I

The logic of retrenchment

The next two chapters provide a framework-for the study of retrenchment, set-
ting the stage for a detailed examinationof the Reagan and Thatcher records.
To be persuasive, accounts of welfare state change must combine microscopic
and macroscopic analysis. They must consider both the goals and incentives of
the central political actors and how the institutional rules of the game and the
distribution of political resources structure their choices.
This chapter outlines the peculiar nature of retrenchment as a political project.
For politicians eager to win reelection, seeking cutbacks in social programs
raises considerable risks. Such cutbacks impose concentrated costs in return for
diffuse benefits, and there is substantial reason to believe that concentrated in-
terests possess marked advantages in political conflicts. To make matters even
more difficult, retrenchment advocates must contend with an imbalance in vot-
ers” reactions to losses and gains; transfers of resources tend to induce more
resentment from losers than gratitude from winners.
An understanding of retrenchment politics must start from an appreciation of
this distinctive political problem. This chapter will focus on the nature of this
distinctiveness and the strategic options available to retrenchment advocates that
may make their problems more tractable. The following chapter will consider
how the broader context — patterns of interest-group representation, institutional
structures, and preexisting policy designs — influences the prospects for imple-
menting these strategies.
As a preface, however, 1 discuss the concept of retrenchment itself. This is
crucial, because fuzzy conceptions of retrenchment have encouraged confusion
about exactly what has happened to these welfare states. Without understanding
what retrenchment is and without specifying how it can best be measured, one
cannot get a firm grip on what has happened, much less determine why.

CONCEPTUALIZING AND MEASURING

Retrenchment is one of those cases in which identifying what is to be explained

13
be measured. As a result, discussions of the recent experience of welfare states vary independently of total expenditure. Consequently, countries can maintain
have often floundered before even reaching the issue of causation. The few high spending levels without producing substantial redistribution or social sol-
available studies of welfare cutbacks have generally used social-expenditure idarity among different groups.* Richard Titmuss captured this situation in his
trends as a proxy for retrenchment success.' Although such trends are important, distinction between ““institutional”” and **residual”” welfare states. Institutional
they provide an inadequate measure. Such figures indicate only short-term welfare states rely mainly on comprehensive programs, utilize public provision
spending patterns, ignoring programmatic changes that produce long-term rather of major services, attempt to reduce distinctions between different classes or
than immediate cutbacks. Furthermore, they focus attention exclusively on status groups, and generally operate to partially restrict the market's impact on
changes in the size of welfare state programs while ignoring changes in program life chances. Residual welfare states are more reluctant to interfere with market
structure. Finally, social-expenditure patterns say little or nothing about broader mechanisms; they reject comprehensive services, and prefer state subsidization
policy changes that may have important consequences for welfare state devel- of private services to public provision. The goal is to provide a **safety net,”
opment. usually based on a means test, while reinforcing market-produced patterns of
These three points establish some basic ground rules for the study of re- social stratification.* a
trenchment: A study of welfare state change must focus on structure as well as size. To
discuss retrenchment rather than cuts is to analyze political conflicts over the
1. Examine long-term as well as short-term spending cuts. This ground rule is character of the welfare state. As Esping-Andersen has put it, **It is difficult to
the most straightforward. Governments interested in curtailing social programs imagine that anyone struggled for spending per se.””* Just as organized labor
may enact policies that cut spending immediately; they may also enact changes and left-of-center parties once pushed not just for higher spending but also for
to be phased in over time, the full effects of which may not be felt for many more extensive modifications of market outcomes, conservatives work to en-
years. For example, a change in indexation rules may only reduce expenditures courage market-oriented reforms as well as lower spending. Retrenchment
by 1 or 2 percent in the first year, but this *“decremental”” cutback will gradually should be seen as a process of shifting social provision in a more residualist
grow in scope. Benefit or eligibility restrictions may exempt current recipients, direction, not just as a matter of budget cuts.
