Resumé

In this article* it is argued that rational choice analysis can help explain social democratic welfare states (SDWS). It explores the common ground between the methodological individualism of rational choice political economy and the "U-shaped curve" central to neocorporatist analysis, the idea of a curvilinear relationship between the level of corporatism in institutional arrangements and societal welfare. Central to the SDWS, it is suggested, is the reduction of information costs due to the configuration of institutional arragements in three key arenas of social action in the SDWS: the informational-cultural; the organizational-economic; and the governmental-political. These arrangements frame the choices under which rational individuals act to maintain the SDWS and the outcomes it produces. Concrete examples demonstrating the usefulness of the approach are offered in terms of explaining differences in policy choices (and therefore outcomes) between the Scandinavian countries and more laissezfaire systems such as the US. The conclusion looks at the prospects for the SDWS, asking just how the institutional arrangements central to the SDWS are being challenged by recent developments.

Introduction

Can rational choice analysis help explain the successes as well as the failures of social democratic welfare states (SDWS)? In this article I seek to show that it can - and how. This effort is especially important now that we have entered the post-socialist world, which - potentially - opens up new theoretical space for thinking about economic and social institutions.

SDWS serves as a shorthand description of a modern industrial society characterized by democratic institutional arrangements leading to relatively egalitarian outcomes. The reference to a post-socialist world does not merely allude to material developments, but also to intellectual ones. While exploitation based on collective attributes, class, race, gender . . . remains real, attempting to build a political economy on such collective interests - as

---

*This paper was completed in the spring of 1995 when the author held a visiting chair in political economy at the University of Western Ontario.

Marxist socialism did - has proven to lead to a dead end. The notion of an objective working-class interest, it turned out, assumed away the fundamental problem to be resolved: namely to explain the choices of the individual worker/trade union member, to identify the circumstances under which he or she will support organizational leaders in their actions.¹

The tendency to subsume individual choices under collective interests was not limited to ideologues of the former Marxist regimes. Let me give an example which will prove important in the discussion of the SWDS below. Basing policy on methodologically collectivist analysis, Swedish Labour in power in the 1970s and 80s enacted measures which replaced pay for sickleave at 90 to 100 percent and failed to monitor employees' actions while removing from employers all financial incentives to do so. The resulting increased tendency to call in sick when healthy proved especially pernicious as labour shortages in the late 1980s exacerbated the already high absenteeism. At 5 percent (in 1988), Sweden led the OECD in percentage of illness-related absences with disastrous effects on productivity.

The methodological collectivism built into inherited intellectual baggage blinded Swedish policy-makers to this reality, unable to see such "decommodification"² as in anything but the best interests of the workers. The reality is that each worker (potentially) knows that though she benefits by being freed of market-driven imperatives inhibiting her from staying home when capable of working, she loses if other workers draw the same benefit. To be useful, our theoretical framework must take into account there being benefits as well as costs for workers in "commodifying" labour market institutions that inhibit them from free-riding on welfare state guarantees.

To put the matter in more general yet concrete terms, we need to take incentives and disincentives into our account of policies and institutions, including those of the SDWS. This means rejecting methodological collectivism in favour of the methodological individualism of mainstream, or rational choice, political economy. Except that we can apply rational choice analysis in ways that most mainstream analysts are not wont to do to enhance our understanding of the performance and prospects of the SWDS. This is the task I set for myself here, developing on my recent work (Milner 1994). I attempt to stake out common ground between the methodological individualism of rational-choice political economy with the insights of the neocorporatist school, and particularly the notion of the "U-shaped curve".

Methodological individualism means that assumptions and explanations about determinants and effects of policies must be consistent with the choices of rational individuals.³ Rational behaviour, by definition, can always be expressed in cost-benefit terms, chosen action entailing higher benefits and/or lower costs to the actor than known alternatives, though the benefits need be neither selfish nor material. Choices can also be mistaken

when relevant knowledge about the relationship between given choices and preferred outcomes is too costly.⁴ Rational choice political economy models the effect of alternative institutions in terms of altered costs and benefits: what would be the effect on behaviour, say, of changes in laws affecting labour force training or collective bargaining rights . . .?

The Contribution of Neocorporatism

Based on such considerations, the central question of concern for political economy - to explain the relationship between institutional arrangements and welfare (defined for now as aggregate well-being) - is posed as follows: how do we identify institutional choices (see below) that result in lowering the costs of (removing disincentives from) individuals choosing to act in such as way as to enhance welfare? To direct this question at the workings of the SDWS, I turn to a reconsideration of the neocorporatist framework.

