.146 0.0375 4.2363533 98.038 1.915 0.0476 4.4549927 97.254 2.674 0.0737 4.6272535 96.569 3.328 0.1048 4.7622814 96.030 3.838 0.1329 4.8671344 95.642 4.205 0.15310 4.9477657 95.383 4.451 0.16611 5.0092686 95.221 4.606 0.17212 5.0559098 95.127 4.698 0.17513 5.0911552 95.076 4.749 0.17414 5.1177424 95.052 4.775 0.17315 5.1377856 95.042 4.787 0.17216 5.1528939 95.040 4.790 0.17117 5.1642808 95.041 4.789 0.17118 5.1728585 95.043 4.786 0.17119 5.1793125 95.046 4.782 0.17320 5.1841599 95.048 4.777 0.17521 5.1877921 95.049 4.773 0.17722 5.1905074 95.050 4.769 0.18123 5.1925336 95.049 4.766 0.18424 5.1940455 95.048 4.763 0.189 Decomposition of Variance for Series INFLACBN 1981-1994 Step Std Error MRR INFLACBN LOGRGDP1 6.1520145 0.071 99.929 0.0002 11.1840983 1.802 95.291 2.907 61
  • 62. Journal of Economics and Sustainable Development www.iiste.orgISSN 2222-1700 (Paper) ISSN 2222-2855 (Online)Vol.2, No.6, 20113 15.0407379 3.934 91.002 5.0634 17.5910999 5.983 87.708 6.3085 19.0795657 7.843 85.209 6.9496 19.8504330 9.436 83.355 7.2097 20.2077581 10.698 82.044 7.2588 20.3628070 11.607 81.174 7.2209 20.4350821 12.201 80.627 7.17210 20.4786903 12.556 80.296 7.14911 20.5120032 12.752 80.092 7.15512 20.5388952 12.857 79.962 7.18213 20.5596652 12.911 79.872 7.21714 20.5748203 12.941 79.807 7.25215 20.5855134 12.958 79.758 7.28516 20.5930675 12.967 79.719 7.31317 20.5985954 12.972 79.689 7.33818 20.6028938 12.975 79.665 7.36019 20.6064990 12.975 79.645 7.38020 20.6097764 12.973 79.627 7.40021 20.6129850 12.970 79.611 7.42022 20.6163087 12.966 79.595 7.43923 20.6198692 12.961 79.580 7.45924 20.6237341 12.957 79.564 7.480 Decomposition of Variance for Series LOGRGDP 1981-1994 Step Std Error MRR INFLACBN LOGRGDP1 0.0606513 2.224 35.541 62.2352 0.1051127 4.838 41.180 53.9823 0.1455233 6.553 43.994 49.4534 0.1817223 7.954 45.165 46.8825 0.2136589 9.201 45.422 45.3766 0.2416622 10.339 45.191 44.4707 0.2662894 11.377 44.717 43.9068 0.2881627 12.320 44.145 43.5359 0.3078669 13.172 43.556 43.27210 0.3258997 13.937 42.997 43.06611 0.3426582 14.624 42.488 42.88812 0.3584433 15.240 42.039 42.72213 0.3734729 15.793 41.647 42.55914 0.3878983 16.293 41.310 42.398 62
  • 63. Journal of Economics and Sustainable Development www.iiste.orgISSN 2222-1700 (Paper) ISSN 2222-2855 (Online)Vol.2, No.6, 201115 0.4018202 16.745 41.019 42.23616 0.4153033 17.157 40.767 42.07617 0.4283879 17.534 40.548 41.91818 0.4410995 17.880 40.355 41.76519 0.4534554 18.201 40.183 41.61620 0.4654687 18.498 40.028 41.47421 0.4771516 18.774 39.887 41.33822 0.4885159 19.032 39.758 41.21023 0.4995741 19.273 39.639 41.08824 0.5103391 19.499 39.529 40.973 Decomposition of Variance for Series MRR 1995-2007 Step Std Error MRR INFLACBN LOGRGDP1 1.1350644 100.000 0.000 0.0002 1.6579561 98.990 0.634 0.3763 1.9518658 98.041 1.594 0.3654 2.1059278 97.576 2.035 0.3895 2.1890956 97.572 1.940 0.4886 2.2471307 97.176 2.068 0.7577 2.3094247 95.447 3.279 1.2758 2.3891806 92.145 5.816 2.0399 2.4856879 87.850 9.202 2.94910 2.5901409 83.418 12.706 3.87611 2.6920664 79.478 15.792 4.73012 2.7834456 76.298 18.235 5.46713 2.8600548 73.886 20.033 6.08114 2.9210218 72.125 21.291 6.58415 2.9676945 70.865 22.140 6.99516 3.0025083 69.967 22.701 7.33217 3.0281459 69.320 23.069 7.61118 3.0470430 68.841 23.314 7.84519 3.0611754 68.473 23.482 8.04520 3.0720291 68.178 23.603 8.21921 3.0806628 67.931 23.695 8.37422 3.0878022 67.716 23.771 8.51423 3.0939296 67.522 23.836 8.64224 3.0993574 67.344 23.894 8.762 Decomposition of Variance for Series INFLACBN 1995-2007 Step Std Error MRR INFLACBN LOGRGDP 63
  • 64. Journal of Economics and Sustainable Development www.iiste.orgISSN 2222-1700 (Paper) ISSN 2222-2855 (Online)Vol.2, No.6, 20111 2.6954031 5.347 94.653 0.0002 5.4978721 7.572 89.068 3.3603 8.1216124 11.195 84.123 4.6824 10.2911556 14.430 80.076 5.4945 11.9097715 17.231 76.848 5.9216 13.0111184 19.547 74.326 6.1277 13.6946354 21.374 72.430 6.1968 14.0795505 22.727 71.083 6.1919 14.2745317 23.651 70.196 6.15310 14.3626479 24.223 69.665 6.11311 14.3984216 24.535 69.380 6.08512 14.4124976 24.679 69.246 6.07513 14.4192880 24.730 69.190 6.08014 14.4241316 24.737 69.168 6.09515 14.4283203 24.729 69.157 6.11416 14.4318686 24.718 69.149 6.13317 14.4346655 24.708 69.140 6.15218 14.4367583 24.701 69.131 6.16719 14.4383088 24.696 69.123 6.18120 14.4394976 24.692 69.116 6.19321 14.4404705 24.689 69.109 6.20322 14.4413283 24.686 69.102 6.21223 14.4421380 24.683 69.097 6.22124 14.4429453 24.680 69.091 6.229 64
  • 65. Journal of Economics and Sustainable Development www.iiste.orgISSN 2222-1700 (Paper) ISSN 2222-2855 (Online)Vol.2, No.6, 2011 New approaches to sustainable forest management: a study of service innovation in conserving forestry resources Verma Rajeev (corresponding author) Fellow Participant in Management Indian Institute of Management, Indore, India Email: f09rajeev@iimidr.ac.inReceived: October 11st, 2011Accepted: October 14th, 2011Published: October 30th, 2011The author would like to thank Prof. P C Kotwal, Indian Institute of Forest Management, ProjectCoordinator, IIFM –ITTO Project and Forest MIS of Chhattisgarh Forest Department, to operationalize thisstudy.AbstractSustainability has been a primary concern for the forestry professionals. This paper is concerned with thecontinuing evolution of approaches to monitor sustainable forest management. It summarises the existingknowledge base and primary techniques and strategies for achieving socially and environmentallyacceptable SFM in various forest formations. Service innovation is one means for the improvedmonitoring of SFM through the introduction of scientifically based criteria and indicators. This set has beendeveloped on the basis of Dry Forest Asia Initiative. In addition, their implementation has played a key rolein interactions with local beneficiaries. Yet, research on the link between service innovation and naturalresource management is scant. The paper identifies innovation orientation, external partner collaboration,and information capability as operant resources along with the operand resources captured under the 8criteria and their respective field level indicators. These antecedents are analyzed to know impact of overallservice innovation on social economic development of the area.Keywords: sustainable forest management, service innovation, C&I, community participation1. IntroductionThe United Nations Conference on Environment and Development (UNCED) in 1992 led the generaladoption of a concept of sustainable development based on the equilibrium between three primecomponents i.e., economic development; conservation of the environment and social justice. Forestryresources have been featured prominently at the conference and have remained high on the internationalagenda for sustainable development. In pursuance of the ‘Forest Principles’ and of Agenda 21 (chapter 11)adopted at UNCED, the notion of sustainable forest management has been adopted in more specific andoperational terms. Criteria and indicators were identified in order of planning, monitoring and assess theforest management practices at the national as well as for the individual forest management unit (FMU)level. The selection and use of suitable criteria and indicators are thus one of the keys to progress in thepractice of sustainable forest management. 65
  • 66. Journal of Economics and Sustainable Development www.iiste.orgISSN 2222-1700 (Paper) ISSN 2222-2855 (Online)Vol.2, No.6, 2011These criteria and indicators are mainly intended for defining objectives and priorities for Sustainable ForestManagement (SFM) and for monitoring progress during their implementation. The objective of SFM is, toincrease the adoption of forest management practices to sustain and enhance the yields of multiple products,services and values for multiple stakeholders, over the long term. However, the intermediate goal is tocontribute, sustain and enhance the benefits from natural tropical forests like India, by increasingopportunities and benefits to various stakeholders. The end benefit of the project includes rural livelihoodpromotion; industrial timber production; and environmental services. In a way, improving the level ofadoption of scientific findings in forest management, leads to adoptive management process and hence use ofbest practices. In India, all the forestry resources are under the direct control of the government. Hence, itprovides a very less scope for the innovative management. However, the latest management practices suchas, Sustainable Forest Management (SFM), Joint Forest Management (JFM) bring the inputs from variousstake holders and a hope to bring the service innovation in the management of the forestry resources.These service innovations in the management of natural resources have been an effective way for theforestry organizations to accelerate its growth rate and profitability as products or services become more orless homogeneous or an original competitive advantage cannot be sustained (Berry, Shankar, Janet, Susanand Dotzel, 2006). Accordingly, researchers and practitioners are interested in explaining and predictingkey antecedents of and outcomes associated with service innovation in managing forestry resources. Muchof the research on service innovation in the last few decades has addressed many considerations,including decisions of service innovation adoption (Frambach, Barkema, Bart and Wedel, 1998; Kleijnen,Ruyter, and Andreassen 2005), typologies (Avlonitis, Papastathopoulou, and Gounaris 2001) of serviceinnovation, service innovation strategy and process (Blazevic and Lievens 2004), and drivers of serviceinnovation (Berry et al. 2006) mainly in the corporate scenario of the western countries. However, thepresent work focuses on the importance of innovation practices in conserving forestry resources andhighlights the need of future research in this area. By shedding light on the short and long-term benefits anddifferent goods/ services for different stakeholders, this forest management practice contributes both to theenvironment as well as to its various stakeholders.As suggested by Berry et al. (2006), service innovation aims to create new markets and hence possibilitiesof extending the organizational service reach. Research seeks to contribute to sustainable forestmanagement by increasing the understanding of the costs and benefits for different stakeholders ofsustaining or replacing, managing well or degrading natural forests, by enhancing the incentives forimproved forest management through contributions for institutional development and policy decisions; andby evaluating harvesting and management recommendations to sustain commodities and environmentalvalues from natural tropical forests.However, this topic has recently attracted increasing interest from academics and practitioners (e.g., Blind2006; Dean 2004; Pavlovski 2007; Verganti and Buganza 2005; Zomerdijk and de Vries 2007), there is littleevidence of significant innovation in managing forestry resources. We argue that there is no full andadequate understanding of the concept of service innovation and its role in managing forestry resources.Recently, the field of marketing has evolved toward a service-dominant (S-D) logic (Vargo and Lusch 2008)through which we can re-examine the role of innovation in service delivery. Compared to traditionalgoods-dominant (G-D) logic, service in S-D logic is the application of specialized competences (knowledgeand skills, i.e., operant resources) to provide through goods (operand resources) that benefit an entity.In the present article, we argue that service innovation is the process of applying specialized competences,consistent with S-D logic. Innovation in this context is the process of applying new ideas or currentthinking in fundamentally different ways, resulting in significant changes. According to S-D logic, 66
