dia 2010 productivity in latin america

DIA 2010Productivity in Latin America

The Political Economy of Productivity

Carlos Scartascini and Mariano Tommasi (based on joint work with Fabiana Machado and colleagues)

Prepared for presentation at DIA’s Working Sessions. May 26th 2009

The chapter

The rest of the report tries to bring light on which policies constitute key obstacles for productivity in Latin America E.g., low credit, low infrastructure, etc

Given that information, a Benevolent Dictator could either implement those that maximize productivity or (if trade-offs existed with other policy objectives) “turn the

knob” and adjust policies precisely according to the different policy objectives

However, policies are the outcome of a political decision process.

The chapter

This chapter will attempt to explain the determinants of public policies and political choices in Latin America

The ultimate objective of the chapter is to help the overall report to produce recommendations that include (rather than ignore) the politics behind the policies E.g., Government officials already know that more

infrastructure is better than less. The key issues are:• why they do not produce more, and • under what conditions they would be willing to provide

more, and • how we could help them to achieve those conditions

Background work

1. Developed a new framework for the analysis of decision-making for productivity-related policies

It builds upon previous work done by the authors on the policymaking process –productivity enhancing policies have a high demand for intertemporal cooperation-

It builds upon the more traditional political economy models on the role of interest groups

2. Conducted country studies in several LA(C) countries AR, BO, BR, CH, CO, CR, ME, and VE

3. Tested some of these intuitions “econometrically” Caveat: in our case, data limitations are even more

pronounced than for the other chapters• We have the same restrictions for measuring policy

outcomes• Political variables are hard to come by. They are usually at

the country level, hardly ever time series are available, and if they are, changes are rare.

Main message

Productivity enhancing policies:1. Are highly demanding in terms of Government

capabilities• A “good” policymaking process is needed

• Evidence we have: country cases and data

2. Are highly dependant on the type of demands of the private sector• The policies they demand depend on who populates the

interest groups, how they are organized, who has the stronger voice, etc. • Evidence we have: unfortunately, mostly anecdotal and

inferences

3. Can only happen if there is a good articulation between the public and the private sector• The aggregation of preferences and its articulation are key

factors. Who enters and how? Who do they talk to? How stable is the relationship? What is the level of aggregation?• Evidence we have: mostly from the country cases

Main message

Latin America is particularly weak in (1) PMP does not encourage intertemporal cooperation (worse than

East Asia) The type of demands that may come out of the private sector

and how it interacts with the public sector may be different to other regions It may be the case that how many have voice, who has voice, and

at what level matters. That is, different sectors may exist, they may be organized differently, there may be different ways in which groups are represented in the political sphere and so on.

Changes in terms of its economic structure, the ways in which groups are organized, and how groups articulate with the public sector has been changing The “time series” of the last 20 years shows that greater

incorporation has been usually at the expense of higher “deconcentration” of decision making

Consequently, we might be able to “explain” the cross-section and part of the time series on why LAC’s productivity is low and deteriorating

Main message

LAC ranks relatively low and is more similar to Sub-Saharan Africa than to East Asia in terms of policymaking capabilities

0 1 2 3 4

Sub-Saharan Africa

South Asia

Latin America/Caribbean

Easter Europe/CentralAsia

Middle East/North Africa

East Asia/Pacific

Developed Countries

Reg

ion

Index

Main message



and how it interacts with the public sector may be different to other regions It may be the case that how many have voice, who has voice,

and at what level matters. That is, different sectors may exist, they may be organized differently, there may be different ways in which groups are represented in the political sphere and so on.

Changes in terms of its economic structure, the ways in which groups are organized, and how groups articulate with the public sector has been changing The “time series” of the last 20 years shows that greater

incorporation has been usually at the expense of higher “deconcentration” of decision making


Main message

0 1 2 3 4 5 6

Sub-Saharan Africa

Latin America/Caribbean

Eastern Europe/CentralAsia

South Asia

Middle East/North Africa

East Asia/Pacific

Developed Countries

Corporate activity is (1=dominated by few business group; 7= not)

0.00 0.05 0.10 0.15 0.20 0.25

Europe and Central Asia

Developed

East Asia and Pacific

LAC

South Asia

Middle East and NorthAfrica

Sub-Saharan Africa

Sector Level Herfindahl Index

1970-1974

1995-1999

LAC seems to be more concentrated than the DEV in terms of manufacturing sectors, and more concentrated within sectors (this is very preliminary and imprecise – BETTER DATA NEEDED!)