In the neocorporatist framework, less fashionable now than its heyday in the 1970s and early 80s, corporatist societies were characterized first and foremost by a "social partnership" between "peak" or "encompassing" representative organizations (Schmitter 1981). The outcome of this partnership over many years corresponds largely to what we have characterized as the SDWS; indeed, one recent work simply terms neocorporatist structures as in fact constituting the social-democratic form of government (Bergounioux & Manin 1989).

The most sophisticated applications of neocorporatist analysis focused on controlling inflation as a specific manifestation of the tendency toward mutually advantageous compromise.⁵ It was postulated that under centralized bargaining shared knowledge that real wage gains are dissipated in higher prices and reduced employment-creating activity leads actors to work

within the corporatist structures to restrain wages, and thus not to undermine economic growth. Of course, this was not the only means of restraining inflation and enhancing growth: as a rule, firms in the non-corporatist (laissez-faire) countries are forced to keep down wages because they cannot raise prices without weakening themselves in the product market. Thus an alternate "laissez-faire strategy for economic growth . . . succeeds in certain countries with dominant neo-conservative parties and where there are weak and fragmented trade unions" (Paloheimo 1984). The relationship between corporatism and economic performance was thus portrayed as U-shaped (Fig- 1).

Under high levels of corporatism, policies reinforcing corporatist institutions can be expected to have a positive effect on economic performance, while under low levels of corporatism, laissez-faire policies have a similar effect. Nations at middle levels of corporatism, with labour markets dominated by medium-sized institutions with sufficient market power to win concessions, but small enough to pass most of the cost of added inflation onto others, end up at the bottom of the "U".⁶

While attractive in its simplicity, economic developments in recent years - especially reverses in some SDWSs which seem to make the "U" look more like an "L" - have rendered the neocorporatist explanation less compelling, less able to withstand the criticism levelled by critics of the welfare state.⁷ The suggested route here is rather than discarding the neocorporatist conceptualization of the SDWS, we endeavour to reconstitute it by

attempting to incorporate a central element of the theoretical framework of
those very critics, namely methodological individualism.

Disaggregating the U-Shaped Curve

I begin by recasting the U-shaped curve, visualizing welfare as composed of
two separable elements. A stylized representation of what this might look
like can be found in Fig. 2.

In Fig. 2, a society's overall welfare rating - its place on the "U" curve - is conceptualized as derived by summing up the values for it on (1): the economic performance curve "P", and (2): the redistributiveness curve "R". "P" corresponds to its overall wealth, measured, say, by per capita GDP. One useful expression of the redistributiveness level, "R", is itself a composite, derived by combining the level of income inequality as revealed by post-tax Gini coefficients,⁸ with the long-term unemployment rate.⁹

"P" is thus understood to contribute to "U" disproportionately on the (left-hand) laissez-faire, non-corporatist side, while "R" does so on the (right-hand) corporatist SDWS side. The direction of "P" and "R" reflects developments in the democratic industrial world over the past forty years: "R" is positively correlated with corporatism;¹⁰ "P" is seen to correlate negatively with corporatism from low to middle levels and then remain roughly steady. The US, universally characterized as situated at the noncorporatist, laissez-faire end among the OECD countries, owes its high welfare rating to its high GDP (especially when calculated in purchasing power parities); the same is true, though more moderately, of Canada and Switzerland. The Scandinavians take up positions at the opposite (top right) position, their high welfare levels attributable especially to their level of redistributiveness. *l

Rational Individual Choice and Redistributiveness

While few would dispute the direction of the "R" curve, the portrayal of the "P" curve in Fig. 2 might be more suspect. Neocorporatists in the 1980s tended to assert that the "P" curve was itself U-shaped (e.g. Keman 1984) based on the apparent solid performance of corporatist states. On the other hand, for neoconservative critics of the SDWS, "P" is conceived of as sloping downward in a continuing trade-off between efficiency and equality: the more equally the pie is shared, the smaller it gets. Clearly, if the latter were an accurate reflection of reality, then neocorporatist analysis would be irreconcilable with methodological individualism. For, over time, a certain threshold would necessarily be crossed with people perceiving the redistributive institutions as necessarily entailing falling economically

behind comparable neighbouring societies. They would thus, if free to do so,
vote against the policies propping up these institutions first with their feet
(by moving resources to the neighbouring society) and then with their hands.

