A Strategy for Producing Hybrid Regional Input-output Tables*

1. INTRODUCTION Hybrid input-output models combine nonsurvey techniques for estimating regional direct requirements tables with superior data, which are obtained from experts, surveys, and other reliable sources (primary or secondary). Such data can be added at any stage of model construction. To date, hybrid model production has been defined only in very general terms. Elsewhere (Lahr, 1993), I have pointed out directions that could prove fruitful in developing a more detailed approach. I start the paper by searching for an approach that recognizes critical portions of input-output models that are ripe for survey work or the collection of other superior data. This search begins with a review of existing measures that identify sectors that are most critical to the accuracy of the Leontief inverse. I next compare these measures for their ability to predict sectors critical to the survey-based 1972 Washington State Input-Output Table. In Section 3, I combine the findings of Section 2 with that of other literature to develop a quasi-complete superior-data search strategy for producing hybrid regional input-output tables. In Section 4, I test the strategy by employing the 1972 Washington State input-output table. Tests are of models both closed and open with respect to households. Following this, I conclude with implications of the findings.