An Analysis of Recent

Productivity Trends in Chile

Chad Syverson
J. Baum Harris Professor of Economics
University of Chicago Booth School of Business

Presentation at “Besides Copper, What? Policies for Productivity

Growth in Chile”
April 23, 2014
 Annual Total Factor Productivity Growth (%)

0
2
3
5
6

1
4

-4
-3
-2
-1
1985
1986
1987
1988
1989
2.8%/yr

1990
1991
1992
1993
1994
1995
1996
1997
1998
1999
2000
2001
2002
2003
2004
What This Is All About

2005
2006
-0.2%/yr

2007
2008
2009
2010
2011
2012
 What This Is All About
6
7 years of negative TFP growth since
5 1997; TFP slowdown accounts for 3.1%
Contribution to Annual Output Growth (%)

of 3.7% drop in average GDP growth

4

0
1988

1993
1985
1986
1987

1989
1990
1991
1992

1994
1995
1996
1997
1998
1999
2000
2001
2002
2003
2004
2005
2006
2007
2008
2009
2010
2011
2012
-1

-2

-3

-4
Labor Growth Capital Growth TFP Growth
 Drop Is Not from Shift to Services

• Economy shifted toward services during the slowdown

– From 53% (1985-97) to 59% (1998-2002)
• Services have lower productivity than industry
– 40% lower output per worker
• But shift is not large enough to explain much of slowdown
• Holding industry and services shares constant at 1998 levels
would have raised labor productivity growth by only 0.2%/yr
– Less than one-tenth of drop during slowdown
 Approach

• Focus on factors that influence productivity at the micro level

– Slowdown occurs mostly within sectors
– Chilean macroeconomic policies and institutions are widely regarded
as sound and well managed
• Use framework in Syverson (JEL 2011) to guide analysis
– Factors that operate directly at the producer level—”levers” that can
be pulled to increase productivity
• Managerial practice or talent; general labor quality; capital quality
and IT capital; R&D, innovation, and learning-by-doing
– Factors that affect producers’ external environments; affect allocation
of activity across producers as well as incentives to pull “levers”
• Productivity spillovers; competition (domestic and international
via trade); regulatory structure; input market flexibility
 Sources of Productivity Level Deficits

• Management practices in Chilean manufacturing lag those in

higher-income countries (source: World Management Survey)
Overall score Monitoring Targets People
United States 3.35 3.58 3.26 3.25
Germany 3.23 3.57 3.22 2.98
Japan 3.23 3.50 3.34 2.92
…
Chile 2.83 3.14 2.72 2.67
Argentina 2.76 3.08 2.68 2.56
Greece 2.73 2.97 2.66 2.58
China 2.71 2.90 2.63 2.69
Brazil 2.71 3.06 2.69 2.55
Average 2.99 3.29 2.95 2.82
 Sources of Productivity Level Deficits

• Lagging innovative activity

FTE R&D researchers R&D spending
(per million pop) (% of GDP)
Chile 350 0.4
Latin America & Caribbean 500 0.8
Euro Area 3200 2.1
U.S. 4700 2.8

– Chile can buy rather than make innovations, but still a very strong
relationship between R&D and product innovation even among
Chilean producers
– Learning-by-doing tied to product improvements appears slower in
Chile than other countries
Potential Sources of TFP Growth Slowdown

• Slowdown in educational achievement growth

– Since 1950, avg. schooling among age 25+ rose about 0.85 years per
decade
– Accelerated to 2X that pace during the 1980s
• Mostly driven by an increase in secondary school attainment
• Considerable share of these gains among low SES groups
– But substantial convergence with other OECD countries had been
achieved by 1990
• Attainment growth inevitably slowed after; less scope for labor
quality to drive productivity growth
• Partially counteracted by entry of more women into labor force
Potential Sources of TFP Growth Slowdown

• Labor market regulations

– Labor market reforms of earlier decades appear to have driven
productivity gains, but now those have been taken
– Some restrictive policies remain
– World Bank/Inter-American Development Bank Enterprise Survey
suggests restrictions becoming more acute
• Between 2006 and 2010 surveys, fraction of respondents reporting
labor regulations as “major” or “very severe” obstacle to current
operations rose from 20% to 28%
• Fraction reporting “no obstacle” fell from 46% to 25%
– Labor regulations serve many useful purposes, but where possible
should strive to help worker rather than job
Potential Sources of TFP Growth Slowdown
• Slackening of domestic competition
– Some recent domestic competition policy reforms, though measured
margins haven’t fallen for non-tradeables
– Firms also reporting fewer competitors (WB-IDB ES)
Number of reported Fraction of respondents, Fraction of respondents,
competitors 2006 2010
0 3% 4%
1 3 5
2 to 5 38 42
6+ 56 49

• Substantial productivity gains from opening to trade in earlier

decades, but hard to reduce tariffs much more
– Nontariff barriers appear modest, so also little scope for more gains
 How Factors Shape Firm-Level Differences

• “Levers” and especially external factors shape allocation of

activity across firms with different productivity levels
• Productivity dispersion within industries is large in Chile (as
everywhere): average 90-10 TFP percentile ratio is 2.5:1
• Hsieh-Klenow misallocation metric suggests Chilean TFP could
be 50% larger under better allocation
• Very largest companies in Chile (the 0.2% largest) have labor
productivity levels within 10% of similarly sized U.S.
companies, but gap quickly widens for smaller firms
 How Factors Shape Firm-Level Differences
200
180
Sales per Employee (millions CLP)

160
140
120
100
80
60
40
20
0
1 10 100 1000 10000
Employees per Firm

U.S. Chile
 How Factors Shape Firm-Level Differences

• Productivity heterogeneity also has an important interaction

with producer dynamics
• Chile has fairly robust firm turnover in manufacturing
– Annual gross firm entry and exit rates around 5 percent
– Evidence suggests this is primary source of reallocation from less to
more efficient producers
• New business formation has been accelerating; doubled over
2004-11
– Important source of reallocation
– While new businesses have high failure rates, those that survive
account for a considerable portion of economic growth
 Where This Leaves Us
• It is unlikely there is any single explanation for the TFP growth
slowdown in Chile after 1997
• This analysis has identified various factors as either:
– Unlikely to be driving forces (e.g., sectoral reallocation, reduced trade
openness)
– Factors that hold productivity levels below potential but are uncertain
sources of drop in productivity growth (e.g, management practices,
lagging innovation)
– Potential candidates for drop in productivity growth (slowdown in
educational attainment, side effects of labor market regulations,
slackening domestic competition)
• Further work is needed to sharpen focus on and quantify
these potential candidates