Who will live and who will die? The determinants of common stock attrition

The delisting of a company's shares from a stock exchange is a major event affecting a variety of stakeholders. Such events are also of importance to anyone relying on historical share price data for investment insights, as standard statistical approaches often call for exclusion of delisted shares, thus leading to a survivorship bias. This paper studies the determits of delisting of ordinary shares on the London Stock Exchange. To do so, we employ hazard regression, and disaggregate delistings into those due to acquisition or merger on one hand, and those explicitly linked with poor operating performance (bankruptcy, liquidation, or failing listing requirements) on the other. The key findings concern the very different behavior of the two delisting categories, and the complex dependence of the delisting likelihood on prior returns. Our characterization of the attrition process for stocks indicates how delisting propensities of individual stocks and portfolios can be evaluated by relying on market data only. Traditional insights about the effects of disregarding "dead" observations (e.g. in the literature on mutual funds performance) are unlikely to be relevant when dealing with data on common stocks