In this thesis, I use two strategies of inquiry to further our understanding of indirect short-selling constraints. First, I interview a series of experienced market practitioners to identify their attitudes towards indirect constraints. I find little support for D’Avolio’s (2002) suggestions that short-selling is inhibited by managers’ fear of tracking error and by the cultural pressures of a society that can vilify short-sellers. However, I am able to introduce a new, social, indirect constraint to the literature – the perception that short-selling is a form of ‘trading’ as distinct from ‘investment’, and the consequent lack of acceptance amongst stakeholders that this engenders. This constraint reveals a divide between the attitudes of the academic community and parts of the institutional practitioner community on the subject of short-selling. However, interviewees argue that this indirect constraint is diminishing over time. This raises the prospect of markets in practice converging in behaviour towards the markets assumed in classical asset pricing models, and has implications for market efficiency. My second strategy of inquiry is to use a large, new stock lending database to explore three risk-related indirect constraints to short-selling. I examine ‘crowded exits’, a general class of liquidity problem, and find that these are associated with statistically and economically significant losses for short-sellers. I also examine ‘manipulative short squeezes’, a liquidity problem arising from predatory trading. Consistent with theory and the literature on the subject, I find that these are rare for larger, more liquid stocks. However, when they do occur, these events generate statistically significant losses for short-sellers. Finally, I build upon the work of Gamboa-Cavazos and Savor (2007) and investigate the response of short-sellers to losses. I find that short-sellers close their positions in response to accounting losses and not simply in response to rising share prices. This is consistent with short-sellers’ use of risk management tools that are designed to crystallize small losses. These serve to limit the risk of potentially unlimited losses and to reduce short positions at times of heightened synchronization risk. Stocks subject to shortcovering in this manner do not subsequently under-perform the market. My findings demonstrate that a sophisticated group of traders, strongly associated with price setting, does not suffer from the bias known as loss realization aversion.