2016年-世界发展银行全球_Prices_and_welfare_72页_2mb

Summary of "Prices and Welfare" by Abdelkrim Araar and Paolo Verme

Core Content

This paper explores the welfare effects of price changes, focusing on how different welfare measures and computational methods yield varying results depending on the magnitude of the price shock and the underlying demand system. The authors aim to provide a general framework for using theoretical contributions in empirical studies, especially in the context of poor people and poor countries.

Main Purpose

The primary objective of the paper is to:

• Review and clarify the essential microeconomic literature on welfare measurement.
• Organize and simplify this literature for researchers and practitioners with diverse backgrounds.
• Identify and measure the differences between welfare measures and computational methods.
• Provide guidelines for using alternative approaches under imperfect information.
• Offer computational codes in Stata for the application of all welfare measures and methods.

Key Welfare Measures

The paper evaluates five welfare measures:

1. Consumer's Surplus Variation (CS)
2. Compensating Variation (CV)
3. Equivalent Variation (EV)
4. Laspeyers Variation (LV)
5. Paasche Variation (PV)

These measures are based on the concept of indirect utility and are used to assess how changes in prices affect consumer welfare.

Key Computational Methods

The authors examine various computational approaches to estimate welfare changes:

• Taylor's approximations (first and higher order)
• Vartia's method
• Breslaw and Smith's method
• Ordinary differential equations methods
• A simple method based on known elasticities

The paper emphasizes the importance of understanding the relationship between these methods and the welfare measures they approximate.

Key Findings

1. Convergence of Welfare Measures: For price changes below 10%, welfare measures tend to converge to similar results, regardless of parameter choices.
2. Divergence for Larger Changes: Above 10%, these measures begin to diverge significantly, with the degree of divergence being influenced more by budget shares than by the choice of demand system.
3. Laspeyers and Paasche as Bounds: Under standard utility assumptions, Laspeyers (LV) and Paasche (PV) variations are always the outer bounds of welfare estimates, while Consumer's Surplus (CS) is typically the median estimate.
4. Decomposition of Effects: The paper clarifies the decomposition of higher-order Taylor approximations into substitution and income effects.
5. New Welfare Approximation: A new simple approximation method is introduced, which is based on known elasticities and is useful for practical applications.

Applications and Relevance

The paper is relevant for:

• Empirical economists analyzing the impact of price changes on welfare.
• Policy makers interested in assessing the effects of economic shocks or policy reforms on social welfare.
• Researchers working with micro data, especially in poor countries and developing economies, where data on consumer behavior may be limited.

Methodological Contributions

• The authors compare the behavior of different welfare measures under various demand systems (Cobb-Douglas, Linear Expenditure System, Almost Ideal Demand System, Quadratic Almost Ideal Demand System, and Exact Affine Stone Index).
• They highlight that the difference in welfare measurement is minimal compared to changes in other parameters like the price change or budget share.
• The marginal approach is used, which assumes that the budget constraint remains fixed under price changes, implying short-term decisions and no inter-temporal effects.

Limitations and Justifications

• The paper excludes income, supply, partial or general equilibrium effects to simplify the analysis.
• These assumptions are justified on the grounds that they are particularly relevant for poor consumers and developing countries, where savings are minimal and consumption is primarily based on current income.
• The authors also note that practitioners and international organizations often overlook the importance of choosing the right estimation method, which can significantly affect the results.

Conclusion

The paper concludes that the choice of welfare measure and computational method can lead to substantial differences in welfare estimation. It recommends that researchers and practitioners be aware of these differences and choose methods accordingly, especially when dealing with nonlinear pricing and limited data availability.

Computational Tools

• Stata codes are provided for all computations and welfare measures discussed in the paper.
• These codes allow for the application of the methods in empirical research and policy analysis.

References and Context

• The paper is part of the Poverty and Equity Global Practice Group at the World Bank.
• It builds on previous contributions by economists such as Hicks, Harberger, Slesnick, and Fleurbaey.
• It also references nonlinear demand systems and recent empirical studies that have attempted to estimate demand systems directly from data in developing countries.

Summary of Key Equations

• Consumer Surplus (CS):
  $$
  CS = \int_{p^a}^{p^b} D(p) dp
  $$
• Compensating Variation (CV):
  $$
  CV = e(p^a, v^a) - e(p^b, v^a) = \int_{p^a}^{p^b} h(p, \nu^a) dp
  $$
• Equivalent Variation (EV):
  $$
  EV = e(p^a, v^b) - e(p^b, v^b) = \int_{p^a}^{p^b} h(p, \nu^b) dp
  $$
• Laspeyers Variation (LV):
  $$
  LV = e(p^b, x^a) - e(p^a, x^a)
  $$
• Paasche Variation (PV):
  $$
  PV = e(p^b, x^b) - e(p^a, x^b)
  $$

Figures and Geometric Interpretation

• Figure 1 provides a geometric interpretation of the five welfare measures.
• It illustrates how each measure corresponds to a specific area or distance in the budget constraint diagram.
• The relationships between the measures are summarized as:
  $$
  LV < CV < CS < EV < PV
  $$
  and
  $$
  2x^a > (x^a + x^c) > (x^a + x^b) > (x^e + x^b) > 2x^b
  $$

This paper serves as a comprehensive guide for understanding and applying different welfare measures and computational methods in the context of price changes and their impact on welfare.