Process Model to Undersatnd Development of
Public Pension Systems in Germany & Japan
著者名(英)
Steven GREEN H
journal or
publication title
Toyohogaku
volume
54
number
2
page range
252-232
year
2010-12-20
URL
http://id.nii.ac.jp/1060/00000798/
Creative Commons : 表示 - 非営利 - 改変禁止《 論 説 》
Macro-Historical Perspectives and the Power Process
Model to Understand Development of Public Pension
Systems in Germany & Japan
H. Steven Green
Public policies are enforced as law and laws are based on claims about both the way the world is and the way it ought to be. This paper presents a research methodology for comparing public policies in two different nations that is both empirical and normative to reflect the fact that claims about both is and ought are inherent in the object of study. In particular, this considers the historical development of state institutions for the pro-vision of pensions for the aged.
Germany and Japan are useful cases for comparing the politics of pension reform because they represent “conservative-corporatist” welfare regime types and they face similar demographic pressures from shrinking workforces and expanding numbers of retirees. However, the politics of pension reform in Germany in the first decade of the 21st Century produced a surprising outcome: the decision by the German government in
2001 to allow workers to divert a portion of their pension contributions to private funds. Why Japan did not create a similar policy is a puzzle. The German case also raises the question of whether welfare systems in general can persist in the face of a changing ratio between workers and retirees. This paper justifies this choice of cases and explains the basis for a particular methodology to study them.
I. Introduction
Can democratic governments continue to extract immediate costs from their elector-ates to pay for a future benefit? At the beginning of the 21st Century changing
demo-graphics in the advanced democracies have created a challenge that imperils the ability of these states to maintain their national pension systems. The ratio of workers to reti-rees is decreasing. Those in the workforce today face cuts in their expected retirement benefits tomorrow and many younger workers wonder if they may not outlive their country s pension system. Different groups develop their own stake in particular pro-grams of the welfare state, but in the wealthy democracies all workers contributes to pension systems, and benefits are nearly universal.（ 1 ） Pension systems therefore offer
an excellent subject and study to answer the question of how, or whether, the elector-ates in democratic stelector-ates can endure pain now for gain later.
Pension systems look like large, immovable objects off of which bounce the efforts of reformers to change them. Declining fertility rates and rising aged-dependency rates -the ratio of retirees to workers - in the advanced democracies have created a demo-graphic challenge that imperils not only the sustainability of national pension systems, but risks tearing the fabric of the so-called social contracts in these nations as well. Rescuing old-age pension systems from insolvency may seem like trying to move an immovable object. Given the popularity of these programs, plus the fact that older peo-ple vote in high numbers, politicians may be loathe to push or to pull too hard lest they get “crushed” by defeat in the next election.
However, an intergenerational conflict may loom on the horizons of the advanced in-dustrialized democracies. States may decide they cannot afford to preserve one of the main pillars of the post-war welfare state. Government pension fund managers must figure out how to get more out of a shrinking workforce, even as they may have to tax it at higher rates to pay for old-age pensions. This unhappy calculation generates the potential for conflict across age cohorts: retirees who expect a return on their working
（ 1 ） Not all payers receive benefits. Non-documented workers, for example, cannot count on a pension benefit and, in some nations a minimum number of years worth of contributions are required in order to receive any benefit. In Japan, for example, workers must pay contributions for at least 25 years in order to qualify for a pension benefit.
life-long payments versus contemporary workers who balk at paying more now for less in the future. Is demography destiny? Not necessarily, but if a society s elderly popu-lation and life-expectancy increase without fertility rates keeping pace to replace retir-ing workers, then pension systems become strained.
The governments of both Germany and Japan have presented demographic shifts in their nations as crises that compel them to exert downward pressures on worker contri-butions to pension programs. From the state s view retrenchment is an inevitable conse-quence of low fertility and aging populations （the two trends tend to move hand-in-hand in the developed world, which enjoys longer life-spans.） Herman Schwartz observes that welfare rights generally have been construed to be social rights when, in fact, he points out, they are property rights.（ 2 ）
Protection of capital from market sources generates income for producers, as does the protection of wages for workers. Contributions to old-age pension programs protect wages in as much as they force workers to earmark a certain amount of them for pay-ment upon retiring. Germany and Japan operate so-called pay-as-you-go, or “paygo” pension systems. Paygo systems mandate income transfers to retirees from earnings by current workers, while promising workers a similar benefit upon their exit from the la-bor force.
