On the 3rd of July 2018, Dr Chuah Lay Lian from the Macroeconomics and Growth team of the World Bankโs Development Research Group hub in Kuala Lumpur presented on the topic of โResource Misallocation and Productivity Gaps in Malaysiaโ.
๐+919953056974 ๐young Delhi Escort service Pusa Road
ย
Malaysia's Productivity Gaps Due to Resource Misallocation
1. Resource Misallocation and Productivity
Gaps in Malaysia
Lay Lian Chuah, Norman V.Loayza and Ha Nguyen
June 7, 2018
2. Motivation
โข Overarching question for development: Why are some countriesโ productivity
high and some othersโ low?
โข A traditional candidate is technology.
- Policy implications include investing in innovation, FDI (technological
spillover), human capital etc
โข New way of thinking:
- Resource misallocation1 (Restuccia and Rogerson (2008) , Hsieh and
Klenow (2009)). Policy implications are very different.
โข Are markets generally good at allocating resources?
- Yes, within a distorted environment
Misallocation1
1See Banerjee and Duflo (2005) and Restuccia and Rogerson (2008)
3. Sources of Growth โ A Review
Misallocation2
โข Capital
โข Labour
Increasing inputs
โข Allocational (resources used in most
productive activities)
โข Technical (more output from same inputs)
Increase efficiency
โข New technologies
โข Innovation
Increasing
productivity
4. Decomposing Sources of Growth
Misallocation3
Production Frontier 2
Production Frontier 1
Y
X
X1 X2
Y1
Y2
E2
E1
Increase in
technology
Y2-Y1 = Growth in output
E1 - E2=Growth from increasing efficiency
Z = Growth from increasing inputs from X1 to X2
T2-T1=Growth from technological change
Z
T1
Y1*
Y2*
T2
Y1โ
Increase in inputs
Y2-Y1=E1-E2+Z+T2-T1
5. Growth Accounting Decomposition and Total Factor Productivity
(TFP)
Growth Accounting
Y=Akฮฑ(hL)1-ฮฑ
where Y is GDP
A is total factor productivity (TFP)
K is physical capital stock
ฮฑ is the output elasticity of physical capital
L is labour (number of workers)
h is the average level of human capital
Total Factor Productivity (TFP)
โข TFP is the portion of output not explained by the amount of inputs used in production
โข Differences are due to differences in technology or efficiency.
Misallocation4
Firm Workers + Machines = Output
Firm A 10 5 25
Firm B 10 5 35
6. Resource Misallocation โ Distortions in Capital Market
Misallocation5
R* R*
Firm 1: Low Productivity Firm 2: High Productivity
Marginalrevenueproduct/
rentalrateofcapital
capital capital
R1
K1*
MRP
K2*
R2
K2
mK1
m
capital
capital
MRP
MRP
Marginalrevenueproduct/
rentalrateofcapital
Marginalrevenueproduct/
rentalrateofcapital
Marginalrevenueproduct/
rentalrateofcapital
๐ ๐
๐ ๐
๐ ๐
๐
๐ ๐
๐
7. Optimal Resource Allocation: Equalising Marginal Revenue Product
โข An efficient economy is one where most firms in the same narrow industry have
the same marginal revenue product:
MRP1 = MRP2 = MRP
โข An economy that's inefficient will be characterised by a large variance or kind of
wide distribution of marginal revenue products in the same industry
MRP1 โ MRP2 โ MRP
โข Large potential gains to be made from reallocating resources more efficiently
Misallocation6
8. Misallocation Inferred from Distributions of โRevenue Productivityโ
Misallocation7
โข Wider dispersions for India
and China compared to the
US. This suggest:
โข Larger variances in โrevenue
productivityโ for India and
China.
- Less efficient allocation
of resources.
- Larger distortions in
Indian and Chinese
economies.
More firms in the
left hand tail
More firms in
the left hand
tail
Source: Hsieh & Klenow (2009)
Distortions in allocation show up in the
variance of โrevenue productivityโ
across firms
๐ต๐๐๐: ๐๐น๐๐ ๐ ๐ โ (๐๐ ๐๐พ๐ ๐) ๐ผ๐
(๐๐๐ ๐ฟ ๐ ๐)1โ๐ผ๐
9. Why Do We Observe Misallocation?
Why do some firms have more inputs than other firms โ even when they have the
same technology?
