On back-testing "zero-investment" strategies

''Zero-investment'' strategies typically involve forming a long portfolio in one set of securities and a short portfolio in another, with the two identified by the use of some trading rule. When back-testing such strategies, the ex post difference in their abnormal returns represents an unbiased estimate of their economic profitability in perfect markets and thus serves as an indication of market efficiency. However, Regulation T's short-selling constraints are a potentially significant imperfection facing retail and institutional investors. After development of a procedure that recognizes these constraints in measuring returns, the impact of these constraints on economic profitability is analyzed.