Gosta Esping-Andersen suggests that, “pensions constitute a central link between work and leisure, between earned income and redistribution, between individualism and solidarity... Pensions, therefore, help elucidate a set of perennially conflictual prin-ciples of capitalism.”（ 3 ） The designers of pension systems sought to facilitate the
de-velopment of industrial capitalism in exchange for promises of a secure, basic standard of living in retirement （including the benefit of being able to leave the toil of work it-self.） As long as they expect to receive benefits in what they perceived as a fair
propor-（ 2 ） Cf. Herman Schwartz, “Round up the Usual Suspect! Globalization, Domestic Politics and Welfare State Change,” in New Politics of the Welfare State, Paul Pierson, ed. Cambridge, UK: Cambridge UP: 2001 （ 3 ） Esping-Andersen, Gøsta 1990: The Three Worlds of Welfare Capitalism. Princeton, NJ: Princeton UP, p.80
tion both to what they contributed over the course of their careers as well as to what their parents and grandparents received in retirement, workers in the wealthy democra-cies have accepted the pension component of the post-war social contract. States threaten not only to transfer property from workers today but also to not return it in equal, much less greater, share tomorrow. Today s workers face the prospect of having to support today s pensioners, as well as extend their time in the labor force- all the while knowing that they will receive less in pension benefits, as a proportion of their current income, and work more years than their predecessors. German and Japanese workers face similar threat to their pension rights most acutely among the G 5 nations as their populations are aging the most rapidly. Yet, their governments have proposed different solutions to the same problem.
II. Comparative Methodology and “Small-n”
Germany and Japan share similar political and economic structures, which have been less liberal and more regulated than the Anglo-American capitalist democracies, be-cause both states developed along nearly identical paths of conservative modernization, or “revolution from above.”（ 4 ） In fact, the Meiji oligarchs who directed Japan s
late-19th Century industrialization modeled elements of both their 1890 constitution and
so-cial welfare policies upon the Prussian constitution and the welfare policies of Otto von Bismarck. “Despite their great differences culturally, Germany and Japan were remark-ably similar in the reasons for their historical aversion to liberalism,”（ 5 ） although
liber-al democracy would take root in both nations after being implanted by the same occu-pying powers - the United States and the Allies - following World War Two.
A study of these two cases produces inferences about a particular universe of cases;
（ 4 ） Cf. Moore, Barrington, Social Origins of Dictatorship and Democracy: Lord and Peasant in the Making
of the Modern World （Boston, MA: Beacon Press Books. 1966）
（ 5 ） Tony Smith, “The International Origins of Democracy,” in Democracy, Revolution and History, Theda Skocpol, ed （Ithaca, NY: Cornell UP 1998）, pp. 191 210
first, about other conservative corporatist welfare regimes and, second, about other wealthy democracies undergoing public pension reform processes. This universe is smaller than the number of variables necessary to test in order to develop a theory be-cause there are not 32 wealthy democracies. The comparative method solves the prob-lem of having an insufficient number of cases to test all independent variables, which is also known as the problem of small-n. The econometric, or quantitative, approach to social science research claims that inferences may not be drawn from research with fewer than 32 observations, or cases. “Is inference fundamentally quantitative?” asks Timothy Mc J. McKeown.（ 6 ） My research design follows the lead of qualitative
re-searchers such as McKeown and answers in the negative. Both qualitative research and quantitative research generate inferences when applied to the research question for which each is best suited.