  • 67. Journal of Economics and Sustainable Development www.iiste.orgISSN 2222-1700 (Paper) ISSN 2222-2855 (Online)Vol.2, No.6, 2011innovations had significantly changed the user preferences and their perception about the service quality(Madhavaram and Hunt, 2008). Therefore, by implementing innovative practices in service delivery,organizations could change their method of creating stakeholders value and hence positively impact theirperceptions. Accordingly, we argue that service delivery innovation in the SFM context, involves an entireorganization viewing and addressing both value creation and environmental services within an S-D logicframework.The purpose of this article is to contribute to the literature on Sustainable Forest Management throughservice delivery innovation by developing and empirically testing a model that attempts to explain whatmotivates service delivery innovation and, in turn, influences performance in terms of optimum utilizationof forestry resources. Here the performance has been measured on the basis of Bhopal-India initiative forSFM, which in itself is an outcome of Dry Forest Asia Initiative. We had three research objectives: (a) tounderstand the role of service innovation in SFM (b) to investigate the antecedents of service deliveryinnovation based on the Local Unit Criteria and Indicator Development (LUCID) for SFM, and (c) toexamine whether proposed service innovation can result in better social economic development in terms ofrural livelihood promotion.This article makes three contributions to the literature. First, we try to identify the nature of service deliveryinnovation in the management of natural resources. By studying innovation within the framework of S-Dlogic, we view service delivery innovation as ability of a firm to create stakeholder value (Lievens andMoenaert, 2000). Second, based on resource advantage (R-A) theory (Lusch, Vargo and Brien, 2007), wetest the links among organizational innovation orientation, service delivery innovation, and performancethrough an empirical survey with samples from 126 JFMC’s. Third, the results provide practical steps formanagers to understand service innovation that can result in better social economic development in terms ofrural livelihood promotion. The article is structured as follows. First, we review Sustainable ForestManagement in terms of S-D logic framework to identify the key operant resources that facilitate servicedelivery innovation. Then, after describing the research framework (Figure 1), we report the results of astudy conducted in the Forest Management Committees aimed at empirically testing the research model(Figure 2). At last, we conclude with a discussion of theoretical and managerial implications and directionsfor future research.2. Theoretical Background and Conceptual FrameworkLooking into the service innovation literature, most of the prior innovation literature has treated serviceinnovation as product innovation. However, extensive literature review includes the interaction (i.e.,co-production with end users) between new service development and service delivery (Zaltman, Duncanand Holbek, 1973). Although strategic innovation theory (Markides, 1997) addresses a new way ofdelivering new products or services to existing or new customer segments and most adequately explainsservice innovation, it focuses mainly on goods (i.e., operand resources) but not on operant resources.As we know, R-A theory is compatible with the S-D logic’s emphasis on competences, value propositions,and operant resources. In this study, we view service delivery innovation from the R-A theory to betterunderstand the relationships among strategic and organizational issues. R-A theory is a process theory ofcompetition, which asserts that firms achieve superior financial performance by occupying marketplacepositions of competitive advantage. Here in terms of forest management, the competency may be achievedin terms of optimum utilization of forestry resources, socio-cultural benefits and better ecosystem functionand vitality. 67
  • 68. Journal of Economics and Sustainable Development www.iiste.orgISSN 2222-1700 (Paper) ISSN 2222-2855 (Online)Vol.2, No.6, 2011Hence it relies on those resources that provide the firm a comparative advantage over its competitors. Thesecomparative resources, in the S-D logic perspective, are mainly operant resources. Figure 1 presents ourresearch framework.Figure 1: The Research FrameworkOperant resources that can be leveraged to develop innovation practices, a source of sustained competitiveadvantage, produce superior performance in terms of improved sustainability indicators. Further, we arealso suggesting a direct effect between innovation practices on sustainability of forestry resources. In ourdiscussion, we refer to the concept of applying R-A theory to support service delivery innovation research.To determine which operant resources facilitate service delivery innovation, however, one needs a modeldescribing the resources/capabilities of a firm and how these enable service delivery innovation. The modelhas been based on the study Verganti and Buganza (2005) that describes service delivery as being facilitatedby organization (internal and external) and technology. We therefore propose a research model and suggestthat innovation practices in service delivery are mainly influenced by organizational, relational, andinformational resources (Hunt 2000). In R-A theory, organizational (e.g., cultures), relational (e.g.,relationships with partners), and informational (e.g., technology) assets are operant resources.We further identify organizational resources as innovation orientation, relational resources as externalpartner collaboration, and informational resources as IT capabilities. The emphasis in the literature is thediscussion of organization, relationships, and technology and their influence on service delivery innovationand monitoring of sustainable management of forestry resources. It shows the relationships that wehypothesized exist among innovation orientation, external partner collaboration, IT capability, servicedelivery innovation, and performance of sustainability indicators. 68
  • 69. Journal of Economics and Sustainable Development www.iiste.orgISSN 2222-1700 (Paper) ISSN 2222-2855 (Online)Vol.2, No.6, 2011Figure 2: The Research Model3. Operationalization of ConstructsInnovation orientation of a Stakeholder: refers to an organizational openness to new ideas and propensityto change through adopting new technologies, resources, skills, and administrative systems (Zhou, Gao,Yang, 2005). Innovation orientation consists of both openness to innovation (Zaltman, Duncan, and Holbek1973) and capacity to innovate (Bolton, 2003). Here it has been measured on the basis of individual effortsfor conserving the forestry resources.External partner collaboration: It’s an important parameter for conserving ecosystem biodiversity as highdegree of networking efforts took place while conserving the forestry resources. It has been measured usingmodified scales of Kalaignanam, Shankar and Varadarajan (2007). Apart from institutional efforts (forestcommittee), the organizational participation in exchanging resources and capabilities with external partnerssuch as universities, research institutions, customers, and suppliers have also been studied.IT capability: It includes the availability of IT infrastructure, human IT resources, and IT-enabledintangibles. It also includes the availability of GPS, palmtops at the beat level. It has been measured thescale by Bharadwaj (2000). It also includes the computerization of land records and computer awareness atthe grass-root level.Service Delivery Innovation: It refers to the actual delivery of a service (Zeithaml, Berry, andParasuraman 1988) and the delivery of services and products to the stakeholders (Lovelock andGummesson, 2004). In forest management terms, it is a process of applying specialized competences(knowledge and skills) to maintain and enhance the ecosystem function and vitality. It has been measuredby using 10 items adapted and modified from research on the S-D logic perspective by Vargo and Lusch(2004).Measuring the forest sustainability: Standards for sustainable forest management typically consist of anumber of principles which are the components of the overall goal and objective, and of criteria andindicators which are meant to enable an assessment as whether or not the objective and its components are 69
  • 70. Journal of Economics and Sustainable Development www.iiste.orgISSN 2222-1700 (Paper) ISSN 2222-2855 (Online)Vol.2, No.6, 2011being accomplished. In the present study, criteria have been formulated to describe a desired state ordynamics of the biological or social system and allow a verdict on the degree of achievement of anobjective in a given situation. A scale consists of 8 criteria and 44 indicators have been used to monitor theforest sustainability at the National Level.4. Hypothesis DevelopmentIn a study conducted by Zhou et al, (2005) the innovation orientation has been determined as one of thefactors for service innovation. It has been noted as a key driver for overcoming hurdles and enhancingorganizational ability to successfully adopt or implement new systems, processes, or products. In case ofconserving forestry resources, it comes in the form of individual efforts to maintain the forest resourceproductivity. Therefore, innovation orientation represents the extent to which (a) an organization is open tonew ideas (i.e., culture) through the adoption of new technologies and integrated resources and (b)encouragement of forest management committee members to consider the adoption of innovation. Hence,we propose that the organizational ability to adapt to changing or existing service depends upon innovationorientation of its individual members. We formulate the following hypothesis:Hypothesis 1: Innovation orientation of individual JFMC members has a positive impact on serviceinnovation in forest conservation scenario at 95 percent confidence level.The Inter-organizational collaboration is important in supplementing the internal innovative activities oforganizations (Deeds and