The Effects of an Increase in Petrol Excise Tax: The Case of New Zealand Households
Creedy, John, National Institute Economic Review
This paper reports estimates of the potential welfare effects of hypothetical increases in the petrol excise tax in New Zealand. Equivalent variations, for a range of household types and total expenditure levels, are obtained along with distributional measures. Household demand responses are modelled using the Linear Expenditure System, where parameters vary by total expenditure level and household type. The effects on inequality were found to be negligible, but the marginal excess burdens typically ranged between 35 and 55 cents per dollar of additional revenue.
This paper examines the potential welfare and redistributive effects of a hypothetical increase in the petrol excise tax in New Zealand. Equivalent variations and excess burdens are obtained for a variety of household types at a range of total expenditure levels. Households are divided into a number of categories according to their composition. Also, overall summary measures of the distributional effects are reported using a money metric utility measure. Changes in indirect taxes give rise to changes in prices, assuming tax shifting to consumers, so the core of the modelling involves a method of examining the welfare effects of price changes. Household demand responses to price changes are modelled using the linear expenditure system (LES), with the modification that parameters are allowed to vary by total expenditure level and household characteristics. (1) The demands are assumed to vary in a partial equilibrium context. Thus, possible changes in production (associated with the changing structure of demands) and factor prices, along with the distribution of income, are ignored. (2)
The present analysis contrasts with studies of indirect taxes that are based on computations of tax paid, and expenditure net of taxes, for various household types and income levels, using the assumption that expenditure patterns remain fixed when taxes, and hence prices, change. It seems worthwhile, despite the use of strong assumptions about consumer demands, to model behavioural responses and, in so doing, to obtain welfare measures such as equivalent variations. Furthermore, a central concept in public finance is that of the excess burden of taxation, reflecting the inefficiency caused by the distortions arising from taxes. Rational policy analysis requires that such inefficiencies are compared with any redistributive (equity) or other effects.
Section 2 describes the relationship between tax rates and price changes. This is complicated in the present context by the fact that the goods and services tax (GST) is imposed on the excise-inclusive price, and the excise is a unit tax rather than an ad valorem tax. The measurement of welfare changes and the way in which household demands are modelled are discussed in section 3, and further details of parameter estimation are given in an Appendix. The approach is used in section 4 to examine the potential implications of a tax reform involving an increase in the petrol excise tax. Conclusions are in section 5.
2. Indirect taxes and price changes
This section describes the New Zealand indirect tax structure and the link between excise tax changes and proportional price changes, on the assumption that tax increases are fully passed to consumers. It is necessary to express all indirect taxes in terms of a tax-exclusive ad valorem tax rate. While this is straightforward for most commodity groups, for which only GST applies, the translation is more awkward where an excise tax is also imposed, since these are typically based on units of the commodity rather than values.
The analysis uses Household Economic Survey (HES) data (described further in section 4). However, it is not possible, mainly because of estimation difficulties, to use all the separate and highly detailed HES commodity categories. Instead, these were consolidated into 22 groups. Table 1 shows the commodity groups used and the effective ad valorem tax-exclusive percentage rates at 2001 (the last year for which HES data are available). …