Forecasting Labor Productivity Using Factor Model

Abstract

This paper uses data from 22 masonry projects to compare two approaches for forecasting labor productivity. The first approach is to divide the current work‐hour total by the percent complete (PC) of the activity. The second method uses the factor model to develop a predicted labor‐productivity curve. Actual deviations from this curve are reflected in the labor‐productivity forecast at completion. The paper details how the predicted curve and the forecast are made. Absolute and relative errors are the basis of comparison. On average, the factor‐model forecasts are within ±5% when only 5% of the work is completed. The range based on the standard deviation is ±25%. The PC approach is equally as accurate when considering the average of the 22 projects. However, the range of error is consistently about 50% larger. It is concluded that the factor model is a more reliable method of forecasting labor productivity.