2006年-世界发展银行全球_Lobbying_and_Agricultural_Trade_Policy_in_the_United_States_40页_446kb

Summary of "Lobbying and Agricultural Trade Policy in the United States"

Core Content

This paper investigates the relationship between political lobbying and agricultural trade policy in the United States, testing the predictions of the Grossman-Helpman political economy model. The study is the first to apply this model specifically to agricultural data, and it aims to resolve a puzzle regarding the model's quantitative implications in the context of U.S. agricultural policy.

Main Objectives

• To assess whether political campaign contributions influence agricultural protection in the U.S. in line with the Grossman-Helpman model.
• To develop a theoretically grounded empirical analysis of agricultural protection, incorporating both subsidies and tariffs.
• To address the issue of endogeneity in the model by constructing appropriate instrumental variables.

Key Findings

• The Grossman-Helpman model predicts that agricultural protection and subsidies are determined by the interaction of lobbying intensity, the import-to-output ratio (inverse of $ z^M $), and the export-to-output ratio (inverse of $ z^X $), along with their respective elasticities.
• The model is qualitatively affirmed by the data, showing that protection increases with the inverse import-to-output ratio and that subsidies are similarly influenced by the inverse export-to-output ratio.
• The parameter $ a $, which represents the government's preference for welfare over contributions, is estimated as the inverse of the coefficient $ \beta_1 $ in the model.
• The study finds that the implied value of $ a $ is surprisingly high, suggesting that U.S. agricultural policy is driven more by welfare considerations than by lobbying contributions, which contrasts with previous findings in other sectors.
• The use of ad valorem equivalents for both tariffs and subsidies ensures that the model is consistent with theoretical assumptions.
• The data on lobbying contributions is drawn from PAC (Political Action Committee) spending between 1993 and 1998, showing that certain agricultural products, such as sugar, dairy, and cotton, are heavily represented by PACs.
• The political organization of agricultural sectors is a critical factor in shaping trade policy, with the most politically active sectors receiving a significant share of PAC contributions.
• The robustness of the results is tested using different measures of political representation, based on percentile cutoffs of PAC spending.
• The study uses instrumental variables to address the endogeneity of the regressors, particularly $ I_i $ (political organization) and $ z_i / e_i $ (import/export ratios and elasticities).

Methodology

• The authors use a structural approach, linking the empirical model directly to the theoretical framework of the Grossman-Helpman model.
• They estimate three cross-sectional econometric models:
  • A combined model (4) that includes both protection and subsidies.
  • Separate models (5) and (6) for imported and exported goods, respectively.
• Instrumental variables are constructed using data from the 1997 agricultural census to account for endogeneity issues.
• The Kelejian method is employed to instrument the composite regressor $ I_i \times z_i / e_i $, using both linear and higher-order terms of the instruments.
• Error correction is applied to the import demand elasticity variable, using the Fuller correction, to account for measurement errors.
• The export supply elasticity is assumed to be constant at 2, due to the relatively undifferentiated nature of most agricultural goods.

Data Description

• Protection measures include:
  • Ad Valorem Equivalent (AVE) of core nontariff measures (Core NTMs).
  • Overall Protection, which combines tariffs, nontariff measures, and domestic subsidies.
  • All Tariff, which includes both ad valorem and specific tariffs.
• Subsidy measures include:
  • Export Subsidies as a proportion of exports (XS).
  • Total Subsidies as a proportion of exports (TS).
• All data is from the year 1999.
• Trade and output data are sourced from the USDA's Production, Supply and Distribution (PSD) database.
• Lobbying data is from PAC contributions between 1993 and 1998, obtained from the Federal Election Commission (FEC).

Conclusion

• The study confirms the qualitative predictions of the Grossman-Helpman model, showing that lobbying activity and the structure of agricultural trade policy are closely related.
• The quantitative results suggest that the U.S. government places a high value on aggregate welfare compared to lobbying contributions, which is inconsistent with the large deadweight losses typically associated with agricultural protection.
• The authors argue that this discrepancy arises from overly simplistic assumptions in the model, particularly the unitary government assumption. By incorporating policy uncertainty and a more realistic representation of the political process, the model becomes more consistent with real-world data.
• The paper provides a novel contribution to the political economy literature by applying the model to agricultural data and resolving the puzzle of its quantitative implications.

Key Variables and Equations

• Equation (1): Government utility function $ G = a W + C $, where $ a $ is the weight given to welfare over contributions.
• Equation (2): The equilibrium condition for protection and subsidies:
  $$
  \frac{t_i}{1 + t_i} = \left(\frac{1}{a}\right) \times \left(I_i \times \frac{z_i^M}{e_i^M}\right), \quad \text{if i’s producers compete with imports}
  $$
  $$
  \frac{s_i}{1 + s_i} = \left(\frac{1}{a}\right) \times \left(I_i \times \frac{z_i^X}{e_i^X}\right), \quad \text{if i’s producers export}
  $$
• Equation (3): Combined model:
  $$
  \frac{g_i}{1 + g_i} = \beta_1 \left(I_i \times \frac{z_i}{e_i}\right) + \varepsilon_i
  $$
• Equation (4): Separate models for imports and exports are also estimated.

Robustness and Instrumental Variables

• The authors test the robustness of their results using different measures of political organization.
• They construct exogenous instruments from the 1997 agricultural census to address the endogeneity of $ I_i $ and $ z_i / e_i $.
• The Kelejian method is used to instrument the composite regressor.
• The first-stage $ F $-statistic is used to assess the strength of instruments, and both 2SLS and LIML estimators are reported.

Implications

• The results imply a political economy equilibrium that is difficult to change.
• The study highlights the complexity of real-world political decision-making and the need for more realistic models that incorporate uncertainty and multiple actors.
• It underscores the importance of lobbying in shaping agricultural trade policy and the role of subsidies in the U.S. agricultural sector.