Quantitative researchers first objection to inferences is that small-n studies cannot contain enough degrees of freedom, because they do not utilize the experimental meth-od, which is to perform statistical regression analyses to find correlations between phe-nomena in large-n data sets. King, Keohane and Verba （KKV） argue that causal infer-ences can only be drawn from data sets with a high degree of freedom, or “a high number of data-set observations vis-à-vis the number of parameters- of which there is usually one.”（ 7 ） Because KKV view all empirical activity in social science as “the
making of discrete observations, which are represented as values assigned to varia-bles”（ 8 ） their recommendation to researchers studying only a few cases is to increase
the number of observations within each case. In fact, even in a design based on many cases and a few variables, KKV still identify as “the fundamental problem of causal in-ference,” that, “we can never hope to know a causal effect for certain.”（ 9 ） Scholars
（ 6 ） Timothy J. McKeown, “Case Studies and the Limits of the Quantitative Worldview,” in Rethinking
Social Inquiry: Diverse Tools, Shared Standards, Henry E. Brady and David Collier, ed. （Rowman & Lit-tlefield Publishers, Inc.: New York, NY）, p.144
（ 7 ） Ibid., p.284 （ 8 ） Ibid., p.145
such as KKV assume that reality is deterministic, but our ability to measure enough variables to demonstrate that fact is limited so we can reach conclusions based only on probabilities. These scholars are suspicious of small-n researchers who claim to search for causality.（10）
How do qualitative researchers respond to these claims? One response is that the re-gional expertise of a small-n researcher allows her develop more from a case than a single, discrete observation. Close knowledge of the cultural, historical and political context of a case allows the researcher to assess whether and to what degree hypotheses apply across cases.（11） Dietrich Rueschemeyer argues that when the case in question is
process of historical development, such a “long sequence” inherent in the “case histo-ry” provides many “theoretically relevant observations that may rule out or suggest the revision of a whole series of propositions.”（12） Comparativists also get around the
small-n problem through “pattern-matching” within the same case, when that case has developed over time.（13） In particular, historically situated cases require hypothesis
gen-eration, testing and revising of propositions within the same set of material. This proc-ess of matching patterns within the same case can actually link empirical evidence to “analytic intent” more closely than most large-N studies because the object of study requires a more intimate knowledge of cases than a quantitative model. Simply put, de-tailed case studies require expert knowledge of context. As Timothy McKeown writes of the matter of the value of one or a few cases, “Here, what guides research is not log-ic, but craftsmanship, and the craft in question is implicitly far more substantively rich
（ 9 ） Gary King, Robert Keohane and Sydney Verba, Designing Social Inquiry （Princeton , NJ: Princeton UP, 1994）, p.79
（10） James Mahoney, “Strategies of Causal Assessment,” in James Mahoney & Dietrich Rueschemeyer,
Com-parative Historical Analysis in the Social Sciences （Cambridge Studies in Compartive Politics） （Cambridge UP: Cambridge, UK, 2003）, pp.339 341
（11） Gerardo L. Munck, “Tools for Qualitative Research,” in Brady & Collier, p.110
（12） Rueschemeyer, “Can One or a Few Cases Yield Theoretical Gains?” in Mahoney and Rueschemeyer, p. 311
than that of ‘social scientist without portfolio. The latter s lack of context-specific knowledge means that the researcher cannot call on information from outside the sam-ple being analyzed to supsam-plement the information gleaned from the quantitative analy-sis.”（14） McKeown advocates for a “folk Bayesian approach” that lets the researcher
make sense of her research results with preconceptions derived from theory and con-textual knowledge.（15）
Researchers may also employ an informal method of Boolean algebra to realize causal homogeneity such as Paul Pierson s concept of path dependency. This approach identifies variables that place similar cases on different trajectories and, in so doing, explains casual heterogeneity.（16） James Maohoney suggests that certain deterministic
methods take us safely around the pitfalls of both determinism and degrees of freedom pointed out by KKV, et al. First, what do researchers mean by determinism? Mill s methods provide tools for “systematically eliminating rival causal hypotheses, even when only a small number of cases are selected.”（17） Mill s method of agreement is a
useful heuristic for eliminating potential necessary causes, and his method of differ-ence allows us to eliminate potential sufficient causes. Second, qualitative methods rely not only on nominal variables, such as are used in Mill s methods, to eliminate rival hy-potheses, but on other strategies of causal assessment, which resolves the degrees of freedom.
Comparing the development of two similar state types, avoids the problems of de-grees of freedom and of selection bias because test hypotheses will develop within cas-es over time. Scholars in the macro-historical have already matched certain patterns relevant to my case studies: Considering the similarity in the paths Germany and Japan took toward economic development, it is not surprising that they adopted similar
wel-（14） Timothy McKeown, “Case Studies and the Limits of the Quantitative Worldview,” in Brady and Collier, p.148
（15） Ibid., p.161
（16） Gerardo Munck, “Tools for Qualitative Research,” in Brady & Collier, p.110 （17） Mahoney, in Mahoney & Rueschemeyer, p.351
fare systems. But, given the similarities in their welfare systems, it is surprising that they have taken such different approaches to reform recently, a hypothesis I could gen-erate only through detailed knowledge of both cases. In Section IV I apply Mill s meth-od of difference to select a particular variable that is likely to provide causation.
