Decomposing of variation in poverty in growth and redistribution components

	Distributive indices and statistical inference
	By distributive indices, we mean indices that quantify poverty, inequality, etc. Statistical inference enables us to: Estimate the confidence interval of an index. When the confidence interval is narrow, the precision of the estimate is high. Test a null hypothesis. For instance, we may want to test if a variation in poverty or inequality is positive. To do statistical inference, we need to: Estimate the value of an index Determine the sampling distribution and the standard error of the estimator of the index. The sampling distribution can usually be shown to be asymptotically normal. Besides the asymptotic distribution, we can also use simulation methods, such as the bootstrap.
	References:
	Duclos, J. and Araar A. (2004) Poverty and Equity: Measurement, Policy and Estimation with DAD Statistical inference with complex sample design Araar A. and Duclos J., (2004) Users Manual for DAD 4.4 Sampling Design Computation of standard errors in DAD Standard deviation, confidence intervals and hypothesis testing
	DAD Application
	The free software DAD developed by our research center can help you to perform statistical inference: Application: For the applications Poverty\|FGT index, Inequality\|Akinson or Gini index, press the SD STD button. For the other applications, also use the application Distribution\|Confidence Interval

	Exercise