Main message

Government officials in LAC seem to be more favorable to well connected individuals than in other regions.

0 0.5 1 1.5 2 2.5 3 3.5 4 4.5 5

LAC

ECA

SAS

SSA

EAP

MNA

DEV

Reg

ion

Favoritism of Public Officials (1=officials favor well connected firms)

Main message



and how it interacts with the public sector may be different to other regions It may be the case that how many have voice, who has voice, and

at what level matters. That is, different sectors may exist, they may be organized differently, there may be different ways in which groups are represented in the political sphere and so on.

Its economic structure, the ways in which groups are organized, and how groups articulate with the public sector has been changing At least, it is true in some of the case studies such as Colombia


This presentation

1. Determinants of public policies1. Supply2. Demand3. Equilibrium

2. Empirical evidence1. GCR data2. Availability of credit3. Other exercises

We are not going to talk much about the evidence from the country cases, but they will be very important for the chapter

Characteristics of policiesY+ vs. Y- Having Y- policies is “easy”

Unaccountable money is easy to distribute Barriers are easy to introduce Inefficient and ineffective spending does not

need much capabilities They usually have impact in the short term They may resonate more easily with some

constituencies Examples:

Increasing the number of public employees Certain social or labor programs Providing tax exceptions Low tax enforcement

Characteristics of policiesY+ vs. Y- Having Y+ policies is hard

“Productivity” as an outcome of policy (Y+) is an object which has a number of features:

• Hard to measure (hard to establish relationship specific policies to productivity)

• Hard to sell politically

• Depends on the articulation of a myriad of policies

• Having it in the radar requires some important policymaking capabilities

• Long horizons and consistency• (Yet) flexibility to adjust to changing world• Capacity to work well with private sector

• Leading or helping in a consistent way over time

• Embedding while avoiding (distorting) rent-seeking tendencies

• Informationally intensive

Characteristics of policiesY+ vs. Y- For Y+ to happen it is necessary:

Ability and willingness to supply Y+• The government has to be able to provide Y+

• E.g., it has to be able to generate a highway construction plan

• The government has to have an interest in providing Y+• E.g., it has to be willing to construct the roads where it is

necessary to have them and not only where is politically advantageous in the short run

A demand for Y+• Voters and groups should be willing to trade short term

returns for higher expected returns in the long run• Interest groups have to be interested in Y+ policies

• E.g., interest groups have to demand infrastructure instead of rents

• E.g., people have to understand that not everybody can be employed in the public sector

Characteristics of policiesY+ vs. Y- In equilibrium, Y+ policies require the existence of:

Resources Planning Commitment Coordination Implementation Accountability Enforcement

While some of these conditions may seem relatively exogenous, they are all determined by the policymaking capabilities of the polity. E.g., the availability of resources depends on the ability

to sustain policies over time, to adapt them according to economic circumstances, to implement policies, etc

Characteristics of policies Example: Infrastructure Improving infrastructure requires:

Resources• To start the projects

• One of the reasons why Iinf is low is “lack of resources” Planning

• So the projects could be efficient and effective• Many projects are badly designed

Long term commitment• So the projects are finished

• Governments change and so do priorities Coordination

• So projects do not overlap or can be completed• Sometimes projects end at the border or they compete with siblings

Implementation• So the projects can go from the paper to the real world

• Sometimes governments are not able to go through the procurement stages Accountability

• So there is willingness for the projects to be efficient and effective Enforcement

• So they actually are efficient and effective A side payment to some firms, a tax exemption, or subsidies to gasoline

are much easier to implement

Characteristics of policies Example: Tax system Having a tax system that does not discriminate requires:

Resources• To accept the cost of being non-discriminatory

Planning• So the tax system is well designed

Commitment• So it is not reformed very easily in favor of short term solutions/individual

demands. Coordination

• Among different agencies Implementation

• Implement the system as designed Accountability

• So there is willingness for a good design Enforcement

• Enforce the system designed A system that intends to maximize resources but is discriminatory

and provides the wrong set of incentives (e.g., high labor taxes, taxes on bank transactions, etc) is much easier

