Sofie Van Osselaer’s research while affiliated with Ghent University and other places

This paper explores a unique data set comprising the transactions executed by a large sample of mutual funds for the period August 2002 - April 2007. This data set qualifies as a worthy counterpart for another (often used) transactional data set provided by the Plexus Group. It will serve as input for various papers, but is interesting in its own right. We discuss the structure of the data set, we enumerate the major asset classes traded by the institutionals and we describe the global representation of the traded securities. Finally, we detail the filtering procedures necessary to obtain a sample of equity transactions executed by actively managed mutual funds.

Using a unique data set of mutual fund transactions, this paper examines two widely acknowledged behavioural biases: overconfidence in trading and disposition behaviour. We test for the first bias by comparing the future profitability of the purchased and sold securities by mutual funds. Our empirical results show that the return difference between the purchased and sold securities is not sufficient enough to cover the incurred transaction costs, suggesting that fund managers exhibit overconfidence in their trading capabilities. The disposition bias, i.e. the reluctance of investors to sell losing stocks, is tested by the widely accepted methodology of Odean (1998). In contrast to Odean`s findings for individual investors, we reject the disposition hypothesis and instead document a propensity of mutual fund managers to cut losses early.

This paper evaluates the performance of the stop-loss, synthetic put and constant proportion portfolio insurance techniques based on a block-bootstrap simulation. We consider not only traditional performance measures, but also some recently developed measures that capture the non-normality of the return distribution (value-at-risk, expected shortfall, and the Omega measure). We compare them to the more comprehensive stochastic dominance criteria. The impact of changing the rebalancing frequency and level of capital protection is examined. We find that, even though a buy-and-hold strategy generates higher average excess returns, it does not stochastically dominate the portfolio insurance strategies, nor vice versa. Our results indicate that a 100% floor value should be preferred to lower floor values and that daily-rebalanced synthetic put and CPPI strategies dominate their counterparts with less frequent rebalancing.