Rothaermel 2003). Therefore, firms need to collaborate to build greaterinnovation practices and lock-in partners for the long term. Due to this firms may improve their ability toengage in process innovation by managing their relationships with suppliers and customers (Kaufman,Wood, and Theyel 2000). Therefore, we propose that innovative service delivery that a firm creates is basedon support from external partners’ collaboration (e.g., stakeholders, research institutions, and universities).Hence, we propose that firms having stronger collaborations with external partners will be better atdeveloping new methods (approaches) of service innovation for conservation of natural resources.Therefore, we hypothesize the following:Hypothesis 2: External partner collaboration has a positive impact on service innovation in forestconservation scenario at 95 percent confidence level.Based on the study on operant resources that bundle basic resources (Hunt 2000), we propose that ITcapability is a hierarchy of composite operant resources (COR) that includes IT infrastructure, human ITresources, and IT-enabled intangibles. Technology may influence organizational ability to create value thatwill transform the way customers interact with an offering. To create a new channel or method of servicedelivery, organization needs to possess this infrastructure. In terms of managing the sustainable resources, itcomes in the form of latest GPS supported infrastructure; Satellite based forest fire fighting equipments etc.Thus, IT capability is the operant resource for a new service that offers an opportunity to provide new andinnovative services. We hypothesize the following:Hypothesis 3: IT capability has a positive impact on service innovation in forest conservation scenario at95 percent confidence level.Prior research has studied business performance from different perspectives, such as financial performance, 70
  • 71. Journal of Economics and Sustainable Development www.iiste.orgISSN 2222-1700 (Paper) ISSN 2222-2855 (Online)Vol.2, No.6, 2011business unit performance, or organizational performance (Wiertz, Ruyter, Keen and Streukens, 1986).Based on R-A theory, once competitors achieve superior performance through obtaining marketplacepositions of competitive advantage, firms attempt to leverage the advantages through major innovationpractices. We therefore propose that if organization is able to innovate in more varied ways to deliverservice, they will achieve superior performance objectives. The criteria and indicators; are used to describea systematic approach to measuring, monitoring and reporting SFM. C&I indicate the direction of changeas regards the forests and also suggest the ways to expedite the process of SFM. Hence, we postulate thefollowing:Hypothesis 4: Service innovation in forest management has a positive impact on sustainability of forestryresources.5. Research MethodologyThe Study Site: To achieve the purposes of the study, the field data has been collected through 126 JFMCmembers under localized community biodiversity programme (L-CBP) at Marwahi (North Bilaspur)Forest Division in the state of Chhattisgarh, India. It is one of the project site where C&I approach forSFM is being implemented with the help of International Tropical Timber Organization (ITTO), Japan.The division is rich in species diversity and the dependency of local community on the forest resources isrelatively higher. The forest division is located between 810 48’ to 820 24’ E longitude and 220 8’ to 230 7’N Latitude. As per the Champion & Seth’s (1968) classification of Forests Types of India, forest of thedivision has been classified under the 5B / C1c Dry Peninsular Sal Forest. The criteria and indicatorapproach has been used to develop a site-specific set of criteria and indicators for Marwahi ForestDivision, North Bilaspur. This set has been developed on the basis of Bhopal-India initiative for SFM,which in itself is an outcome of Dry Forest Asia Initiative.Design/methodology/approach: The research plan is deductive in nature as the theory building processprecedes the data collection process. The data collection has been done using a 24 item scale (ForestSustainability Index) refined from previous studies. To measure the scale internal reliability consistencies,alpha value has been found out between 0.76 and 0.87, exceeding the 0.70 benchmark suggested byNunnally (1978). A set of hypotheses has been developed pertaining to potential predictors of two distinctfacets (Operant and Operand resources) of service innovation and the impact of the latter on the measuresof forestry indicator performance in terms of, increase in bio-diversity, soil and water conservation, forestproduction of NTFP’s and socio-economic development of the village. Path analysis using SEM techniquehas been used for the data analysis.Measure: All constructs in the study has been measured using multiple items. A five-point likert scale hasbeen used to capture the variables and indicator items. The scale has been adopted from previous studiesand checked for scale reliabilities (coefficient α). It consists of total 30 (6+24) items to operationalize 5construct level variables. The 6 item scale has been used to measure service innovation (SI) and 24 itemFMU level Criteria and Indicator scale has been used to measure the construct of Innovation Orientation,External Partner Collaboration, Information Technology Capability and Forest Sustainability Index.However, the final questionnaire for IO, EPC and IT has been adopted from various authors apart fromITTO scale. S. No Measure Construct and indicator Scale Questionnaire adopted from 71
  • 72. Journal of Economics and Sustainable Development www.iiste.orgISSN 2222-1700 (Paper) ISSN 2222-2855 (Online)Vol.2, No.6, 2011 variables Reliability (α) 1 IO Innovation Orientation 0.76 Burns and Stalker (1977) 2 EPC External Partner 0.79 Faems, Looy, and Debackere Collaboration (2005) 3 IT Information Technology 0.81 Bharadwaj (2000) Capability 4 SI Service Innovation 0.76 Vargo and Lusch (2004) 5 SFM Forest Sustainability Index 0.87 Kotwal P.C. et.al (2006)Table 1: Construct indicator variables and scale reliability valuesThe Descriptive statistics for the selected variables are, Variables IO EPC IT SI SFM Sample Size 126 126 126 126 126 Mean 3.563 4.453 2.113 3.652 4.120 Standard Deviation 1.467 0.564 0.478 1.510 0.893 Skewness -0.913 -0.343 -0.221 -0.212 -1.513Table 2: The Descriptive statistics for the studied variablesThe correlation table among the variable and cronbach’s alpha along the diagonal is as, 1 2 3 4 5 1 Innovation Orientation (0.76) 2 External Partner 0.231** (0.79) Collaboration 3 Information Technology 0.198** 0.301** (0.81) 72
  • 73. Journal of Economics and Sustainable Development www.iiste.orgISSN 2222-1700 (Paper) ISSN 2222-2855 (Online)Vol.2, No.6, 2011 capability 4 Service Innovation 0.495** 0.201** 0.256** (0.76) 5 Forest Sustainability Index 0.219** 0.136** 0.436** 0.524** (0.87)** Correlation is significant at the 0.01 level (2 –tailed), * correlation is significant at the 0.05 level (2–tailed), Note: alpha values in the parenthesis along with the diagonalTable 3: The correlation table among the variable and cronbach’s alpha along the diagonal6. Data Analysis and ResultsFigure 3: The path coefficient values in the studied modelWe used structural equation modelling technique for the data analysis. AMOS 18.0 has been used for thepath analysis. A review of the literature indicated that empirical testing of service innovation in forestryperformance is quite less, it worked as a pioneer study in this domain. Based on the 4 proposed hypotheses,the effect of mediator variables has been studied. Result shows the significant relationship of innovationorientation and external partner collaboration with service innovation. Both together explained 15.6% of the 73
  • 74. Journal of Economics and Sustainable Development www.iiste.orgISSN 2222-1700 (Paper) ISSN 2222-2855 (Online)Vol.2, No.6, 2011total variance. Collaborative competence has found to have a positive and significant effect (r=0.403, p<0.001).The main effect model has been explained using the structural equation modelling. The structural modelwith relevant path coefficients has been mentioned in the Figure 3. SEM takes a confirmatory approach totest the dependence relationships and account for measurement errors in the process of testing the model.The assessment of model fit has been done using the various fit indices. The testing of moderator effect hasbeen done using a interaction variable. The results of the SEM for main effects are shown in the table, χ2/df AGFI PGFI NFI TLI CFI PNFI RMSEA RMR 3.115 0.751 0.812 0.962 0.855 0.862 0.813 0.721 0.026Table 4: SEM model fit summaryThe chi-square/ df ratio of 2 to 3 is taken as good or acceptable fit (Bollen, 1989; Gallagher, Ting andPalmer, 2008). The various incremental fit indices include the Normal Fit Index (NFI), comparative FitIndex (CFI) or the Tucker-Lewis Index (TLI), with suggestions for a cut of 0.90 for a good fitting model(Hu and Bentler, 1999). Further the absolute fit index of Adjusted Goodness of Fit Index (AGFI) is greaterthan the minimal 0.75 cutoff (Gallagher, Ting and Palmer, 2008). The multiple R square for the model is0.612.The first hypotheses (H1) focus on the interrelationships innovation orientation and extent of serviceinnovation among the stakeholders. Similarly second and third hypothesis (H2 and H3) focused on theexternal partner collaboration and IT support with extent of service innovation respectively. In the presentmodel the direct effect and indirect effect (mediated by burnout) has been summarized as, Hypothesized Relationship Estimate P-value Service Innovation Innovation orientation 0.371 0.002 Service Innovation External Partner Collaboration 0.291 0.000 Service Innovation Information Technology 0.072 0.004 support Forest Sustainability Service Innovation 0.471 0.050 IndicesTable 5: Path coefficients from the SEM analysisIt can be seen from the SEM results that all the direct and indirect relationship has been found significant(p<0.05). The direct estimate of Job demand and Job resource on performance and turnover intentions havebeen found significantly higher than the indirect effect (mediated by burnout). It signifies that the burnoutpartially mediates the overall effect (lowering the estimate value with a significant relationship). 74
