International Journal of Management (IJM), ISSN 0976 – 6502(Print), ISSN 0976 –6510(Online), Volume 4, Issue 2, March- April (2013)245standard of living, higher purchasing power, greater opportunities for employment and overall development, better, proficient and optimum use of natural and agricultural resources arevital. Punjab is predominantly agricultural state and economy mainly depends uponagriculture. Now there is a time to report that Punjab economy is not only known for itsagriculture production rather industrial sector is also playing an important role in the overalldevelopment of the Punjab. Therefore, the need of the hour is to devote greater attentiontowards the development of industries in the state. Only then, Punjab will be able to maintainits flourishing and strong economy. The main aim of the paper to bring the notice that Punjabeconomy also has the industrial potential and with help of industrial exports, the economycan achieve the higher rate of growth.Punjab has highly developed small scale industries and has surplus of various smallscale and other industrial and manufactured products such as bicycles, sewing machines,hosiery goods, sports goods, leather goods, hand tools and machine tools etc. Intensive andcommercial agriculture has generated surplus income in Punjab and thousands of migrant andNRI Punjabi’s are sending large amount of money back to their homes in Punjab. This hasresulted in higher purchasing power and there has developed demand for luxury andconsumer goods in Punjab. Therefore, Punjab has a large flourishing trade. This trade ofPunjab is internal or inter-state or international. This paper consider only international i.e.goods which are exported to other countries from Punjab and contribution of Punjab state inIndia foreign trade. Punjab is an agriculture dominant state. It has surplus of agriculturalproduce. With a population of 27.7 million (Data based on 2001 Census), the two-thirds(66.05 per cent) of the population is dependent on agriculture. Though Punjab is only 1.53per cent of the geographical area of India, but its contribution to Indian agriculture isremarkable. In 2009-10, the total production of food grains in the state was around 26.9million metric tonnes. In 2009-10, the total fruit production was 1.3 million metric tonnes. Incase of food grains, wheat is the major crop. It was followed by rice and maize. Punjab is thesecond-largest producer of wheat in the country, with a share of around 20 per cent of thetotal wheat production. Besides, Punjab has tremendous potential to develop food-processingindustry of citrus fruits, grapes and potatoes. Potato production in the state was around 2.1million metric tonnes in 2009-10. (Statistical Abstract of Punjab-Various issues)The principal export items were yarns and textiles, hosiery and readymade garments,rice and machine tools/hand tools in the year 2009-10.Ludhiana, Jalandhar and Amritsaraccount for around 92 per cent of the total exports of Punjab. Clusters identified for bicyclesand bicycle parts (Ludhiana), steel re-rolling (Mandi Gobindgarh), textiles (Ludhiana), sportsand leather goods (Jalandhar), and woollens (Amritsar). (ibef.org)A large part of industrial exports of Punjab originated from its three major industrialdistricts namely Ludhiana (51 per cent), Amritsar (18 per cent) and Jalandhar (21.7 per cent)in 1999-2000 and in the year 2009-10 total exports from Jalandhar were Rs. 2729.46 crore,Amritsar Rs. 2306.53 crore and from Ludhiana Rs. 9730.73 crore. Total exports fromPunjab in 2009-2010 were worth Rs. 15972.48 crore. (Department of Industries & CommercePunjab)
