Search

The author replicates three wage equations suggested by McCallum (1974a,b), by Henry, Sawyer and Smith (1976), and by Johnston (1972), using quarterly British data 1951 (I)--1977(IV). The latter pair of equations yields fair results when tested by the usual methods. However, the application of a test of post fitting period parameter stability leads to their rejection. (The first equation performs so badly by conventional criteria that the stronger test is not required.)
The results point to the need (a) for more replication and (b) for testing the power of models to predict outside the fitting period. [ABSTRACT FROM AUTHOR]

Sargan's (1964) wage equation is respecified within the feedback control framework to which it belongs and confronted with quarterly British data. The starting point for the work reported in this paper is the interesting and important puzzle that despite over twenty years of fitting wage equations to British post-war data, no great success can be claimed either in explanation or in prediction. Indeed Sargan in the original paper used the model to illustrate the use of autoregressive least squares and instrumental variable estimators. The implication is that the equation is misspecified in some undetected manner. The objective of the research reported herein was to attempt the removal of the specification error by a careful reconsideration of the theory, then to test the appropriateness of the suggested reformulation by econometric methods. Some writers e.g. Sargan (1964) and Johnston (1972) have attempted to break away from the conventional excess demand/expectation models and to take account of bargaining in a labour market characterized by bilateral monopoly. To a Keynesian this appears as a potentially more fruitful conception of the labour market, both for prediction and explanation.

Parkin et all (1976) present a set of results for the post-war Great Britain economy which (they claim) confirm inter alia the role of excess demand for labour in the determination of the rate of change of money wage rates. Their most interesting behavioural hypothesis is that money wage rates are so adjusted by transactors that the change in excess demand, given expectations of price and tax changes, eliminates (in whole or in part) excess demand during some previous period. The disequilibrium framework introduced by Fair and Jaffee (1972) permits the estimation of excess demand quantities, and the method has been applied to the postwar Great Britain labour market by Dawson (1979) in an attempt to identify the Phillips Curve. Independent evidence lends some support to Parkin et al.'s market clearing hypothesis. However, the support is rather weak, since the explanatory power is very low. As one might anticipate, the implied speed of adjustment is slow. Given that economics abounds with unresolved issues (many of them with non-trivial policy implications) rather more effort needs to be applied to model-building in order to obtain identification prior to estimation.