The open politiconomy: a dynamic analysis of social democratic popularity and economic policies in Scandinavia.

In what has been identified as the 'second wave' of the vote-popularity literature.(1) the dynamic interrelationship between voting and government policies has become a topic for analysis.(2) The dynamic model has two components: a vote (or popularity) function delineates how economic performance determines voters' evaluation of the government, while a reaction (or policy) function describes how the government steers the economy to fulfil ideological and strategic objectives. The essence of the model is succinctly summarized by Tufte: 'As goes politics, so goes economic policy and performance. This is the case because as goes economic performance, so goes the election'.(3)

Parallel to the development of politico-economic models, the impact of international forces on domestic conditions has received more attention both in economics and political science. Today all macroeconomic textbooks emphasize the importance of the transmission of shocks from the international environment to the national economy.(4) It has become (painfully) obvious that policy making is decisively affected by economic openness. There is econometric evidence that for a typical European country as much as 40 per cent of the total income effect from an isolated fiscal stimulus is absorbed by higher imports.(5) From the perspective of political science, pioneering studies by Cameron and Katzenstein demonstrate that corporatist and welfare state arrangements hinge on the openness and the size of the economy.(6) Indeed, it is now standard procedure to include international control variables in economic policy studies.(7)

This article seeks to combine, and to expand on, the politico-economic model and the international integration perspective. In addition to the traditional focus on how the domestic economy affects the domestic polity, and vice versa, the dynamic links between politico-economic systems, or 'politiconomies', are analysed. The analysis shows that the popularity of the social democratic parties in Denmark, Norway and Sweden is related not only to economic and political variables in each of the respective countries but also to the corresponding variables in the two neighbouring countries. The statistical analysis develops a vector autoregressive model using quarterly data from 1967:1 to 1992:2.

THE THEORETICAL BACKGROUND

There is a distinct lack of consensus among politico-economic theorists as to the precise nature of the relationship between macroeconomic indicators and party support. The tradeoff between the pursuit of ideological goals and the need to secure re-election has been portrayed as a particular social democratic policy-dilemma.(8) Yet every party that seeks government responsibility faces the same quandary,(9) and the literature on the economic impact of governments describes the tradeoff as a general dichotomy between strategically and ideologically motivated policies.(10) In brief outline, strategic policies directed primarily towards the capturing of floating voters yield economic fluctuations synchronized to the electoral calendar.(11) Changes in government popularity can also affect economic policies independently of the election schedule. If so, it is a matter of dispute whether a popularity deficit yields inflationary or deflationary policies.(12) As for ideological policies devised to satisfy core voters, the usual assumption is that left-wing governments prefer lower unemployment to lower prices. Right-wing governments are supposed to give priority to low inflation.(13)

The dichotomy between re-election strategy and ideology appears in the economic voting literature as well.(14) The most common hypothesis, dating from Kramer, says that governments - irrespective of their ideological colour - are penalized for every kind of economic misfortune.(15) As an explicit alternative to this 'responsibility hypothesis', Rattinger builds on the partisan policy model to advance a 'clientele hypothesis'.(16) The crux of his argument is that economic indicators affect political parties independently of whether they are in government or opposition. Even if the government fails to achieve a particular economic target, voters may suspect the opposition would do even worse. By the same token, when the government is doing well the opposition may be expected to do even better. As a general tendency, social democratic governments are perceived as more trustworthy when it comes to reducing unemployment, whereas 'bourgeois' governments are held to be particularly concerned with curbing inflationary pressure. Accordingly, when unemployment rises, so should the popularity of social democratic parties, no matter whether they govern or not. Conversely, when prices go up, the popularity of the social democrats - even when in opposition - should go down. Rattinger's general argument does not, however, square altogether with Madsen's comprehensive analysis of Scandinavian social democratic parties.(17) Madsen acknowledges that initially, due to the depression legacy, the two terms 'full employment' and 'social democracy' became almost synonymous. However, as time has passed a growing share of the electorate consists of post-war generations that look upon employment more as a constant than a variable.(18) As a result, most voters have come to see the depression years more as a historical oddity rather than an immediate threat. This attitude has been reinforced by the policy stance of the bourgeois parties that in opposition promised to continue the policy of full employment - a promise that they followed when eventually assuming power. By implication then, if the Scandinavians 'are all social democrats', as the slogan has it, the clientelist hypothesis leaves something to be desired.

It comes as no surprise, therefore, that the search for a Scandinavian model of economic voting has followed other routes. In a Danish analysis, Paldam and Schneider argue that multi-party systems are characterized by a cleavage between the larger and 'responsible' parties, on the one hand, and the smaller and 'exciting' parties, on the other.(19) As long as the economy is going well, voters feel they can afford to support the more exhilarating alternatives. However, when the economy deteriorates, it is time to turn to the 'dull workhorses' .20 Then the exciting party is over, so to speak. If we relate this argument to the popularity of the three social democratic parties, each of which has to be classified as large and responsible, higher unemployment and higher inflation - counterintuitively - should be electorally beneficial.

INTRODUCING OPENNES

Openness raises both substantive and analytical questions. 'Galton's problem' sheds some light on the analytical aspect.(21) The essence of the problem is that a correlation between two aggregate variables in a cross-national analysis may stem from two different sources: either functional forces within a country or diffusion mechanisms between countries. In cross-national analyses there is a tendency to attribute all of the explained variance to the former and to ignore the possibility of the latter. To the extent that diffusion mechanisms are observed at all, the elimination of this bias usually involves 'controlling' for the indirect diffusion effects before concentrating exclusively on the direct functional effects. However, as Ross and Homer point out, it would be better to 'test for the effects of diffusion and function simultaneously'.(22) This is indeed the approach adopted in the empirical analysis below.

In Tufte's seminal study the diffusion mechanism …