A finite horizon model of industrial exit is developed. After an initial lag, most exits are by young firms. The duration of the lag is positively related to sunk entry costs, but not due to the fallacy of sunk costs. The conception of entry differs from previous research; as a result, not all entrants are identical; and firm size affects the rate of learning. On average, larger new firms last longer. Entrepreneurs in declining firms act more lazily as the firm declines. A number of empirical observations about declining firms are organized by the model.