Computers & Firm Size

From: Robin Hanson (hanson@hss.caltech.edu)
Date: Fri Aug 15 1997 - 16:47:15 MDT


For those tied to the notion that computers will make firms smaller,
the following paper argues that the situation is lots more complex. Robin

     "Information Technology's Impact on Firm Structure: A Cross-
      Industry Analysis"

      BY: DAVID N. BEEDE
            U.S. Department of Commerce
          SABRINA L. MONTES
            U.S. Department of Commerce

     ELECTRONIC DOCUMENT DELIVERY: available at no charge from:
     http://papers.ssrn.com/paper.qry?abstract_id=15569

          Paper ID: Department of Commerce WP ESA/OPD 97-2
          Date: March 1997

          Contact: David N. Beede
          E-Mail: MAILTO:DBeede@doc.gov
          Postal: U.S. Department of Commerce, Economics and
                    Statistics Administration, Office of Policy
                    Development, Office of Business and Industrial
                    Analysis, Mail Stop H4878, Washington, DC
          Phone: (202) 482-1225
          Fax: Not Available
          Co-Auth: MAILTO:SMontes@doc.gov
          ERN Ref: I/O:WPS97-159

     Since the 1970s, economists have speculated on the effects of
     the proliferation of new computer and communications
     capabilities on business structure and performance. The
     present analysis explores information technology's (IT)
     relationship to employment and firm structure by examining
     how IT affects the relative size of employment at auxiliary
     units. The analysis treats auxiliary units--establishments
     where employees provide support services (mainly
     administrative) to production establishments--as a proxy for
     the highest administrative levels of the organizational
     hierarchy. Changes in the relative size of auxiliary
     employment give a broad indication of IT-related changes in
     firm structure. Statistical analyses of 46 industries show
     large variations across industries in the size, sign, and
     statistical significance of the elasticities of auxiliary
     unit employment shares with respect to IT capital stock
     shares. We find no economy-wide trends associated with IT.
     There is too much variation among industries to rely on
     estimates obtained from pooling industry data. For the most
     part, sectorial trends are scarce. Only in the transportation
     sector do the sign and statistical significance suggest that
     IT related changes are similar. Ultimately, the enormous
     variation revealed by our results suggests that one cannot
     make economy-wide generalizations about the effects of IT.

     Nevertheless, our results, combined with other evidence,
     suggest that economies of scale--gained from using IT to
     reduce coordination and monitoring costs--influence firm size
     and structure. One reason why the effects of IT are so
     different across industries is variation in the firm size
     distribution across industries prior to the IT revolution:
     1) For industries with a predominance of small firms, IT-
     related economies of scale may encourage growth in firm size
     and lead to an increase in the relative size of centralized
     back office establishments across the industry. This appears
     to have occurred in the retail trade industry. 2) In some
     industries where large firms predominate, IT may induce
     greater efficiency in back-office jobs, enabling firms to
     reduce back office employment relative to total employment.
     This appears to have occurred in some of the transportation
     industries. 3) In industries where IT primarily substitutes
     for production workers, auxiliary unit employment share is
     likely to rise because central administration office
     employment tends to change less than proportionately in
     response to changes in overall employment. This appears to
     have occurred in the primary metals industry.

     JEL Classification: L22, L23, D23, D24, L72, L91
     __________________
------- End of forwarded message -------



This archive was generated by hypermail 2.1.5 : Fri Nov 01 2002 - 14:44:43 MST