III. State Building and Welfare Regime-Type in Germany and Japan
Historical similarities between political and social policy development in Germany and Japan justify their selection as cases in a small-n study. I apply Mill s indirect method of difference to a timeline of nominal variables representing state and welfare regime development in Germany and Japan （see Figure 1 ）. In the method of most similar cases （modeled after J.S. Mill s method of difference） cases are matched for as many variables as possible, but modeled to differ on one. The unmatched variable sup-posedly establishes the different outcomes across the two cases.（18） In the method of
most different cases, all variables between cases differ except one, which is assumed to explain similar outcomes. No two cases can possibly match in all variables but one, nor can they differ in all but one variable. The indirect method of difference combines the methods of agreement and difference. In assigning causation, the indirect method ex-amines what variables are present in a given outcome and disqualifies variables that are not present whenever a particular outcome occurs. （However, this method is unable to explain multiple causation.）
Figure 1 presents an indirect method of difference model to illustrate which varia-bles may explain the different outcomes in these similar cases. Variavaria-bles A- F include defining features of the German and Japanese states and welfare regimes, and variable G represents a shared quality of their populations, namely its ratio of retirees to work-ers. In figure 1 variables A-F match in both cases. However, variables H, I, J and K differ: H and J, which are marked in the columns for Japan and Germany respectively,
indicate differences in organized labor and I and K indicate different responses to the oil shocks of the 1970s. Variable I follows from H and K follows from J: the presence of strong, horizontal unions （J） in Germany is a necessary condition for social policy （K）.
The absence of J in the case of Japan facilitated the government s choice of fiscal policy in response to the oil shocks. To alleviate rising unemployment the German gov-ernment encouraged early retirement through a negotiated settlement with labor union leaders that promised increased pension benefit levels to workers in return for an earlier exit from the labor force.（19） The indirect method frames my inquiry and implies that
（19） Nobuhiro Hiwatari, Economic Downtown, Unemployment and Policy Reform in Germany and Japan:
Wage Moderation and Compensation Policies in Coordinated Market Economies, （Paper presented at the 2002 American Political Science Meeting, Boston, August 2002）, p. 1
Figure 1 Similar Cases, Different Outcomes: Using J.S. Mill s Indirect Method of Dif-ference to Understand the Divergent Paths to Pension Reform Outcomes in Japan and Germany.
Japan Germany
A. Era of Industrialization From mid-19th century: “Revolutions from Above”
B. Regime type leading development Bureaucratic, centralized agrarian states C. Outcome of developmental path Expansionism leading to WWII
D. Origins of liberal democracy Imposed from without: US-led democratization E. Type of democracy Parliamentary: corporatist-capitalist
F. Welfare regime type Corporative state-dominated G. Aged dependency rate by 2030 Highest among affluent democracies
Labor H. Intra-firm; weak
na-tionally
J. Inter-firm; strong na-tionally
Response to Oil Shocks I. Fiscal policy K. Social policy
the combination of state-labor relations and the impact of those relations on economic policy will explain different outcomes. In short, the German government s decision to create a privatized component to the pension program is the result of a path taken in the 1970s, which emphasized using social policy, not fiscal policy, to cope with fluctua-tions in employment and welfare state costs.
Work in the macro-historical tradition by Skocpol and Moore explains the structural conditions that determined the points of departure for economic and state development in Germany and Japan. Industrialization occurred in both nations during the latter half of the nineteenth century through a coalition of landed elites and a rising bourgeoisie. Barrington Moore has classified the Meiji Restoration of 1868 as well as the Stein-Hardenburg Reforms between 1807 and 1814, and Bismarck s Unification in the 1860s as “revolutions from above.” For Moore, these revolutions were facilitated by a coali-tion forged between the labor-repressive upper class and the bourgeoisie. Moore classi-fies the form of industrialization these states followed as “conservative moderniza-tion”（20） because it fostered rapid economic development without infringing on the
power of the landed aristocracy. Even as efforts were made to establish semi-parlia-mentary institutions, elites in neither nation allowed constitutional changes, such as universal male suffrage, that would have given more rights to both rural and urban la-bor, or reduced the holdings of the landlord class. Industrialization began and could only proceed upon a foundation of suppressed labor and a state captured by the inter-ests of landed elites.
Theda Skocpol faults Moore for ignoring the role of the state and she urges that Moore s “Marxist political sociology [be] supplemented by an understanding of the po-tential of bureaucratic political structures for sustaining cohesive elites” （italics in original）.（21） Skocpol illustrates that the 19th Century agrarian states of Japan and
Ger-many are better understood as bureaucratic and centralized agrarian states. Neither
（20） Barrington Moore, Social Origins of Dictatorship and Democracy: Lord and Peasant in the Making of
state was staffed solely either by members of the aristocracy or from private capital; both states were staffed by trained bureaucrats, many of whom owned neither land nor capital, who created policy from options determined by the structures of their society and with trained militaries at their disposal.
