Potential Pilot Problems: Treatment Spillovers in Financial Regulatory Experiments

Abstract

The total effect of a regulatory change consists of direct effects and indirect effects (spillovers), but the standard difference-in-difference approach measures only direct effects and ignores potential indirect effects. By examining the short-sale aggressiveness during the 2007 full repeal of the uptick rule by the SEC, we find that short sellers become much more aggressive across the board, even in control stocks where the uptick rule is already suspended, which is consistent with positive and significant indirect effects on control stocks. In contrast, for the 2005 partial uptick repeal, short sellers become more aggressive in treatment stocks without an uptick rule, and less aggressive in control stocks with an uptick rule in place, which is consistent with negative indirect effects. We provide supportive evidence that the positive indirect effects in 2007 might be driven by aggressive broad list-based shorting, which includes both control and treatment stocks, and the negative indirect effects in 2005 might result from substitutions between control and treatment stocks. We conclude that regulatory pilot designers should carefully consider potential spillovers.

Keywords: interference, short sales, short interest, tick test, Regulation SHO

JEL Classification: G14