This paper examines the empirical performance of alternatives to the monetary aggregates currently published by the Bank of Canada. The results show that real M1 and real M1a perform about equally well in providing leading information about real output at short horizons. However, on theoretical grounds, M1a is a more attractive aggregate, since it excludes federal government deposits held at trust companies (M1 excludes only such deposits held at banks). Also, the broad aggregates—M2+, M2d (the sum of M2+, treasury bills, provincial savings bonds, and Canada Savings Bonds), M3b (the sum of M2d, mutual funds, 1- to 3-year government bonds, mortgage-backed securities, foreign holders of Canadian dollar deposits, and foreign currency deposits of residents booked in Canada), and LLb (the sum of M3b, bankers' acceptances, and commercial paper issued by non-financial corporations)—generally are the best in providing leading information about inflation at long horizons (one to two years) across the various models that were considered. Also, in the context of P* models of inflation, M1 performs well as a leading indicator of inflation with long leads. Based on the analyses carried out in this paper, the aggregates M1, M1a, M2+, M2d, M3b and LLb are found to be deserving of further attention.