The Age of

Productivity:
Transforming economies from
the bottom up.

Development in the Americas (DIA)

2010
The region suffers from chronic low growth

1.05
Evolution GDP p
per capita
p relative to US. 1960=1 1
1.00

0.95

0.90

0.85
ndex 1960=1

0.80

0.75
In

0.70 0.69

0.65

0.60
1966

1970

1976

1980

1984

1996

2000

2004
1962

1972

1978

1982

1988

1992

2002
1960

1964

1974

1986

1990

1994

2006
1968

1998
GDP pc LAC/ GDP pc US

Source: Authors’ calculations based on Heston, Summers and Aten (2006), World Bank (2008), and Barro and Lee (2000).

2
What is behind the region’s
g poor
p economic performance?
p

• The study breaks against the commonly held notion that the
main cause of the region’s
region s low growth is insufficient
investment

• The region suffers from chronic slow growth because it

suffers from low productivity growth

• The region could greatly accelerate its economic growth and

close the income per capita gap with policies that promote
better ways of using existing resources.
 Loss in relative productivity is the main cause of loss
in relative GDP p
per capita
p

1.20
Evolution of relative GDPpc,
p TFP, and Factor accumulation, vs US
1.10

1
1.00
Loss of factor
Loss of factor 
accumulation 0.91
0.90
ndex 1960=1

Loss of 
productivity
0.80
0 76
0.76
In

0.70 0.69

0.60
1966

1970

1976

1980

1984

1996

2000

2004
1962

1972

1978

1982

1988

1992

2002
1960

1964

1974

1986

1990

1994

2006
1968

1998
GDP pc LAC/ GDP pc US TFP LAC/ TFP US Factor Accumulation LAC vs. US

Source: Authors’ calculations based on Heston, Summers and Aten (2006), World Bank (2008), and Barro and Lee (2000) .

4
 Only in one country, Chile, productivity increased relative to
th U
the United
it d St
States
t (1960
(1960-2005)
2005)
China 219%
136%
Hungary 132%
115%
5
Sri Lanka 103%
103%
Thailand 86%
68%
Japan 55%
41%
Korea, Republic of 40%
36%
Papua New Guinea 36%
32%
Egypt 32%
31%
Italy 30%
25%
India 24%
21%
Chile 19%
19%
Nor a
Norway 17%
17% Note: Figures correspond to gains in productivity above
Lesotho 17%
16% and beyond gains of productivity in the United States.
United Kingdom 16% Source: Authors calculations
15%
Australia 12%
11%
Denmark 9.4%
6 6%
6.6%
Mali 6.4%

0% 50% 100% 150% 200% 250%

Percentage Gain Relative to the United States

Source: Authors’ calculations based on Heston, Summers and Aten (2006), World Bank (2008), and Barro and Lee (2000).
In the rest, productivity decreased relative to the United
St t (1960
States (1960-2005)…
2005)
Portugal, -0.6%
Panama, -2.0%
Indonesia, -2.1%
Brazil, -2.5%
Turkey, -2.5%
Canada, -5.7%
Zambia, -6.0%
Benin, -6.2%
South Africa, -6.5%
Dominican Republic, -8.2%
Fiji, -8.2%
Netherlands, -9.2%
Malawi, -11%
Ecuador, -12%
New Zealand, -13%
Uruguay, -14%
Mozambique, -16%
Bolivia, -17%
Colombia, -17%
Syria, -22%
Sierra Leone, -22%
Cameroon, -25%
Peru, -26%
Jamaica, -28%
Philippines, -28%
Kenya, -30%
Mexico, -31%
Uganda, -33%
Algeria, -35%
Argentina,
g -35%
Costa Rica, -36%
Paraguay, -37%
Nepal, -37%
Iran, -40%
El Salvador, -42%
Senegal, -45%
Venezuela, -47%
Honduras, -48%
Togo, -55%
Nicaragua, -63%
Niger, -63%
Jordan, -70%

-80% -70% -60% -50% -40% -30% -20% -10% 0%

Percentage Drop Relative to the United States

Source: Authors’ calculations based on Heston, Summers and Aten (2006), World Bank (2008) and Barro and Lee (2000).
What is the cost of low productivity?
• Productivity is closely linked to per capita income

• A typical Latin American country could have increased

income per capita by 54 percent since 1960 if its
productivity had grown like the rest of the world.

• Income per capita in this typical country would almost

double if its productivity were close to its full potential.

