EMPIRICAL IMPLEMENTATION OF THE TRIANGLE MODEL

Time-series econometrics lives a life of its own. When I started running Phillips curve regressions, the sample period was 1954–69, a mere 64 quarters of data, and many fewer degrees of freedom once current and lagged explanatory variables are included. My latest work, still in progress, covers the sample period 1954–2002, fully three times as many quarters of data. Why should any empirical paper be included in this volume other than the latest, since the latest overlaps the time coverage of the earliest and provides far more scope to explore stability and changes in parameters?

Part Four includes as Chapter Seventeen the latest (1998) published empirical paper on the dynamics of inflation behavior. However, more is involved in understanding empirical work than archiving the old framework in order to pursue the new framework. The first two papers included in Part Four (Chapter Fourteen, published in early 1977, and Chapter Fifteen, published in 1982) taken together helped to reorient the econometrics of the Phillips curve toward the specification and format that has become standard since the early 1980s. This specification has been fruitfully used to study the sources of low inflation in the 1990s not just by myself, but also by Douglas Staiger, James Stock, Mark Watson (1997, 2001), and others.

The framework introduced in Chapters Fifteen and Sixteen contrasts sharply with the standard approach of inflation studies dating back to the earliest econometric models of the late 1950s and early 1960s.