revisiting integration and productivity discussion for ipes 2010 october 14, 2008

35
Revisiting Integration and Productivity Discussion for IPES 2010 October 14, 2008

Upload: beverly-young

Post on 05-Jan-2016

216 views

Category:

Documents


0 download

TRANSCRIPT

Page 1: Revisiting Integration and Productivity Discussion for IPES 2010 October 14, 2008

Revisiting Integration and Productivity

Discussion for IPES 2010 October 14, 2008

Page 2: Revisiting Integration and Productivity Discussion for IPES 2010 October 14, 2008

Outline of presentation

Discussion on background papers

Tentative outline for the chapter

Page 3: Revisiting Integration and Productivity Discussion for IPES 2010 October 14, 2008

Motivation The policy debate has moved on from whether or not countries

should open their economies (most of them have already done it and there is no clear police reversal in sight) to how can they take better advantage of trade.

There is a clear need to better understand the channels through which trade can boost productivity and growth. Gains from trade cannot be taken for granted.

Research has proceed in two directions:1) Macro level: Dollar (1992), Sachs and Warner (1995), Edwards

(1998), Frankel and Romer (1999), Rodriguez and Rodrik (2000), Taylor and Estevadeordal (2008)

2) Micro level: Lopez-Cordova, Mesquita (2003), Muendler (2002); Pavcnik (2002), Fernandes (2003), IDB (2002)

Shortcomings and omissions:1)Revenue bias2)Lack of clear theoretical framework for firm

heterogeneity3) Exclusive focus on policy barriers (tariffs)

Background papers will revisit this second and more robust strand of the literature. Seek to address these weaknesses

Page 4: Revisiting Integration and Productivity Discussion for IPES 2010 October 14, 2008

1) Revenue bias Proxy for firm’s output: firm’s revenues deflated by

industry price index biased productivity measure

Page 5: Revisiting Integration and Productivity Discussion for IPES 2010 October 14, 2008

1) Revenue bias Proxy for firm’s output: firm’s revenues deflated by

industry price index biased productivity measure

How serious is this problem? Loecker (2007); Haltiwanger and Syverson (2005); Katayama, Lu and Tybout (2006)

Page 6: Revisiting Integration and Productivity Discussion for IPES 2010 October 14, 2008

1) Revenue bias Proxy for firm’s output: firm’s revenues deflated by

industry price index biased productivity measure

How serious is this problem? Loecker (2007); Haltiwanger and Syverson (2005); Katayama, Lu and Tybout (2006)

Solutions? Levinsohn and Melitz (2002); Loecker (2007), Katayama, Lu and Tybout (2006)

Page 7: Revisiting Integration and Productivity Discussion for IPES 2010 October 14, 2008

2) Weak theoretical framework

Until recently, trade theory had not much to say about heterogeneity of firm performances observed within industries

Page 8: Revisiting Integration and Productivity Discussion for IPES 2010 October 14, 2008

2) Weak theoretical framework

Until recently, trade theory had not much to say about heterogeneity of firm performances observed within industries

Melitz (2003); Bernard, Eaton, Jensen and Kortum (2003)

Page 9: Revisiting Integration and Productivity Discussion for IPES 2010 October 14, 2008

2) Weak theoretical framework

Until recently, trade theory had not much to say about heterogeneity of firm performances observed within industries

Melitz (2003); Bernard, Eaton, Jensen and Kortum (2003)

New models generate clear predictions about the effects of trade at the micro level

Page 10: Revisiting Integration and Productivity Discussion for IPES 2010 October 14, 2008

2) Weak theoretical framework

Until recently, trade theory had not much to say about heterogeneity of firm performances observed within industries

Melitz (2003); Bernard, Eaton, Jensen and Kortum (2003)

New models generate clear predictions about the effects of trade at the micro level

New models also make more explicit the role of other factors (competition, foreign entry) in the linkage between trade and productivity

Page 11: Revisiting Integration and Productivity Discussion for IPES 2010 October 14, 2008

3) Excessive focus on tariff barriers

Previous research looked exclusively at tariff barriers as the only obstacle to trade

Page 12: Revisiting Integration and Productivity Discussion for IPES 2010 October 14, 2008

3) Excessive focus on tariff barriers

Previous research looked exclusively at tariff barriers as the only obstacle to trade

After years of trade liberalization, policy barriers have lost relevance vis-à-vis other trade costs, like transport costs

Page 13: Revisiting Integration and Productivity Discussion for IPES 2010 October 14, 2008

3) Excessive focus on tariff barriers

Previous research looked exclusively at tariff barriers as the only obstacle to trade