Characteristics of policiesY+ vs. Y- In equilibrium, Y+ policies require the existence of:

Resources Planning Commitment Coordination Implementation Accountability Enforcement

While some of these conditions may seem relatively exogenous, they are all determined by the policymaking capabilities of the polity. E.g., the availability of resources depends on the ability

to sustain policies over time, to adapt them according to economic circumstances, to implement policies, etc

Policymaking capabilities

The policymaking capabilities can be measured with the following indicators: Policy Stability

• Ability to sustain most policies over time. Policy Adaptability

• Ability to be able to adapt policies to changing economic conditions and to change policies when they are clearly failing.

Policy Coordination and Coherence• Ability to coordinate multiple actors in the policymaking process.

Policy Implementation and Enforcement• Ability to implement and enforce public policies

Policy Efficiency • Ability to allocate its scarce resources to those activities where they have the

greatest returns. Public-Regardedness of Policies

• Extent to which policies promote the general welfare and resemble public goods.

The Policymaking Capabilities Index• Composite index

country Policy Index

Adapt Implement Coordinate Efficiency Public Regard

Stab

Chile high high high mid-hi high high high Brazil mid-hi high mid-hi mid-hi mid-hi mid-hi mid-hi Uruguay mid-hi high mid-hi mid-hi high mid-hi

Mexico mid-hi mid-hi mid-hi mid-hi mid-hi mid-hi mid-low Colombia mid-hi mid-hi mid-hi mid-low mid-low mid-low mid-low

Costa Rica mid-hi mid-hi mid-hi mid-low low mid-hi mid-hi Jamaica mid-hi mid-hi mid-hi mid-low mid-low mid-low

Trinidad and Tobago mid-hi mid-low mid-hi mid-low high Peru mid-low mid-hi mid-low low mid-low mid-hi mid-low Argentina mid-low mid-hi low mid-low mid-low low mid-low

Panama mid-low mid-low mid-low mid-hi mid-low mid-low low Bolivia low mid-low mid-low mid-low low low low

Honduras low mid-low mid-low low low low low

Dominican Republic low mid-low low mid-low low mid-low low Ecuador low mid-low low mid-low low low low

Nicaragua low low low mid-low mid-low low low Venezuela low low low mid-low low mid-low low

Haiti low low low low low low mid-hi Paraguay low low low low low low low

Guatemala low low low low low low low


As we will see, they are important because They may affect private sector allocations

• E.g., if policies are highly volatile (which increases uncertainty) firms may be unwilling to incur certain investments –such as investing on R&D

They may affect the demand of public policies• E.g., if policies are highly volatile there would be a

lower demand for those policies that provide benefits in the long run –such as infrastructure- and higher for those that produce short term return –such as rents

They also help us to evaluate the impact of the political environments through the outcomes

• E.g., whether there is enforcement; whether some particular individuals are represented better than the population at large, etc.


0 1 2 3 4 5 6

Sub-Saharan Africa

LatinAmerica/Caribbean

South Asia

EasternEurope/Central Asia

Middle East/NorthAfrica

East Asia/Pacific

DevelopedCountries

Legal and political changes over the last five years have (1=severely undermined your firm's planning capacity; 7=had no effect)


Having better policymaking capabilities depends on: the impact of political institutions on the

workings of the political institutions, and the willingness of the actors to invest in the

capabilities of the polity.

These investments do only take place if the conditions for intertemporal cooperation are present.

Intertemporal cooperation

Intertemporal cooperation is more likely if, Key actors have long time horizons There are good “aggregation technologies” There are well-institutionalized arenas for political

exchange There are credible enforcement technologies

These conditions are associated with some characteristics of key players and arenas, such as: Tenure and reelection rates of politicians Arenas in which transactions take place: Congress? The

street? Backrooms? Do parties play any role in aggregating preferences?

Government capabilities

Some variables that we have been able to collect on the workings of the political institutions that approach the theoretical concepts are: Congress capabilities Political party system institutionalization Cabinet stability Bureaucratic capacity Judiciary independence

We don’t focus here on the direct impact of political institutions and regulations. For example –even though never proved- it may be the

case that parliamentary regimes work better than presidential, etc.