affecting only new beneficiaries. Cutbacks in housing-construction programs
may have little impact on the availability of low-income housing for a decade. 3. Study systemic retrenchment as well as programmatic retrenchment. Pro-
An analysis of current spending levels will completely miss some of the most grammatic retrenchment results from spending cuts or a reshaping of welfare
important cuts in social programs. This is a particularly important point in light state programs. However, policy changes that alter the broader political economy
of the propensity, discussed later in this section, of retrenchment advocates to and consequently alter welfare state politics may also promote retrenchment.
pursue strategies that hide the magnitude of cuts by minimizing short-term neg- Welfare state programs do not exist in a vacuum. Their shape is determined by
ative consequences. the complex interplay of such factors as budgetary pressures, the structure of
political institutions, and the strength and priorities of interest groups. Policy
2. Examine program structure as well as program spending. Long-term expen- changes that alter the context for future spending decisions — what can be termed
diture trends still provide an insufficient basis for evaluating welfare state systemic retrenchment — may be as important for the welfare state as changes
change. Gpsta Esping-Andersen has offered a telling critique of the preoccu- in spending or program structure **within”” the welfare state itself.
pation of comparative welfare state research with expenditure levels.? Spending Systemic retrenchment can take four forms. First, a government can attempt
levels are simple, quantifiable indicators that are ideal for cross-national statis- to defund the welfare state by constraining the flow of revenues to future ad-
tical analyses. However, because expenditures reveal only size and not content, ministrations. Revenues provide the underpinning for a vast range of government
they provide an incomplete description of a country's welfare state. Welfare activity. Welfare states require revenues: Where there is no money there can be
states intervene in markets to different degrees and in different ways. They also no programs. Ronald Reagan recognized this when he compared big government
diverge on a variety of important dimensions, including their use of state- to an unruly child, arguing that the way to discipline children's *“extravagance””
provided services, reliance on means-tested benefits, tax progressivity, commit- was **by simply reducing their allowances.”**
ment to full employment, promotion of private alternatives to state benefits, A number of mechanisms can produce defunding. The simplest is to cut taxes
segmentation of recipients by status group, and willingness to loosen workers” or reduce the ease with which governments can raise taxes in the future. Just
dependence on wages. as 1 will argue that a government pursuing retrenchment will seek to diminish
The social and political role of the welfare state depends as much on these the visibility of unpopular program cuts, it may try to increase the visibility of
various dimensions of policy choice as it does on spending levels. Each can taxation as a way to restrict the resource base for welfare state initiatives. For

1205992
example, inflation-induced ““bracket creep,'” which moves people into higher political arenas. Finally, decentralizing efforts may allow national advocates of
tax brackets, was traditionally a quiet way to increase revenues in the United retrenchment to shift the blame for cutbacks to local officials.
States. Eliminating bracket creep would thus remove one of the most effective The fourth type of systemic retrenchment would be a weakening of pro-
techniques for generating government revenues without creating a political out- welfare state interest groups. The impact of an administration dedicated to re-
cry. Alternatively, a government could diminish future financial capacities by trenchment is likely to depend in part on the political strength of welfare state
shifting to unsustainable sources of finance such as asset sales. Heavy deficit supporters. Groups offering such support include organizations of beneficiaries,
financing is another option. Although maintainable for limited periods of time, producer interests with a stake in either the provision of specific services (e.g.,
such financing would eventually leave governments with the unpalatable choice housing or education) or in a pattern of public intervention in the marketplace
of raising taxes or further squeezing public expenditure. (such as labor unions), and advocacy organizations dedicated to advancing pro-
Finally, a government might seek to increase non-welfare state expenditures grams for the underprivileged and underrepresented.
(e.g., those for defense). Like the other approaches, this tactic is based on the Governments seeking cutbacks will try to weaken these opponents. Whether
reasonable expectation that any restriction of resources available for social pro- governments have the capacity to achieve this goal is likely to depend signifi-
grams will eventually result in increased pressures to check expenditures. A cantly on the characteristics of particular groups. Those dependent on gov-
government that succeeded in these defunding efforts would have created a ermment financing are likely to be highly vulnerable. In some instances,
significant level of systemic retrenchment, even if the structure of social pro- governments may be able to bring about reforms that make it more difficult for
grams remained relatively constant in the short run. groups to organize or pursue political action. For example, rules covering col-
A second type of systemic retrenchment would be a policy-induced change lective bargaining have a significant impact on labor-union strength. Tax
in public opinion, weakening popular attachments to public social provision. changes, for that matter, may impede a group's ability to raise money.