Firms will hire labour and use capital until MPL = w and MRK = r. If that does not
happen (causing resource misallocation), it is possible that firms face different
prices.
How is that so?
โข Government policies ๏จ subsidies; taxes on inputs or outputs based on certain
characteristics.
o Discretionary government provisions eg. subsidized access to land or capital
for some firms; tax breaks or low interest rate loans
o Unfair bidding practices for government contracts, preferential treatment
(โcrony capitalismโ)
โข Firm characteristics
o Location of firms which adds more to cost
o Firm size:
- Differential access to finance and labor market
- Differential pricing power
Misallocation8
10. Resource Misallocation and Productivity Gaps in Malaysia
(i) Empirical Framework - Hsieh & Klenow (2009)
โข Main assumptions
โข Empirical model โ key equations
(ii) Data
โข Description of firms in the census data
(iii) Results
โข The degree of misallocation in Malaysian manufacturing sector
โข Comparison of misallocation across selected countries
(iv) Policy Implications & Conclusion
11. Questions
โข How important is misallocation as a source of productivity differences in
Malaysia? What is the extent of Malaysiaโs productivity gaps compare to its
peers? Which sectors experienced the biggest distortions?
โข If there is misallocation, what are the potential sources causing misallocation?
โข What is the direct cost of misallocation on economic growth?
โข What are the policy implications?
Misallocation10
12. Summary of Findings
โข Productivity gains arising from a reallocation of resources is estimated to be
95% between 2010. Moving towards the US level of efficiency in 1997 shows
productivity gains of about 36%.
- Productivity gains of 60-95%, in 2000, 2005 and 2010 โ suggest the โcatch-
upโ process remains a challenge.
โข By sector, misallocation seems more severe in selected resource-based
industries (food and beverage and petroleum and chemical industries) and less
so, in the textile, machinery, electrical and electronics and computers industries.
โข Both distortions in the capital and output markets are important sources of
productivity losses.
โข Productivity gains if realised, adds to +0.4 to +1.3 ppt to overall growth.
Misallocation11
13. An Overview of Hsieh and Klenow (2009)
Why is aggregate China and Indiaโs TFPs are lower than US?
- Traditional explanations focus on barriers to technology diffusion
- Misallocation explanation focuses on inefficient use of resources
(licensing regulations, size-dependent policies, SOEs)
Main findings
- Larger gaps in China and India than US
- Can account for about half of aggregate TFP differences
- Shrinking gaps in China but not India
- Large plants have large marginal products in China and India
Misallocation12
14. Hsieh and Klenow (2009) Empirical Framework
Assumptions:
CES demand function, monopolistic competition and Cobb Douglas production
function
โข Absence of distortion, firms in the same sector have the same โrevenue
productivityโ.
โข Differences in โrevenue productivityโ across firms must be due only to
distortions.
Misallocation13
15. Hsieh and Klenow (2009): Empirical Framework (cont.)
Each firm in the sector s has a production function:
๐๐ ๐ = ๐ด ๐ ๐ ๐พ๐ ๐
๐ผ๐
๐ฟ ๐ ๐
1โ๐ผ๐
(1)
Each establishment maximises current profits:
๐ ๐ ๐ = 1 โ ๐ ๐๐ ๐ ๐๐ ๐ ๐๐ ๐ โ ๐ค๐ฟ ๐ ๐ โ 1 + ๐ ๐พ๐ ๐ ๐ ๐พ๐ ๐ (2)
where ๐๐ ๐ ๐๐ ๐ is the firm's value added, w and R are the cost of one unit of labour
and capital respectively. ๐ ๐๐ ๐ denotes firm-specific output distortions and ๐ ๐พ๐ ๐ is
"capital" distortions.
Misallocation14
FOC:
๐พ ๐ ๐
๐ฟ ๐ ๐
=
1
๐
๐ผ ๐
1โ๐ผ ๐
๐ค ๐ ๐
1+๐ ๐๐ ๐
(3)
As a result of the wedges, firms produce different amounts than what would have been
dictated by their different capital-labour ratios.