Government capabilities

Government Capabilities

0 0.5 1 1.5 2 2.5

LatinAmerica/Caribbean

EasternEurope/Central Asia

Middle East/NorthAfrica

South Asia

Sub-Saharan Africa

East Asia/Pacific

Developed Countries

Demand and provision of public policies Policymaking and government capabilities

provide the conditions for policies to happen. However, only those effectively demanded will be provided.

Demand for Y+ will depend (among others) on:1. Who the interest groups are

• Are they concentrated or not?• Who are the biggest or more influential?• What type of industries does the country have?

• Focused on local or external markets• Have local or external sources of inputs

• Are they concentrated in the territory? Is the concentration aligned with political divisions?• Seems to be very important in Brazil

Demand and provision of public policies2. How they are organized and the ways in which

they are articulated Organization is relevant:

• Because most actors would not be heard if left alone• So policies do not cancel each other out (e.g.,

vertical policies affecting the quality of the horizontal)

• So they are able to demand policies that need coordinated demands (e.g., large infrastructure projects)• In Colombia (and many other countries in LAC),

businesses used to be represented by only a few number of very sophisticated associations such as Fedecafe. Nowadays, the number of associations has exploded

• In Chile, businesses are part of peak associations that belong to the right wing coalition

Demand and provision of public policies

Which demands will be heard depend on how private interest are articulated with the public sector?

• Do they have a common and formal point of entry or do they enter through different channels? Different channels generate different outcomes• Each channel provides different bargaining possibilities and

tokens of exchange. In some arenas it may be easier to provide pork, in another regulations, and so on.• Colombia has very formal channels but political and

economic fragmentation has increased the number of entry points (which may be the factor behind productivity results)

• Does everybody have access or only subsets of the groups, associations, firms, or voters?

• How stable are these relationships?• More stable relationships may provide possibilities for

internalizing certain demands• In Chile, businesses belong to the right wing coalition

and labor to the left-wing coalition. This stable engagement explains why parties internalize their demands so well.

Demand and provision of public policies3. Relative power

Who has the strongest voice matters because demand will be tilted towards their preferences

• Whether declining or growing industries have better representation;

• Whether voters or labor have the ability to resolve collective action problems

4. Demands are not independent of credibility. How much they believe the government will be able to

provide the policies affects what they would demand• E.g., a particular sector would invest on requesting certain

infrastructure projects only if they believe the public sector will be able to deliver

Demand and provision of public policies For brevity (and as more intuitive focal point) we have

concentrated the examples on businesses The analysis can (and should) be replicated for labor

organizations, civil society groups, and voters as they are as relevant for policymaking as businesses

How much each groups weights in the decision process depends on things such as:

• Accountability (of the judiciary type)• Accountability (of the electoral type)

• Some political systems may allow politicians to cater more freely to businesses. Others, may generate greater incentives for clientelistic connections and catering to party followers or certain groups of the population, etc.

• Visibility of government actions• Role of perceptions, formation of expectations, etc (See DIA

2009)

Provision of public policies Sector specificities

Not everything matters equally for every type of policy

Individual policies might differ because Some need more capacities to be developed

(e.g., productivity related need a more capable public sector than rents)

Some sector policies are particularly at risk in some countries. For example, because the government chooses it as the bucket for making payouts / compensating losers in other areas (residual policies).

Summarizing

The previous framework can be represented in a series of baseline models that provide some testable hypotheses

Example: 2 sectors (firms, interest groups, etc) Policy space Y(Y+,Y-) Probability distribution P that the public sector can provide Y+,

Y-• The public sector can default on its promises (lack of capacity or

volatility of policies) α : parameter on how much does Y+,Y- affect each sector

returns Probability distribution Q on sector’s probability of being

decisive γ : parameter on the externality of the public good

• Captures separation on the production or geographical space If voters are introduced, then the elasticity of votes to campaign

contributions do also matter• Different electoral systems (articulation of voters with politicians)

can affect how much voters weight compared to firms

Summarizing

As indicated before, how much Y+ and Y- a country may have seems to depend on

Policymaking and government capabilities The government should be able to provide better policies if