Public attitudes may affect the long-term position of the welfare state. If gov- Governments may also pursue more indirect strategies. Restrictive economic
ernment policies increase public preferences for private provision, growing hos- policies that raise unemployment rates, for instance, tend to undermine organized
tility toward public social programs could be expected to facilitate retrenchment. labor's political position. Thus, governments have a range of possible tools for
Thatcher's efforts to foster an **enterprise culture'” by encouraging private social attacking the organized interests supporting social programs.
provision, and through other privatization initiatives (substitution of private for The failure to develop a convincing conceptualization of retrenchment has
public employment in industry or increased share ownership, for example), pro- greatly impeded research on the recent history of the welfare state. Far from
vide a good example of this kind of strategy.? Whether such strategies actually being simply a matter of immediate cuts in public spending, retrenchment is a
have a substantial impact on popular attitudes toward public programs will be complex, multifaceted phenomenon. This discussion provides a basis for iden-
considered in Chapter 6. tifying the relevant dimensions of policy change.
Systemic retrenchment might also take the form of modifications in political Retrenchment can be defined to include policy changes that either cut social
institutions, changing the way decision making about the welfare state is carried expenditure, restructure welfare state programs to conform more closely to the
out, and thus potentially changing policy outcomes. The **rules of the game”” residual welfare state model, or alter the political environment in ways that
can have a tremendous impact on the construction of political interests and on enhance the probability of such outcomes in the future. To determine the success
the capacities of competing interests to exert influence within the political proc- or failure of Reagan and Thatcher's efforts all three of these dimensions must
ess. Although political scientists have recently shown renewed interest in the be considered.
impact of political institutions, the consequences of institutional design for so-
cial-policy development remain a subject of controversy.* This issue is taken up
in detail in Chapter 2. For now, it is necessary to note only that retrenchment THE HAZARDS OF IMPOSING LOSSES

advocates possess two strategic options that might make institutional conditions Government leaders want to advance their policy agendas, and they want to be
more conducive to the pursuit of cutbacks: They may try to centralize political reelected. If at all possible, they will seek to make these two goals mutually
authority, hoping to increase their capacity to implement their own policy pref- reinforcing. There is no need to assume that governments consider only electoral
erences; altematively, they might pursue a decentralizing strategy, transferring implications in formulating policies; it is enough for this analysis that such
authority over social policy to local governments. Decentralization could facil- concerns are a central consideration, if only because failure to consider electoral
itate retrenchment, primarily because economic competition among local juris- consequences can jeopardize policymakers” long-term prospects for implement-
dictions often makes it difficult for them to pursue redistributive policies. Interest ing their preferred policies.
groups supporting the welfare state may also be weaker in more decentralized Given the desire to reconcile policy initiatives with electoral considerations,
governments favoring retrenchment are likely to face a dilemma. Cutbacks in acter 1s critical to an analysis of retrenchment politics. he struggles over social
social programs usually raise the risk of electoral retribution. There are two policy discussed in the following chapters all reveal the Reagan and Thatcher
distinct reasons that retrenchment is generally an exercise in blame avoidance governments” sensitivity to the perilous nature of their reform agendas. Marked
rather than ““credit claiming.””? First, the costs of retrenchment are concentrated, by this consciousness of potential political hazard, retrenchment politics has a
whereas the benefits are not. Second, there is considerable evidence that voters quality quite different from that of the period of welfare state expansion, when
exhibit a ““negativity bias,”” remembering losses more than gains. As a result, political rivals often eagerly outbid each other in the quest to receive credit for
retrenchment initiatives are extremely treacherous. The unpopularity of almost government action.'*
all efforts to curtail public social provision creates a sizable danger that policy
goals and electoral ambitions will conflict.
