The specification of econometric strike models: A VARMA approach

Abstract

In the case of macro-econometric strike models, classical regression techniques do not deal with important specification issues: (1) the question of causality between strikes and other variables; (2) the choice of appropriate lag structures; and (3) omitted variable bias. Moreover, since rational bargainers would tend to avoid previous bargaining mistakes, the determinants of strike activity in a given period may not apply in subsequent periods. This study attempts to model post-war US strike behaviour. Beacuse of the spurious regression problems, vector autoregressive moving average (VARMA) techniques are preferred to standard regression methods. For the early post-war period, strikes are successfully modelled in terms of unemployment and wage and price inflation. There is no feedback from strikes to the other variables. Contrary to the predominant view that US strikes are more frequent when real wage increases are small, increasing rates of wage change (both real and nominal) lead to more strikes. For the later post-war period, strikes cannot be modelled by these variables; supporting the idea that learning by rational bargainers means that the same strike model will not be universally applicable to all periods.

Cite this work

No public copy is stored for this record, so only the metadata is cited.

Stored only when both papers are in the corpus — 10 citations exist upstream.

Papers that cite the same references — the relational fallback while embeddings are still backfilling.