L'auteur passe en revue les principaux travaux récents sur la relation d'arbitrage et souligne les incertitudes qui entourent cette notion et son utilisation pour la prise de décision. Suite à une brève présentation de résultats obtenus à partir d'un modèle plus complexe que les traditionnelles relations bi-dimensionnelles d'arbitrage, l'auteur s'attache aux conséquences pour la politique économique de la mise en question de la stabilité et de l'existence même de cette relation. Il souligne alors l'importance des politiques de gestion de la demande et compare les coûts respectifs de l'inflation et du sous-emploi.
Trade-Offs and Economic Policy
Ever since its introduction in 1958 by Phillips, the trade-off curve has been the focus of attention in both academic and policy making circles. Indeed, some years ago, this relationship seemed to have become one of the most useful and accepted tool of policy analysis.
Recently however, new doubts have arisen concerning both the stability and the usefulness of the traditional two-dimensional trade-off.
One of the most widely known contributions to this debate has come from Professors Friedman and Phelps. They argue that no historically established trade-off curve can be stable and serve as a useful policy guide. According to them, any policy action designed for instance to reduce unemployment at the expense of price stability is self-defeating since changes in price expectations brought about by this policy move will shift the whole Phillips curve. Partial and indirect confirmation of this hypothesis can be found, in Canada, in Vanderkamp's study which shows that since 1965 price changes are more rapidly reflected in wages than during the 1958-65 period.
The shifts or worsening of the trade-offs, in the U.S., have been attributed by Perry to changes in the composition of the labour force. According to him the increase and the relative importance of the secondary labour force has deprived the unemployment rate of its significance as an indicator of labour market tightness. Other explanations have been supplied by Archibald and Schultze, who maintain that the instability in the trade-off relationship has been caused by structural shifts in demand both from the industrial composition and regional point of view. Along the same lines, and following the lead given by Lipsey, Benjamin Higgins believes that there is a relationship between the trade-off and regional disparities, the trade-off being worse where disparities are larger.
All in all then, it seems doubtful that the static trade-off curve can be considered as a very useful policy tool. It is still not clear whether this apparent instability is attributable to expectations, structural changes in demand or supply, etc.
A more sophisticated approach to the trade-off relationship is possible through policy simulations within the context of large econometric models. Such an exercise has been carried out at the Economic Council in relation with the CANDIDE project. One of the simulations involves comparison between the control solution of the model for 1958-1970 with a solution into which was incorporated a $200 million a year increase in government purchases of goods and services.
The results tentatively indicate that during most of this period there was indeed some trade-off relationship between inflation and unemployment but that this trade-off was not very severe.
This approach to the trade-off is able to avoid many of the pitfalls of the two-dimensional model in that it can take into account changes in many of those variables held constant in the simple Phillips curve.
The uncertainty surrounding the trade-off has many policy implications.
First, if the constraints expressed in the traditional trade-off curve are far from rigid and absolute, the role of demand management becomes even more important.
Second, the numerous programs devised in the last decade to improve the trade-off relationship (manpower policies, regional policies, etc.,) will have to be evaluated on their own merit, not with respect to their contribution to an hypothetical trade-off.
Third, since inflation and unemployment will continue as major problems in our economy, more attention should be devoted to programs specially designed to better live with either of those two evils. Programs such as the Local Initiatives Program and Opportunity for Youth are auspicious beginnings in that direction. As for wholesale compensation of inflation victims, it does not appear at the present stage in our knowledge to be an appropriate policy move. This is so mainly on equity grounds since large numbers of so-called victims of inflation could have already been compensated through market adjustments.
The mention of programs such as L.I.P. and O.F.Y., move us strictly into a second-best universe. Within it, and given the uncertainties of trade-off relation-ships, it is preferable to slant economic policies in the direction of inflation and away from unemployment. To err in that direction would involve lower real costs for society since foregone earnings through unemployment cannot be made up while the redistributive aspects of inflation, albeit unfortunate, are at least partially mitigated through market mechanisms and, for the rest, could be controlled through ad hoc measures.