While there is always room for differing interpretations of the data, a gleaning of the vast quantities of empirical work more or less in the neocorporatist tradition carried on over the past 25 years reveals the "P" curve as presented here to be highly plausible (see, for example, Kenworthy 1995). Everything hinges on its flattening out somewhere in the middle. It means that, with the exception of the distinctly non-corporatist laissez-faire societies on the left side of Fig. 2, any trade-off between equality and efficiency is sufficiently mitigated by corporatist institutional arrangements that the rational individual, given the choice, would reinforce these arrangements in his or her actions as voter, earner, organizational member and consumer.¹² Concretely, this means that citizens of a country with institutional arrangements at the laissez-faire end, like the United States, act rationally in reinforcing arrangements that keep taxes and levels of redistribution relatively low, quite the opposite of the Scandinavians¹³ who look to tax-supported public services to meet their needs. For example, the latter send their children to state schools and take an informed interest - if they do not actively participate - in the activities of the education subcommittee of their municipal council. In contrast, to meet the same needs, a similar American household seeks an urban "condo" with private school nearby or an exclusive suburb with good schools, choices restricted by paying taxes or otherwise contributing to public amenities.

The logical consequence of the flattening of the "R" curve is thus conceptualized in rational choice terms. We can express the underlying logic as follows: if given the opportunity democratically to choose between living in one of two roughly equally rich societies, a rational individual can be expected to choose the society where wealth is distributed more equally and poverty less extensive over the one with greater extremes between rich and poor. This is based on the expectation that a rational individual can be expected to prefer the more egalitarian distribution to the combination of decreasing marginal utility combined with uncertainty. People buy insurance since the marginal utility of a given unit of income is greater when one has less than when one has more. Though some individuals are risk takers for themselves, they can nonetheless be expected to seek to reduce risk once the well-being of dependents and descendants is at stake since they cannot be certain of their future position. Thus, even excluding the effect of altruism among the better off toward the worse off, rational individuals prefer outcomes at the right-hand side over those at the centre of Fig. 2. A recent, highly sophisticated experiment tending to confirm this argument can be found in Frohlich & Oppenheimer (1992).¹⁴

The Question of Institutional Choice

Answering the question in this way opens another question. At this point in history, the SDWS shows little sign of making progress among societies in the middle of the "U". While factors of international political economy beyond the scope of this analysis enter here, we cannot leave it at that. If our analysis is correct, then people in countries in the middle should make choices the effect of which is to move their society toward the more corporatist/redistributive position (on the right in Fig. 2). We can avoid the problem by downplaying the salience of rational individual choice in favour of "path dependency", adopting the language of certain "new" institutionalists (e.g., March & Olsen 1989) by regarding individual preferences as, for the most part, endogenous to institutional arrangements.

I do not take that approach, but rather that of mainstream political
economy treating preferences (at any given point in time) as exogenous:
institutions do frame individual choices, but through incentives and

information dissemination. Moreover, as illustrated in Fig. 3, the configurations
of those institutions are themselves understood to result, in the long
run, from what are termed "institutional choices".

From within this perspective we can look for an explanation for corporatist societies (on the right side of Fig. 2) fostering rational individual choices resulting in SDWS policies and outcomes, but not those in the middle? Figure 3 identifies the key factors. For the purposes of the analysis of the SDWS, I divide institutional arrangements (at the top of Fig. 3) into three clusters, corresponding to three dimensions of social action: (1) informational-cultural, (2) organizational-economic, and (3) governmentalpolitical. The first dimension focuses on the institutions affecting the dissemination of the information individuals bring to bear in making choices, especially institutional choices. The second concerns the degree of corporatism in relationships between organizations, especially those operating in the labour market. The third concerns political structures, especially the electoral system and the horizontal and vertical relationships between political and administrative structures at the different levels.

I suggest that the institutional arrangements of the SDWS - as they have developed especially in Scandinavia - are such as to enhance the capacity of individuals to make (institutional) choices that, on balance, reinforce the capacity of the SDWS to redistribute but without undermining its capacity to produce. To simplify a complex argument, we can conceive of this relationship in terms of knowledge dissemination: built into the nature of SDWS institutions is a tendency to reduce the cost of information and knowledge. More informed individuals are less prone to be mistaken in the policies they support, and whom they choose to implement those policies, to achieve the desired welfare outcomes, as well as about the effect their individual choices will have on the institutional arrangements themselves. This last point can be expressed as an assertion that individuals in a more informed society are less likely to free ride.¹⁵

To illustrate this, let me describe two related concrete examples from among those arrangements associated with first dimension. These are, first, the requirement in Sweden and other Nordic countries for individual and corporate tax-return information to be public, and, second, that public institutions are required to open their books to interested citizens. Such information makes individuals more knowledgeable, and therefore less likely to err in their choices when it comes to the costs and benefits of public services. Moreover, knowing what they can know about others - and others about them - they are less likely to undermine institutions by free-riding.¹⁶

Elsewhere (Milner 1994), I extend this argument to the wider cultural dimension by stressing the availability and quality of adult education in its widest sense in the Scandinavian SDWS, not only in the education system, but through the communications media, where the dissemination of

knowledge is still seen as inversely related to the pervasiveness of
commercialism.

