Do the Disposition and House Money Effects Coexist? The differential impact of realized and unrealized gains and losses Songyao YAO a, Darren DUXBURY a,*, Robert HUDSON b and Kevin KEASEY a a Centre for Advanced Studies in Finance (CASIF), Leeds University Business School, The University of Leeds, Leeds, LS2 9JT, UK b Newcastle University Business School, Ridley Building, Queen Victoria Road, Newcastle upon Tyne, NE1 7RJ Abstract – short Contrary to the assumptions of rational economic behavior, there is substantial evidence that prior outcomes impact on financial behavior. Two important behavioral biases illustrating the impact of prior outcomes have been documented in the financial markets; namely the disposition effect and the house money effect. The two effects, however, are premised on contradictory behavior, with the disposition effect requiring decreased risk taking behavior following prior gains (selling winners) and increased risk taking following prior losses (holding losers), while the house money effect requires increased risk taking following prior gains and reduced risk taking following losses. Even after aggregating the evidence from numerous empirical studies looking at one or other of these behavioral biases, given the disparate datasets employed it is not possible to comment on whether the two effects can contemporaneously coexist in a single financial market, whether individual investors might succumb to both types of behavior or whether the two effects might interact with one another. To address these important questions requires a simultaneous examination of both biases using the same dataset, but this is missing in the literature currently. This paper fills this gap in the literature and is the first to show that the disposition and house money effects can contemporaneously coexist in a single stock market, that investors can simultaneously succumb to both types of behavior and that the house money effect moderates the disposition effect. In addition, in reconciling such behavior we show the importance of distinguishing prior outcomes along two dimensions; 1) unrealized or realized and 2) assessed at the stock or portfolio level. * Corresponding author: email dd@lubs.leeds.ac.uk, telephone +44 (0) 113 343508 1 Abstract – long For decisions based on rational economic principles (subjected expected utility theory) the carrier of value is terminal wealth and individuals are assumed to be indifferent as to how they arrive at this terminal wealth (i.e. changes of wealth do not factor in the determination of utility). Thus prior outcomes should be irrelevant to individuals’ future decisions. However, there is a growing body of research suggesting that prior outcomes impact upon decision making via their influence on individuals’ risk attitudes. In a finance context there is empirical evidence that investors are influenced by prior outcomes (i.e. gains and losses), displaying a disposition to sell winning shares too soon and hold losing shares to long, thus appearing to be risk averse in the presence of prior gains and risk taking in the presence of prior losses (consistent with prospect theory). There is also evidence, albeit less substantial, of contradictory behavior whereby investors exhibit risk taking behavior following prior gains and risk averse behavior following prior losses. So how can such seemingly contradictory risk taking behavior, termed the disposition effect and the house money effect, respectively, be reconciled? Neilson (1998) proposes a three-argument utility function that can account for the coexistence of both types of behavior. Two implications of Neilson’s (1998) approach are that it must be possible for the two effects to contemporaneously coexist in a single financial market and that a single investor could simultaneously exhibit behavior consistent with both effects. Even after aggregating the evidence from numerous empirical studies looking at one or other of these behavioral biases, given the disparate datasets employed it is not possible to comment on whether the two effects can contemporaneously coexist in a single financial market, whether individual investors might succumb to both types of behavior or whether the two effects might interact with one another. To address these important questions requires a simultaneous examination of both biases using the same dataset, but this is missing in the literature currently. This paper fills this gap in the literature using detailed transaction and position information for all investors registered with a stock brokerage in Beijing. We find strong evidence that the disposition and house money effects do contemporaneously coexist in a single stock market and that a substantial proportion of investors simultaneously succumb to both types of behavior. 2 We draw an important distinction between the differing roles played by unrealized and realized prior outcomes, with the former associated with the disposition effect and the latter the house money effect. Moreover, whether the impact of the prior outcomes effect on observed risk attitudes is investigated at the individual stock level or at the aggregate portfolio level is an important determinant of whether we observe behavior consistent with the disposition effect or house money effect, with the former only observed at the individual stock level. Furthermore, there is evidence that the house money effect mediates the disposition effect, as we observe lower values of the disposition effect measure in the presence of prior realized gains than in the presence of prior realized losses. Investor characteristics also have a role to play, with institutions, more sophisticated investors, investors with higher wealth levels, and investors holding more diverse portfolios, all being less prone to the behavioral biases than their counterparts. 3