2018年-世界发展银行全球_Analysis_of_the_Mismatch_between_Tanzania_Household_Budget_Survey_and_National_Panel_Survey_Data_in_Poverty_and_Inequality_Levels_and_Trends_50页_1mb

Summary of the Mismatch between Tanzania Household Budget Survey and National Panel Survey Data in Poverty & Inequality Levels and Trends

Core Content

This study investigates the discrepancies in poverty and inequality levels and trends between the Tanzania Household Budget Survey (HBS) and the National Panel Survey (NPS). The research aims to identify the methodological and data collection differences that may contribute to these mismatches and their implications for policy analysis.

Main Findings

• Poverty Levels:

  • The difference in poverty levels between HBS and NPS is primarily due to the different methods used to estimate the poverty line.
  • Using the Consumer Price Index (CPI) to adjust consumption variation across years, both surveys show a decline in poverty over the past five years.
  • The mean household consumption levels in the last rounds of the surveys are comparable when CPI is used for adjustment.
  • HBS 2011/12 and NPS 2010/11 show similar poverty rates, while NPS 2012/13 indicates a higher poverty rate (around 10 percent higher than HBS 2011/12), driven by higher consumption among better-off groups in the NPS.

• Poverty Trends:

  • The HBS data show a declining trend in poverty from 2007 to 2012, while NPS data indicate an increasing trend from 2008/09 to 2012/13.
  • At the regional level, the poverty rate estimated using NPS is significantly lower than that using HBS, particularly in rural and urban areas.
  • HBS suggests a decline in poverty across all regions, whereas NPS only shows a decline in Dar es Salaam and Zanzibar.

• Inequality Trends:

  • HBS 2011/12 shows a decline in inequality (Gini index from 39 to 36), while NPS data indicate a stagnation or increase in inequality.
  • The Gini index for NPS 2012/13 rises to around 41, suggesting a deterioration in welfare distribution.
  • Inequality is higher in urban areas than in rural areas according to both surveys.
  • The consumption share of the poorest quintile declined by around 12 percent nationally and 11 percent in rural areas, while it increased in urban areas, especially in Dar es Salaam.

• Pro-Poor Growth Patterns:

  • The HBS shows a downwardly sloped Growth Incidence Curve (GIC), indicating higher growth among the poorest.
  • In contrast, the NPS GIC is upwardly sloped, suggesting that richer groups benefited more from growth.
  • This divergence in growth patterns between the two surveys is a subject for further analysis.

Key Survey Methodological Differences

• Data Collection Methods:

  • HBS:
    • Uses a diary method for 28 days to capture consumption data.
    • Includes more detailed items (193 food items and around 250 non-food items).
    • Distinguishes between expenditure and consumption.
  • NPS:
    • Uses a recall method for 7 days.
    • Includes fewer items (around 59 food items and 46 non-food items).
    • Does not distinguish between expenditure and consumption.

• Consumption Aggregates and Price Deflators:

  • The discrepancy in poverty trends is mainly attributed to differences in inter-year temporal price deflators.
  • Spatial price deflators also play a minor role.
  • The use of CPI to adjust for inter-year price variations results in similar poverty levels in the last rounds of the surveys.

• Sample Size and Coverage:

  • HBS has a larger sample size and is restricted to mainland Tanzania.
  • NPS has a smaller sample size but covers Zanzibar.
  • The population estimates from NPS are close to the actual census figures, and the differences in population size between the surveys do not significantly affect poverty and inequality levels.

• Attrition and Representativeness:

  • NPS has attrition rates of 3% and 4%, which are low enough not to bias poverty and inequality estimates significantly.
  • Propensity score matching is used to address attrition in NPS.

Conclusion

The mismatch in poverty and inequality levels and trends between HBS and NPS is attributed to methodological differences in poverty line estimation, price deflators, and data collection approaches. Despite similar mean consumption levels in the last rounds, the divergence in inequality trends and pro-poor growth patterns remains unresolved and requires further investigation. The study highlights the importance of understanding these differences to ensure accurate policy analysis and effective poverty reduction strategies.