Among political institutions, I stress proportional electoral parliamentary systems, and integrated (non-discontinuous) relationships - and therefore less costly communication of information - between local and national levels of politics and administration. Compared to the alternatives, such arrangements reduce the costs and increase the benefits of political actors at the local and state-national levels sharing and communicating requisite information. Proportional systems have far lower incentives for politicians to distort reality in communicating with (or in designing institutions that disseminate knowledge to) the population (Milner 1993, 1995), as well as fostering conditions under which parties and actors at the same level share knowledge among themselves. In sum, information passes more easily both vertically and horizontally.

A similar logic applies to the institutional arrangements promoting corporatist relationships on the labour market that reduce the cost and increase the benefit of sharing information relating to technological research, human resource development, and so on. (Milner 1989). A concrete illustration is to be found in the role of the labour consultancy firms set up under Sweden's co-determination law (MBL). Typically, the consultants have trade-union backgrounds as well as specific professional skills - managerial, accounting, public finance, law. In the case of a firm planning serious changes affecting working conditions, the unions are entitled to engage such a firm at the employer's expense. The company is required to open its books to the consultant, knowing its confidentiality will be respected. In the end, the consultant's compromise recommendations stand a good chance of being accepted as based on a realistic assessment that has the workers' best interests at heart.

These kinds of institutional arrangements writ large being relatively absent in the countries in the middle, individuals there generally lack the information on which to make (institutional) choices favouring SDWS outcomes, discouraging them from free-riding on existing corporatist arrangements.

The Future of the SDWS

The tendency of the corporatist SDWS to reinforce rather than undermine basic institutional arrangements does not safeguard it against inappropriate policy choices that place perverse incentives on individuals. Our approach in looking for explanations for a given society's attaining a level of aggregate welfare lower than its level of corporatism (falling below the "U" curve) would lead one to look for policies which reduce the benefits and increase

the costs of reinforcing (rather than undermining) the capacity of existing institutional arrangements to optimize welfare. In the laissez-faire (noncorporatist) context, there is a wide rational choice literature on incentives for given groups to draw unproductive rents, a famous recent example of which was the hundreds of billions of American taxpayers' dollars that disappeared into corrupt and wasteful savings and loan operations (S&Ls).¹⁷

Similarly, poor incentives help explain the billions lost by Scandinavian banks in the same period. The Social Democratic governments of the 1980s opened up financial markets enabling the banks, who had no real experience in this, to speculate on foreign and domestic property under a taxation system that rewarded borrowing money to invest in real estate. The result was a bubble the recent bursting of which forced the Swedish, Finnish and Norwegian governments, at great cost, to bail out several banks. To some extent, the decentralization of the 1980s had a similar effect: with controls lifted, municipal and regional councils could not easily resist the temptation to add services and hire staff.¹⁸ The other example, already noted above, concerns the perverse built-in disincentives of Swedish sick-leave and long-term disability policy in the 1980s. The policy was finally changed by the Conservative-led government elected in 1991, with the grudging acquiescence of the Social Democrats.

On the whole, such perverse incentives are the exception rather than the rule in the Scandinavian SDWS. There are clear counter-examples such as the deterrent fees charged when receiving medical services, prescription drugs, etc., or the Swedish tendency, cited by Olson (1990), to discourage subsidies and tariff protection for cartels. Critics maintain that the SDWS is unable to adjust to changing conditions due to inflexible corporatist labour market arrangements. I contend rather that such inflexibility is better understood as the result of avoidable and, in principle, reversible perverse policy choices. In that case, as long as the institutional arrangements identified as directly or indirectly reducing information and knowledge costs are reasonably secure, individuals remain in a position to seek out appropriate actions, and, where needed, change policies that place incentives on individuals to undermine corporatist institutional arrangements.

What has been the experience of Sweden and the other Nordic countries in this regard? I noted that Swedish sick-leave policy exacerbated labour shortages, contributing to inflation and a lag in productivity. Combined with poorly timed deregulation of financial markets, it contributed to a massive outflow of capital in the latter 1980s. These developments gave credence to the critics' expectations. Threats to corporatist institutional arrangements followed when SAF, the Swedish Employers' Federation, rejected centralized bargaining and participation in corporatist bodies in the late 1980s,¹⁹ all coinciding with the coming to power of a Conservative-led government (Swenson 1991).