16. Differences between physical TFP (TFPQ) and revenue TFP (TFPR)
15 Misallocation
Hsieh and Klenow (2009) define their productivity measures as follows:
๐๐น๐๐๐ ๐ = ๐ด ๐ ๐ =
๐ ๐ ๐
๐พ๐ ๐
๐ผ ๐ ๐ฟ ๐ ๐
1โ๐ผ ๐
(4)
๐๐น๐๐ ๐ ๐ = ๐๐ ๐ ๐ด ๐ ๐ =
๐ ๐ ๐ ๐ ๐ ๐
๐พ๐ ๐
๐ผ ๐ฟ ๐ ๐
1โ๐ผ ๐
(5)
โข TFPQ captures โphysical productivity" while TFPR measures โrevenue
productivity".
โข In the framework, if there are no distortions, TFPR does not vary across
plants (firms) within an industry:
๐๐ ๐ โ ๐ด ๐ ๐ = ๐๐ ๐ โ ๐ด ๐ ๐ = โฏ (6)
- More capital and labor should be allocated to plants with higher TFPQ up to the point
where their higher output results in a lower price. Hence their TFPRs equal to TFPRs of
smaller plants.
โข In the framework: ๐๐น๐๐ ๐ ๐ โ (๐๐ ๐๐พ๐ ๐) ๐ผ๐
(๐๐๐ ๐ฟ ๐ ๐)1โ๐ผ๐
17. Firm-specific TFPR with Distortions
โข For each firm, TFPR can be calculated as:
๐๐น๐๐ ๐ ๐ =
๐
๐โ1
๐
๐ผ ๐
๐ผ ๐ ๐ค
1โ๐ผ ๐
1โ๐ผ ๐ (1+๐ ๐พ๐ ๐) ๐ผ ๐
1โ๐ ๐๐ ๐
(7)
where ๐ ๐๐ ๐ is output โwedgeโ ; ๐ ๐พ๐ ๐ is capital โwedgeโ.
- If ๐ ๐พ๐ ๐=0 and ๐ ๐๐ ๐=0, TFPR is equalised for all firms in sector s
๐๐น๐๐ ๐ ๐ =
๐
๐ โ 1
๐
๐ผ ๐
๐ผ ๐
๐ค
1 โ ๐ผ ๐
1โ๐ผ ๐
- If ๐ ๐พ๐ ๐>0 and ๐ ๐๐ ๐>0, TFPR is larger (i.e. the firm is โtaxedโ more)
Misallocation16
MC
18. TFP Loss from Resource Misallocation
๐ ๐๐๐ก๐ข๐๐
๐ ๐๐๐
= ฯ ๐ =1
๐ ฯ๐=1
๐๐ ๐ด ๐ ๐
๐ด ๐
๐๐น๐๐ ๐
๐๐น๐๐ ๐ ๐
๐โ1
๐ ๐
๐โ1
(9)
If ๐ ๐พ๐ ๐=0 and ๐ ๐๐ ๐=0, ๐๐น๐๐ ๐ =๐๐น๐๐ ๐ ๐. Hence, ๐๐๐๐ก๐ข๐๐= ๐๐๐๐: no TFP loss from
misallocation:
Misallocation17
19. Data Treatment
Malaysia
โข We use census data in 2000, 2005 and 2010 to assess the changes in productivity
gains from better allocation of resources across time.
- Census of manufacturing firms
- Approximately 17,500 firms in 2000; 27,300 in 2005; and 33,800 in 2010
- 55 industries (3 digit MSIC codes) โ before trimming
- Truncate data by setting >10 employees as a threshold
โข Other data treatment issues
- Labour compensation (wages+ benefits), employment, value-added
- Capital stock = constructed from investment
- Drop firms with no capital stock or labour compensation data
- Drop industries which are represented by less than 9 firms
- Trim 1% tails to get rid of outliers in the sample
Firm Level Productivity Study18
20. Distribution of firms by size
Misallocation19
.4
.3
.2
.1
0
Density
0 2 4 6 8 10
Log employment distribution
2010 2005
2000
.4
.3
.2
.1
0
Density
2 4 6 8 10
Log employment distribution
2010 2005
2000
Distributions of firms in the three census
periods
Distributions of firms at the threshold of
>10 employees
21. Construction of Capital Stock
Initial capital stock, ๐พ0 calculated as:
๐พ0 =
๐ผ ๐
(๐+๐)
(10)
Where
๐ผ0 =
๐๐ฃ๐๐๐๐๐ ๐๐๐ฃ๐๐ ๐ก๐๐๐๐ก
๐๐ฃ๐๐๐๐๐ ๐ฃ๐๐๐ข๐ ๐๐๐๐๐
x value added at time t
๐ is the depreciation rate assumed to be 5%
๐ the growth rate of investment which is calculated from the data to be zero, in
other words, investment rate is constant across periods within a firm.