• It has higher capabilities • it can deliver

• More incentives for intertemporal cooperation • it internalizes the incentives for long term policy effects

The type of demands The private sector should be able to demand more Y+ if

• They can coordinate better• They receive more externalities from the Y+• There is a lower probability of default on Y+• There is a consensus on Y+ (voters myopia, biased perceptions)

How the private and public sector articulate Who enters, how, and where matters for Y+. More Y+ if

• Those that benefit the most from Y+ tend to have more voice • Private interests are aggregated, they internalize private sector

demands, and they are articulated in stable relationships• Decision-making is more transparent

Empirical evidence(PRELIMINARY)

So far, we have conducted three types of exercises:

1. Used our own database on the role of policymaking capabilities

2. Used GCR data to proxy for the quality of policies and some characteristics of economic structure and the like

3. Looked at specific exercises using similar data to that in the previous chapters.

Policymaking and TFP Growth

Using our database on policymaking and government capabilities we find:A positive association between

• Policymaking capabilities and TFP growth• Government capabilities and TFP growth

Very week associations with TFP levels• Only the measure of efficiency seems to

work relatively wellAmong the government capabilities,

those of the bureaucracy and congress seem to be particularly relevant.

Policymaking and TFP Growth-2

02

4T

FP

Gro

wth

0 1 2 3 4Policymaking Capabilities Index

LAC Other

Policymaking and Government Capabilities and ProductivityTFP Level TFP Growth

Policymaking Capabilities Index 0.101 0.013***(0.074) (0.003)

Adaptability 0.066 0.009***(0.055) (0.002)

Stability -0.052 0.007**(0.079) (0.003)

Public Regardedness 0.071 0.008***(0.062) (0.002)

Implementation and Enforcement 0.065 0.007***(0.061) (0.002)

Coordination 0.003 0.009***(0.052) (0.002)

Efficiency 0.124** 0.007***(0.052) (0.002)

Cabinet Stability 0.22 0.013**0.136 0.006

Bureaucratic Capabilities 0.047 0.007***(0.052) (0.002)

Party Institutionalization 0.028 0.006*(0.080) (0.003)

Congress Capabilities 0.091 0.006**(0.058) (0.002)

Judicial Independence 0.066 0.003*(0.046) (0.002)

Standard errors in parentheses; *** p<0.01, ** p<0.05, * p<0.1Each cell indicates the coefficient of interest from a multivariate regressionthat includes the following controls: Ln (GDPpc 1990), Regional DummiesDependent variables: Average 1998-2006

Policymaking and Government Capabilities and Productivity

(1) (2) (1) (2) (3) (4) (5) (6)Adaptability 0.046 0.008*** 0.006** 0.008***

(0.073) (0.003) (0.002) (0.003)Stability -0.171 0.001 0.002 0.003 0.002

(0.107) (0.004) (0.004) (0.004) (0.004)Public Regardedness 0.041 0.005 0.003 0.006

(0.107) (0.004) (0.004) (0.004)Implementation and Enforcement -0.051 -0.007** -0.002 -0.005

(0.103) (0.004) (0.003) (0.004)Coordination -0.032 0.006*** 0.006*** 0.007***

(0.062) (0.002) (0.002) (0.002)Efficiency 0.169** 0.000 0.001 0.001 0.001

(0.083) (0.003) (0.003) (0.003) (0.003)Cabinet Stability 0.271* 0.004

(0.159) (0.006)Bureaucratic Capabilities 0.002 0.004 0.004**

(0.065) (0.003) (0.002)Party Institutionalization -0.036 -0.001

(0.097) (0.004)Congress Capabilities 0.052 0.003 0.004*

(0.073) (0.003) (0.002)Judicial Independence -0.011 0.001

(0.059) (0.002)Observations 69 68 69 69 69 69 68 68Standard errors in parentheses; *** p<0.01, ** p<0.05, * p<0.1Both regressions include the following controls: Ln (GDPpc 1990), Regional DummiesDependent variables: Average 1998-2006

TFP Level TFP Growth

GCR Exercises

GCR is the only source from which we could get a broad number of policy variables similar to those being considered in the DIA.

Because this data comes from subjective surveys, it can not be taken very seriously.

To avoid capturing mood, we are using the average for all the available years.

