chapter 2: reallocation and productivity
DESCRIPTION
Chapter 2: Reallocation and Productivity. Outline IPES. I . Setting the stage Chapter 1: Productivity level and trends Chapter 2: Productivity and Reallocation Chapter 3: Within firm productivity growth. Outline of IPES. II. Policy levers for higher productivity - PowerPoint PPT PresentationTRANSCRIPT
Chapter 2: Reallocation and Productivity
Outline IPES
I. Setting the stage– Chapter 1: Productivity level and trends– Chapter 2: Productivity and
Reallocation– Chapter 3: Within firm productivity growth
Outline of IPES
II. Policy levers for higher productivity– Social policies (taxes on social security) – Taxes & uneven enforcement – Trade, FDI and transportation costs – Small firm policies– Barriers to the entry and exit of firms– Access to credit & capital– Innovation policies– Industrial and cluster-based policies
Outline of IPES
III- Going forward– Understanding the political economy of
productivity policies– Unlocking productivity in LAC
Outline
• Hypothesis• Main questions• Reallocation across sectors• Reallocation across firms• Reallocation across formal and informal
firms
Hypothesis
• Lack of dynamism in, and barriers to the reallocation of resources in the economy across sectors and firms is a key factor explaining the low productivity levels and growth in the region.
Main questions
• How much room there is to improve productivity by reallocating existing factors?– Across sectors– Across firms– Across formal and informal firms
• How much entry/exit of firms contribute to productivity growth?
• What drives low reallocation?
Heterogeneity
• Productivity in the aggregate is the weighted productivity across activities.
• High heterogeneity; Productivity in manufacturing and even business sector accounts for fraction of TPF
• Gains through reallocation are possible if heterogeneity in marginal products
There is seemingly a big share of employment in very low productivity activities
We plan to do this for more countries (Brazil, Chile?)Ideally, we would like to estimate marginal product across activities
Reallocation across sectors
Relative labor productivity, Manufacturing=100
38
59
54
110
70
0 50 100 150 200 250 300
Agriculture, hunting, forestry,
fishing
Mining and Utilities
Construction
Wholesale, retail trade,
restaurants and hotels
Transport, storage and
communication
Other Services
824
Average for ARG, BRA, CHL, COL, MEX, PER, VENOutput from UN, Employment from LAC Statistical offices
• Across history, large gains in TFP from getting people out of agriculture and into manufacturing
• Latin America has still a relative large share of agriculture. It also has a relative low share of manufacturing compared to other economies.
Sector distribution of output across countries
Agriculture
0 20 40 60 80
Trini. and Toba.
Panama
Nicaragua
Jamaica
Honduras
Guatemala
El Salvador
Dominican Rep.
Cuba
Costa Rica
Barbados
Uruguay
Paraguay
Ecuador
Bolivia
Venezuela
Peru
Mexico
Colombia
Chile
Brazil
Argentina
Mining and Utilities
0 20 40 60 80
Manufacturing
0 20 40 60 80
Services
0 20 40 60 80
Trini. and Toba.
Panama
Nicaragua
Jamaica
Honduras
Guatemala
El Salvador
Dominican Rep.
Cuba
Costa Rica
Barbados
Uruguay
Paraguay
Ecuador
Bolivia
Venezuela
Peru
Mexico
Colombia
Chile
Brazil
Argentina
Big Countries
South Am. -Small-
Average OECD 24Central Am.
