This thesis is primarily dedicated to the increasingly popular new area of econometrics: the formulation and estimation of efficiency frontiers. This research estimates a multi-factor frontier model that measures nursing home efficiency using data from the Residential Health Care Facility (RHCF-4) nursing home survey, the Facility Profile Report and the Resource Utilization Group (RUG) Case Mix Classification System data provided by New York State Department of Health.

This thesis uses a frontier cost function that is based on a model first introduced by M. J. Farrell in his paper on measuring productive efficiency. The general form of this frontier function was developed by Aigner, Lowell and Schmidt and further modified by William Green.

The methodology breaks the error term $varepsilonsb□rm i □$into two parts v$sb□rm i □$+ u$sb□rm i.□$ This is practical because it facilitates the estimation of inefficiency from the frontier. Each nursing home was examined as to their deviation in u$sb□rm i □$(inefficiency) from the frontier.

The first part of the thesis examines alternate specifications of the frontier cost function for nursing homes. The results show that type of ownership plays a significant role in determining cost levels. Homes with a profit motive had lower inefficiency levels than non-profit homes. Over 87% of the variation in the data was explained by the frontier function. The inclusion of ownership type in the frontier reduced inefficiency by 40%.

The second part of the analysis used regression analysis to explain inefficiencies among nursing homes. The explanatory variables used were regional indicators, ownership types, unionization and available resident activity indices. About 29% of the variation in inefficiency was explained. The findings show that only management per bed was a contributing factor to inefficiency.