  • 75. Journal of Economics and Sustainable Development www.iiste.orgISSN 2222-1700 (Paper) ISSN 2222-2855 (Online)Vol.2, No.6, 20117. Discussion and ConclusionWith the rapid pace of structural change in the forest management practices, service innovation and itsrelationship with forest sustainability performance have increasingly attracted the attention of bothresearchers and practitioners. In this study, we investigated the role of service delivery innovation as amediator in a causal framework concerning the link with service innovation antecedents and the impact ofinnovation on SFM practices. The Assessment of model fit in SEM using the model fit indices (χ2 = 15.04,p = .087, RMSEA= .072, NFI = .96, NNFI = .95, CFI = .86, SRMR= .026) explains 62% of the totalvariance.The primary findings suggest that (a) innovation orientation and external partner collaboration are the keydrivers that lead to service delivery innovation, (b) service delivery innovation leads to improvedmanagement of forestry resources and in turn sustainability of forestry resources. Further, result showsthat incorporating service innovation in sustainable forest management techniques significantly contributeto the rural livelihood generation and hence socio-economic development of the area.8. Implications for ResearchOur results have three significant implications for research. First, our study highlights the S-D logicperspective to link service delivery innovation, with the overall performance of the sustainable forestmanagement programme. The study has been found pioneer in this direction as no previous other study hasbeen made in this direction till date. Further, the study has investigated the role of operant resources in theservice innovation on overall performance of the programme.This study provides encouraging evidence for the service innovation modes that integrate service providers’competences, services, service channels and stakeholders. Delivery innovation reshapes the user’sbehaviour and helps organizations innovate service value with end users (co-creation value with the endusers). Further, this study develops robust insights into the effects of innovation orientation and ITcapability on system performance. More importantly, we propose that this research model is a more suitableframework of service innovation than others in the literature because it includes not only intraorganizational components (i.e., innovation orientation and IT capability) but also inter organizational ones(i.e., external partner collaboration).9. Implications for PracticeThis study is having has various significant implications for practice. If an organization can create anadvantage in operant resources, it not only can gain competitive advantage in the marketplace but alsosustain its resources for the longer term. In the preview of sustainable forest management practice, forestmanagers as well as forest conservation group members need to foster creativity in conservation practices,train employees to accept or adopt any radical new ideas for putting up into the indicator list and develop aninnovation environment or culture of openness within the organization. With regard to IT capability, ITplays a critical role in the implementation of service innovation practices, especially in corporate firms. 75
  • 76. Journal of Economics and Sustainable Development www.iiste.orgISSN 2222-1700 (Paper) ISSN 2222-2855 (Online)Vol.2, No.6, 2011However in the current study, with regard to the sustainable forest management practices, it includes use ofGPS in forest monitoring and satellite based equipments in forest fire fighting.As for external partner collaboration, inter organizational collaboration is important for supplementing theinternal innovative activities of organizations (Deeds and Rothaermel 2003; Kalaignanam, Shankar, andVaradarajan 2007). This means that forest development authorities need to collaborate with agencies like,soil and water conservation, horticulture and other biodiversity conservation authorities that offer differentoperant resources to facilitate service innovation in forest conservation. Second, considering the differenttypes of service innovation already popular with the forest department, we recommend that organizationevaluate the risks/benefits of offering new service for both existing and new stakeholders. Third, theseservice innovations in forestry sector will play a critical role in facilitating superior forest managementpractices. Once successfully implemented in one forest management unit, the same can be replicated todifferent other units and hence larger forest area may come in the preview of SFM practice of criteria andindicator approach.10. Limitations and Future ResearchEven though this study offers valuable insights into service innovation in forest management practices, itstill has some limitations. First, the study was conducted with only 126 JFMC’s of Marwahi Forest Division,so the generalization of the results may not be applicable to other forest areas especially belonging to otherforest types. Second, the research model is based on the cross sectional data and is thus essentially a staticperspective. It may be worthwhile to study the relationship between service innovation and sustainability offorestry resources over time to explain the effects of innovation on performance indicators and also takingthe time lag effect in picture. This consideration is especially important because of the central role ofinnovation in this study. The effects of service innovation on forestry indicator performance may not beimmediately apparent. Third, the three operant resources may not be sufficient to cover the entire scope ofthe study. Here we did not address the properties of the partner collaboration (e.g., governance structure,power, trust, etc.).Further, as this empirical study included only self-reported data, future research should capture the points ofview of external partners. The in depth case studies might add to our knowledge in the same subjectespecially in the preview of national and localised indicators of forest sustainability measurement. 76
  • 77. Journal of Economics and Sustainable Development www.iiste.orgISSN 2222-1700 (Paper) ISSN 2222-2855 (Online)Vol.2, No.6, 2011Appendix 1: Refined set of Criteria and Indicators at FMU Level (2006) based on Bhopal- India Process Criteria Indicators Criterion 1: 1.1 Forest area under encroachment Increase in the extent of forest and 1.2 Area of dense, open and scrub forests tree cover 1.3 Tree cover outside forest area 2.1 Area of protected eco-systems (Protected Areas) Criterion 2: 2.2 Status of locally significant species Maintenance, conservation and (a) Animal and (b) Plant species enhancement of biodiversity 2.3 Status of non-destructive harvest of wood and Non-Wood Forest Produce Criterion 3: 3.1 Status of natural regeneration Maintenance and enhancement of 3.2 Incidences of forest fires ecosystem function and vitality 3.3 Incidences of pest and diseases Criterion 4: 4.1 Area under watershed treatment Conservation and maintenance of 4.2 Area prone to soil erosion soil and water resources 4.3 Soil fertility/Site Quality 5.1 Growing stock of wood Criteria 5: 5.2 Increment in volume of identified species of wood Maintenance and Enhancement of 5.3 Efforts towards enhancement of forest productivity: Forest Resource Productivity (a) Technological inputs 77
  • 78. Journal of Economics and Sustainable Development www.iiste.orgISSN 2222-1700 (Paper) ISSN 2222-2855 (Online)Vol.2, No.6, 2011 6.1 Recorded collection of Non-Wood Forest Produce Criteria 6: 6.2 Direct employment in forestry and forest based industries Optimization of forest resource utilization 6.3 Contribution of forests to the income of forest dependent people 7.1 (a) Number of JFM committees and area(s) protected by them (b) Degree of people’s participation in management and Criteria 7: benefit-sharing Maintenance and enhancement of (c) Level of participation of women social, cultural and spiritual 7.2 Use of indigenous technical knowledge: Identification, benefits Documentation and Application 7.3 Extent of cultural/sacred protected landscapes: forests, trees, ponds, streams, etc. (a) Type and area of landscape (b) Number of visitors 8.1 Existence of policy and legal framework Criteria 8: 8.2 Number of forest related offences Adequacy of Policy, Legal and 8.3 Forest Resource Accounting Institutional framework (a) Contribution of forestry sector to the GDP (b) Budgetary allocations to the forestry sectorReferencesAvlonitis, George J., Paulina G. Papastathopoulou, and Spiros P. Gounaris (2001), “An Empirically-BasedTypology of Product Innovativeness for New Financial Services: Success and Failure Scenarios,” Journalof Product Innovation Management, 18 (5), 324-342.Berry, Leonard L., Shankar V., Janet T. Parish, Susan Cadwallader, and Thomas Dotzel (2006) “CreatingNew Markets through Service Innovation,” Sloan Management Review, 47 (2), 56-63.Bharadwaj, Anandhi S. (2000), “A Resource-Based Perspective on Information Technology Capability andFirm Performance: An Empirical Investigation,” MIS Quarterly, 24 (1), 169-196.Blazevic, Vera and Annouk Lievens (2004), “Learning during the New Financial Service InnovationProcess: Antecedents and Performance Effects,” Journal of Business Research, 57 (4), 374-391.Blind, Knut (2006), “A Taxonomy of Standards in the Service Sector: Theoretical Discussion and EmpiricalTest,” Service Industries Journal, 26 (4), 397-420.Bollen, K.A. (1989), Structural Equations with Latent Variables, Wiley, New York, NY.Bolton, Ruth N. (2003), “Marketing Challenges of e-Services,” Communication of the ACM, 46 (6), 43-44.Burns, Tom and George M. Stalker (1977), The Management of Innovation, 2nd ed. London: Tavistock.Dean, Alison M. (2004), “Links between Organizational and Customer Variables in Service Delivery:Evidence, Contradictions and Challenges,” International Journal of Service Industry Management, 15 (4),332-350Deeds, David L. and Frank T. Rothaermel (2003), “Honeymoons and Liabilities: The Relationship betweenAge and Performance in Research and Development Alliances,” Journal of Product InnovationManagement, 20 (6), 468-485.Frambach, Ruud T., Harry G. Barkema, Bart Nooteboom, and Michel Wedel (1998), “Adoption of a Service 78