International Journal of Management (IJM), ISSN 0976 – 6502(Print), ISSN 0976 –6510(Online), Volume 4, Issue 2, March- April (2013)246Table-1.1: Annual Growth of Export State of Punjab and Nation (1990-2010)YearIndiasExports(in cr)PunjabsExports(in cr)IndiasAnnualExportGrowthPunjabsAnnualExportGrowth90-91 32558 769 17.7 18.891-92 44042 901 35.3 17.192-93 53688 1215 21.9 34.993-94 69751 1815 29.9 49.494-95 82674 2082 18.5 14.795-96 106353 2565 28.6 23.296-97 118817 3641 11.7 42.097-98 130101 4205 9.5 15.598-99 139753 3629 7.4 -13.799-2000 159561 4063 14.2 11.9Average 19.5 21.42000-01 203571 4015 27.6 -1.22001-02 209018 4408 2.7 9.82002-03 255137 7014 22.1 59.12003-04 293367 8933 15.0 27.42004-05 375340 7914 27.9 -11.42005-06 456418 9656 21.6 22.02006-07 571779 11798 25.3 22.22007-08 655864 11267 14.7 -4.52008-09 840755 13888 28.2 23.32009-10 845534 15972 0.6 15.0Average 18.6 16.2Source: Govt. of Punjab, Statistical Abstract of Punjab, (various Issues)Table No: 1 reveals the exports from Punjab during 1990-2010. Period 1990-2010 has been divided into two decades i.e the first decade (1990-2000) and second decade(2000-2010). However the average of annual growth rate of exports in first decade was 21.4per cent, which decreased to 16.2 per cent in the second decade. It clearly shows decrease inthe annual compound growth rate exports from Punjab. It substantiates the fact that exportsfrom Punjab were declined during the second decade. On the whole, it can be said that thegrowth of exports from Punjab was not good. There are many factors responsible for this.2. TIME SERIES MODELING USING ARIMA MODELSThese are special type of regression model where dependent variable is considered tobe stationary and independent variable is lags of dependent variable and lags of errors. AnARIMA process is a combination of an Auto regressive and a Moving Average Process. Boxand Jenkins (1976) first introduced ARIMA models. A time series can follow an ARIMAprocess only when it is stationary. A time series is said to be stationary only when it exhibitsmean reversion around a constant long run mean, has a finite variance and decreasingcorrelogram as lag length increases. Stationarity is important because if the series is non-stationary then all the typical results of the classical regression analysis are not valid.
International Journal of Management (IJM), ISSN 0976 – 6502(Print), ISSN 0976 –6510(Online), Volume 4, Issue 2, March- April (2013)2472.1 Autoregressive ModelAn autoregressive model of order p is represented as:tptpttt uYYYY +++= −−− φφφ .....2211 ----------------------------- (1)Where, 1<φ and ut is a gaussian (white noise) error term. For the AR (p) model to bestationary is that the summation of the p autoregressive coefficients should be less than 1:11<∑=piiφ ----------------------------------------- (2)If the observations are generated by an AR (p) process then the theoretical partialautocorrelations will be high and significant for up to p lags and zero for lags beyond p. Thisrule is generally utilized to define which process the series is following and is incorporated inthe ARIMA model.2.2 Moving Average ModelA moving average model of order q can be written asqtqtttt uuuuY −−− ++++= θθθ ...2211 -------------------- (3)Moving Average MA (q) process is an average of q stationary white noise process, hence it isalways stationary as long as q has a finite value. A time series is said to be invertible if it canbe represented bya finite order MA or convergent autoregressive process. Invertiblity is animportant property for identifying the order of MA process using Autocorrelation and PartialAuto Correlation Function as in this case it is assumed that tY sequence is well approximatedby auto regressive model. An MA(1) process can be inverted to an infinite order AR processwith geometrically declining weights if the necessary condition 1<θ is met. The mean ofthe MA process will be clearly equal to zero as it is the mean of white noise terms. For a MA(q) model correlogram (ACF) is expected to have q spikes for k = 0 and then go downimmediately. Auto covariance of a MA process is equal to zero.2.3 ARMA ModelsThese models are combinations to two processes and usually represented byARMA(p,q). The general form of ARMA (p,q) models is represented by :qtqtttptptttuuuuYYYY−−−−−−++++++++=θθθφφφ......22112211---------------------------------------- (4)The equation can be rewritten as:jtqijtpiitit uuYY −==− ∑∑ ++=11θφ --------------------------------- (5)For stationarity of ARMA process only AR part of the model need to be stationary as MApart by default is stationary.