and the Caribbean
Source: UN
Whereas many other developing countries have larger shares of manufacturing
Agriculture
0 20 40 60 80
Thailand
Singapore
Malaysia
Korea
China
Turkey
South Africa
Morocco
Mauritius
Malta
Egypt
Cyprus
Bahrain
Mining and Utilities
0 20 40 60 80
Manufacturing
0 20 40 60 80
Services
0 20 40 60 80
Thailand
Singapore
Malaysia
Korea
China
Turkey
South Africa
Morocco
Mauritius
Malta
Egypt
Cyprus
Bahrain
East Asia
Average OECD 24
Source: UN
I. Reallocation across sectors
0 0.05 0.1 0.15 0.2 0.25 0.3 0.35 0.4
OECD: English Speaking
OECD: Continental Europe
Africa - Middle Income
East Asia
LAC: Central America and theCaribbean
LAC: Small-South America
LAC: Big Countries
Employment GDP
Reallocation of employment and output Lilien Index. Early 1990s – Early 2000
Note: GDP series by economic activity at one digit level of disaggregation at constant 1990 prices in US Dollars from the United Nations Statistics
Division. Lilien Index measures the dispersion in the growth rate across industries and is defined as follows: 2/1
2)lnln(
stst
t
stt Xx
X
x
0 0.1 0.2 0.3 0.4 0.5 0.6
Trinidad and Tobago
Panama
Nicaragua
Jamaica
El Salvador
Dominican Republic
Costa Rica
Barbados
Uruguay
Bolivia
Venezuela
Peru
Mexico
Colombia
Chile
Brazil
Argentina
Employment GDP
Big Countries
Central Am. and the
Caribbean
Lilien index of reallocation. LAC
Sources: GDP: United Nations Statistics Division; EMP
Measuring gains from reallocation
s
ststkts
kitstits
ksttpPppP )(
• Issue: Marginal versus average products;
Total Growth
-0.03
-0.02
-0.01
0
0.01
0.02
0.03
0.04
0.05
Venezuela Nicaragua Peru Paraguay Brazil Mexico El Salv ador Colombia Panama Costa Rica Argentina
Dominican
Republic Chile
Between Term
-0.03
-0.02
-0.01
0
0.01
0.02
0.03
0.04
0.05
Cross Term
-0.03
-0.02
-0.01
0
0.01
0.02
0.03
0.04
0.05
Venezuela Nicaragua Peru Paraguay Brazil Mexico El Salv ador Colombia Panama Costa Rica Argentina Dominican
Republic
Chile
Within Term
-0.03
-0.02
-0.01
0
0.01
0.02
0.03
0.04
0.05
Total Growth
Counterfactual exercises
• How much productivity growth would we have seen in LAC had we observed the same sector reallocation than in East Asia? – If LAC had same sector allocation than EAST ASIA.
• Compare sector reallocation in LAC with other resource intensive nations (New Zealand, Australia, Finland).
(using GGDC total economy 10-sector database 1950-2005)
Contribution of each sector to productivity• Which sectors are in relative terms more
productive in each country?• Which sectors have contributed more to
growth in LAC?• Which sectors have contributed more to
productivity growth in EAST ASIA and how do they compare to LAC?
• How does the sector productivity evolution in LAC compare to other resource-rich economies?
Reallocation and reforms
• Data presented in Timmer and de Vries (2008) suggest that contribution of “between” term in post reform period negative, in all countries but Mexico.
• We would like to relate between and within growth to reforms
• Did the destruction of jobs in manufacturing and shift towards low productivity services contribute to this phenomenon?
cjtj
obetween
ct formsP Re1
cjtj
obetween
ringtcmanufactu formsP Re1
II. Reallocation across firms
• Decompositions of productivity growth
)()(
)(
ktkitExits
kitktitEntries
it
Continuersititkt
Continuerskititit
Continuerskit
t
PpPp
pPppP
Issues:•Marginal versus average products•Revenue versus quantity productivity
Firm Entry and Exit
• An important margin of how reallocation of resources affects productivity growth is through the entry/exit of firms and establishments
• Policies that prevent exit of unproductive firms (i.e. low competition, high transportation costs) or prevent entry of productive firms (high barriers of entry)
Labor Productivity Growth Decomposition – ManufacturingFive years differencing, real gross output -1990s
-.5
0.5
11
.5
ArgentinaChile
ColombiaUSA
Within BetweenCross EntryExit Firm Turnover
Labor Productivity Growth Decomposition – ManufacturingFive years differencing, real gross output -1990s
-.5
0.5
11
.5
ArgentinaChile
ColombiaUSA
Within BetweenCross EntryExit Firm Turnover
Very little importance of entry/exit in Argentina
Decompositions for LAC countries
• Will do with TFP decomposition• With a few exceptions only in manufacturing• But interpretation problems with this
methodology.– It says how much of the productivity growth
occurred because of the different dimensions.– It does not say how much room for
productivity gains of reallocation.
Other methodological issues
• What does it mean to transfer resources to more productive firms?
• What determines allocation of resources across firms in a world where firms are heterogeneous?
Hsieh-Klenow (2008) approach
• It quantifies the extent of resource misallocation across firms within sectors.
• Static model with heterogenous firms (Ai) and imperfect competition
• In equilibrium, pi is inversely related to Ai and piAi is constant for all firms. Firms hire L and K to the point where MPL =MRL. High Ai firms are larger.
Hsieh-Klenow (2008) approach (II)
• It allows for departures of optimal: firm specific taxes and subsidies imply that MPL≠MRL.
• Taxes are more distorting the more correlated they are with productivity: they make more productive firms smaller than otherwise.
• Reallocation gains occur when resources flow from low to high marginal product firms.
• Hsieh-Klenow 2008 find that if resources were reallocated as in US benchmark, productivity in manufacturing will increase:– INDIA: 40-60%– CHINA: 30-50% – LAC: ?