  • 79. Journal of Economics and Sustainable Development www.iiste.orgISSN 2222-1700 (Paper) ISSN 2222-2855 (Online)Vol.2, No.6, 2011Innovation in the Business Market: An Empirical Test of Supply-Side Variables,” Journal of BusinessResearch, 41, 161-174.Gallagher, D., Ting, L. and Palmer, A. (2008), “A journey into the unknown; taking the fear out ofstructural equation modeling with AOMS for the first-time user”, The Marketing Review, Vol, 8 No. 3, pp.255-27.Hunt, Shelby D. (2000), “ General Theory of Competition: Too Eclectic or Not Eclectic Enough? TooIncremental or Not Incremental Enough? Too Neoclassical or Not Neoclassical Enough?” Journal ofMacromarketing, 20 (1), 77–81.Hu & Bentler (1999), “Cut-off criteria for fit indexes in covariance structure analysis: Coventional criteriaversus new alternatives”, Structural Equation Modeling, 6(1), 1-55.Ja-Shen Chen, Hung Tai Tsou and Astrid Ya-Hui Huang (2009), “Service Delivery Innovation: Antecedentsand Impact on Firm Perfomance”, Journal of Service Research, 12 (1), 36-55Kalaignanam, Kartik, Venkatesh Shankar, and Rajan Varadarajan (2007), “Asymmetric New ProductDevelopment Alliances: Win-Win or Win-Lose Partnerships?” Management Science, 53 (3), 357-374.Kaufman, Allen, Craig H. Wood, and Gregory Theyel (2000), “Collaboration and Technology Linkages: AStrategic Supplier Typology,” Strategic Management Journal, 21 (6), 649-664.Kleijnen, Mirella, Ko de Ruyter, and Tor W. Andreassen (2005), “Image Congruence and the Adoption ofService Innovations,”Journal of Service Research, 7 (4), 343-359.Kotwal P.C., Verm Rajeev, Gairola Sanjay, Kumar A., Dugaya D, Rawat T.S.(2006), Bhopal ‐ India Process:Operational Strategy and Reporting Formats on Criteria and Indicators for Sustainable Forest Management in India: Training Manual, ISBN: 81‐7969‐038‐5, IIFM, Bhopal.Lievens, Annouk and Rudy K. Moenaert (2000), “New Service Team as Information Processing Systems,”Journal of Service Research, 3 (1), 46-65.Lovelock, Christopher and Evert Gummesson (2004), “Whither Services Marketing? In Search of a NewParadigm and Fresh Perspectives,” Journal of Service Research, 7 (1), 9–20.Lusch, Robert F., Stephen L. Vargo, and Matthew O’Brien (2007), “Competing through Service: Insightsfrom Service-Dominant Logic,” Journal of Retailing, 83 (1), 5-18.Madhavaram, Sreedhar and Shelby D. Hunt (2008), “The Service- Dominant Logic and a Hierarchy ofOperant Resources: Developing Masterful Operant Resources and Implications for Marketing Strategy,”Journal of the Academy of Marketing Science, 36 (1), 67-82.Markides, Constantinos (1997), “Strategic Innovation,” Sloan Management Review, 38 (3), 9-23.Nunnally, Jum C. (1978), Psychometric Theory, 2nd ed. New York: McGraw-Hill.Pavlovski, Christopher J. (2007), “Service Delivery Platforms in Practice,” IEEE CommunicationsMagazine, 45 (3), 114-121.United Nations (1993), Agenda 21: Earth Summit - The United Nations Programme of Action from Rio, UN,NY.Vargo, Stephen L. and Robert F. Lusch (2008), “Why Service?” Journal of the Academy of MarketingScience, 36 (1), 25-38.Verganti, Roberto and Tommaso Buganza (2005), “Design Inertia: Designing for Life-Cycle Flexibility inInternet- Based Services,” Journal of Product Innovation Management, 22 (3), 223-237.Wiertz, Caroline, Ko de Ruyter, Cherie Keen, and Sandra Streukens (2004), “Cooperating for ServiceExcellence in Multichannel Service Systems: An Empirical Assessment,” Journal of Business Research, 57(4), 424-436. 79
  • 80. Journal of Economics and Sustainable Development www.iiste.orgISSN 2222-1700 (Paper) ISSN 2222-2855 (Online)Vol.2, No.6, 2011Zaltman, Gerald, Robert Duncan, and Jonny Holbek (1973), Innovation and Organizations. New York:John Wiley.Zhou, Kevin Z., Gerald Y. Gao, Zhilin Yang, and Nan Zhou (2005), “Developing Strategic Orientation inChina: Antecedents and Consequences of Market and Innovation Orientations,” Journal of BusinessResearch, 58 (8), 1049-1058.Zomerdijk, Leonieke G. and Jan de Vries (2007), “Structuring Front Office and Back Office Work inService Delivery Systems,” International Journal of Operations & Production Management, 27 (1),108-131.Zeithaml, Berry and Parasuraman (1988), "Communication and Control Processes in the Delivery ofService Quality," Journal of Marketing, 4, pp. 35-48. 80
  • 81. Journal of Economics and Sustainable Development www.iiste.orgISSN 2222-1700 (Paper) ISSN 2222-2855 (Online)Vol.2, No.6, 2011 Determinants of Technical Efficiency of Rose Cut-Flower Industries in Oromia Region, Ethiopia Mohammed Aman School of Agricultural Economics and Agribusiness, Haramaya University P.O.Box +251 138, Dire Dawa, Ethiopia Email: rosemuhe@gmail.com or muhecristy@gmail.com Jema Haji School of Agricultural Economics and Agribusiness, Haramaya University P.O.Box +251 233, Haramaya, Ethiopia Email: jemmahaji@gmail.com or jema.haji@ekon.slu.seReceived: October 11st, 2011Accepted: October 16th, 2011Published: October 30th, 2011AbstractThe objective of this study is to measure and identify input use efficiency level of 28 rose cut flowerindustries in three districts of Oromia Regional state (Ethiopia) using a two stage approach. . In the firststage, a non-parametric (DEA) method was used to determine the relative technical, scale and overalltechnical efficiencies. In the second stage, a Tobit model was used to identify sources of efficiencydifferentials among industries. The results obtained indicated that the mean technical, scale and overalltechnical efficiency indices were estimated to be 92%, %61 and 58%, respectively for the cut flowerindustries. This Implies, major source of overall technical inefficiencies was scale of operation rather thanpure technical inefficiency. Besides, the estimated measures of technical efficiency were positively relatedwith Farming experience, formal schooling years of manager’s and negatively related with age of farms. Noconclusive result was obtained for the relation between size and efficiency.Key words: Technical Efficiency, Scale Efficiency, DEA, Tobit, Rose cut flowers, farming experience,Oromia, Ethiopia.1. IntroductionDiversification of agricultural production is seen as a priority for least developing countries to reducedependence on primary commodities. The main reason is, despite high dependence on these commoditiesfor their livelihood, declining trend of prices for primary agricultural commodities (Humphrey 2006).Accordingly, floriculture sector is chosen for enhancing farm incomes and reducing poverty in developingcountries. In particular, African countries have a comparative advantage in rose flower varieties production.Fewer economies of scale and labour-intensive nature of production in cut flower industries are majorsources comparative advantage for these countries (Labaste 2005).Due to suitable climatic conditions and natural resources; high level of support by the government;favorable investment laws and incentives; proximity to the global market (mainly Europe) and availabilityof abundant and cheap labour, Ethiopia is the second largest producer of rose cut flowers in Africafollowing Kenya (Habte 2007). Moreover, the cut-flower industry in Ethiopia has emerged as one of thebiggest sources of foreign exchange earning in recent years. This is mainly because of increased demandfor cut-flowers, in the world market, by countries like Netherlands, Germany, Italy, United States, United 81
  • 82. Journal of Economics and Sustainable Development www.iiste.orgISSN 2222-1700 (Paper) ISSN 2222-2855 (Online)Vol.2, No.6, 2011Kingdom and Switzerland (Belwal & Chala 2007).Despite above opportunities, the high costs of technology, knowledge intensity of production, lack of accessto capital, strict market regulations and standards; and demanding infrastructural requirements along withnon existence of diversity in cut flower exports i.e. more than 80 % are a single rose variety, made thecountry not to benefit much (Melese 2007). To achieve simultaneous cost reduction and higher yield levelof rose cut flowers, improving the resource use efficiency of these industries is relevant. And hence, thisstudy is designed to estimate technical efficiency level and to identify its main determinants in theproduction of rose cut flowers in Awash Melkassa, Bishoftu and Ziway districts, respectively. These studyareas are chosen due to relative similarity in terms of their geographic characteristics, market conditions,production practices and type of rose cut flowers grown.In particular, this study tries to answer following questions; what is the existing level of efficiency of rosecut flower industries in the study areas? Is there any room for improvement in the level of efficiency forrose cut flower industries? What are the main causes for the existing level of efficiency? What are the mainpossible solutions to improve the existing level of efficiency in rose cut flowers production? By what levelwill input(s) be reduced to obtain the existing yield level of rose cut flower stems?The rationales for this study are, identifying the technical efficiency level of the rose cut flower industries,will help business owners as to what extent they can reduce scarce resource use while maintaining currentyield level of rose cut flower stems. Moreover, due to recent development of the sector in Ethiopia,checking for technical efficiency of these farms will also help policy makers in future policy design.2. Materials and Methods2.1. Sources and Types of DataPrimary data on the industry features, characteristics and production processes are collected through aninterview with farm managers using a semi-structured questionnaire from the 28 rose cut flower industries.Whereas, the secondary cross sectional data on input and output for one growing season (45 days) werecollected from each daily input use records in order to calculate the variables required for the empiricalanalysis. The data used in this study was drawn from a survey conducted from December 7, 2010-January,2010/11, in the three districts.2.2. Methods of Data AnalysisGiven decision making units (DMUs) producing products (outputs) using inputs, input andoutput vectors may be represented by and , respectively. For each DMU, all data may be written interms of as input matrix (X) and as output matrix ( ). Under the assumption of , thelinear programming model for measuring the technical efficiency of rose cut flower farms can be given asfollows according to Coelli et al. (1998)is: θSubject toWhere, - Vector of rose cut flower stems by the industry - Vector of inputs of the rose cut flower industry. - Rose cut flower stems output matrix for cut flower industries. - Rose cut flower production input matrix for rose cut farms. - The input oriented technical efficiency score having value . If , the industry will be technically efficient; otherwise inefficient. And - Vector of weights which defines the linear combination of peers of the industry.The specification of is only suitable when all work at optimum scale. Otherwise,measures of technical efficiency can be mistaken for scale efficiency. Therefore, the model isreformulated by imposing a convexity constraint. Technical efficiency measure obtained with VRS model is 82