STRATEGIES FOR MINIMIZING COSTS
Cutbacks generally impose immediate pain on specific groups, usually in
return for diffuse, long-term, and uncertain benefits.'” That concentrated interests Confronting these harsh political realities, conservative governments might be
will be in a stronger political position than diffuse ones is a standard proposition expected to back away from attacks on social programs. Instead of throwing up
in political science.'' As interests become more concentrated, the prospect that their hands, however, retrenchment advocates may try to find ways to overcome
individuals will find it worth their while to engage in collective action improves. the obstacles they face. A government determined to pursue a retrenchment
Dairy farmers whose livelihood depends on government subsidies have far more agenda could try to ease the resulting dilemma by maximizing its electoral mar-
reason to organize than the consumers who may pay a few cents more at the gin of safety. The less danger there is of losing the next election, the more
store. Furthermore, concentrated interests are more likely to be linked to organ- political capital governments can afford to spend in pursuit of desired but elec-
izational networks that keep them well informed of what their interests are, and torally costly policies.'*
how policymakers may affect them. These informational networks also facilitate There are, however, a number of problems with this approach. Many of the
political action. 1 will have more to say about the role of information in these factors influencing a government's electoral position are largely beyond its con-
matters shortly. trol. Furthermore, government popularity often fluctuates unpredictably, making
The well-documented imbalance between the way that voters react to losses calculations of safety margins a hazardous undertaking at best.'”
and gains further enhances the political position of retrenchment opponents. The second option is to try to reduce the political repercussions of retrench-
Extensive experiments in social psychology have demonstrated that individuals ment actions by minimizing the mobilization of the opposition. Strong opposi-
respond differently to positive and negative risks; individuals tend to be risk- tion to particular political reformers is not a given; on the contrary, opponents
averse with respect to gains, but they are risk-seeking with respect to losses.'” must be successfully mobilized to pose a real threat to politicians.
That is to say, individuals will take more chances — seeking conflict and ac- There are three broad strategies that retrenchment advocates can use to min-
cepting the possibility of even greater losses — to prevent any worsening of their imize political resistance: obfuscation, division, and compensation. They may
current position. Studies of electoral behavior confirm the findings of psychol- seek to manipulate information flows to decrease public awareness of their ac-
ogists. Negative attitudes toward candidates are more strongly linked with a tions or of the negative consequences of them. Alternatively, they may endeavor
range of behaviors (e.g., turnout or a desertion from the voter's usual party to divide their potential opponents. Finally, retrenchment advocates may offer
choice) than positive attitudes are.'* ““side payments”” to compensate some of those adversely affected by proposed
Why this ““negativity bias”” exists is unclear.!* For current purposes what is changes. This represents a limited but potentially powerful repertoire. The Rea-
important is the constraint that the asymmetry in reaction to losses and gains gan and Thatcher records reveal instances in which each strategy was used to
creates for policymakers. When added to the general imbalance between con- considerable effect, allowing substantial retrenchment to occur with surprisingly
centrated and diffuse interests, the message for retrenchment advocates is clear. modest political costs.
A simple “'redistributive”” transfer of resources from program beneficiaries to
taxpayers, engineered through cuts in social programs, is likely to be a losing
proposition. The concentrated beneficiary groups are more likely to be cognizant
Strategies of obfuscation
of the change, are easier to mobilize, and because they are experiencing losses Of these three strategies, obfuscation, which involves efforts to manipulate in-
rather than gains will be more likely to consider the change in their voting formation concerning policy changes, is the most important. As James Kuklinski
calculations. In short, retrenchment advocates face a difficult clash between their has recently written, *“The idea of information has overtaken political scien-
policy preferences and their electoral ambitions. tists.*!* In research on welfare state development, for example, this theme has
A recognition that the political goal of loss imposition has a distinctive char- been prominent in Hugh Heclo's analysis of how “*political learning” occurs
as policymakers adapt their understandings and design new initiatives based on benefits frozen, social provision is shifted increasingly toward the private sector,
the knowledge gained from past policy initiatives.'? Heclo argues that policy and because change thus occurs very slowly it is less likely to attract attention.
development is as much a matter of puzzling over what to do in a complex Decrementalism also hinders the development of opposition because retrench-
social environment as it is a matter of power struggles among competing inter- ment advocates need only achieve a single policy shift (e.g., a changed index-
ests. ation method) to produce a flow of annual cutbacks.