Skocpol s explanation allows us to understand why the world s first universal pension policy was born not in a democratic polity but in the Bismarck regime. During the 1880s Kaiser Wilhelm I and Chancellor Otto von Bismarck “utilized a large and effi-cient bureaucracy of neo-absolutist origins to deflect a socialist challenge to Imperial rule.”（22） In the package of national health, industrial-accident and old-age insurance,
Bismarck addressed, in his words, “those socialist demands which seem justified and which can be realized within the present order of society.”（23） The pension policy
craft-ed by Bismarck and Wilhelm I resembles the one in place in contemporary Germany: Bismarck established a mandatory retirement age, and tied benefits to contributions, which were collected from payroll taxes.
That the first universal welfare policy came out of a non-democratic setting high-lights two important facts about how welfare policy develops. First, the role played by labor was, and remains, central in Germany and has been central to European welfare programs. （Older democracies removed a primary impetus to labor unions having es-tablished universal male suffrage prior to industrialization.） This role receives further attention in the literature review below. For now, it should be noted that, while labor s role has remained strong in Germany, in Japan is has never been nearly as influential. Japan did not develop a pension program until 1941, and it was not until 1961 that the pension system was organized into its current form. The Japanese state more success-fully repressed labor during industrialization and, therefore, did not face a
well-organ-（21） Theda Skocpol, “Review of Barrington Moore s ‘Social Origins of Dictatorship and Democracy ,” in
Social Revolutions in the Modern World （Cambridge, UK: Cambridge UP, 1994）, p.41
（22） Alexander Hicks, Social Democracy and Welfare Capitalism: A Century of Income Security Politics （Ith-aca, NY: Cornell UP, 1999）, p.43
ized group demanding welfare programs or threatening of revolution.
Second, welfare policy, and by extension, pension policy, is intertwined with notions of citizenship rights. Bismarck understood well that placating workers would earn their support. Welfare provides certain rights, and also binds people to certain duties: Brian Downing even connects pension policy to the drive for imperialism in 19th-Century
Germany（24）. Nearly sixty years after Bismarck s policies were introduced, Hitler and
the National Socialists understood the relationship between welfare and citizenship and offered a social insurance package to the citizens of the newly-occupied territories of Western Europe. Between 1940 and 1942 Nazi occupiers in Norway, Denmark, Bel-gium and the Netherlands modified pre-existing social insurance policies, in accord-ance with German practices. Furthermore, they extended these benefits to those labor-ers expatriated to Germany and these “agreements restated the principle of treatment on an equal footing with German nationals” as concerned social insurance laws.（25） By
granting non-Germans similar pension rights as German nationals, Nazi officials solved the principle-agent problem: conquered subjects had a personal stake in the for-tunes of the Third Reich s economy and would be less-likely to try to undermine the Reich.
The specific features of the pension systems described in the previous section reside as part of broader strategies for citizenship rights and duties of both the German and Japanese state. We do not fully comprehend any pension system s efficacy in terms measured solely by benefits paid. Esping-Andersen s observes, “Expenditures are epi-phenomenal to the theoretical substance of welfare states.”（26） As Germany s Nazis
un-derstood, pension policies bind people to the economic order in which they work, pro-viding both an incentive to work within that order and a disincentive to challenge it.
（24） Brian Downing, The Military Revolution and Political Change （Princeton, NJ: Princeton UP, 1992）, p.105 （25） Erna Magnus, “Social Insurance in Nazi-Controlled Countries,” Political Science Quarterly, Volume 59,
Issue 3 （Sep.,1944）, pp. 388 419 （26） Ibid., p.19
Different types of “pension regimes”（27） reflect different types of state development as
well as different goals of states.
Gosta Esping-Andersen s The Three Worlds of Welfare Capitalism facilitated new cross-national research in the field of welfare state studies. Unlike previous scholarship that studied expenditure levels as units of analysis across different industrialized na-tions, Esping-Andersen compares the different structures - founded upon different sys-temic principles - of welfare states and discovered three ideal-typical welfare regimes: liberal, conservative-corporatist and social democratic. The conservative-corporatist re-gime helped to foster rapid reconstruction and development in the post-war decades of the 1950s through the 1970s. It freed capital for industrialization, instead of large-scale redistribution, and encouraged families in general and women in particular to take care of children and the elderly at home.