๐พ๐ก+5 = ๐พ๐กโ4 1 โ ๐ฟ + ๐ผ๐ก+5 (11)
โข The initial capital stock, ๐พ0 will be at the period which positive investment is
observed for a firm and therefore, the capital stock for the other periods will
either be iterated forward or backwards.
Equation (12) iterated backwards through repeated substitution of ๐พ๐กโ4
๐พ๐ก+5 = ๐พ0(1 โ ๐ฟ)5
+ าง๐ผ (1 + ๐)(1 โ ๐ฟ)4
+ โฏ (12)
Misallocation20
22. Measurement/Calibration
Assigned parameters
โข R=0.1 real rate 5%+ depreciation rate 5%
โข ฯ=3 elasticity of substitution across producers in the same industry
โข 1-ฮฑ labour share in the corresponding US industry (*scaled up)
Inferred distortions
โข 1+๐ ๐พ๐ ๐=
๐ผ ๐
1โ๐ผ ๐
๐ค๐ฟ ๐ ๐
๐ ๐พ ๐ ๐
capital wedge
โข 1-๐ ๐๐ ๐ =
๐
๐โ1
1
(1โ๐ผ ๐ )
๐ค๐ฟ ๐ ๐
๐ ๐ ๐ ๐ ๐ ๐
output wedge
Misallocation21
23. A Large Proportion of Firms are in Non Resource Based Industries
and in Operations for > 6 years.
Misallocation22
Textile and apparel accounts for the largest
proportion of the firms (44%); followed by basic
metal and fabricated metal (17%).
The majority of firms are in operations for 6-15
years (38%); old firms accounted for 33%.
17.4
44.3
4.7
4.4
17.0
1.9
0.1
10.2
F&B Textile Wood
Petchem Metal M&E&C
Transport equip Others
Percent of firms by Industry, 2010* Percent of firms by age, 2010*
โค5 years,
29.1
โฅ6 to โค15 years,
38.4
โฅ16 to โค50
years, 31.9
>50 years,
0.7
โค5 years โฅ6 to โค15 years โฅ16 to โค50 years >50 years
*Note: applying >10 employees threshold
24. The Hypothetical Productivity Gains from Better Reallocation are
Estimated to be about 94%
23 Misallocation
๏ท Productivity gains from the better allocation remained between 60-95% in 2000 -
2010. A declining productivity gains would suggest an improvement in the
distribution of resources over time.
๏ท Distortions to capital improved while distortions to output worsen.
๏ท Moving to the efficiency levels of the US 1997 was about 36% in 2010.
Year No. of
obs.
TFPQ TFPR Wedge P.H
Gain 1
P.H.
Gain 2
SD 75โ25 90โ10 SD 75โ25 90โ10 Capital
SD
Output
SD
2000 7,698 1.0537 1.5096 2.7399 0.7564 0.9414 1.8666 1.6637 0.5713 61.22 12.82
2005 9,511 1.0738 1.4774 2.7769 0.7813 0.9685 1.9440 1.6427 0.5681 80.81 26.53
2010 7,412 1.1385 1.5394 2.9679 0.7843 1.0017 1.9591 1.6220 0.6002 93.95 35.72
25. TFPQ and TFPR Dispersions
24 Misallocation
๏ท TFPQ dispersions are getting slightly wider in 2010, less productive firms are getting even less
productive (left tail). The SD for TFPQ increased slightly to 1.14 in 2010, from 1.07 in 2005.
๏ท TFPR dispersions also show a similar pattern, slightly wider dispersions in the left tail.
0
.2
.4
.6
Density
-3 -2 -1 0 1 2 3
TFPR distribution
2010 2005
2000
0
.1
.2
.3
.4
Density
-5 -4 -3 -2 -1 0 1 2 3
TFPQ distribution
2010 2005
2000
26. Malaysia has Similar Levels of Resource Misallocation as China
25 Misallocation
Source: Authorsโ calculations (2017), Cirera et al (2017), Nguyen et al (2016) and Hsieh and Klenow( 2009)
% TFP gains % TFP gains
TFP Gains from Moving to US Efficiency LevelsTFP Gains from Efficient Allocation of Resources
42.9
78.0
86.6
93.9
127.5
162.6
0
20
40
60
80
100
120
140
160
180
US
1997
TUR
2014
CHN
2005
MYS
2010
IND
1994
KEN
2010
24.5
30.5
35.7
59.2
83.8
0
10
20
30
40
50
60
70
80
90
TUR
2014
CHN
2005
MYS
2014
IND
1994
KEN 2010
27. Malaysiaโs Estimated TFPQ and TFPR Dispersions are Comparable to
Turkey and India
26 Misallocation
Source: Authorsโ calculations; Nguyen, Taskin, and Yilmaz (2016); Hsieh and Klenow (2009).
