Location, growth and productivity in the manufacturing and service sectors

In the course of our project, we have produced nine papers using the Census data sets of the Census of Manufacturers, the Annual Survey of Manufacturers, the Census of Services, the Business Expenditures Survey, the Company Auxiliary Organization Survey, the Large Company Survey and the Standard Statistical Establishment List. In this Technical Report we present six sets of findings. First, outsourcing of business services is prevalent by establishments of all sectors of the economy, and their magnitudes confirm their importance for data collection programs. Second, heterogeneity in functions performed by central administrative offices suggests NAICS industry 55 Management of Companies and Enterprises should be expanded to include more 6-digit NAICS sub-categories from the current three. Third, better classification of NAICS 55 could help with response rates, which are increasing in the data. For firms, we find in the data that often either all central administrative establishments respond, or none do, indicating firm level follow-up could improve response rates. Fourth, firm level collection of data introduces severe bias for aggregation into industry series. Establishment based data collection should be used for industry or geography series publications. Fifth, the value of building rent expenditure data would be significantly enhanced if a companion question on floor space were introduced to questionnaires. Finally, Census’ current strategy of geocoding economic census establishments at the block level, thus conforming to standard Census geography definitions in both internal data files as well as external publication of aggregates is the right direction to proceed.