NATIONAL BUREAU OF ECONOMIC RESEARCH
NATIONAL BUREAU OF ECONOMIC RESEARCH

Aggregation Effects And Panel Data Estimation Problems: An Investigationof the R&D Intensity Decision

Ariel Pakes

NBER Working Paper No. 344
Issued in May 1979
NBER Program(s):   PR

This paper considers why the determinants of the inter- and intra-industry variance in R&D intensity in U.S. manufacturing differ markedly even though response parameters are similar across industries. A similar aggregation effect is noted by Grunfeld and Griliches (1960), and this paper gives that effect operational content in terms of grouped data estimation procedures. Observationally equivalent aggregation results can be generated by errors in variables models (see Aigner and Goldfeld [1974]).A later section considers specifications which identify the empirical importance of both these problems. Finally, a summary of the empirical results on the determinants of R&D intensity is provided.

download in pdf format
   (388 K)

email paper

Machine-readable bibliographic record - MARC, RIS, BibTeX

Document Object Identifier (DOI): 10.3386/w0344

Published: Pakes, Ariel. (1) "On the Asympotic Bias of Wald-Type Estimates of a Straight Line When Both Variables are Subject to Error." Inter. Econ. Rev. June 82, pp. 491-497. (2) "On Group Effects & Errors in Variables in Aggregation ." Rev. of Econ. & Statistics. February 1983, pp. 168-173.

Users who downloaded this paper also downloaded* these:
Griliches t0039 Data Problems in Econometrics
Pakes and Schankerman An Exploration into the Determinants of Research Intensity
Merwin, Jr. American Studies of the Distribution of Wealth and Income by Size
Heckman and Smith w6105 The Sensitivity of Experimental Impact Estimates: Evidence from the National JTPA Study
Wei Some Consequences of Temporal Aggregation in Seasonal Time Series Models
 
Publications
Activities
Meetings
NBER Videos
Themes
Data
People
About

National Bureau of Economic Research, 1050 Massachusetts Ave., Cambridge, MA 02138; 617-868-3900; email: info@nber.org

Contact Us