Despite its shortcomings, it seems to capture similar evidence to the chapters: The policy variables seem to explain

productivity growth quite well

Public Policies and TFPTFP Level TFP Growth

Infraestructure Coefficient 0.034 0.580*** Standard Error 0.03 0.147 R-squared 0.75 0.21 Observations 66 69Credit index Coefficient 0.014 0.736*** Standard Error 0.04 0.180 R-squared 0.75 0.21 Observations 66 69Extent distortive tax system (1=limits incentives …) Coefficient 0.103 0.272* Standard Error 0.04 0.139 R-squared 0.78 0.07 Observations 66 69Extent distortive government subsidies (1=keeps uncompetitive alive…) Coefficient 0.106 0.886*** Standard Error 0.04 0.209 R-squared 0.76 0.27 Observations 51 54Administrative burden for startups Coefficient 0.012 0.767*** Standard Error 0.04 0.160 R-squared 0.73 0.34 Observations 62 65Organized efforts to improve competitiveness (1=non-existent …) Coefficient 0.057 0.751*** Standard Error 0.04 0.143 R-squared 0.73 0.27 Observations 60 63Informal Sector (% unofficial) Coefficient 0.023 -0.641** Standard Error 0.06 0.261 R-squared 0.75 0.18 Observations 64 67Standard errors in parentheses; *** p<0.01, ** p<0.05, * p<0.1Each cell indicates the coefficient of interest from a multivariate regressionthat includes the following controls: Ln (GDPpc 1990)Dependent variables: Average 1998-2006GCR data: average for available years

GCR Exercises

We also find a positive relationship between policymaking and government capabilities and the policies.

We would expect that the impact might be different across policies. For example,

Some policies may demand a longer planning horizon than others

Some policies may demand more capacity from the bureaucracy than from congress

Policymaking Capabilities and Public PoliciesOLS controlling for Log GDPpc 1990, Openess, and Regional Dummies

Infraestructure Credit Extent Distortive Extent Distortive Administrative Efforts to Improve InformalIndex Index Tax System Govt Subsidies Burden Startups Competitiviness Sector

Policymaking Capab. Coefficient 0.845*** 0.508*** 0.336* 0.795*** 0.702*** 0.994*** -0.676*** Standard Error 0.13 0.10 0.17 0.12 0.16 0.12 0.12 R-squared 0.87 0.80 0.39 0.54 0.60 0.64 0.64 Observations 115 114 114 80 98 95 111Stability Coefficient 0.316** 0.288*** 0.064 0.775*** 0.519*** 0.620*** -0.266* Standard Error 0.13 0.10 0.15 0.15 0.16 0.16 0.16 R-squared 0.85 0.78 0.40 0.54 0.56 0.45 0.57 Observations 105 104 104 80 95 92 101Adaptability Coefficient 0.296*** 0.147* 0.146 0.299*** 0.122 0.505*** -0.285*** Standard Error 0.11 0.08 0.13 0.11 0.12 0.10 0.09 R-squared 0.82 0.76 0.45 0.39 0.54 0.53 0.58 Observations 110 109 109 76 94 91 106Coordination Coefficient 0.364*** 0.359*** 0.079 0.332*** 0.530*** 0.358*** -0.377*** Standard Error 0.12 0.07 0.14 0.12 0.12 0.11 0.08 R-squared 0.83 0.79 0.46 0.40 0.63 0.46 0.60 Observations 102 101 101 70 87 84 98Imp & Enforcement Coefficient 0.659*** 0.332*** 0.131 0.477*** 0.454*** 0.720*** -0.503*** Standard Error 0.11 0.08 0.15 0.12 0.14 0.12 0.11 R-squared 0.87 0.79 0.38 0.44 0.57 0.57 0.62 Observations 114 113 113 80 98 95 110Efficiency Coefficient 0.569*** 0.315*** 0.496*** 0.567*** 0.498*** 0.668*** -0.400*** Standard Error 0.11 0.08 0.14 0.10 0.12 0.13 0.11 R-squared 0.86 0.79 0.45 0.52 0.59 0.57 0.60 Observations 115 114 114 80 98 95 111Public Regardedness Coefficient 0.737*** 0.393*** 0.163 0.541*** 0.557*** 0.611*** -0.528*** Standard Error 0.12 0.10 0.18 0.12 0.13 0.14 0.10 R-squared 0.87 0.80 0.38 0.47 0.59 0.50 0.62 Observations 115 114 114 80 98 95 111

