The Early Years of Panel Data Econometrics

This essay focuses on the early years of panel data econometrics and two seminal papers by Yair Mundlak (1961) and Pietro Balestra and Marc Nerlove (1966). The issue of latent individual heterogeneity was identified as a central problem. One main ambition was to get a clear understanding of how differences in behavior across individuals and/or through time could and should be modeled. Another main aspect was the assumed inherent dynamic pattern of economic behavior, which requires the use of specific models and data collection. These articles attempted to build relevant models and statistical methods, mainly based on the distinction between fixed effects (analysis of covariance) and random effects. They brought back identification issues in the context of estimating a production function of sixty-six farms in Israel for 1954-58 (Mundlak 1961) and the demand for natural gas in the residential and commercial market of thirty-six U.S. states observed for 1957-62 (Balestra and Nerlove 1966). © 2011 by Duke University Press.