These two conservative modernizers developed what Gosta Esping-Andersen classi-fies as “conservative-corporatist” social welfare regimes whose pension systems rely on contributions paid during a citizen s years of employment and are segmented along occupational lines.（28） The conservative-corporatist regime also relies on means-tested
assistance and family resources for the unemployed and elderly. Despite the breadth and depth of Occupation-led changes in both societies during the early post-war years, the pension systems would be allowed to retain the same basic structure and intention as they had before the war and as far back as the 19th Century. The rise of a new middle
class in both nations might have weakened state provisions for pensions, since this group historically has been the most capable of relying on private means. However, given that industrial development had been led by legislators, bureaucrats and industri-alists uninterested in promoting free markets, Germany and Japan “institutionalized a middle-class loyalty to the preservation of both occupationally segregated social-insur-ance programs, and ultimately to the political forces that brought them into being.”（29）
（27） Esping-Andersen, p.85 （28） Ibid.
The welfare regimes in Japan and Germany are examples of what Esping-Andersen labels “corporative state-dominated insurance systems,” which segregates pensions along occupational lines, while linking benefits to contributions. With the rise of the size of the white-collared workforce in all G- 5 nations following the war, this new middle class should have weakened state provisions for pensions since they historically have been the most capable of relying on private means. However, given that develop-ment was led in both states by legislators, bureaucrats and industrialists uninterested in promoting free markets, Germany and Japan “institutionalized a middle-class loyalty to the preservation of both occupationally segregated social-insurance programs, and ultimately to the political forces that brought them into being.”（30）
Germany and Japan share common state-building histories and have developed simi-lar pension regimes. They differ in that the former contains a strong left-wing party with a popular base of support in labor unions organized horizontally, across firms. In Japan the parties of the left lack such a base for two reasons. First, Japan s unions re-main “in-house;” they are organized within firms and their membership includes mem-bers of managements as well as the rank and file floor workers. Labor in Japan evolved along different lines than in Germany, because the Japanese state could more effective-ly repress labor than could the Prussian and German states. Marius Jansen notes that the development of labor fraternities in Japan differed from in the West because the fragmentation of the Tokugawa order persevered into the Meiji era. By the time of the advent of state-led industrialization, labor groups were confined to guilds, and enter-prises had an easy time containing these guilds. In response to growing disputes be-tween labor and management during the Sino-Japanese war, the state passed the Peace Police Law, Article 17, which made illegal inciting others to join unions, engaging in collective bargaining or striking.（31） Second, socialist and communist parties have been
（29） Ibid., p.85 （30） Ibid., pp.31 32
defined mostly by their intellectual programs and less by their political platforms. The JSP came about in this environment, not as a true workers party, but as an intellectuals party, which concentrated on the minutiae of Marxist ideology. In addition, the elector-al rules in place between 1955 and 1996 go a long way in explaining the perennielector-al fail-ure of the Japan Socialist Party （JSP） to capture more than a token share of the vote, as does their decades-long insistence on clinging to a rigidly Marxist ideology in the face of repeated rejection at the ballot box.（32）
Using Mill s indirect method of difference to arrange nominal variables related to the history of state- and welfare state development in Germany and Japan allows me to eliminate most of them as potential independent variables. Once I control for the fac-tors listed in A through F, I am left with a key difference on variables H and J and the processes that followed them, indicated as I and K, lead to the different outcomes that form the puzzle I am trying to solve. In the next section I present a model for explain-ing the process of recent pension reforms.
IV. The Power-Process Model
Robert J. Thomas s power-process model provides a useful tool for the task of under-standing the context in which the German and Japanese governments renegotiate pen-sion policies with particular groups in their societies. As shown in the introduction above, officials and politicians assume, and try to convince their electorates, that re-form is not only necessary in light of demographic trends, but that their rere-form packag-es were necpackag-essarily determined by a population-related “crisis,” such as the “aging-so-ciety “ and “low birth-rate problems” in Japan. Workers balk at the prospect of increased contributions and later retirements, and retirees probably will not vote for an-yone who threatens their promised benefit levels. However, neither the state s choice set for policy change nor the electorate s preferences are exogenous to the welfare
re-（32） Herbert Kitschelt,: The Transformation of European Social Democracy （Cambridge UK: Cambridge UP, 1996 （1994））, pp.294 295
gime status quo. The power-process model captures the path-dependent nature of pub-lic popub-licy reforms.
