A problem with constructing regional economic models from secondary sources is that of undisclosed data. For example, the RIMS II technique for developing nonsurvey input-output models, requires disaggregated industry earnings data. County Business Patterns employment and earnings data are suppressed for industries with few firms or with a dominate firm, requiring that earnings be estimated for some industries. We examine eleven representative methods for estimating undisclosed earnings. A simple earnings ratio technique using more aggregated data from the same state is shown to provide reasonable estimates of industry earnings.
The authors would like to thank Otis Gilley and Bill Workman for their helpful comments. We are greatful for financial support from the Alaska Department of Commerce and Economic Development.