  • 83. Journal of Economics and Sustainable Development www.iiste.orgISSN 2222-1700 (Paper) ISSN 2222-2855 (Online)Vol.2, No.6, 2011also named as ‘pure technical efficiency’, as it is free of scale effects. Thus, technical efficiency with in DEA model could be obtained from following linear programming: θSubject toAndWhere, is a convexity constraint ( is an vector of ones) and other variables are asdefined in the model.When there is a difference in efficiency score values between and models, scale inefficiency isconfirmed, indicating that the return to scale is variable, i.e. it can be increasing or decreasing (Färe &Grosskopf 1994). Scale efficiency values for each analyzed can be obtained by the ratio betweenthe scores for technical efficiency with and . If the ratio value equals to 1, it indicatesare scale efficient and a value less than 1 imply scale inefficiency. Furthermore, operating withDecreasing Returns to Scale is operating under super-optimal condition while those operating withincreasing Returns to scale are assumed to operate under sub-optimal conditions.After efficiency scores are obtained, to identify the determinants of efficiency, a Tobit model is used for thesecond-stage relationship between efficiency measures and suspected correlates of inefficiency (Binam et al.2003; Iráizoz et al. 2003; Chavas et al. 2005; Barnes 2006). The reason for using Tobit model for DEAefficiency scores is the bounded nature of efficiency level between 0 and 1. In this case, estimation withOLS would lead to biased parameter estimates (Green 1991; Dhungana et al. 2004). Rather a two-limitTobit regression is estimated using commonly used statistical software STATA version 10.2.3. Definition of VariablesThe dependent variable, in first stage, is given by total number of rose cut flower stems produced. While,the independent variables included in this stage are: land measured by total hectare of land undergreenhouse, the total labour (total number of temporary and permanent), water (the total amount of water(m3) used in the greenhouse rose cut flower farms), the rose plant seedlings (estimated by the total numberof rose flower plant seedlings stems used by the farms.), nitrate, sulphate and acid fertilizers measured by(kg) used.In the second stage, however, the dependent variable is the technical efficiency score level of the rose cutflower industries. Then this dependent variable is regressed over the following farm specificsocio-economic variables. Average area per greenhouses; location of the farm in KM (distance from theBole International airport); Ownership (measured by dummy values of 1 if the rose cut flower farm isdomestically owned and 0 if owned by foreign investors.); Age of the farm (years since the establishmentof the rose cut flower farm); Manager’s education level (formal schooling years spent by the farm managers)and Manager’s farming experience (total years of farming experience by the farm manager’s in same orrelated farming).The output and inputs data, from the twenty-eight rose cut flower industries, are used to estimate thetechnical efficiency levels in the production of rose cut flowers by using DEAP version 2.1 with an inputorientation option i.e. since the industries are targeted at minimization of input use.3. Results and Discussion3.1. Descriptive statisticsThe mean land size holding, under greenhouses, in the study areas was 19.17 hectares, with minimum andmaximum sizes of 4.98 and 42 hectares, respectively. While, the mean employment level was 536 workers.The mean amount of nitrate, sulphate and acid fertilizers used in greenhouses were 12.82 kg, 26.46 kg and1.98 kg, respectively. Furthermore, the mean, water, rose flower plant seedlings and cut flower yield levelswere 47,300 m3, 1,066,500 and 9,671,800 stems, respectively.3.2. Efficiency (DEA) ResultsIn this section, district as well as industry level technical efficiency results are discussed. For the sake of 83
  • 84. Journal of Economics and Sustainable Development www.iiste.orgISSN 2222-1700 (Paper) ISSN 2222-2855 (Online)Vol.2, No.6, 2011comparison, technical efficiency indices are estimated both under and along with scaleefficiency level and type of returns to scale.The district level technical efficiency results indicate that, industries at Awash Melkassa, Bishoftu andZiway districts could reduce their input use by 24%, 64% and 22% without any loss of rose cut flowerstems. While, the mean technical efficiencies, using , were 100, 90 and 94 percent for Awash Melkassa,Bishoftu and Ziway districts, respectively (Table 1).Table 1. Descriptive statistics of the technical efficiency scores of districts Case Districts N Mean S.dev Min Max Range Awash Melkassa 1 0.76 - 0.76 0.76 0.00 CRS Bishoftu 13 0.36 0.22 0.14 1.00 0.86 Ziway 14 0.78 0.34 0.15 1.00 0.85 Total 28 0.58 0.35 0.14 1.00 0.86 Awash Melkassa 1 1.00 - 1.00 1.00 0.00 VRS Bishoftu 13 0.90 0.11 0.71 1.00 0.29 Ziway 14 0.94 0.10 0.64 1.00 0.36 Total 28 0.92 0.11 0.64 1.00 0.36Source: Author survey, 2011.Furthermore, results obtained indicate statistically significant difference in mean technical efficiency levelfor rose cut flower industries in case of CRS i.e. farms in Ziway district were performing well followed byfarms in Awash Melkasa and Bishoftu districts. However, the difference was not statistically significant incase of VRS. This may be due to similar technologies of production used by rose cut flower industries.Frequency distribution of the technical efficiency scores in both CRS and VRS are given in Table 2.Accordingly 8 industries under CRS and 16 under VRS are technically efficient and the remainingtechnically inefficient.Table 2. Frequency distribution of technical efficiency scores (VRS, CRS and SE) Frequency of Technical EfficiencyTE scores CRS Percent VRS Percent SE Percent1.00 8 28.57 16 57.14 8 28.570.91- 0.99 2 7.14 2 7.14 3 10.710.81- 0.90 1 3.57 5 17.86 -0.71- 0.80 1 3.57 4 14.28 1 3.570.61- 0.70 - 1 3.57 1 3.570.51- 0.60 - - -0.41- 0.50 3 10.71 - 2 7.140.31- 0.40 5 17.86 6 21.430.21- 0.30 4 14.28 5 17.860.11- 0.20 4 14.28 2 7.14Mean 0.58 0.92 0.61Minimum 0.14 0.64 0.19Maximum 1.00 1.00 1.00S.dev 0.35 0.11 0.33Source: Author survey, 2011.3.3. Returns to Scale of IndustriesThe majority,19 (67.85%), of scale inefficient rose cut flower industries were operating under IRS withonly one farm operating under CRS. Those operating under IRS are small industries that need to increasetheir size of operation. While, those operating under DRS are large industries operating above their optimalscale and thus could be better-off by reducing their size of operation. Accordingly, most of the scale 84
  • 85. Journal of Economics and Sustainable Development www.iiste.orgISSN 2222-1700 (Paper) ISSN 2222-2855 (Online)Vol.2, No.6, 2011inefficient industries, in the three districts, need to expand their size of holding to efficiently utilize theirresources (Table 3).Table 3. Efficiency and returns to scale distribution of rose cut flower industries Technical Efficiency Returns to Scale CRS CRS SE IRS CRS DRS Industries Efficient 8 16 8 Inefficient 20 12 20 Total 28 28 28 19 8 1 Mean 0.58 0.92 0.61Source: Own (Authors) calculationThe mean SE was 0.61 This result imply that , the average size of the greenhouse rose cut flower industriesin the study areas is far from the optimal scale and an additional 39 % productivity gain could be feasible,provided they adjusted their farm’s operation to an optimal scale.The causes of inefficiency for the industries could be either inappropriate scale or misallocation ofresources. Inappropriate scale suggests that industries are not taking advantage of economies of scale inrose cut flower production process. While, misallocation of resource refers to inefficient input combination.As shown in Table 3, the mean SE and TE (VRS) score were 0.61 and 0.92, respectively. This relativelylow scale efficiency mean value indicates the main cause of technical inefficiency, for the rose cut flowersindustries, is inappropriate scale (scale inefficiency) rather than misallocation of resources (pure technicalinefficiency).In this study, input slacks are also estimated. The input slacks, using the VRS technical efficiency measure(pure technical efficiency), indicate excess use of that input(s) relative to other peer farms. The mean slackvalues for land, labour, water, nitrate, sulphate and acid fertilizers and rose plant seedling inputs are 0.76hectares, 478 workers,1,810 m3,1.3 kg,1.76 kg,0.22 kg and 130041 stems, respectively. Among fertilizers,smallest mean input slack value is obtained for acids followed by nitrates. The reason is that, acid fertilizersare used to clean the drip pipes and less frequently applied than the two fertilizers. The rose cut flowerindustries can reduce costs, incurred on inputs, by the amount of slacks without reducing production levelof rose cut flower stems.3.4. Determinants of Technical EfficiencyIn order to examine the effect of relevant technological, farm specific and socio-economic factors ontechnical efficiency of rose cut flower farms, the input oriented VRS technical efficiency scores areregressed on the selected explanatory and farm specific variables using a two-limit Tobit model sinceefficiency scores are bounded between 0 and 1 (Table 4). Table 4. Determinants of technical efficiency in rose cut flowers production The log likelihood estimates The Robust standard estimates Variables Parameters Efficiency effect Parameters Efficiency effect Constant( ) b0 -0.976 b0 -0.976 (0.667) (0.258) b1 -0.020 b1 -0.020 (0.023) (0.016) b8 -0.004 b8 -0.004 (0.012) (0.009) b9 0.259 b9 0.259 (0.101) (0.093) 85
  • 86. Journal of Economics and Sustainable Development www.iiste.orgISSN 2222-1700 (Paper) ISSN 2222-2855 (Online)Vol.2, No.6, 2011 b10 -0.009* b10 -0.009** (0.035) (0.022) b11 0.104** b11 0.103*** (0.039) (0.028) b12 0.048** b12 0.048*** (0.023) (0.014) Log likelihood 0.937 Log 0.937 Pseudo likelihood Sigma δ 0.133 δ 0.133*** Significant at 1%; ** significant at 5%; * significant at 10%. Standard errors were shown inparenthesis, (x1= Average land area, x8 =location, x9 = ownership, x10 = age of industry, x11= managerseducation and x12= managers experience).The result obtained for age of the industry shows a negative and significant effect on technical efficiency ofrose cut flower production implying older rose cut flower farms are less technically efficient than new ones.While, the formal education schooling years and experience in same or related business of farm manager’shas positive and significant effect on technical efficiency level of the farms.Furthermore, the marginal effects for the determinants of technical efficiency were also estimated. Andhence, for a unit percentage increase in years rose cut flower farms, technical efficiency decreases by 1 %.While, a one percent additional formal schooling years of the farm mangers will improve technicalefficiency of the industry by 10.3 %. Finally, a one percent additional farming experience of the farmmanager will improve technical efficiency by 4.8 % (Table 5).Table 5. Marginal effects of efficiency variables after Tobit regression The log likelihood estimates The Robust standard estimates Variables Parameters dy/dx Parameters dy/dx b1 -0.020 b1 -0.020 (0.023) (0.016) b8 -0.004 b8 -0.004 (0.012) (0.009) b9 0.259 b9 0.259 (0.101) (0.093) b10 -0.01* b10 -0.01* (0.035) (0.022) b11 0.103*** b11 0.103*** (0.039) (0.028) b12 0.048** b12 0.048*** (0.023) (0.014)Source: authors own calculation4. ConclusionResults obtained indicated that there is a room to improve technical efficiency level of rose cut flowerindustries in the study areas. For instance, the mean scale efficiency value of 0.61 implying, on average,rose cut flower industries in the three study districts are not operating at their optimal farm size. Among thefactors that are assumed to affect technical efficiency level, experience in same or related farming activitiesas well as more years of formal schooling by the farm manager increased technical efficiency levelWhereas, age of the farm, along with rose cut flowers grown inside, decreased the technical efficiency level.As far as marginal gain in technical efficiency is concerned, formal years of schooling dominates that offarm manager farming experience in same or related farming activities.ReferencesBarnes, A.P. (2006), “Does multi-functionality affect technical efficiency? A non-parametric analysis of theScottish dairy industry”, Journal of Environmental Management 80,287-294. 86