Thomas developed the power-process model as a response to contradictions in con-ventional explanations for the ways in which production processes are built around new technologies. Previous explanations had divided the independent variable between
ei-ther technological determinism or social choice. The former assumes that the structure
of the manufacturing firm is constructed to conform to existing technologies - a good may be produced only in a manner made self-evident from by the limits of manufactur-ing technology. Technological determinism subordinates the social system around pro-duction to the technical system.（33） The analog to political science perspectives includes
arguments that reduce all outcomes to structural arrangements, which “manufacture” policies. So, in this view, demographic changes demand the creation of new policies for pension systems and these policies will be produced within the framework of exist-ing policy guidelines.
Social choice models, on the other hand, insist that “rather than being objective, technology is infused with objectives” （italics in original）.（34） These models analyze
the manner in which interests become attached to, and expressed through physical processes. In short, this perspective subordinates the technical system to the social sys-tem around it. The analog in the social sciences is social constructivism and seen from its perspective, pension reform depends entirely upon the subjective meanings attached to existing policy programs by state and society.
The shortcomings of the technological determinist perspective begin with the fact that “new technologies do not fall from the sky. Inattention and ignorance may lead technology to be experienced as an independent force, but inattention and ignorance are themselves the outcomes of social processes.”（35） Thomas notes that many new
（33） Cf. Robert J. Thomas, What Machines Can’t Do: Politics and Technology in the Industrial Enterprise （Berkeley, CA: U of California Press, 1994）, Chs. 1 , 6
technologies are often created by one set of actors but used by another, so that the product appears as an established fact. Furthermore, the process by which a product is manufactured also appears as an unalterable fact to those who were not involved in the initial design of a manufacturing system. Reducing all outcomes to institutional matri-ces generally ignores the fact that these matrimatri-ces are of human design, and may be al-tered. Certainly the state and those who benefit most from status quo arrangements per-petuate this perspective. Institutional analyses of the welfare state, as noted above, do not allow much space in which actors may maneuver for change. Existing policies do, indeed, limit the number of reform options, particularly as these policies remain so deeply enmeshed in rights, as well as by existing “rules of the game.” However, this analysis cannot illuminate policy differences across similar pension regime types, for example, or explain why the Japanese state produced generous benefits in the absence of an encompassing institution such as labor. Finally, the use of welfare policies to fa-cilitate state-building by Bismarck, for example, while made explicit by him to the Re-ichstag （he insisted it was necessary to nip socialist revolution in the bud）, was not publicly known: pension policies provide certain benefits, but also come about in a manner useful to states.
Thomas s critique of social choice theory rings true for our discipline: social choice theorists disregard that “choices in structures may originate…at some distance re-moved in space and time） from” policy makers and electors in the present.（36） Thus,
they do not consider that framing choice-sets prior to a formal decision actually repre-sents an important element in the process of technological, or social, change. Values are built into the political system. Ideas-driven analyses, then, disregard the manner in which institutional design reflects particular interests and that over time this design perpetuates these interests.
Thomas presents his power-process perspective as an antidote to the shortcomings of
（35） Ibid., p. 8 （36） Ibid., pp.214 215
the two perspectives. Rather than disregarding either, however, this perspective “bridg-es determinism and choice. “ Thomas introduces his approach in the following passage, in which I have substituted the word “policy” for every instance of the words “technol-ogy” and “technological:”:
The power-process perspective is distinguished by its intense concern with [policy] change as a process…before [policy] change is possible, before any
impacts occur, many choices have to be made. These choices are not limited to
the features of a given [policy] or to the formal decision to adopt given [policy]. In fact, the choices of greatest interest are the choices about how choices will or
can be made; these are manifestations of history （prior choices）, that take the form of structure, a set of procedures, and a set of norms, precedents, and insti-tutionalized understandings about the range of possible [policy] the social sys-tems and organization is willing to consider, much less implement. The way an organization structures and conducts its search for alternatives offers valuable insights as to the effects of history not only on its current form and functioning but on its possible futures as well （italics in original）.（37）
The power-process model describes the normative, intentional foundations of any wel-fare regime. Rather than infer outcomes from structures or infer choices from out-comes, this perspective examines the policy-making process as it unfolds. However, this model does not treat current options and opinions around those options outside of the histories that produced them. The pension policy options advanced by the German and Japanese state will not likely break sharply from tradition. Other reforms in both nations are constrained by a combination of existing choice-sets, fears of losing at the ballot box and voter preferences, which seem to be endogenous to the welfare regime.
Inherent in the power-process model is a dualistic conception of organizational structure. On the one hand, structure is a constant in social action. As cited numerous times above, it reflects historically derived rules, but structure also reflects understand-ings and social patterns. So, on the other hand, the power-process model acknowledges that structure is also the product of social action; it is a function of the repetition of rules, and patterns for human interaction.
