Alcohol consumption declined by about 10.5% in Quebec and 2.5% in Ontario but increased by 9.5% in Alberta in the years 1974-1983. This created a 'natural experiment' in which the effects of various changes in alcohol consumption on alcohol problems could be assessed. Declines in rates of most alcohol problems were found in Quebec and Ontario despite their different rates of decrease in consumption. However, there were declines in alcoholism rates, and deaths from liver disease and the alcohol dependency syndrome in Alberta. Some problems such as impaired driving and toxic or accidental deaths from alcohol increased in Alberta, probably because of its relatively youthful population compared to Ontario and Quebec. Changes in alcohol consumption appeared to be an unreliable indicator of how various problems are changing.