  • 87. Journal of Economics and Sustainable Development www.iiste.orgISSN 2222-1700 (Paper) ISSN 2222-2855 (Online)Vol.2, No.6, 2011Belwal, R., & Chala, M. (2007), “Catalysts and barriers to cut flowers export: A case study of Ethiopianfloriculture industry”, International Journal of Emerging Markets 3(2), 216-235.Binam, JN., Sylla, K., Diarra, I., & Nyambi G. (2003), “Factors affecting technical efficiency among coffeefarmers in Cote d’Ivoire: Evidence from the centre west region”, African Development Review 15, 66-76.Chavas, J-P., Petrie, R., & Roth, M. (2005), “Farm household production efficiency: Evidence from theGambia”, American Journal of Agricultural Economics 87(1), 160-179.Coelli, T.J., Rao, D.S.P., O’Donnell, C.J., & Battese, G.E. (1998), “An introduction to productivity andefficiency analysis (2nd edition)”, Springer Science, New York.Dhungana, B.R., Nuthall, P.L., & Nartea, G.V. (2004), “Measuring the economic inefficiency of Nepaleserice farms using data envelopment analysis”, The Australian Journal of Agricultural and ResourceEconomics 48 (2), 347-369.Färe, R., Grosskopf, S., & Lovell, C.A.K. (1994), Productivity frontiers, Cambridge: Cambridge UniversityPress.Green, W. (1991), LIMDEP: User’s manual and reference guide, New York: Econometric software, Inc.Habte, S. (2007), Ethiopia cut flower export industry and international market: Investment opportunityprofile in Ethiopia.Humphrey, J. (2006), “Horticulture: Responding to challenges of poverty reduction and global competition”,Acta Horticulture 69(9), 19–38.Iráizoz, B., Rapùn, M., & Zabaleta, I. (2003), “Assessing the technical efficiency of horticulturalproduction in Navarra, Spain” Agricultural Systems78, 387-403.Labaste, P. (2005), The European horticulture market. Opportunities for Sub- Saharan African exporters.’Working Paper NO. 63, Washington: The World Bank.Melese, A. (2007), Triple role of the Dutch in the growth of the cut-flower industry in Ethiopia.Unpublished Thesis for partial fulfillment of masters of arts in development studies, The Hague, TheNetherlands. List of TablesTable PageTable 1. Descriptive statistics of the technical efficiency scores of districts ............................................................... 84Table 2. Frequency distribution of technical efficiency scores (VRS, CRS and SE) .................................................. 84Table 3. Efficiency and returns to scale distribution of rose cut flower industries ...................................................... 85Table 4. Determinants of technical efficiency in rose cut flowers production ............................................................ 85Table 5. Marginal effects of efficiency variables after Tobit regression ..................................................................... 86 87
  • 88. Journal of Economics and Sustainable Development www.iiste.orgISSN 2222-1700 (Paper) ISSN 2222-2855 (Online)Vol.2, No.6, 2011 AppendicesAppendix Table A 1.Name and location of DMUs (rose cut flower farms) DMUs Name of the rose cut flower farm District (location ) 1 AQ Roses Plc Ziway 2 Exp. Incorporated Chibo Flowers Ziway 3 Sher flowers 8 Ziway 4 Braam Flowers PLC Ziway 5 Ziway Roses PLC Ziway 6 Sher Ethiopia PLC Ziway 7 Rainbow Colors PLC Bishoftu 8 Yassin Legesse J. Flower Farm Bishoftu 9 Dugda Floriculture Devt PLC Bishoftu 10 Joytech PLC Bishoftu 11 Bukito Agro Industry Bishoftu 12 Friendship Flowers Bishoftu 13 ZK Flower Bishoftu 14 Eyasu Sirak Workineh Flowers PLC Bishoftu 15 Olij Flowers PLC Bishoftu 16 Minaye Flowers PLC Bishoftu 17 Roshanara Rose PLC Bishoftu 18 Super Arsity Flower PLC Awash Melkasa 19 Sher flowers 1 Ziway 20 Sher flowers 2 Ziway 21 Sher flowers 3 Ziway 22 Sher flowers 4 Ziway 23 Experience Flowers PLC Ziway 24 Evergreen Roses PLC Bishoftu 25 Zubka General Business Flower Farm Plc Bishoftu 26 Sher flowers 5 Ziway 27 Sher flowers 6 Ziway 28 Sher flowers 7 Ziway 88
  • 89. Journal of Economics and Sustainable Development www.iiste.orgISSN 2222-1700 (Paper) ISSN 2222-2855 (Online)Vol.2, No.6, 2011 Socio-Economic Impact through Self Help Groups D.Amutha Asst.Professor of Economics, St.Mary’s College (Autonomous), Tuticorin Email: amuthajoe@gmail.comReceived: October 14th, 2011Accepted: October 19th, 2011Published: October 30th, 2011AbstractThe overall objective of the present study is to analysis the economic empowerment of women thoughSHGs in three villages of Tuticorin District of Tamilnadu. This study is compiled with the help of theprimary data covered only in a six month period (2011). Totally 238 respondents were selected from 18SHGs of three villages by using simple random sampling method. Empowerment signifies increasedparticipation in decision-making and it is this process through which people feel themselves to be capableof making decisions and the right to do so. Women’s participation in decision-making in family is importantindicator for measuring their empowerment. The analysis shows that 66 percent beneficiaries reporteddecisions are being taken by their husbands, yet, more than 34 percent respondents accepted that they doparticipate in decision-making process. Thus, the socio-economic conditions of women have demonstratedthat their status has improved since the joining of SHG’s and availing microfinance. The result ofchi-square- test revealed that there is significant difference between participation in decision-making infamily and SHG women members in Tuticorin District.Keywords: Self-Help Groups, women empowerment, percentage analysis, averages, chi-square tests1. IntroductionEmpowerment can serve as a powerful instrument for women to achieve upward social and economicmobility and power and status in society. Women’s empowerment would be able to develop self-esteem,confidence, realize their potential and enhance their collective bargaining power. A SHG is a smalleconomically homogeneous affinity group of the rural poor voluntarily coming together to save smallamount regularly, which are deposited in a common fund to meet members emergency needs and to providecollateral free loans decided by the group. (Abhaskumar Jha 2000). They have been recognized as usefultool to help the poor and as an alternative mechanism to meet the urgent credit needs of poor through thrift(V. M. Rao 2003) SHG is a media for the development of saving habit among the women (S. Rajamohan2003). SHGs enhance the equality of status of women as participants, decision-makers and beneficiaries inthe democratic, economic, social and cultural spheres of life. (Ritu Jain 2003). The basic principles of theSHGs are group approach, mutual trust, organization of small and manageable groups, group cohesiveness,sprit of thrift, demand based lending, collateral free, women friendly loan, peer group pressure inrepayment, skill training capacity building and empowerment (N.Lalitha). Some estimates put these atcurrently 2.5 million SHGs in India. (Economic Survey of India, p.67). In Tamil Nadu the SHGs werestarted in 1989 at Dharmapuri District. At present 1.40 lakh groups is a function with 23.83 lakhmembers. Many men also eager to form SHGs, at present. Tuticorin District having 19 town panchayatsformed 1230 SHGs, and their achievement is 259%. In the process, it aims to commission women withmultiform forms of power; hence a study was conducted on empowerment of women by SHGs in TuticorinDistrict, Tamil Nadu.1.1 ObjectivesThe main objectives of the study are mentioned below: • To study the socio economic background of the members of Self Help Groups of Tuticorin District • To know the reasons for joining SHGs. • To examine the activities of Self Help Groups in the study area. • To evaluate the political and entrepreneurial empowerment of SHG members. 89
  • 90. Journal of Economics and Sustainable Development www.iiste.org ISSN 2222-1700 (Paper) ISSN 2222-2855 (Online) Vol.2, No.6, 2011 1.2 Methodology The present study has covered three villages from Tuticorin District viz, Meelavittan, Mullakkadu and Korampallam. These three villages were selected for this because of the SHGs in these villages is functioning in a very successful manner. This study is compiled with the help of the primary data covered in only six month period (2011). The primary data collected with the help of specially prepared interview schedule. Totally 238 respondents were selected from 18 SHGs of three villages by using simple random sampling method. The sample size was 1/3 of the total members of the SHGs. This is purely a descriptive study. Percentage analysis, averages, standard deviation, variance, chi-square tests, Cramer’s V and probability analysis were used for the analysis. 2. Discussion and analysis 2.1 Age Age is an important factor in determining the empowerment of SHG members. In the present study an effort has been made to know the age group of the respondents. In the study area, 238 respondents in three villages from Tuticorin District viz, Meelavittan, Mullakkadu and Korampallam were selected for the study and average number of the respondents in Tuticorin District study areas was about 48 members and standard deviation was about 16. Table 1 Age wise classification of the respondentsParticulars(years) Frequency Percentage Results31-40 39 16.4 Mean (Average):59.541-50 162 68.1 Standard deviation: 70.0214351-60 33 13.9 Variance (Standard deviation): 490361 and Above 4 1.6 Population Standard deviation: 60.64033 14.2632460.64033Total 238 100.0 Variance (Population Standard deviation):3677.25 Source: Primary data The table 1 shows the age wise classification done into four different categories. From the table we find that the respondents mainly fall under the age group of 41-50 where the highest frequency occur i.e. 162 and in total sample size it constitutes to 68.1% followed by the age group of 31-40 and its frequency is 39 which constitute 16.4% in the total sample, followed by the age group of 51-60 which the frequency is 33 and its percentage of the total sample size is 13.9%. This indication would be relevant to the study because most of the earning members in a family would be in the age group of 41-50. 2.2 Community The caste system was introduced in ancient India on the basis of occupation. Even now to some extent people of a particular caste or community stick on to a particular trade. Even though the caste behaviour can be moulded with the help of education, exposure and multi-media development, it plays its own role relating to empowerment of SHG members. Table 2 Community-wise Classification of the respondents Particulars Frequency Percentage Backward Classes 85 35.7 Most Backward Classes 22 9.2 MBC SC/ST 131 55.1 Total 238 100.0 Source: Primary data 90