But how profound was this threat? A second glance reveals a more nuanced picture. Sick-leave policy has been changed, and the Social Democrats were able easily to win the September 1994 election. At the local level, cooperation among business, labour and public agencies dealing with employment, training, and research and development has remained very much alive, while sectorial bargaining combined with coordination of demands on the labour side has meant that labour market outcomes have not been all that different from those under centralized bargaining.²⁰ And, when the economy worsened, Swedish business was prepared to look to corporatist solutions. In the "crisis deal" of September 1992, the political allies of business and labour reached a wide-ranging agreement on economic and social policies to overcome financial problems, as they did on a complete overhaul of the pension system in January 1994.

The real threat lies in Sweden's macroeconomic situation. With high unemployment and very large deficits, the Swedish Social Democrats face major difficulties in maintaining the SDWS programmes they were brought back to power to preserve. The recent budget made major moves in that direction but also brought in tax increases that business condemned as impeding growth. Moreover, interest rates are higher than they should be because the financiers have been sceptical of the government's fiscal rectitude despite significant spending cutbacks. This is partially because the SAP government, in order to keep its supporters from deserting it for parties to its left, is seen to have underplayed the severity and duration of the problem.²¹

The crisis is by no means over. But Sweden is the outlier, in a far worse position in terms of public debt and deficit than Denmark, Norway, Austria - and even Finland. Sweden was the slowest growing OECD country one from 1987-93, {OECD Outlook, June 1993).Yet Norway ranked third and Denmark a respectable eleventh in economic growth among the 24 OECD countries between 1990 and 1994 (The Economist, January 28, 1995, 100), and both moved up, Norway to second place and Denmark to seventh in 1995 (The Economist, July 1, 1995, 96). Even Finland, despite its astronomical unemployment, is growing rapidly.

In sum, as long as cultural and political institutional arrangements remain largely intact, I suggest that the continuing adjustments, painful as they may be, will reinforce more than undermine corporatist institutional arrangements. Sweden will go through the kinds of difficult adjustment that Denmark went through in the early 1980s inside the EC, in particular having to learn to live with unemployment figures comparable to those of other OECD countries.

But even such mitigated optimism needs to be guarded. How strongly will
national systems of public electronic broadcasting stand up to the
commercial satellite networks? And could existing political structures

remain intact within a federal EU (Milner 1995)? For now, these prospects still seem far off. Now that they are members, Sweden and Finland will in all likelihood join Denmark as opponents of a federal Europe, either blocking the project, or, at worst, keeping to an outer second tier of the EU, leaving the inner tier to the Eurofederalists in Germany and the Benelux countries.

Thus, if this analysis is right, Sweden like Denmark, Norway, and even Finland,²² should prove likely to withstand the present challenge, reinforced by the choices of people with the knowledge to preserve long-standing institutional arrangements and the outcomes they have produced.

---

NOTES

1. "Marxism is basedon ... one crucial assumption - the proletariat's natural affinity for socialism - for which it has never been able to supply even a vaguely plausible rationale and for which the historical evidence is overwhelmingly negative" (Van der Berg 1988, 501). The explanation is hardly mysterious. Even if class interest accounts for a position espoused by an individual, her actual behaviour cannot be induced from that position, for it fails to take "opportunity cost" into account: whether the net benefit derived from compliance outweighs the cost paid in the form of resources (time, energy, money . . .) no longer available for other purposes.

2. The term is found in the work of Esping-Andersen, who, with Korpi, developed a highly influential analysis of the welfare state as the expression of the interests of the working class. Key programmes seek to "decommodify" labour (Kolberg & Esping 1992), an outcome assumed to be intrinsically in the interests of the workers.

3. "The choice theoretic approach is essential because a logically consistent, potentially testable set of hypotheses must be built on a theory of human behavior" (North 1990, 5).

4. Rational choice theory assumes that under the same circumstances and with the same knowledge, individuals with the same preferences will make the same choice. Thus, while allowing for culturally based differences in preferences and knowledge, it starts from the only possible common ground for transcending cultural relativism. Similarly, the rational choice approach breaks down the barrier between the producer and consumer of knowledge, "persuading the reader that the action undertaken was optimal and that she would have adopted the same course of action under the circumstances" (Tsebelis 1990, 46-47).

5. "Because of their individual power and small number . . . agents have the incentive to behave strategically, taking explicit account of how their actions will affect others - and what the resulting feedback to themselves will be. ... As the theory of repeated games shows, such [ongoing and repeated interactions] are conducive to compromise and to forms of cooperative behavior which are to the mutual benefit of all concerned" (Rowthorn 1992, 84-85; see also Calmfors & Drifill 1988, and Freeman 1988).