In contrast to Heclo”s stress on the uncertainties facing government decision The second link in Arnold's causal chain is between negative events and
makers, 1 wish to emphasize the ways in which knowledge contributes to the public policies. Even when voters are aware that their circumstances have wors-
exercise of power. Information is a scarce and valuable political resource. Po- ened, they may have trouble connecting that to shifts in government policy.
litical actions must be based on understanding, but our understandings are nec- Sometimes, of course, the link is obvious. Someone who learns that his or her
essarily constrained by the sheer complexity of the social world and our own job has been eliminated, or that this month's pension check has been slashed,
limitations of time and cognitive capacity. That policymakers can make it more is unlikely to be puzzled for long. Nevertheless, the connections between policy
difficult for the electorate to get information may make retrenchment initiatives and adverse events can be obscure. Politicians can reduce the visibility of their
less politically hazardous. All social actors possess imperfect information about actions by making the effects of reforms indirect. As economists have long
issues relevant to their interests. Furthermore, the distribution of information is observed, there is often a difference between the de jure and de facto incidence
usually highly unequal.?” In this context, it may be possible for policymakers of taxes. Political scientists have observed that because the use of indirect taxes
to lower the political costs of retrenchment actions by making it more difficult obscures the links between policy and negative events, it leads to greater public
for possible opponents to obtain relevant information about policy reforms. tolerance for high taxes.?** To take an example from the politics of retrenchment,
Douglas Arnold has argued persuasively that voters endeavoring to reward cutbacks in health care might be imposed on hospitals or doctors, who then pass
or punish politicians are engaged in an effort to reconstruct “causal chains”” the cost on to consumers through higher prices. The visibility of effects can also
linking negative or positive events to particular policy choices, and those choices be diminished by increasing the complexity of reforms. The consequences of
in turn to the actions of specific politicians.? How voters actually construct simple cuts are easy to identify, but even though elaborate rule changes may
these causal chains remains somewhat unclear. Arnold acknowledges that it is ultimately have the same or even greater effects, that impact is often harder to
a subjective process, the details of which are likely to be highly complex and detect. Washington lobbyists term this the **Dan Rather test””: Reforms are less
culturally contingent.”? Despite this complexity, the implications for policymak- likely to generate a popular outcry if television reporters cannot explain the
ers are relatively straightforward. Those engaged in efforts to initiate unpopular implications of the new policies in fifteen seconds or less.
policies will try to lower the visibility of their reforms by complicating the Finally, as an alternative or supplement to efforts to lower the salience of
reconstruction of causal chains that would allow voters to exact retribution. negative consequences and their connection to policy change, policymakers can
Three possible sites of obfuscation, related to different parts of these causal try to diminish public awareness of their own responsibility for those effects.
chains, should be distinguished. First, retrenchment advocates can try to lower Arnold calls this part of the chain the **traceability** of policy change. Since it
the salience of negative consequences. Not all negative events are equally ap- is the fear of being held accountable for unpopular actions that constrains pol-
parent. Anything that makes it harder to detect the negative consequences of icymakers, they are likely to seek means of covering their tracks. R. Kent Wea-
policy reforms will decrease political mobilization against them. Negative con- ver has discussed in detail some possible strategies of blame avoidance.”* One
sequences are less likely to be observed if they are spread widely rather than option is burden shifting, that is, passing responsibility for imposing cutbacks
concentrated, and if they are diffused over time rather than delivered in a single to local officials, who may then attract some of the blame. This tactic obviously
shot. requires a program structure that allows central government to separate itself
Thus, a frequent tactic used to lower the visibility of negative effects is from the actual imposition of cutbacks. An alternative is to make cutbacks au-
decrementalism. Social programs operate within a changing economic context tomatic. Changes in indexation rules, for example, lead to annual reductions
marked by rising prices and (generally) rising incomes. Retrenchment advocates without requiring repeated, visible actions on the part of policymakers. Deficit-
may be able to achieve their goals with limited political exposure, then, by reduction mechanisms, like the Gramm-Rudmann budget reforms in the United
freezing a program within a growing economy. States, may be designed to operate the same way.”