International Journal of Management (IJM), ISSN 0976 – 6502(Print), ISSN 0976 –6510(Online), Volume 4, Issue 2, March- April (2013)2482.4 Integrated processes and the ARIMA modelsARMA models can only be applied on a stationary time series. If a series is notstationary then stationarity need to be induced into it by differencing it such that differencedtime series tY∆ is represented by:1−−=∆ ttt YYY -------------------------------------- (6)Generally time series need to be difference atleast once to make them stationary. Afterdifferencing once the series hence obtained is said to integrated to order one and denoted byI(1). Hence a series which needs to be differenced d times to make it stationary and thenfollows ARMA(p,q) model then the series is said to be following ARIMA(p,d,q) process.3. METHODOLOGYMoving Average structure as explained by ARIMA models. Punjab’s export ofindustrial goods will be modeled as ARIMA process. Identification of the values ofparameters p,d and q is done on basis of ACF and PACF analysis. Data analyzed in the studyis yearly exports from Punjab in Crore Rupees from 1991-1992 till 2009-2010. Data from1990-91 till 2009-10 is used to train the structural models while next 10 years data is used totest the accuracy of the model forecast. Table (1) describes the data used in the analysis. Firstand foremost step before fitting the model is making the time series stationary. If time seriesis not stationary then it has to be transformed to make it stationary. Generally time series isdifferenced to make it stationary. Plots of ACF and LBQ test statistics will be used to checkthe stationarity of the model.Table1.2 AUTO-ARIMA (Autoregressive Integrated Moving Average)ModelsAdjustedR-SquaredAkaikeInformationCriterion(AIC)SchwarzCriterion(SC)Durbin-WatsonStatistic(DW)NumberofIterationsModelRankP=1, D=0, Q=0 0.9457 15.7671 16.0771 2.4824 0 1P=2, D=0, Q=0 0.9408 16.6282 17.1100 2.2465 0 2P=0, D=0, Q=2 0.8423 17.6791 18.1285 0.3550 32 3P=2, D=2, Q=0 0.6337 16.4837 17.0035 1.6495 0 4P=0, D=0, Q=1 0.5715 18.7356 19.0351 0.5412 29 5P=0, D=2, Q=0 0.0000 17.5143 17.6748 2.8611 0 6P=0, D=1, Q=0 0.0000 15.8895 16.0445 1.9995 0 7P=2, D=1, Q=0 -0.0155 15.7450 16.2450 1.5883 0 8P=0, D=1, Q=1 -0.0532 15.8845 16.1944 1.8398 12 9P=1, D=1, Q=0 -0.0599 16.8016 17.1228 1.9645 0 10
International Journal of Management (IJM), ISSN 0976 – 6502(Print), ISSN 0976 –6510(Online), Volume 4, Issue 2, March- April (2013)249Table 1.3 Regression StatisticsR-Squared (Coefficientof Determination)0.9487 Akaike InformationCriterion (AIC)15.7671Adjusted R-Squared 0.9457 Schwarz Criterion(SC)16.0771Multiple R (MultipleCorrelation Coefficient)0.9740 Log Likelihood -149.79Standard Error of theEstimates (SEy)4512.76 Durbin-Watson (DW)Statistic2.4824Number ofObservations19 Number of Iterations 0Table 1.4 Regression ResultsIntercept AR(1)Coefficients 283.9372 1.0945Standard Error 414.6082 0.0617t-Statistic 0.6848 17.7309p-Value 0.5027 0.0000Lower 5% 1005.1924 1.2019Upper 95% -437.3180 0.9871Table 1.4 Analysis of VarianceSums ofSquaresMeanofSquaresF-Statisticp-ValueHypothesis TestRegression347764392.9347764392.9314.380.0000Critical F-statistic(99% confidencewith df of 1 and 17)8.3997Residual 18805041.671106178.92Critical F-statistic(95% confidencewith df of 1 and 17)4.4513Total 366569434.5Critical F-statistic(90% confidencewith df of 1 and 17)3.0262
International Journal of Management (IJM), ISSN 0976 – 6502(Print), ISSN 0976 –6510(Online), Volume 4, Issue 2, March- April (2013)252Fig 1.2 Comparison of actual and forecasted ExportsProjections have been made for the industrial exports of Punjab at current prices onthe basis of their actual performance during 1991-92 to 2009-10. Table 1.7 shows theseprojections. Punjab can export goods worth Rupees 43814 crore in 2020-21. Thus, based onPunjab’s actual exports, there exists a scope for her exports in future. Therefore, efforts at theinternational level are required to be made to increase the exports to earn a fair name forPunjab in the world trade.REFERENCES1. Statistical Abstract of Punjab, Government of Punjab, various issues.2. Economic Survey of Punjab, Government of Punjab, various issues.3. Economic Survey of India, Government of India, various issues.4. http://www.ibef.org/, accessed on 12thMay 20125. Nanda (1988),”Forecasting: Does the Box-Jenkins Method Work Better thanRegression?” Vikalpa, Vol. 13, No. 1, January-March 1988.6. www.rbi.org7. www.pbindustries.gov.in/