Government Capabilities and Public PoliciesOLS controlling for Log GDPpc 1990, Openess, and Regional Dummies


Party Institutionalization Coefficient 0.061 -0.012 -0.095 0.049 0.076 0.539*** 0.072 Standard Error 0.143 0.096 0.159 0.155 0.191 0.165 0.156 R-squared 0.82 0.75 0.28 0.30 0.47 0.40 0.54 Observations 113 112 112 79 97 94 110Congress Capabilities Coefficient 0.680*** 0.330*** 0.323** 0.450*** 0.563*** 0.677*** -0.253** Standard Error 0.101 0.094 0.127 0.147 0.119 0.119 0.128 R-squared 0.87 0.79 0.35 0.42 0.60 0.54 0.56 Observations 112 111 111 80 98 95 111Bureaucratic Capabilities Coefficient 0.300** 0.183** -0.023 0.227* 0.250** 0.396*** -0.286*** Standard Error 0.134 0.086 0.131 0.125 0.103 0.098 0.102 R-squared 0.84 0.77 0.38 0.38 0.53 0.45 0.58 Observations 112 111 111 80 96 93 108Cabinet Stability Coefficient 0.620** 0.307 0.778** 0.118 0.723*** 0.320 -0.188 Standard Error 0.289 0.222 0.305 0.290 0.261 0.331 0.261 R-squared 0.82 0.76 0.35 0.30 0.50 0.34 0.54 Observations 114 113 113 79 97 94 110Judicial Independence Coefficient 0.461*** 0.286*** 0.106 0.433*** 0.539*** 0.647*** -0.288*** Standard Error 0.096 0.079 0.129 0.116 0.102 0.097 0.094 R-squared 0.84 0.79 0.35 0.46 0.62 0.61 0.58 Observations 114 113 113 80 98 95 111Intertemporal Index Coefficient 0.274*** 0.152*** 0.112 0.201*** 0.271*** 0.333*** -0.122* Standard Error 0.053 0.051 0.072 0.060 0.058 0.053 0.065 R-squared 0.86 0.78 0.27 0.38 0.57 0.55 0.56 Observations 108 107 107 79 95 92 107

Government Capabilities and Public PoliciesOLS controlling for Log GDPpc 1990, Openess, and Regional Dummies


Party Institutionalization Coefficient 0.061 -0.012 -0.095 0.049 0.076 0.539*** 0.072 Standard Error 0.143 0.096 0.159 0.155 0.191 0.165 0.156 R-squared 0.82 0.75 0.28 0.30 0.47 0.40 0.54 Observations 113 112 112 79 97 94 110Congress Capabilities Coefficient 0.680*** 0.330*** 0.323** 0.450*** 0.563*** 0.677*** -0.253** Standard Error 0.101 0.094 0.127 0.147 0.119 0.119 0.128 R-squared 0.87 0.79 0.35 0.42 0.60 0.54 0.56 Observations 112 111 111 80 98 95 111Bureaucratic Capabilities Coefficient 0.300** 0.183** -0.023 0.227* 0.250** 0.396*** -0.286*** Standard Error 0.134 0.086 0.131 0.125 0.103 0.098 0.102 R-squared 0.84 0.77 0.38 0.38 0.53 0.45 0.58 Observations 112 111 111 80 96 93 108Cabinet Stability Coefficient 0.620** 0.307 0.778** 0.118 0.723*** 0.320 -0.188 Standard Error 0.289 0.222 0.305 0.290 0.261 0.331 0.261 R-squared 0.82 0.76 0.35 0.30 0.50 0.34 0.54 Observations 114 113 113 79 97 94 110Judicial Independence Coefficient 0.461*** 0.286*** 0.106 0.433*** 0.539*** 0.647*** -0.288*** Standard Error 0.096 0.079 0.129 0.116 0.102 0.097 0.094 R-squared 0.84 0.79 0.35 0.46 0.62 0.61 0.58 Observations 114 113 113 80 98 95 111Intertemporal Index Coefficient 0.274*** 0.152*** 0.112 0.201*** 0.271*** 0.333*** -0.122* Standard Error 0.053 0.051 0.072 0.060 0.058 0.053 0.065 R-squared 0.86 0.78 0.27 0.38 0.57 0.55 0.56 Observations 108 107 107 79 95 92 107