Decrementalism can take two forms. Most obviously, a failure to adjust for A powerful tactic to diminish traceability is to delay the implementation of
higher prices lowers the real value of benefits. Less widely noted but more often cutbacks. Opposition is easier to mobilize when retrenchment policies impose
pursued is the tactic of **implicit privatization,”” in which benefits retain their immediate losses on program recipients. Postponing such effects decreases the
real value but play a diminishing role in an expanding economy. With public prominence of the cutbacks. This tactic has proven particularly effective in hous-
 sistance in both countries, and both Aid for Families with Dependent Children
Table 1.1. Obfuscation strategies and the reconstruction of causal chains
(AFDC) and food stamps in the United States, New eligibility rules may also
exclude specific categories of recipients from benefits (e.g., excluding strikers
Stage of causal chain Obfuscation technique
from unemployment benefits in Britain and from food stamps in the United
Identification of negative effects Decrementalism States, and barring British students from income support during school vaca-
Linking of negative effects to policies Indirect incidence tions).
Linking of policies to decision makers Burden shifting, automaticity, lagged cutbacks Programs that provide services create an additional potential cleavage be-
tween service consumers and service producers. Retrenchment advocates may
design reforms in such a way that they divide consumers of a particular service
ing policy in both the United States and Britain, where a cut in public-housing from producers. Again, housing provides a prominent example, where construc-
construction has no impact on current recipients. The effects are concentrated tion cutbacks had immediate repercussions for producer groups but not for con-
on potential future recipients, who are unlikely to see their current interests as sumers. Similarly, divisions between health-care producers and consumers in the
closely linked to the fate of public housing and are by their very nature uni- United States have facilitated retrenchment in Medicare.
dentifiable and therefore impossible to organize. In any event, by the time the Where retrenchment is widely anticipated, targeting cutbacks on particular
negative repercussions of policy changes begin to be felt, the political decisions subgroups within a beneficiary population will minimize the size of the potential
responsible will be buried in the distant past. opposition to proposed reforms.”” Benefit recipients who are spared the axe may
“Never be seen to do harm.'” Often cited as the first maxim of electoral feel quietly grateful that the burden has fallen on someone else. In any event,
politics, this folk wisdom has not been well integrated into theories of policy they are unlikely to mobilize effectively against such retrenchment efforts. Of
development.?* By obscuring negative outcomes, the connection between them course, next time it may be their group that finds itself singled out for attention.
and public policies, or one's responsibility for those policies, politicians can As a result, a common dynamic of retrenchment struggles involves competing
dramatically lower the political price they pay for pursuing a retrenchment efforts of governments to play one group off against another while program
agenda. Table 1.1 offers a summary of some techniques that retrenchment ad- supporters attempt to **circle the wagons.””
vocates can use to hide potentially unpopular activities. The chapters to follow
are full of instances in which retrenchment advocates used obfuscation to
achieve their goals. Information is a crucial resource, distributed in a highly Strategies of compensation
unequal way. Its importance in political struggles has been vastly underrated. Offering something positive to the victims of retrenchment policies will diminish
prospects for heated opposition. Given the existence of *'negativity biases”” it
is often very helpful to have some plausible way of making the case that ben-
Strategies of division
eficiaries will not be hurt. Compensation may be offered to groups most likely
Although the potential for organized opposition in a particular policy area may to mobilize against retrenchment, or to those most likely to garner public sym-
be large, it is sometimes possible to isolate subgroups within that opposition. pathy. 'Grandfather clauses'? may be introduced so that current recipients are
“*Divide and conquer”” is an obvious political ploy, but that need not render it excluded from the impact of policy changes. Losses are then restricted to an
any less effective. The constituencies of all public programs are to some extent unspecified, and probably unorganizable, group of future recipients.