The power-process model roots its ontology firmly in Weberian notions of the rela-tions between structure and social action. First, Weber accepted the Kantian epistemol-ogy that an empirical discipline cannot logically define what “ought to be.” Value judgments cannot be validated, but they may be treated as empirically observable phe-nomena by a social scientist. Weber suggests that society contains irreducibly compet-ing ideas. In short, what people believe affects what they expect from a social policy, so preferences matter. However, Weber also observes that social science may compre-hend human action when people manipulate a given means toward an end. By defini-tion, people engage in political action in order to realize particular ends and Weber notes that when the motivation for action accords with established norms it makes sense to the observer. But, as Anthony Giddens explains, it has been the basic flaw of idealists “to identify subjective adequacy with causal adequacy. Similar actions…may be the result of a diversity of motives, and conversely, similar motives can be linked to different concrete forms of behaviour.”（38）
Second, although institutions are of human designs, they are no less structures, for their perseverance is predicated upon their legitimacy and the threat of penalties to those who would disobey them. Weber writes that, when “a civil servant appears in his office daily at a fixed time…it is not determined by custom or self-interest alone…As a rule such action is determined by his subjection to an order, the rules governing the de-partment which imposes obligations on him, which he is usually careful to fulfill, partly
（38） Anthony Giddens, Capitalism & Modern Social Theory: An Analysis of the Writings of Marx, Durkheim,
because disobedience would carry disadvantageous consequences but usually also in part because it would be abhorrent to the sense of duty, which…is an absolute value to him.”（39） Institutionalists infer courses of action based upon the fact that actors will
likely play by established rules, because, Weber notes, although “the orientation of an action to an order involves a wide variety of motives, “so long as the order “is also held by at least part of the actors to define a model or to be binding, naturally increases the probability that action will in fact conform to it.”（40）
Finally, Weber emphasizes the role of bureaucratic structures in enabling capitalism, stressing that the predictability that allows the latter to guarantee contracts and invest-ments is predicated upon the record-keeping skills of the former. As Theda Skocpol demonstrates in her response to Barringtion Moore, political outcomes play out in the arena of political bureaucratic structures and the importance of these structures to the power-process model is two-fold. First, Weber shows that technical knowledge, which is indispensable for the business of the state and of economic actors, is housed within bureaucratic structures. Second, Weber argues that “the question is always who con-trols the existing bureaucratic machinery,” and proposes that career bureaucrats will, in the long run, wrest control of agencies from elected officials who lack specialized training.（41）
Conventional approaches to explaining pension reform have in common what Her-man Schwartz casts as “a professional deformation of reality in which academics focus closely on the formal welfare state as an instrument of redistribution only towards workers.”（42） Schwartz reminds us that workers have not been the only actors to benefit
from redistribution and state-managed economic stability. Throughout the post-war pe-riod states facilitated trade protection, centralized bargaining, and market regulation as
（39） Max Weber, The Theory of Economic and Social Organization, （New York, NY:Free Press, 1997）, p.124 （40） Ibid.
（41） Ibid., 337 339 （42） Schwartz, p10
well as the welfare state itself and in so doing they decommodified not only workers by sheltering them from market forces, but also large amounts of capital. Furthermore, Schwartz notes that welfare rights generally have been construed to be social rights when, in fact, it they are more appropriately understood as property rights. Protection of capital from market forces generates income for producers just as protection of wag-es prwag-eservwag-es workers livelihoods.
Extricating pension policies from the list of social protection policies cited by Schwartz, illuminates the relevance of considering them as a property right, albeit one with an especially political basis. In the first place, “pensions constitute a central link between work and leisure, between earned income and redistribution, between individ-ualism and solidarity,…Pensions, therefore, help elucidate a set of perennially conflict-ual principles of capitalism.”（43） To consider pensions solely as a social right places
employers on one side of the aforementioned pairings and employees on the other. While transfers such as unemployment benefits or sick leave may affect employers ad-versely, all adults of working age contribute to pension funds and count on receiving benefits when they retire. Second, therefore, the conflict over restructuring pension re-form is not simply a conflict over transfers that benefit labor primarily: once working age is reached most every member in society has a stake in pension policy. That stake begins, arguably, before one enters the workforce for pension policies rely on future projections. The next generation of retirees depends upon the next generation of work-ers and will not likely settle for benefits less than what they provided their eldwork-ers.
―H. Steven Green・法学部専任講師―