• As a result, low productivity has hampered efforts to lift

millions out of poverty
Need to understand what is behind productivity
and
dhhow it grows
• Productivityy as the main engine
g of g
growth;; but how does
aggregate productivity increase?

• It used to be considered a measure of technology with

technological growth credited as the main engine of growth.

• Studies either analyze aggregate or manufacturing firm

level data… unfortunately, manufacturing constitutes a
small share of economic activity.

-8-
What’s
What s new?

• In addition to improvements in firm level productivity, TFP reflects

inefficiencies about how markets work. (Banarjee and Dufflo, 2005;Hsieh and
Klenow, 2010)

– Improving the efficiency of markets is an important route towards

higher productivity and income growth for developing economies

• Productivity is given by the weighted average of productivity across all

firms and sectors.

– Look at what is happening at all sectors of the economy, not only

manufacturing;
– Look
L k att all
ll fi l llarger ones.
firms, nott only

-9-
 Where is the problem?

• Productivity growth is highest in agriculture, but

still below world average
g

• Productivity growth is seriously lagging in

manufacturing
a u actu g and
a d specially
spec a y in the
t e service
se ce sector.
secto
 Productivity growth is lagging in
manufacturing and more importantly in services
Average Annual Labor Productivity Growth
Agriculture,
g , Industry,
y, and Services,, 1951–2005

5.0% Latin America East Asia High Income Countries

3.8%
3.51% 3.6% 3.5%
3.2%
2.8%
2.5% 2.6% 2.4% 2.5%
2.2%
2.0%
1.8% 1.8%
1.3% 1.3% 1.4%

0.1%

-0.9%

-1.8%

1951─75 1975─90 1990─2005 1951--75 1975─90 1990─2005 1951─75 1975─90 1990─2005

Agriculture Industry Services

Source: Authors' calculations based on Timmer and de Vries (2007).
Productivity gaps are highest in the service
sector…
t

Labor Productivityy byy Sector typical

yp LAC countryy relative
to US=100
70
60
50
40
30
20
10
0

Agriculture

Industry

Services
2002 1973

Source: Authors’ elaboration based on Duarte and Restuccia (2009)

12
Yet, because services are 60% of employment,
they are increasingly dragging down aggregate
productivity…
1970 2005

18.4
38.1 38.8
Agriculture Agriculture
Industry Industry
Services 60.9 20.7 Services

23.2

Source: Authors
Authors' calculations based on Timmer and de Vries (2007)

13
In consequence, sector reallocation plays now a
minor
i role
l iin overallll productivity
d ti it growth….
th
Productivity Decomposition Across Periods in Latin America, 1950–2005
(
(annual
l growthth rates)
t )

1990 2005
1990─2005

1975─90

1950─75

-0.01 -0.005 0 0.005 0.01 0.015 0.02 0.025 0.03 0.035 0.04 0.045

Within Between Cross

Source: Authors' calculations based on Timmer and de Vries (2007)

Overall productivity growth would at least double if
productivity
d ti it ini services
i grew like
lik in
i E
Eastt A
Asia…
i

Annual Labor Productivity Growth (1990

(1990-2005)
2005)
Different Scenarios

3 5%
3.5%

3.0%

2.5%

2.0%

1.5%

1.0%

0.5%

0.0%

Actual If industry grew as in East Asia If services grew as in East Asia

Source: Authors’ calculations based on Timmer and de Vries (2007)

15
In sum…
• It has become commonplace to focus on boosting
exports as a strategy to improve competitiveness
and income levels.

• However, this study suggests that increasing

productivity
p oduct ty in the
t e vast
ast service
se ce sector
secto iss key
ey for
o
boosting economic growth and reducing poverty in
LAC in coming years.
Firm level analysis: The data

• Results of individual country studies based on establishment level data

produced by National Statistical offices.

• Only a subsample of countries in the region collects these types of data,

data
and of those, even a smaller number make these data available to
researchers.

• Despite such constraints, data for ten Latin American countries have
been assembled: Argentina, Bolivia, Brazil, Chile, Colombia, Ecuador, El
Salvador Mexico
Salvador, Mexico, Uruguay and Venezuela
Venezuela.

- 17 -
The data

• Coverage varies, but generally censuses of largest establishments with

a random sample of smaller ones.

g y, all sectors of economy.

• Mexico and Uruguay, y In the rest,, onlyy
manufacturing.

• Processed with the goal of ensuring cross

cross-country
country comparisons
comparisons. For
many applications restricted to establishments with 10 or more workers
(mostly registered)…

• Yet when available, we also document how measurement change when

adding the smaller establishments (registered and not registered)

- 18 -
Do countries lose on productivity due to:

• …too many unproductive firms? Not enough

productive firms?
• Which ones?

