INTRODUCTION

  Our figure for the loss of indirect tax revenue was the one that was the most recently available from Customs and Excise at the time the paper was written. However, the measure of indirect revenue lost through cross-border shopping is merely illustrative. Our analysis looks at the relationship between domestic tax rates and the domestic tax base (measured by domestic sales in the Family Expenditure Survey). Estimating the effect of a tax change on indirect tax revenue does not require us to have an estimate of the scale of legitimate cross-border shopping or smuggling since both will be captured by changes in the domestic tax base.

LONG-TERM TRENDS IN ALCOHOL SPENDING

  It is a well-known fact that the Family Expenditure Survey does not capture aggregate spending on alcohol. Some of this is due to survey design—the survey does not sample groups with very high levels of alcohol consumption. Some of the under-recording may be due to respondents failing to report their true level of spending. However, what matters for our analysis of changes over time is that the degree of under-reporting is constant over time and we have good reason to believe this to be the case. The grossed-up FES data has been shown to capture a constant proportion of total spending on alcohol over the period 1978-92[4]. A preliminary analysis of consumer spending for the period 1986-95[5] confirms that the trends in the aggregate data[6] for spending on beer and spirits are not significantly different from those in the FES. This is shown by the fact that the percentage change in aggregate spending over the period 1986-95 lies within the confidence interval[7] for the change in average spending over the same period in the FES. For wine, the FES does not capture all of the increase in total spending over the period (see Table 1).

  Our analysis focuses on the relationship between domestic tax rates and the domestic tax base, measured by domestic spending on alcohol in the FES. To the extent that both legitimate cross-border shopping and smuggling reduce the domestic tax base our analysis takes account of both of these. Our potential concern was that the FES data might not be able to take proper account of commercial smuggling—ie goods that were smuggled and re-sold domestically—and we therefore thought it appropriate to include footnote 5 as a caveat to our analysis.

  By a similar argument, if the change in the duty-free allowances on spirits had an impact on the domestic tax base this effect would have been taken into account in our analysis.

TAX RATES AND TAX REVENUES

  We do not think that our table of typical tax rates and prices is heavily biased against beer. We use the on-licence price for beer since that is where most beer is sold (a point made later on in the All-party Parliamentary Beer Group's briefing).

  Our discussion of the externalities of alcohol consumption outlines a theoretical argument why policy-makers might impose additional taxes on alcohol. Of course, there is a further set of arguments about the size of those potential social costs (or benefits) which we do not address here (and which health experts might be better placed to address than economists).

  Similarly, our argument about how different types of alcohol should be taxed is also a theoretical one. If the social costs are directly related to the amount of alcohol consumed than all alcohol should be taxed in the same way. If the concentration of alcohol also matters then there is a case for taxing alcohol in different forms at different rates according to the degree of alcohol concentration. Again, this is an issue for health experts not economists. However, we did contact Alcohol Concern and their view was that amount of alcohol mattered, not the degree of concentration.

  The study by Crawford and Tanner did take account of the impact of cross-border shopping. The paper looked at the link between domestic tax rates and the size of the domestic tax base and the size of the domestic tax base is clearly affected by cross-border shopping. If after the Single Market consumers respond to tax changes by increasing their cross-border shopping there will be a corresponding increase in the observed responsiveness of the size of the domestic tax base to changes in the tax rates. However the study found no evidence of a significant increase in demand responsiveness for beer following the Single Market.

  The analysis does not include foreign prices in the regression directly since an official price series was not available to use. However, omitting foreign prices will not affect the measured relationship between domestic tax rates and the domestic tax base so long as changes in domestic and foreign prices are not significantly correlated. In any case it is not possible to say, a priori, the direction and extent of any bias that would follow from such a correlation.

  We focus on the relationship between indirect tax rates and indirect tax revenues since we do not have reliable data on the effects of changes in demand for alcohol on other revenues from income tax or corporation tax. However, we have extended our analysis to make preliminary estimates of how much "other revenue" would have to be gained for each additional unit sold to make a cut in the indirect tax rate overall revenue-enhancing. Our preliminary estimates showed that for beer the amount of "other revenue" per pint would have to be greater than the price of each pint.

  We have argued in the past that demand responsiveness might differ between on- and off-sales and that there is a reason for thinking that demand for off-licence sales may be more price responsible since they are a closer substitute for cross-border shopping. However, we did not have a sufficiently long time-series of data to allow a separate analysis of on- and off-sales.

  Our most recent estimates of the own-price elasticities of demand are not statistically significantly different from the earlier SPIT estimates.

  Our results for whether the current tax rates are revenue-maximising take the standard errors on the elasticity estimates into account—and should therefore be seen as a fairly robust set of results. The results that you get a positive revenue effect from increasing the rates of duty on beer and wine are statistically significant even though the standard errors on the elasticity estimates are fairly large, implying that you would get the same revenue results for a fairly wide range of elasticity estimates.

  Our analysis measures the relationship between indirect tax rates and tax revenues controlling for other factors which may affect demand for alcohol (such as income). Of course there may have been changes in demand due to factors other than the price which caused a change in revenues.

5   A period for which we have aggregate spending figures on beer, wine and spirits readily available. Back

6   Taken from Consumers' Expenditure, ONS Back

7   Since the FES is a sample of the population, average spending is only an estimate of the underlying population spending. The confidence interval places an upper and lower bound around the sample averages to show where the true population figures lie. Back