Estimation of Working Capital Turnover Measures on Interval Data

This work estimates two measures of working capital (including inventory) turnover on interval-censored data: the turnover and the turnover coefficient. The interval-censored observations quantify working capital fluctuations during pre-determined reporting periods (e.g., months). The upper and lower boundaries of these interval observations record the minimum and maximum amounts of working capital during each of the reporting periods. Under the assumption that these interval observations are independent, the distribution of working capital is estimated nonparametrically with the estimator suggested by Bruce Turnbull in 1974. This estimated distribution is used for calculating the mean working capital across the whole observation period and the measures of working capital turnover. The suggested methodology is illustrated on an artificially censored dataset with interval observations mimicking real-life investment levels in inventory. In this example, the new method shows different values of average inventory cost. The averaged investment level was higher, the turnover was longer, and the turnover coefficient was lower. These findings were also confirmed by non-parametric bootstrap.