After years of trade liberalization, policy barriers have lost relevance vis-à-vis other trade costs, like transport costs

For most LAC countries, transport costs are significantly higher than tariffs –for imports and exports (IDB, 2008)

Page 14: Revisiting Integration and Productivity Discussion for IPES 2010 October 14, 2008

3) Excessive focus on tariff barriers

Previous research looked exclusively at tariff barriers as the only obstacle to trade

After years of trade liberalization, policy barriers have lost relevance vis-à-vis other trade costs, like transport costs

For most LAC countries, transport costs are significantly higher than tariffs –for imports and exports (IDB, 2008)

Need to acknowledge the role of transport costs as a barrier to trade and assess their impact on productivity

Page 15: Revisiting Integration and Productivity Discussion for IPES 2010 October 14, 2008

Background papers Objective: revisit the relationship between

integration and productivity using plant (firm) level data for ARG, BRA, CHL, MEX and VEN, addressing the previous weaknesses

Page 16: Revisiting Integration and Productivity Discussion for IPES 2010 October 14, 2008

Background papers Objective: revisit the relationship between

integration and productivity using plant (firm) level data for ARG, BRA, CHL, MEX and VEN, by addressing previous weaknesses

Aspects the papers will examine:i) Reallocation effects from changes in trade costsii) Plant-productivity effects from changes in trade costsiii) How competition and foreign entry (FDI) interact with (i)

and (ii)

Page 17: Revisiting Integration and Productivity Discussion for IPES 2010 October 14, 2008

i) Reallocation effects

Page 18: Revisiting Integration and Productivity Discussion for IPES 2010 October 14, 2008

i) Reallocation effects Some testable hypotheses (from Melitz, 2003). A

decrease in trade costs:

1) leads to an aggregate productivity gain2) raises the probability of firm exit3) increases the number of exporting firms4) increases export sales of the existing exporters

Page 19: Revisiting Integration and Productivity Discussion for IPES 2010 October 14, 2008

i) Reallocation effects Some testable hypotheses (from Melitz, 2003). A

decrease in trade costs:

1) leads to an aggregate productivity gain2) raises the probability of firm exit3) increases the number of exporting firms4) increases export sales of the existing exporters

Aim: to explore the extent of the reallocation effects and the potential channels by which they take place

Page 20: Revisiting Integration and Productivity Discussion for IPES 2010 October 14, 2008

ii) Plant productivity effects

Page 21: Revisiting Integration and Productivity Discussion for IPES 2010 October 14, 2008

ii) Plant productivity effects Previous studies found only modest effects of

exporting on plant productivity (Bernard and Jensen, 1999; Clerides, Lach and Tybout, 1998; Aw, Chung and Roberts, 2000)

Page 22: Revisiting Integration and Productivity Discussion for IPES 2010 October 14, 2008

ii) Plant productivity effects Previous studies found only modest effects of

exporting on plant productivity (Bernard and Jensen, 1999; Clerides, Lach and Tybout, 1998; Aw, Chung and Roberts, 2000)

But there are many possible channels: scale efficiency, more and better inputs, import discipline, Schumpeterian incentives

Page 23: Revisiting Integration and Productivity Discussion for IPES 2010 October 14, 2008

ii) Plant productivity effects Previous studies found only modest effects of

exporting on plant productivity (Bernard and Jensen, 1999; Clerides, Lach and Tybout, 1998; Aw, Chung and Roberts, 2000)

But there are many possible channels: scale efficiency, more and better inputs, import discipline, Schumpeterian incentives

Aim: to explore the role of different channels

Page 24: Revisiting Integration and Productivity Discussion for IPES 2010 October 14, 2008

iii) The role of competition and foreign entry in the trade-

productivity link

Page 25: Revisiting Integration and Productivity Discussion for IPES 2010 October 14, 2008

iii) The role of competition and foreign entry in the trade-

productivity link Barriers to entry (including foreign entry) could

limit the reallocation gains from a fall in trade costs

Page 26: Revisiting Integration and Productivity Discussion for IPES 2010 October 14, 2008

iii) The role of competition and foreign entry in the trade-

productivity link Barriers to entry (including foreign entry) could

limit the reallocation gains from a fall in trade costs

Barriers to entry could also limit the plant-productivity gains from a fall in trade costs

Page 27: Revisiting Integration and Productivity Discussion for IPES 2010 October 14, 2008

iii) The role of competition and foreign entry in the trade-

productivity link Barriers to entry (including foreign entry) could

limit the reallocation gains from a fall in trade costs

Barriers to entry could also limit the plant-productivity gains from a fall in trade costs