Note: Number of firms: Malaysia (7,412); Turkey (22,148); China (211,304); India (41,006).
1.1
0.8
1.1
0.8
1.0
0.6
1.2
0.7
1.5
1.0
1.5
1.0
1.3
0.8
1.6
0.8
3.0
2.0
2.9
1.9
2.4
1.6
3.1
1.6
T FP Q T FP R T FP Q T FP R T FP Q T FP R T FP Q T FP R
M A LA Y S I A ( 2 0 1 0 ) T U R K E Y ( 2 0 1 4 ) C H I N A ( 2 0 0 5 ) I N D I A ( 1 9 9 4 )
N = 7 4 1 2 N = 2 2 1 4 8 N = 2 1 1 3 0 4 N = 4 1 0 0 6
SD 75-25 90-10
28. Higher Productivity Gains from Better Reallocation are Estimated for
the Resource-Based Industries
27 Misallocation
% TFP gains
Source: Authorsโ calculation
0
50
100
150
200
250
Average TFP gains
F&B Petchem Metal M&E&C Wood Transport equip Textile Manufacturing
30. Distortions in both Markets Appear to Affect Productivity
29 Misallocation
Productivity and Factor Market
Distortions
Productivity and Output Market
Distortions
Source: Authorsโ calculation
31. Eliminating Distortions will Hypothetically Increase Growth by +0.4
to+1.3ppt
30 Misallocation
Source: Authorsโ calculations and Eleventh Malaysia Plan (11MP).
Note: Top row = percentage point contribution to growth. Bottom row in brackets = percent contribution. 11MP = Eleventh Malaysia Plan; MA(3) = moving average
3; S1, S2, and S3 = Scenarios 1, 2, and 3, respectively; TFP = total factor productivity.
The assumption is that the manufacturing sectorโs share of approximately 25% of the economy is also reflected in its share contribution to the aggregate TFP.
Calculated from the baseline growth of 5.8%.
Calculated from the baseline growth of 5%.
Baseline projections Baseline using 11MP Baseline using MA(3)
11 MP
(2016โ20)
ARMA(p,q)
(2016โ20)
S1 S2 S3 S1 S2 S3
Capital 2.6
(44.8)
2.2
(44.8)
2.6
(41.6)
2.6
(39.0)
2.6
(36.8)
2.2
(41.0)
2.2
(38.1)
2.2
(35.7)
Labor 0.9
(15.5)
0.8
(15.5)
0.9
(14.4)
0.9
(13.5)
0.9
(12.7)
0.8
(14.2)
0.8
(13.2)
0.8
(12.4)
Total factor
productivity
2.3
(39.7)
2.0
(39.7)
2.7
(44.0)
3.2
(47.5)
3.6
(50.4)
2.4
(44.8)
2.9
(48.7)
3.3
(51.9)
Gross domestic
product (%)
5.8 5.0 6.2 6.7 7.1 5.5 5.9 6.3
Additional
growth
+0.4a +0.9a
+1.3a +0.5b +0.9b +1.3b
32. Conclusion
31 Firm Level Productivity Study
โข Malaysia shows relatively high dispersion in productivity of firms in the same
narrowly defined industry. This can be explained by misallocation of resources
across firms.
โข Our estimates show that:
- Reducing distortions to the US level of efficiency could result in a
hypothetical aggregate productivity gain of 36%.
- A hypothetical productivity gain of 94% is estimated if distortions are
completely removed.
- Productivity gains are quite comparable with other upper-middle income
countries.
- Food and beverage firms operate in the most distorted market conditions.
- Both the input and output market distortions constrain productivity
โข Broadly, there should be efforts to increase domestic competition and reduce
non-performance-based incentives.
โข But, further investigation is needed to pin the sources of distortions.
โข Results highlight the need to maintain the efforts to further ingrain productivity at
the core of policy making in Malaysia.