6. Note that extreme inegalitarianism as well as extreme egalitarianism are impediments to efficiency. For example, income inequality, as measured by the multiple of the income of the richest to the poorest 20 percent of households, is lowest in the three fastest growing industrializing economies of the 1980s (China, South Korea and Taiwan), while it is highest in the slowest growing ones (a multiple of 4 to 6.5 compared with one of 11 to 15 (The Economist, 16 November, 1991, Survey 17-19). Thus the U-shaped curve is understood to apply only within boundaries on each side set in effect by the policies acceptable to a majority of the population in the economically advanced, liberal democratic regimes.

7. This critique was articulated, for example, in the report of the Commission set up in 1992 to investigate Swedish economic policies and political institutions and chaired by leading economist, Assar Lindbeck. The Commissioners, though asserting that their goal was to preserve the essentials of the welfare state and reduce unemployment, were not kind to the Swedish corporatist model. In the name of flexibility they called for further decentralization of collective bargaining and dismantling of corporatist institutions. A summary of the Lindbeck Report as well as several critical assessments are to be found in Economic Policy (October 1993, 220-263).

---

7. This critique was articulated, for example, in the report of the Commission set up in 1992 to investigate Swedish economic policies and political institutions and chaired by leading economist, Assar Lindbeck. The Commissioners, though asserting that their goal was to preserve the essentials of the welfare state and reduce unemployment, were not kind to the Swedish corporatist model. In the name of flexibility they called for further decentralization of collective bargaining and dismantling of corporatist institutions. A summary of the Lindbeck Report as well as several critical assessments are to be found in Economic Policy (October 1993, 220-263).

8. Gini coefficients are derived by dividing the area between the Lorenz curve and the diagonal by the total area under the diagonal. With good, comparable, up-to-date Gini data not available for many countries, an often-used substitute is the ratio of shares of national income of top and bottom quintiles. In addition, a rating can be based on a summing of more than one indicator. See Freeman (1989) for examples of "blends of distributional" equity. Note that we are interested in outcomes in so far as they affect individuals. This is why we select such comparative measures as wage dispersion, level of employment and post-tax Gini coefficients, rather than level of public or social spending.

9. A more sophisticated approximation of "R" might be the score on the Rowthorn composite scale derived by subtracting wage dispersion (the inter-industry coefficient of variation in earnings) from employment (the proportion of the population aged 15-64 that is employed, with part-time jobs tabulated as half units - Rowthorn (1992)).

10. On the Rowthorn composite scale, the clearly corporatist Nordic countries occupy the top four places, followed by Japan - which defies locating on the corporatism axis (see Dore 1990). The relative "R" ratings for most states can be expected to remain constant under the different combinations using such a composite scale since - with devaluation effectively ruled out by European economic integration - corporatist societies increasingly face a trade-off between keeping unemployment low and reducing wage-income dispersion.

11. The Luxembourg Income Survey (LIS) has carried out the most thorough and genuinely comparable studies of income inequality in Western countries using Gini coefficients. Using households as units and assigning different weights to the first adult and the remaining members of the family, and including transfers and excluding taxes, LIS studies reveal a sharp contrast between the US and Canada on one side, and Sweden, Norway and Finland on the other, with Gini levels of 32.6 and 29.9 versus 20.5, 22.5 and 24.3 respectively (Uusitalo 1989, 80).

12. A leading rational choice theorist who is willing to take this proposition seriously is Mancur Olson. Olson is suspicious of neocorporatism, which he characterizes as counting on powerful groups to act in the interests of the weak in society rather to siphon the benefits of redistributive policies to their members. Yet he is open to the possibility of arrangements under which very large organizations cooperate for the common good, something he suggests is manifested in the relative absence of subsidies and tariff protection for Swedish cartels (Olson 1990).

13. This interpretation corresponds with the results of attitudinal surveys. The vast majority of Swedes and Americans interviewed by Steinmo answered "yes" when asked: "if you could be guaranteed that increased government spending would be efficiently and effectively used to address society's problems, would you agree to an increase in your taxes?", but the Americans far more often added that "they do not believe that revenue from higher taxes would be used efficiently or effectively and therefore would not approve tax increases if given the choice" (Steinmo 1993, 199). The explanation is also helpful in making sense of different approaches in corporatist and laissez-faire societies to the same policy choices since these may very well differently affect welfare in a corporatist as opposed to a non-corporatist society. For example, Iversen (1994) finds that corporatist societies have been significantly better at keeping unemployment low under a flexible monetary policy regime, while noncorporatist ones require a non-accommodating regime to do so. Or consider immigration policy. In the context of non-corporatist institutional arrangements, high immigration levels contribute not only to the welfare of the new immigrants but also, potentially, to that of the society as a whole. On the other hand, the arrival of a large number of immigrants with very different cultural backgrounds can undermine the (cultural expectations underpinning) redistributive institutional arrangements of the SDWS, thus reducing welfare. To improve the lives of people in poor countries, the citizens rationally support generous foreign aid policies rather than the relatively open immigration policies in North America.