Aim: assess to what extent barriers to competition and FDI erode the effects of trade on productivity

Page 28: Revisiting Integration and Productivity Discussion for IPES 2010 October 14, 2008

DataManufacturing Data Argentina: Encuestas Nacionales sobre la Conducta Tecnológica de

las Empresas Industriales Argentinas Brazil: Pesquisa Industrial Anual –PIA Chile: Encuesta Nacional Industrial Anual –ENIA Mexico: Encuesta Industrial Anual –EIA Venezuela: Encuesta Industrial Anual

Trade Costs Data Foreign Trade Statistic System (from ALADI) US Imports Merchandise (from US Census Bureau)

Both sources report data on import duties and on freights at the 6-digit harmonized system

Coverage Potentially 1995-2005

Page 29: Revisiting Integration and Productivity Discussion for IPES 2010 October 14, 2008

Summary of empirical strategy

On reallocation effects Strategy based on Bernard, et al (2006)

Tests of hypotheses based on appropriate specification. For example:

Industry effects (hypothesis 1): industry-level regressions of changes in industry TFP on lagged changes in trade costs and industry and year fixed effects

Firm exit (hypothesis 2): logistic regressions of the rate of plant death on lagged changes in trade costs, plant controls and industry and year fixed effects

Page 30: Revisiting Integration and Productivity Discussion for IPES 2010 October 14, 2008

Summary of empirical strategy

On reallocation effects Strategy based on Bernard, et al (2006)

Tests of hypotheses based on appropriate specification. For example:

Industry effects (hypothesis 1): industry-level regressions of changes in industry TFP on lagged changes in trade costs and industry and year fixed effects

Firm exit (hypothesis 2): logistic regressions of the rate of plant death on lagged changes in trade costs, plant controls and industry and year fixed effects

On plant effects Strategy based on Lopez-Cordova and Mesquita (2004)

General idea is to construct proxies to test different channel using plant-level regressions

Page 31: Revisiting Integration and Productivity Discussion for IPES 2010 October 14, 2008

Summary of empirical strategy

On the role of competition and foreign entry The general idea is to exploit the industry variation on barriers to entry

(including foreign entry) to analyze whether the reallocation effects, or the plant effects, differ in industries with various degrees of entry

We will use different proxies for barriers to entry High-entry versus low entry industries (Herfindal indexes, US natural rate of entry,

etc) High-FDI versus Low-FDI industries Tradable versus non-tradable industries Industries close and far from technology frontier (RCA indexes, shares of exports)

Page 32: Revisiting Integration and Productivity Discussion for IPES 2010 October 14, 2008

Summary of empirical strategy

On the role of competition and foreign entry The general idea is to exploit the industry variation on barriers to entry

(including foreign entry) to analyze whether the reallocation effects, or the plant effects, differ in industries with various degrees of entry

We will use different proxies for barriers to entry High-entry versus low entry industries (Herfindal indexes, US natural rate of entry,

etc) High-FDI versus Low-FDI industries Tradable versus non-tradable industries Industries close and far from technology frontier (RCA indexes, shares of exports)

Note: In each of the above aspects we will try to differentiate the impact

of tariffs barriers from the impact of transport costs

Page 33: Revisiting Integration and Productivity Discussion for IPES 2010 October 14, 2008

Tentative OutlineIntroduction Literature review: links between integration and productivity Main empirical findings already available for LAC Why it is important to revisit this topic again?

Revenue bias Recent theoretical developments Excessive focus on traditional barriers to trade

Page 34: Revisiting Integration and Productivity Discussion for IPES 2010 October 14, 2008

Tentative OutlineIntroduction Literature review: links between integration and productivity Main empirical findings already available for LAC Why it is important to revisit this topic again?

Revenue bias Recent theoretical developments Excessive focus on traditional barriers to trade

Main Empirical Findings (based on in-house background papers plus papers for Colombia and

Uruguay) Reallocation effects, plant effects What’s new, what’s confirmed, what’s not confirmed

Page 35: Revisiting Integration and Productivity Discussion for IPES 2010 October 14, 2008

Tentative OutlineIntroduction Literature review: links between integration and productivity Main empirical findings already available for LAC Why it is important to revisit this topic again?

Revenue bias Recent theoretical developments Excessive focus on traditional barriers to trade

Main Empirical Findings (based on in-house background papers plus papers for Colombia and Uruguay) Reallocation effects, plant effects What’s new, what’s confirmed, what’s not confirmed

Policy implications The role of trade barriers on productivity: tariffs versus transport costs How to move forward in reducing transport costs (IDB, 2008; Special

boxes) The role of complementary policies: addressing barriers to entry and FDI