INTRODUCTION

  The IFS analysis "looks at the relationship between domestic tax rates and the domestic tax base (measured by domestic sales in the Family Expenditure Survey". While domestic sales in the FES account for the vast proportion of the domestic tax base they do not account for the total nor even a constant proportion of this over time. For example the growth in incomcong foreign visitors, which have more than doubled to 25 million since 1983, means that they now account for an appreciable proportion of total excise duty paid on alcohol in the UK but would not be reflected in the FES. Similarly the FES does not cover significant spending on beer, say, by students living in halls of residence—the number of full-time students has more than doubled to 2.1 million since 1984—which may also affect trends.

  We agree with the IFS that estimating the effect of a tax change on indirect tax revenue does not require an estimate of cross-border shopping and that the "measure of the indirect revenue lost . . . is merely illustrative". Our purpose in quoting the more recent data was merely to appraise the Committee of the most up to date data. In this respect we are pleased to further update our earlier numbers with new data estimates issued by HM Customs & Excise. These show that tax losses from legitimate cross-border shoppping for alcohol grew to £290 million in 1998 versus the £185 million quoted by IFS for 1996.

LONG-TERM TRENDS IN ALCOHOL SPENDING

  We agree with the IFS that "what matters for their analysis of changes over time is that the degree of under-reporting (of the FES data) is constant over time". However with rapid changes in the beer market in recent years we are not convinced that this remains true and indeed we believe that the proportion may be falling rather than constant. The following quotes from the 1993 and 1998 FES reports are also relevant:

1993  ". . . the estimated average expenditure . . . in the FES on beer is about two-thirds of corresponding estimates from statistics produced by HM Customs & Excise".

1998  ". . . the estimated average of all households on beer is rather over half of corresponding estimates produced by HM Customs & Excise".

  We also quote from the 1996 FES report "there is some evidence that the non-repsonse bias in alcohol expendidture increased in 1995-96 due to higher non-response. Changes to the wording in the expenditure diary may have . . . lead to an increase in the amount recorded as `other drinks' rather than specific items eg wine and beer". Again this does not suggest the FES is a sufficiently robust series for the purpose in hand.

  We also note that the ONS do not use the FES series to calculate Household Spend on beer since this is less accurate than using HM Customs & Excise data and market survey data.

  We therefore repeat the statement from our original commentary that real consumer spend on beer fell by—20 per cent between 1979 and 1996 and this does not reconcile with the IFS FES based analysis showing "an average real increase . . . of less than half a per cent each year" in this period. This difference between the actual growth and the IFS calculation is well outside the c +/-6 per cent confidence intervals calculated by the IFS.

  We also note that the latest IFS analysis for 1986-95—a different time period from their original analysis—produces results which are on the very limit of confidence for beer. The IFS show an average growth of -4 per cent (within a range of +2 per cent to -10 per cent) in comparison to the actual decline of -10 per cent in Consumer Spend. The position is similar on spirits where the actual decline is -17 per cent versus IFS average of -23 per cent (in a range between -16 per cent and -30 per cent). The wine data from the IFS is significantly different from the Consumer Spend at +32 per cent (IFS average at +15 per cent in a range between +5 per cent +25 per cent).

TAX RATES AND TAX REVENUES

  We accept that most beer is sold in pubs. We do not accept that it is valid to compare the duty plus VAT proportion of the retail price in the on-trade channel for beer against off-trade wine/spirits. It is valid to compare the tax proportions on different drinks within each channel but not across channels.

  We agree completely with the IFS economists that the externalities of alcohol consumption are "an issue for health experts and not economists". If economists do venture in to such specialised areas then they should do so with trepidation and take care to ensure that they present a balanced view of opinion.

  We are pleased to note the IFS accept that other tax revenues—eg income tax and corporation tax—have a bearing on the overall analysis but note that they did not have the date available to take this in to account. These factors were considered in an earlier analysis carried out by Oxford Economic Forecasting on behalf of the BLRA. This showed that a reduction in beer duty yield, as a result of a duty cut, would be substantially offset even in the first year by additional beer sales and other tax income.

  We also note that the IFS estimate the "other tax revenue" needed to offset a cut in beer duty "would have to be greater than the price of each [additional] pint". The implication here is that this is so unrealistic as to be unbelievable. We would point out that other economic analysis has shown that each £1 of beer duty earned in a pub supports c £4 in other taxes, most of which and are proportional to beer volumes sold. Thus the requirement to earn more in extra tax than the apparent revenue generated by extra beer sales is not necessarily as outlandish as it seems.

  We agree with the IFS that their most recent own-price elasticities of demand are not significantly different—in the statistical sense—from the SPIT figures ie the "fairly large" standard errors encompass both estimates. We merely note that they give different interpretations of which product is the more price elastic and thus, perhaps, suggest that any such interpretations need to be treated cautiously.