---

13. This interpretation corresponds with the results of attitudinal surveys. The vast majority of Swedes and Americans interviewed by Steinmo answered "yes" when asked: "if you could be guaranteed that increased government spending would be efficiently and effectively used to address society's problems, would you agree to an increase in your taxes?", but the Americans far more often added that "they do not believe that revenue from higher taxes would be used efficiently or effectively and therefore would not approve tax increases if given the choice" (Steinmo 1993, 199). The explanation is also helpful in making sense of different approaches in corporatist and laissez-faire societies to the same policy choices since these may very well differently affect welfare in a corporatist as opposed to a non-corporatist society. For example, Iversen (1994) finds that corporatist societies have been significantly better at keeping unemployment low under a flexible monetary policy regime, while noncorporatist ones require a non-accommodating regime to do so. Or consider immigration policy. In the context of non-corporatist institutional arrangements, high immigration levels contribute not only to the welfare of the new immigrants but also, potentially, to that of the society as a whole. On the other hand, the arrival of a large number of immigrants with very different cultural backgrounds can undermine the (cultural expectations underpinning) redistributive institutional arrangements of the SDWS, thus reducing welfare. To improve the lives of people in poor countries, the citizens rationally support generous foreign aid policies rather than the relatively open immigration policies in North America.

14. Frohlich & Oppenheimer (1992) conducted just under 100 group experiments among university students in the US, Canada, and Poland. Almost 80 percent of the groups presented with a series of possible distributions of the earnings derived from the efforts of the individual members arrived at a consensus over a principle of distribution which taxed all members to guarantee a specific income floor as a safety net for those with the lowest earnings. Moreover, this floor was respected in choices taken in subsequent experimental practice with positive consequences upon productivity.

15. Tax evasion is the most common manifestation of free-riding. I evade on the premise that everybody else does not, thus leaving the existing distribution unaffected since my tax contribution is negligible. But whatever motivates me to free ride likely motivates my neighbour with whom, in the language of game theory, I am in a Prisoners' Dilemma, for which the solution is institutional, especially in the form of stringently enforced tax laws.

16. "Tax ethics are negatively and significantly correlated with a general distrust of other people and the suspicion that others cheat; educational level, knowledge of fiscal matters and belief in perceived fairness of the tax burden are all positively correlated with tax ethics" (Lewis 1982, 153-154).

17. Such considerations lead Tullock (1975) to favour the continental European "inquisitorial" court over the adversarial one in the US, which promotes rent-seeking on the part of lawyers. Americans pay exorbitant medical fees to cover insurance premiums for physicians and other professionals at the mercy of a system of litigation which, by allowing lawyers to be paid on a contingency basis, places no disincentives on taking court action.

18. According to detailed studies of productivity in the public sector conducted by the Swedish Board of Auditors (Statskontoret), the general decline in productivity was slowed or halted in the state sector but continued in the municipal and regional sector in the 1980s.

19. In 1990, SAF officially abolished its central bargaining unit. In 1991, it removed its representatives from the boards of central decisions-making public bodies, and called for the abolition of state contributions to voluntary bodies, another key feature of Nordic corporatist institutional arrangements. Employers were responding to the decline of Swedish competitiveness, but also to suspicions of Labour's willingness to maintain the old corporatist compromises, suspicions stimulated by a certain Marxist rhetoric surrounding labour-market initiatives of the early 19705, especially the wageearner funds (Milner 1989, 129-138).

20. The results of a 1992 Swedish survey demonstrate this. The sample was made up of the personnel directors of 300 representative manufacturing firms (employing 40 percent of the sector's employees), of which 179 responded. One question asked how they thought their company's wage level would be affected if all negotiations were conducted at the level of the firm. "There was no clear consensus on the benefits from a move from the current system with both central and local bargaining to a system of only local bargaining. It is indeed hard to escape the conclusion that firms are much less convinced about the advantages of decentralized bargaining than their own central federation" (Agell & Lundborg 1993, 18).

21. Sweden's Social Democratic government found it especially difficult to cut back on expenditures affecting public sector workers. One explanation is the fact that, unlike its Nordic sister parties, the SAP has not had to depend on support from parties of the centre to govern. (Its traditional support from almost 45 percent of the electorate made it an exception in Scandinavia and elsewhere.) Like external pressure from the EC upon Denmark, dependency on the cooperation of a party of the centre can be a means of reducing the political costs of economic restraint. It is possible that the loss of more than 5 percent of its vote to the Left People's Party and the Greens in early 1995 will spell the end of an era: that heretofore the SAP's position and strength will parallel that of its three sister parties.