GCR Exercises

There may be some evidence on the role of the characteristics of the economic structure and markets on decisions and demand of public policies

Market Characteristics and Public PoliciesOLS controlling for Log GDPpc 1990, Openess, and Regional Dummies

Infraestructure Credit Efforts to Improve Company Spending Firm TechnologyIndex Index Competitiviness on R&D Absortion

Concentration Corporate 0.259*** 0.0784 0.323*** 0.312*** 0.281**(0.0986) (0.0889) (0.107) (0.118) (0.107)

Foreign Ownership 0.174** 0.203*** 0.253** -0.149* 0.128*(0.0768) (0.0566) (0.100) (0.0855) (0.0752)

Efficiency Legal Fwk 0.262** 0.135*(0.108) (0.0787)

Observations 113 113 94 113 113R-squared 0.840 0.793 0.534 0.535 0.716Standard errors in parentheses; *** p<0.01, ** p<0.05, * p<0.1Regressions include the following controls: Ln (GDPpc 1990), Trade/GDP, Regional Dummies

Other exercises

We want to have policy level exercises

Two policy areas in which we can do relatively more sophisticated exercises areCreditEntry regulations

Credit

The argument we are going to test is whether credit is higher or not in those countries in which the sectors more dependent on credit are larger.

There are two potential reasons why we would expect a positive relationship:

1. Government realizes the potential for the economy of providing credit when the larger sectors need it

2. When the credit dependent sectors are large, they lobby the government to get it.

Consequently, we will test if credit is a function of the

Sector share weighted RZ Policymaking and government capabilities

Credit

Following (1) and (2) above, the ability of the government to provide good policies matter. For (1) –“benevolent government”- the effect is

direct. • Governments with lower capacities might be less able

to provide the policies that affect credit directly (e.g., creditor rights)

• Because some of these policies, such as financial liberalization, imply a tradeoff between short term costs and long term benefits, government with shorter horizons would likely do not attempt them

• Similarly, shorter term horizon governments may be less likely to accept the higher risks –but potential higher benefits- that financial liberalization entails

Credit

For (2) –“lobby”-, it is indirect.• In good policy environments, firms ask for

good policies because they know that the government is able to provide it.

• In bad policy environments, no matter how much firms lobby, there is a high chance that the government can not deliver. Then, in those cases, instead of asking for credit they might ask for other types of compensations such as tax breaks, directed credit assistance, lax tax enforcement, subsidies, etc.

-20

24

% o

f GD

P

0 1 2 3 4General Policy Index

Marginal effect Confidence bands (2 std deviations)

Note: Dashed lines represent the quartiles of General Policy Index.

Impact of R & Z (t-1) on Private credit/GDP

-10

12

3%

of G

DP

0 2 4 6ICRG Bureaucratic quality index (t-1)

Marginal effect Confidence bands (2 std deviations)

Note: Dashed lines represent the quartiles of ICRG Bureaucratic quality index (t-1).

Impact of R & Z (t-1) on Private credit/GDP

Entry Regulations

Another exercise we may try is to explain the level of entry regulations in the economy.

The underlying model would be relatively easy. Politicians’ decisions regarding regulations depend on the relative weight of interest groups –incumbent firms- and voters.

Consequently, Reg=F(S, ES, Xs, Cl) where S = variables such as electoral systems ES = economic structure variables such as:

Xmanuf/Xtotales; X/GDP; Diversification (sectors); HH of size (lagged); {see also GCR variables on influence}

Xs = judicial independence, bureaucracy, quality of regulatory bodies (GCR?)

Cl = proxies for clientelism such as corruption; particularism; party institutionalization (or programmatic parties)

DIA 2010Productivity in Latin America

The Political Economy of Productivity

Carlos Scartascini and Mariano Tommasi (based on joint work with Fabiana Machado and colleagues)

Prepared for presentation at DIA’s Working Sessions. May 26th 2009

dia 2010 productivity in latin america

Documents