Ungdoms etterspørsel etter alkohol. En empirisk analyse basert på intervjudata 1990-2004

NORSK SAMMENDRAG: I rapporten studeres ungdoms forbruk av alkohol. Forhold som påvirker om og eventuelt hvor mye ungdommer drikker blir analysert. Sammenhengen mellom ungdommers alkoholforbruk og priser, og mellom forbruk og aldersgrenser blir studert spesielt.\ud \ud Undersøkelsen bygger på svar fra 32 000 ungdommer fra 15 til 20 år samlet inn fra 1990 til 2004. I denne perioden har ikke andelen som noen gang har drukket alkohol endret seg. Debutalderen har også vært stabil på ca 15 år. Alkoholbruken blant ungdom som har drukket har imidlertid økt betydelig. Det samlede konsumet har økt med 50 prosent, fra et gjennomsnitt på 0,18 liter ren alkolhol per måned i 1990 til 0,27 liter ren alkohol i 2004.\ud \ud Rapporten konkluderer med at både aldersgrensen og avgiftspolitikken ser ut å virke inn på ungdoms alkoholforbruk. Prisen har sannsynligvis større betydning for hvor mye som drikkes enn om ungdom velger å drikker i det hele tatt. Aldersgrensen ser ut til å ha sterkest effekt for kjøp av øl og vin, og der er også for øl at man finner klare effekter av avgiftspolitikken.ENGLISH SUMMARY: Official policy in Norway seeks to restrict the consumption of alcohol, especially by young people. This report looks at youth consumption patterns with regard to beer, wine and spirits. We probe factors affecting the decision to drink over the past month and how much is drunk. We also investigated other aspects of young people’s drinking patterns, such as changes in the proportion claiming to have touched alcohol at some time; whether there was an increase in the number of young people claiming to have been drunk; whether alcopop’s share of young people’s total consumption had grown; dealings with moonshining etc.\ud \ud There has been little scientific focus on the impact of excise duties on youth drinking patterns either by the Norwegian or indeed international research community. But given that alcohol taxes are considered a prime policy instrument because they directly affect retail prices, we were particularly interested to explore the relationship between the price of alcohol and youth alcohol consumption patterns. We investigated the importance of price changes on young people reporting having drunk in the past month; the extent to which alcohol consumption of those who drink is affected by price changes; we studied drinking patterns in relation to age limits enacted by the authorities on the purchase of different alcoholic beverages. Age limits make it more difficult to obtain alcohol, and represents for potential under-age shoppers an added transaction cost.\ud The body of the data used in this analysis was collected by SIRUS. Between 1990 and 2004 a total of 32,000 young people (15–20), resident in Norway, were questioned about drugs and drink. The surveys included questions about tobacco, alcohol and drug use, as well as obtaining information on standard variables like sex, age, education, domicile etc.\ud We found no change in the period covered by this study (1990–2004) in the number of respondents reporting having ever drunk alcohol, but we did discover a change in young people’s drinking preferences. The number reporting having ever drunk beer, wine or spirits fell; we found however high alcopop consumption levels in 2003–04 (the years for which we have data on alcopop), significantly higher in fact than for the other types of drink. The alcohol debut age remained stable over the period (at about fifteen), whether the drink in question was beer, wine or spirits.\ud \ud On the other hand, alcohol consumption by youngsters who drink rose significantly, as much as 50 per cent overall, and ranging from an average 0.18 litres pure alcohol per month in 1990 to 0.27 litres pure alcohol in the past couple of years. Spirits and alcopop saw a particularly steep rise. Those reporting having drunk wine, spirits and alcopop recently increased in number, as did the frequency of drinking events. The percentage reporting having drunk in the past four weeks rose from about 60 per cent at the start of the period to about 66 per cent by 2003/04.\ud \ud Alcopop was introduced to the Norwegian market in 1998, but did not really take off until high street retailers were allowed to sell it over the counter. By 2004, alcopop represented 40 per cent of young girls’ (eighteen) reported alcohol consumption; the percentage was lower for slightly older girls and boys in the same age-group. Beer was the most popular drink among boys by a good margin. About 75 per cent of our respondents had been drunk at least once (68 per cent within the past six months), and nearly 20 per cent more than fifty times. Frequency of drunkenness reported by the girls in our sample is no different from the boys.\ud \ud As expected, the data show a rise in the average consumption of beer, wine, spirits and alcopop with age. We find the sharpest rise in beer, wine and alcopop consumption when the youngsters turn eighteen; beer consumption increases most. Beyond the age of eighteen, alcohol consumption continues to rise, but not as rapidly as with the lower age-groups. The over-eighteens drank on average twice as much beer as the under-eighteens. Surprisingly, we found the sharpest rise in the consumption of spirits across the fifteen–sixteen divide.\ud \ud Two methods are used in the report to analyse the data. We wanted separate analyses of whether the youngsters had drunk over the past month and how much they (i.e. those reporting having drunk at all) had drunk. As expected, several variables, such as sex and age for instance, were shown by the two methods to affect both the decision to drink beer, wine and spirits, and quantities, and in the same direction. One factor which influences the decision to drink then also affects how much alcohol is consumed. The effects of other variables – such as place of residence – were not as transparent.\ud \ud We used method 1 to analyse factors affecting the decision to drink beer, wine and spirits in the month before completing the questionnaire. The data do not include information on every factor known to influence decisions to drink (we lack data on, for instance, peer pressure, disposable income etc.), and for that reason we used an analytical method (multivariate probit) which takes at least some account of such non-observed factors. The impact of price on the likelihood of drinking is not very great according to our findings. The low effect of price could mean other factors play a greater role in young people’s decisions to drink, or that the prices provided by outlets and the Vinmonopol failed to represent the prices young people actually pay. We did find a positive correlation between high price and decreased likelihood of beer drinking, but it was not significant. We also found a positive correlation between the likelihood of drinking wine and the price of wine, probably explained by Norway’s changing drinking habits in recent years, giving wine an increasingly dominant place despite increasing prices. In the model for spirits, the relationship between price and likelihood of drinking spirits was not significant. This is probably because many of our respondents were not paying normal Vinmonopol prices for the spirits they consumed. There is evidence to suggest that a significant share of the spirits consumed by young people is either paid for by others, purloined from parental stocks or illegally manufactured/traded.\ud \ud From the range of available alcohol beverages, boys are most likely to drink beer and spirits, while girls tend to prefer wine. Respondents with part-time work were more likely to drink beer, wine and spirits than those without a paid job. And respondents holding a full-time job were more likely to consume spirits. Living with one or both parents resulted in a lower likelihood of drinking beer, wine and spirits, in contrast to living apart from their parents.\ud \ud Using method 2 we analysed alcohol consumption of respondents claiming to have consumed alcohol during the month prior to the questionnaire. We made allowances for the fact that these respondents make up a distinct sub-group of the sample per se, and adopted an analytical approach (Heckman’s two stage estimation procedure) which allowed us to generalise our findings across the entire youth population. The estimates on price sensitivity and our reading of them differ according to the type of beverage in question