heterogeneous. In the case of income-transfer programs, a wide range of dis- Alternatively, a government may attempt to offer compensation for public-
tinctions can be exploited, including differences in household composition, in- sector retrenchment by expanding private benefits. Attractive private-sector Op-
come level, age, geographic location and gender, to name but a few. Although tions may mute opposition to the curtailment of public provision. Given the
race rarely constitutes a legally permissible basis for distinguishing among re- preference of retrenchment advocates for market solutions, privatization options
cipients, there may be ways of activating this division as well. The consumers —- even when they require some kind of government subsidy — are likely to
of publicly provided services (e.g., housing and health care) will have the same receive serious consideration wherever they appear to be viable.
internal divisions. Successfully mobilized, program supporters represent a threat that politicians
Cutbacks may be designed so that they affect some benefit recipients but not seeking reelection are unlikely to ignore; the gains from retrenchment then ap-
others. The easiest way to do this is by pursuing retrenchment through tightened pear too limited and uncertain to justify the political risk. Nevertheless, suc-
eligibility rules. Both the Reagan and the Thatcher administrations lowered in- cessful mobilization is the key, and there may be ways to prevent it. Obfuscation,
come ceilings for some means-tested programs. Examples include housing as- division, and compensation: These three techniques have given conservative
 preceding policymakers **feed back”” into contemporary politics, constraming
Table 1.2. Limitations of retrenchment strategies
the options of retrenchment advocates.
Even where strategies can be applied, they may have negative consequences
Strategy Range of application Drawbacks
that offset their advantages for retrenchment advocates. Obfuscation strategies,
Obfuscation for example, often have considerable drawbacks. The techniques that allow gov-
Decrementalism non- or poorly indexed Reversible emments to limit their political exposure may also weaken their control over
transfer programs policy. Decremental approaches are extremely slow. Although this lowers visi-
Division bility, it also lowers immediate payoffs (which may be of greatest interest to
Tightened eligibility Service and transfer programs Policy irrationalities the ruling government) and increases the opportunities for future administrations
Split consumer- Service programs Alienation of supporters to reverse the direction of policy. Burden shifting lowers the central govern-
producer coalitions ment's political exposure, but it also reduces its control over policy. Thus, the
Compensation Thatcher government generally rejected opportunities to shift responsibilities to
Provision of transitional Service and transfer programs Cost, policy irrationalities local governments because of a desire to enhance its own power.
benefits Efforts to divide potential opponents also have drawbacks. If an administra-
Expansion of private Most universal programs Cost, policy irrationalities tion succeeds in splitting consumers and producers, it is also likely to: anger
benefits producer groups (e.g., health-care providers, or construction and development
interests) that are often supportive of conservative politicians. In some cases,
strategies of division can generate significant program irrationalities. Dividing
critics reason to think that the welfare state”s strong political position might not opposition by tightening eligibility rules, for example, can worsen the work-
be unchallengeable. incentive problems prevalent in means-tested programs and increase the com-
plexity of bureaucracy. Even a government unconcernmed about producing
effective policies must pay attention to such implications, because they can
Limitations of retrenchment strategies
provide powerful ammunition for critics. As a result, this approach tends to
Although a range of opportunities exist to hinder the development of a strong yield quickly diminishing returns.
opposition, applying such strategies is no simple solution to the precarious en- The final strategy, compensating losers, offers the greatest political protection
terprise of retrenchment. Each approach also possesses important drawbacks. to retrenchment advocates. However, it too entails substantial costs. The main
Their range of application may be limited, and they produce political costs as limitation of compensation strategies is that they tend to cannibalize the policy
well as benefits. To understand the role of retrenchment strategies in welfare results that retrenchment advocates seek. If substantial resources are poured into
state politics these limitations need to be considered. efforts to buy off opponents, potential budgetary savings dry up. Offering tran-
The most obvious limitation is that even if these tactics can be used, they sitional benefits and expanding private alternatives is generally expensive. For
provide no guarantee of success. The reduction of political opposition must be administrations committed to short-term cost cutting as well as long-term re-
sufficient in relation to the potential returns to the government to justify pro- trenchment, this has generally proven to be a fundamental impediment. Cost
ceeding. Thus, retrenchment is most likely to occur when a variety of these constraints helped derail the **New Federalism”” initiative in the United States
approaches can be combined. (see Chapter 5) and proposals for radical pension reform and the expansion of
In addition, each tactic poses distinct problems. Table 1.2 provides an over- private health care in Britain (Chapters 3 and 6). As a result, compensatory
view of the limitations of the various retrenchment strategies just outlined. Al- initiatives work best as a supplement to other strategies. Skillfully applied, they
most all of the possible approaches have restricted ranges of application. may effectively blunt the opposition while leaving the bulk of a retrenchment
Programs have to possess particular characteristics to make each tactic appli- initiative in place. Compensation can be used alongside a strategy of division,
cable. For example, delaying cutbacks is possible where future expenditures are with losses being offset for only some of those affected. lt can also be part of
linked to current capital outlays (housing) or entitlements are based on long an obfuscation strategy, with protection offered to those most immediately dam-
periods of contributions (pensions). Such strategies are inapplicable where there aged, but not to those for whom costs are more distant.