---

22. Unique factors make both Finland and Norway exceptional - in opposite directions. Finland was especially hard hit when the recession at the beginning of the 1990s coincided with the collapse of the Soviet market, which accounted for up to 25 percent of its exports. Norway was sheltered from the recession by its vast oil and gas production which had come on stream in the preceding decade.

REFERENCES

Agell, J. & Lundborg, P. 1993. Theories of Unemployment: Survey Evidence from Swedish
Manufacturing Firms. Stockholm: lUI.

Bergounioux A. & Manin, B. 1989. Le Regime Social-Democrate. Paris: Presses Universitaires
de France.

Calmfors, L. & Drifill, E. J. 1988. "Bargaining Structure, Corporatism and Macroeconomic
Performance", Economic Policy 6, 13-47.

Dore, R. 1990. "Japan: A Nation Made for Corporatism", in Crouch, C. & Dore, R. eds.,
Corporatism and Accountability. Oxford: Clarendon.

Freeman, R. 1988. "Labor Market Institutions and Economic Performance", Economic Policy
6, 64-80.

Freeman, J. R. 1989. Democracy and Market. Ithaca, N.Y.: Cornell University Press.

Frohlich, N. & Oppenheimer, J. A. 1992. Choosing Justice: An Experimental Approach to
Ethical Theory. Berkeley: University of California Press.

Iversen, T. 1994. "Macroeconomic Policy Regimes and the Politics of Redesigning Wage
Bargaining Institutions". Presented at the American Political Science Association, Annual
Meeting, September 1994, New York, NY.

Keman, H. 1984. "Politics, Policies and Consequences: A Cross-national Analysis of Public
Policy-formation in Advanced Capitalist Democracies (1967-1981)", European Journal of
Political Research 12, 147-169.

Kenworthy, L. 1995. "Equality and Efficiency: the Illusory Tradeoff", European Journal of
Political Research 27, 225-254.

Kolberg, J. E. & Esping-Andersen, G. 1992. "Welfare States and Employment Regimes", in
Kolberg, J.-E., ed., Between Work and Social Citizenship. Armonk N.Y.: M.E. Sharpe.

Lewis, A. 1982. The Psychology of Taxation. New York: St-Martin's.

March, J. G. & Olsen, J. P. 1989. Rediscovering Institutions: The Organizational Basis of
Politics. New York: The Free Press.

Milner, H. 1989. Sweden: Social Democracy in Practice. Oxford: Oxford University Press.
Milner, H. 1993. "In Defense of the European Model". Presented at the Joint Workshops of
the European Consortium for Political Research, Leiden, The Netherlands, April, 1993.

Milner, H. 1994. Social Democracy and Rational Choice: The Scandinavian Experience and
Beyond. London: Routledge.

Milner, H. 1995. "Rational Choice, Political Participation and Institutional Arrangements".
Presented at the March 1995 meeting of the Public Choice Society, Long Beach, California.

North, D. C. 1990. Institutions, Institutional Change and Economic Performance. New York:
Cambridge University Press.

Olson, M. 1990. How Bright are the Northern Lights? Some Questions about Sweden. Lund:
Lund University Press.

Paloheimo, H. 1984. "Political Struggle and Economic Development in Developed Capitalist
Countries", European Journal of Political Research 12.

Rowthorn, B. 1992. "Corporatism and Labor Market Performance", in Pekkarinen, J., Pohjola,
M. & Rowthorn, 8., eds., Social Corporatism: A Superior Economic System? Oxford:
Clarendon Press.

Schmitter, P. C. 1981. "Interest Intermediations and Regime Governability in Contemporary
Western Europe and North America", in Berger, S. et al., eds., Organizing Interests in
Western Europe. Cambridge: Cambridge University Press.

Steinmo, S. 1993. Taxation and Democracy. New Haven: Yale University Press.

Fall and Rise of Labor Market Dualism". Presented at the SASE/lAREP Conference.
Stockholm, June, 1991.

Tsebelis, G. 1990. Rational Choice in Comparative Politics. Berkeley: University of California
Press.

Tulloch, G. 1975. "On the Efficient Organizations of Trials", Kyklos, 28, 745-62.

Uusitalo, H. 1989. Income Distribution in Finland. Helsinki: Central Statistical Office of
Finland.

Van der Berg, A. 1988. The Imminent Utopia: From Marxism on the State to the State of
Marxism, Princeton NJ: Princeton University Press.