Objectives
In developing countries, pricing policy plays the same central role in fiscal policy that income tax and social security plays in developed countries (Deaton, 1989, World Bank Economic Review). For example, some Asian developing countries imposed export restrictions during the recent food price crisis, to protect domestic consumers at the expense of local farmers – 39% of world rice trade was covered by such restrictions (Giordani et al, 2012, Voxeu). In order to examine the effect of such policies, and of indirect taxes more generally, price elasticities of demand are needed. But many developing country time series are too short to estimate demand systems so variation in prices over space is used instead to identify demand responses.
A crucial but largely untested assumption needed to estimate these demand responses is Hicksian separability, whereby all relative prices within a group remain constant (e.g. different cuts of beef within the beef group). Under this assumption, only one price is needed per commodity group since all prices move together. This assumption is especially required when demand analysis uses unit values (household expenditures on a grouped good divided by the quantity purchased) as proxies for market prices. Demand parameters estimated from these unit values are often used in evaluations of indirect tax reforms. But except for a recent study by McKelvey (2011, Journal of Development Economics) this key assumption of Hicksian price separability over space has not been tested. The weaker assumption of stochastic Hicksian separability (or the generalized composite commodity theorem) of Lewbel (1996, American Economic Review) has also not been tested. With this weaker assumption, when the changes in the relative prices of goods within a group are unrelated to the movements of a group price index, the aggregation can be allowed in some specific demand specifications. The goal of the proposed paper is to test those assumptions in the context of demand estimation for analyzing indirect tax reform.
Data/Methods
We use unique data which combines a conventional household survey, the 2010 Vietnam Household Living Standards Survey (VHLSS), and market price data from a spatial cost of living survey fielded in conjunction with the VHLSS. Multiple specifications for the same food group allow a direct test of the fixed price relativities assumption. We also form a price index for each food group for a test of stochastic Hicksian separabilty theorem and its consequence in a tax reform application and an appropriate specification for the analysis will be investigated.
Expected Results
The expected outcome is to improve our general understanding of the validity of Hicksian price separability theorems over space with test results for two theorems. Specifically, our results will suggest appropriate data types and estimation methods in the methods used for estimating price elasticites from household survey data in a tax reform application.
JEL Category Number and Title:
D12 Consumer economics
Secondary JEL Category:
R10 Urban, Rural and Regional Economics