  • 91. Journal of Economics and Sustainable Development www.iiste.org ISSN 2222-1700 (Paper) ISSN 2222-2855 (Online) Vol.2, No.6, 2011 From the table 2 it is understood that the respondents are mainly from the SC/ST community which is 55.1% of the total respondents. Following this next stands the backward class community with 85 members and 35.7% of the total sample size. Here most backward class is only a meager amount in the sample i.e. 22 in numbers and 9.2% in total percentage. 2.3 Religion The religion of the family is a major influence on the empowerment of SHG members. Table 3 Religion wise Classification of the Respondents Particulars Frequency Percentage Hindu 205 86.1 Muslim 2 0.8 Christian 31 13.1 Total 238 100.0 Source: Primary data The table 3 reveals that 86.1% of the total respondents belong to the Hindu religion and the actual number is 205 of the total 238 samples selected. The remaining 13.1% are Christians. This might be about the fact that the field area chosen should be a Hindu religion dominated area. 2.4 Occupation The SHG members were engaged in various occupations such as agriculture operations and private jobs. The occupation of the members helps them to avail themselves of credit from the banks and invest it in their irrespective occupations to earn more. Table 4 Occupation of the RespondentsParticulars Frequency Percentage ResultsUnemployed 9 3.8 Mean (Average): 59.5Agriculture 221 92.9 Standard deviation: 107.69556Industry 3 1.2 Variance (Standard deviation): 11598.33333Collie 5 2.1 Population Standard deviation: 93.26709Total 238 100.0 Variance (Population Standard deviation): 8698.75 Source: Primary data From table 4 we conclude that more number of the respondents is engaged in agriculture for their means for livelihood i.e. 221 respondents are agriculture earning people. 2.5 Reasons for joining SHGs The major aim of SHG is to promote savings, generate income and credit for the productive and consumptive purposes. This is true because in the study area the sample women joined the SHGs for getting loan and promoting their savings and income, in addition to attaining of social status. Table 5 Reasons for joining SHGs Reasons Frequency Percentage Family Income 203 85.3 Bored at home 21 8.8 To give good life to children 14 5.9 91
  • 92. Journal of Economics and Sustainable Development www.iiste.org ISSN 2222-1700 (Paper) ISSN 2222-2855 (Online) Vol.2, No.6, 2011 Total 238 100.0 Source: Primary data (Mean: 79.33333; Standard deviation: 107.15565) The table 5 reveals that 85.3 percent of women stated that the most important motivating factor to join the SHG was to supplement their family income. It also reveals that there is definite growing awareness in society and women in particular that if the family has to maintain a reasonable standard of living, women should supplement to family income with whatever skill they have 8.8 and 5.9 percent of them stated that they join the SHG was bored at home and to provide good life for their children, respectively. 2.6 Satisfactions with Family The satisfaction of respondents with family members means equal status, participation and powers of decision making of women in household level. The satisfaction of respondents with family members has been reported to be quite high. Table 6 Satisfactions with FamilySatisfaction Meelavittan Mullakkadu Korampallam TotalVery Happy 58(60.4) 65(79.3) 39(65) 162 (68.1)Not Happy 38(39.6) 17(20.7) 21(35) 76 (31.9)Total 96(100.0) 82(100.0) 60(100.0) 238 (100.0) Source: Primary data (χ2=7.58, P=0.022596, df=2, Cramer’s V=0.1785), Significant at 1% probability level. As shown in table 6 most of them were found satisfied (68%) with the family members, while a significant proportion was reported to be burdened (32%). The chi-square analysis result (χ2=7.58, P=0.022596), shows no significant relationship of satisfaction of respondents with family members and SHG members of group in Tuticorin District. 2.7 Activities of SHG SHG’s have created positive attitude of community towards functioning of SHG’s, micro-financing as well as being effective on social problems. Table 7 Activities of SHG Activities of SHGs Frequency Percentage Expands services area 54 22.6 Communication skills& marketing techniques updated 24 10.1 Contact with personnel from government & public organizations 63 26.5 NGOs’& other knowledge of how to get things done in public life 58 24.3 Increase in self confidence and risk bearing capacity 39 16.4 Total 238 100.0 Source: Primary data (Mean: 47.6; Standard deviation: 15.94679; Variance: 254.3) Generally SHGs include various activities. The important activities are contact with personnel from government and public organizations (26.5%) and NGOs’ and other knowledge of how to get things done in public life (24.3%) and average number of the respondents in Tuticorin District study area was about 48 members and standard deviation was about 16. 92
  • 93. Journal of Economics and Sustainable Development www.iiste.org ISSN 2222-1700 (Paper) ISSN 2222-2855 (Online) Vol.2, No.6, 2011 2.8 Decisions Making in Family Empowerment signifies increased participation in decision-making and it is this process through which people feel themselves to be capable of making decisions and the right to do so. Women’s participation in decision-making in family is important indicator for measuring their empowerment. Table 8 Decisions Making in FamilyDecision Making Meelavittan Mullakkadu Korampallam TotalHusband 49(54.4) 72(83.7) 36(58.1) 157 (65.9)Yourself 41(45.6) 14(16.3) 26(41.9) 81 (34.1)Total 90(100.0) 86(100.0) 62(100.0) 238 (100.0) Source: Primary data (χ2=19.12, P=0.000070, df=2, Cramer’s V=0.2834), Significant at 5% probability level It is obvious from table 8 that 66 percent beneficiaries reported that decisions are being taken by their husbands, yet, more than 34 percent respondents accepted that they do participate in decision-making process. Thus, the socio-economic conditions of women have demonstrated that their status has improved since the joining of SHG’s and availing microfinance. The result of chi-square- test (χ2=19.12, P=0.000070), revealed that there is significant difference between participation in decision-making in family and SHG women members in Tuticorin District. 2.9 Political Empowerment Participation in Panchayatraj institution, understanding the political environment and accessing political power provide empowerment. So the opinion of the respondents were collected and shown in the following table. Table 9 Political Empowerment Particulars Meelavittan Mullakkadu Korampallam Total Participation in Panchayatraj 68(76.4) 80(90.9) 23(37.7) 171 (71.8) Institution Understand the Political Environment 21(23.6) 8(9.1) 38(62.3) 67 (28.2) Total 89(100.0) 88(100.0) 61(100.0) 238 (100.0) Source: Primary data (χ2=51.88, P=<0.0001, df=2, Cramer’s V=0.4669), Significant at 1% probability level Participation in Panchayatraj institution and understanding the political environment provide empowerment. So the opinion of the respondents were collected and shown in the following table. 72% respondent were expressed that participation in Panchaytraj institution and 28% respondents were of the view that understanding the political environment showed significant difference (P < 0.001). 2.10 Entrepreneurial Empowerment Entrepreneurs are those persons who seek to generate value, through the creation or expansion of economic activity by identifying and exploiting new products, process or markets. Entrepreneurial activity is the enterprising human action in pursuit of the generation of value, through the creation or expansion of economic activity, by identifying and exploiting new products, process or markets. Entrepreneurial activity includes the entry of new products, the creation of new products or service, and the innovation associated with different business activated. Entrepreneurial activity can therefore be associated with organic as well as acquisitive decision. 93
  • 94. Journal of Economics and Sustainable Development www.iiste.org ISSN 2222-1700 (Paper) ISSN 2222-2855 (Online) Vol.2, No.6, 2011 Table 10 Entrepreneurial EmpowermentParticulars Meelavittan Mullakkadu Korampallam TotalIncrease desire to learn more 17(10.2) 4(9.1) 23(82.1) 44 (18.5)professional skillsIntensifies desire to earn and better 149(89.8) 40(90.9) 5(17.9) 194 (81.5)livingTotal 166(100.0) 44(100.0) 28(100.0) 238 (100.0) Source: Primary data (χ2=85.36, P=<0.0001, df=2, Cramer’s V=0.5989), Significant at 1% probability level From the table, 82% respondents were of the view that SHG intensifies desire to earn more and make better living and only 19 percent of women expressed that SHG increases desire to learn more professional skills and the difference was statistically significant (χ2=85.36, P=<0.0001). 3. Conclusion SHGs started functioning all over Tamilnadu, in some areas they are functioning effectively whereas in other areas they face problems. Since SHGs help women to achieve economic empowerment, these policy measures can contribute a lot to the nation. To conclude, the economic activities of SHGs in Tuticorin District are quite successful. References Abhaskumar Jha (2004), “Lending to the Poor: Designs for Credit”, EPW, Vol. XXXV, No.8 and 9. Chiranjeevulu T. (2003), “Empowering Women through Self Help Groups – Experiences in Experiment”, Kurukshetra, March. Chopra Kanchan (2004), “Social Capital and Development Processes – Role of Formal and Informal Institutions “, Economic and Political Weekly, July, 13. Economic Survey of India, 2007-08. Jeyanthi Gayari, R (2002), “SHGs in Kanyakumari District”, M.Phil., dissertation submitted to Alagappa University, Karaikudi, T.N. Jeyaraman, R. et al., (2004), “Role of Self help Groups in Fisher Women Development,” Peninsular Economist, Vol. XII, No.2, pp. 197-200. K. Usha (2003), “Gender, Equality and Development”, Yojana Keishnaraj, Maithreyi (2005), ‘Growth and rural Poverty”, Economic and Political Weekly, September 21. Lalitha, N. “Women Thrift and Credit Groups- Breaking the Barriers at the Gross Roots”, Peninsular Economist, Vol. XII No. 2, pp. 188-195 Manimekalai, N. et. al., “Gross-root Women Entreprenurship through SHGs”, Peninsular Economist Vol. XII, No.2, pp. 181-187 Rajamohan, S. (2003) , “Activities of Self Help Groups in Virudhunagar District-A Study, TNJC, pp. 25-29 Rao, V.M. (2003), Women Self Help Groups, Profiles from Andhra Pradesh and Karnataka”, Kurukshetra, Vol. 50, N0.6, pp. 26-32 Ritu Jain, (2003), “Socio-Economic Impact through Self Help Groups”, Yojana, Vol. 47, No.7, pp.11-12 Sabyasachi Das, (2003), ‘Self Help Groups and Micro Credit Synbergic Integration”, Kurushetra Vol. 51, No.10, pp. 25-28 Sreeramulu, G (2006), Empowerment of Women through Self-Help Group, Eastern Book Corporation. Vandana K. Jena, (2007), Literacy for Women’s Empowerment, Indian Journal of Population Education, IAEA, New Delhi, No.38, July-Sep. www.expressindia.com www.flyhighonline.com www.indianngos.com www.knowledgeallianz.com www.microfinancegateway.org 94

×