摘要:The aim of the study was to estimate the trend in unregistered alcohol consumption in Sweden during the period 1960–1994. The estimation was based on the discrepancy between the observed trend in alcohol related harms (alcohol-related mortality, assault, homicide, and drunk driving) and the trend expected from registered alcohol sales. The impact of alcohol sales on these indicators were estimated by means of ARIMA-analyses, and then the error term series was calculated for each indicator, meaning the difference between the observed and the expected trends in harm rates. A factor analysis of the four error-term series yielded one single factor which was interpreted as the unregistered alcohol consumption. According to the outcome this would have increased linearly by about 80% during the study period. Some comparisons are made with other data; for example it is noted that the trend in travelling abroad reasonably matches the trend in estimated unregistered consumption.