• Computations assume distribution of productivities across firms as given. Gains occur by reallocating L and K across these firms.
HK (2008) Methodology in LAC
• Through Research Network Project (& other sources) will obtain results for:– Argentina– Bolivia– Brazil– Chile– Colombia– Ecuador– Mexico– Uruguay– Venezuela
– El Salvador? – Peru?
Some preliminary results (1)
TFP Gains in Manufacturing
Years Relative to the US Full LiberalizationIndia 100-128%China 86-115%Argentina 1998-2002 35-45%Bolivia 1988-2001 31-94%Brazil 2000-2005 58‐ 65%Colombia 1982-1998 25-46% (1997) 79-109%Ecuador 1996-2005 252-585%Mexico 1994, 1999, 2004 140%Uruguay 1998-2005 52-63%China 30-50%India 40-60%
The data does not show gains in allocative efficiency
Colombia-TFP gains from Equalizing TFPR relative to 1997 U.S. Gains
0
5
10
15
20
25
30
35
40
45
50
82 83 84 85 86 87 88 89 90 91 92 93 94 95 96 97 98
The data does not show gains in allocative efficiency
Colombia-TFP gains from Equalizing TFPR relative to 1997 U.S. Gains
0
5
10
15
20
25
30
35
40
45
50
82 83 84 85 86 87 88 89 90 91 92 93 94 95 96 97 98
Ecuador: 1995-2005
Uruguay: 1998-2005
URUGUAY. TFP gains from equalizing TFPR within industries
Manufacturing Commerce Hotel &Restaurants
Transport &Communications
Real estate &renting
Education Health Others
1998 58,9% 215,4% 76,4% 288,1% 48,1% 14,3% 16,3% 42,1%1999 52,5% 204,9% 105,4% 199,6% 47,7% 17,9% 12,6% 33,5%2000 49,2% 204,5% 94,6% 162,8% 67,1% 13,6% 7,8% 9,5%2001 60,3% 233,9% 36,8% 207,6% 73,0% 19,3% 11,4% 41,2%2002 53,0% 321,5% 41,2% 222,8% 73,1% 20,3% 8,9% 34,7%2003 59,8% 243,2% 36,5% 114,4% 87,1% 28,0% 10,6% 62,3%2004 47,6% 258,1% 59,4% 123,6% 83,5% 19,4% 9,8% 55,5%2005 63,4% 266,4% 164,0% 167,0% 101,4% 20,7% 12,9% 12,1%Total 55,6% 243,5% 76,8% 185,7% 72,6% 19,2% 11,3% 36,4%
TFPR (Distortions) and SizeBrazil
Smaller firms have subsidies (they arelarger than they should be)
Smaller firms are constrained
In Ecuador..
Large firms shouldexpand moreActual versus Efficient size
More firms should expand
In Colombia
Many small should shrink
In Bolivia, many large firms are too large and many small are too small.
12% of largest Plans should shrink
10% of smallest plansShould grow by more
Than 200%
Correlation between distortions and productivityBrazil 2000-2005
More productive firms tend to be more distorted (they should be larger than they are)
Very preliminary results suggest that:
• There is a large correlation between size and distortions in Brazil, Ecuador, and Colombia (as in India). Many large firms are too small; many small firms are favored and therefore are too large given they inherent productivity;
• Things look quite different in Bolivia where many small firms look too small and distortions make large firms too large.
Some remarks
• Picking winners through industrial policy requires important “within” gains in productivity to make it worthwhile, particularly if picking firms of low productivity to begin with. – Low relative to sector average?– Low relative to world average?
• Can those “break even” gains in “within” productivity be computed?– Cost-benefit analysis should account for
‘between’ distortion as well.
III. Reallocation across formal and informal firms: the importance of high taxes and imperfect enforcement
• Are there too many small, informal, and unproductive firms in Latin America?
• How much of the distortions in productivity are explained by taxes+imperfect enforcement?
• Hsieh and Klenow application to Mexican Census Data in 1994, 1999, and 2004
• Gain from equalizing TFPR = 140%
Still more work to do:
• How much due to IMSS non‐compliance?• How much in other sectors?• How has this changed over time?• Why so many firms with low TFPQ?
Mexico
Informality and productivity
• Will compare productivity distributions of formal versus informal firms (Brazil, Chile?).
• Case studies and focus groups: Do formal and informal firms compete?-Supermarkets in Argentina-Lingerie industry in Colombia
Rest of the report: examine in detail reallocation (and within firm) productivity effects of:
• Social policies (taxes on social security) • Taxes & and uneven enforcement • Trade, fdi and transportation costs • Access to credit & capital• Small firm policies• Barriers to entry and exit of firms• Innovation policies• Industrial policies