• …poor resource allocation across firms, due to

market and/or government failures?
Productivity is higher in larger firms
Productivity (TFP) by firm size, relative to firms with 10–19 workers,
Manufacturing Firms

20-49 50-99 100-249

ge increase in Productivity

180

120

60
Percentag

0
Chile
Chil 2006 Uruguay
U 2005 Colombia,
C l bi El S
Salvador
l d Venezuela,
V l Bolivia
B li i 2000
1988 2005 2001

Source: Based on individual countries, firm-level manufacturing surveys.

 Extreme productivity heterogeneity within firms in
the same sector (4 digit), smallest least productive
Manufacturing Firms

US distribution
a. Mexico 2004
b. El Salvador 2005
.3
.4
.33

.2
.2

.1
.1

0
0

1/1024 1/256 1/16 1 4 16

1/256 1/32 1/4 1 4
Productivityy relative to the average
g sector. Average
g sector=1 Productivityy relative to the average
g sector. Average
g sector=1
Firm Size
All firms Firm Size [1-9] All firms [0-10] [11-100] [101-500] [+500] Aprox USA 1997 Dist
[10+] Aprox USA 1997 Dist

Source: Authors’ elaboration based on Economic Census data

21
Mexico: the Census misses 13.6 million workers (33% LF)
1% in 0 to2   6% in 0 to 2
12% 
Own account 2% in 3 a 5
13.6 millons of 3% in 3 to 5 
people
p p 2% in 6 to 10 
1% in 6 to 10
15%  5% in 11 to50
in 2 to 5
1% in 11 to 50 

18% in 51 or more
3%  
3%
in 5 or more
BLUE + RED= urban
3% in 5 or more
GREEN= rural

16% in rural activities 1% in 51 or more
10%  Public
1% in rural activity  government
. employment
DARK Informal
LIGHT Formal  Source: Economic Census, 2003, IMSS and ENE 22
Two counterfactuals:

1. Changing the proportion of small/medium/large

firms to make it coincide with those in US.

2. Increasing productivity at the lower end of the

distribution
Increasing share of medium and large firms,
i
increases productivity
d ti it iin..

• 90% in El Salvador

• 120% in
i Mexico
M i
Instead, increasing productivity at the lower end
has little effect on overall productivity
Effect on Aggregate Productivity
of Raising Productivity in the Least Productive
ms
oductive firm

60
om raising
increase in aggregate

El Salvador 2005
50
Mexico 2004
east to level X

40
productivity resulting fro
ductivity in tthe least pro

30

20
at le
Percentage

10

0
0 1/10 1/5 3/10 2/5 1/2 3/5 7/10 4/5 9/10 1
P

prod
p

Mi i
Minimum productivity
d ti it level
l l X (as
( fraction
f ti off sector
t average. Sector
S t
average=1)
Source: El Salvador: Atal, Busso, and Cisneros (2009), Mexico: INEGI (2004, 2005).
Note: This figure plots the resulting increase in aggregate productivity from raising productivity of all firms with productivity below X to a

- 25 -
Substantial losses due to misallocation..
Cost of Misallocation, Manufacturing sector
10 or more workers
Mexico (2004) 95.0

China (1998-2005) 86.6

Venezuela (1995-2001) 64.7

Bolivia (1988-2001) 60.6

El Salvador (2005) 60.6

Uruguay (1997-2005) 60.2

Argentina (1997-2002)
(1997 2002) 60 0
60.0

Ecuador (1995-2005) 57.6

El Salvador (2004) 56.7

Chile (1996-2006) 53.8

Colombia (1982-1998) 50.5

US (1977-1997) 42.9

0 20 40 60 80 100

- 26 -
…which seem to be much higher outside manufacturing

% in TFP reallocating factors, México

267%
Comerce

246%
Services

183%
Aggregate

95%
Manufacturing

0% 50% 100% 150% 200% 250% 300%

Source: Hsieh y Klenow, 2009 and Economic Census, 2004

- 27 -
What drives misallocation?

• Credit market constraints: Productive firms cannot expand due to credit

constraints; MRP to decline with size.

• Tax heterogeneity
g y: Productive firms are taxed while unproductive
p firms
have lower taxes due to informality or special tax regimes. MRP to
increase with size.

• Social security evasion: Productive firms pay SS charges, less

productive firms evade them. MRP to increase with size.

- 28 -
 Marginal Product of Factors by Firms’ Size
duct

Manufacturingg Firms: 10 or more workers

arginal prod

140
120
100
on to the Ma
Of firms 10--19 workerss

80
60
40
20
uct in relatio

0
-20
-40
-60
60
O
Marrginal produ

-80
Chile Uruguay Venezuela Bolivia Ecuador Argentina Colombia El México
2006 2005 2001 2000 2005 promedio promedio Salvador 2004
1997-02 1982-98 2004
20-49 50-99 100-249 250-499

Fuentes: Workers
Fuentes:  Workers’ elaboration based on individual country studies. 
elaboration based on individual country studies

- 29 -
 Marginal Product of Factors by Firms’ Size
duct

Manufacturingg Firms: 10 or more workers

arginal prod

140
120
100
on to the Ma
Of firms 10--19 workerss

80
60
40
20
uct in relatio

0
-20
-40
-60
60
O
Marrginal produ

-80
Chile Uruguay Venezuela Bolivia Ecuador Argentina Colombia El México
2006 2005 2001 2000 2005 promedio promedio Salvador 2004
1997-02 1982-98 2004
20-49 50-99 100-249 250-499

Fuentes: Workers
Fuentes:  Workers’ elaboration based on individual country studies. 
elaboration based on individual country studies

- 30 -
 Marginal Product of Factors by Firms’ Size
duct

Manufacturingg Firms: 10 or more workers

arginal prod

140
120
100
on to the Ma
Of firms 10--19 workerss

80
60
40
20
uct in relatio

0
-20
-40
-60
60
O
Marrginal produ

-80
Chile Uruguay Venezuela Bolivia Ecuador Argentina Colombia El México
2006 2005 2001 2000 2005 promedio promedio Salvador 2004
1997-02 1982-98 2004
20-49 50-99 100-249 250-499

Fuentes: Workers
Fuentes:  Workers’ elaboration based on individual country studies. 
elaboration based on individual country studies

- 31 -
In sum…
• Latin American Countries miss out on productivity due to a
low share of medium and high productivity (size) firms and
because higher resource misallocation than in developed
economies.

• Misallocation problems larger in the large service sector,

which could explain why the productivity gap is higher
there.

• R
Restt off th
the study
t d analyzes
l sources off llackk off medium
di and
d
high productivity firms and misallocation and finds that
they are likely to be the same…

- 32 -
The book identifies the following factors as
i
important
t t drivers
di off low
l and
d stagnant
t t productivity
d ti it

• High
g transportation
p costs
• Shallow credit markets
g and complex
• High p taxes to firms coupled
p with
widespread evasion.
• Well intended, but poorly designed social
policies.
li i
• Insufficiently informed policies for small firms
• Insufficient innovation,
innovation particularly at firms.
firms
• Coordination problems.
• A difficult political economy

33
Obstacles to development: How do they
operate?
t ?
• High transportation costs Prevent productive firms
• Low access to credit from growing
• High taxes coupled with evasion (missallocation) and/or
that less productive firms
• Insufficient innovation become more productive
• Coordination failures

• Simplified tax regimes for small firms

• Evasion of social security contributions and taxes
• Social programs for informal workers
• Micro and Small firm policies focused only on size
Foster the survival and
expansion of the least
dynamic and productive firms.

- 34 -
 How to improve productivity?
• Reducing transportation costs:
– Improving the efficiency of the transportation sector, the
regulatory framework of ports and airports, and improving
infrastructure.
• Deepening credit markets:
– Improving property registries and creditor rights protection, and
with better supervision and financial regulation.
• Improving tax regimes:
– Simplying tax regimes for all firms and reducing evasion.
• Improving social security:
– Looking for less distortive ways of providing universal access,
cutting links with type of employment and avoiding parallel
programs only for informal.

- 35 -
 How to improve productivity?

• Better Micro and Small Firms policy:

p y
– Aiming to help productive firms to grow, or low productivity firms
to become medium productivity firms and evaluating programs.
• Promoting
P ti innovation:
i ti
– Fostering innovation at firms and improving the links between
firms and research centers.

• With proactive but restrained industrial policy (PDPs)

- 36 -
 Conclusion
• Low growth caused by low productivity growth. Countries
need to make p
productivityy the cornerstone of their
economic policies in coming years

• Focusing on improving competitiveness of export sector is

not enough, higher productivity in non-tradable sector is
keyy to increasing
g the standards of living
g and eliminate
poverty in the region.

• A number of policies can unlock productivity growth by

promoting more innovation and a better allocation of
existing resources in the economy.