are no long lags between current policies and the actual distribution of benefits. Furthermore, compensation strategies also can precipitate program irration-
The role of program characteristics raises a theme that will be pursued at length alities. Offering transitional benefits softens the blow to current recipients, but
in Chapter 2, namely the ways in which the specific design choices made by it also creates bureaucratic complications as parallel systems have to be devel-
oped for old and new beneficiaries.* If the divergence between the two groups 2
is clearly visible, the result can be an outcry against unequal treatment that
heightens rather than lowers the level of political opposition. In the United Interests, institutions, and policy feedback
States, transitional benefits were offered to pensioners to soften the impact when
a flawed indexation system was revised in 1977. Those just below the cutoff
for the transitional benefits (dubbed the ““notch babies””) have become an in-
creasingly militant group favoring an extension of these transitional benefits.??
Similarly, expanding private alternatives can produce policy irrationalities.
Because regressive tax expenditures usually subsidize privatization initiatives,
the distributional consequences are difficult to justify. Tax breaks also produce
economic distortions, artificially stimulating investment in subsidized activities.
In some cases, private alternatives appear to be clearly inferior. For example, Politicians pursuing retrenchment face a difficult challenge; just how difficult
the Thatcher government's retreat from private health care seems to have partly will depend on the political context in which they operate. These actors seek
reflected a recognition that the private sector was no match for the National their goals within specific environments. The distribution of political resources
Health Service in providing '*value for money.”” and the institutional rules of the game help to determine the prospects for suc-
In short, identifying retrenchment strategies offers no panacea for conserva- cessfully pursuing various strategies, as well as the benefits and costs associated
tive governmments. All strategies to diminish the political vulnerability of re- with particular outcomes. This chapter considers the aspects of this broader
trenchment advocates have significant limitations. Governments must carefully context that are most relevant for retrenchment politics.
weigh the advantages of these techniques for limiting the mobilization of op- Existing scholarship offers only limited help in this effort. The recent shift
in the welfare state's fortunes has been widely noted, but so far there have been
ponents against their considerable weaknesses.
This chapter has outlined the microfoundations of retrenchment politics. Ad- few attempts to explain retrenchment outcomes. This absence stands in stark
vocates of retrenchment confront a difficult task: The goals they pursue are likely contrast to the flourishing literature on thé origins and development of welfare
to be unpopular, yet at some point they must subject their records to public states. A wealth of detailed research has generated clear analytical perspectives
scrutiny. They also possess some tactical options that may make this dilemma on the main factors contributing to or retarding welfare state expansion.' Two
more tractable. Nevertheless, although it is useful to stress that retrenchment such theories have been particularly influential. One emphasizes the power re-
advocates face certain challenges and possess definite resources, this tells us sources of labor movements, whereas the other focuses on the role of institu-
little about what these policymakers are likely to do in specific situations, or tions.
how successful their efforts may be. The preferences and strategic repertoires There is a natural inclination to turn directly to those theories to explain
of these actors must be linked to the constraints within which they operate. contemporary welfare state politics. In a study of retrenchment, however,
What determines whether those seeking retrenchment can successfully apply these arguments need to be carefully reappraised. As 1 have already noted,
these strategies? Chapter 2 takes up this question. there is little reason to assume that theories designed to explain outcomes in
a particular context and involving the pursuit of particular goals will still ap-
ply once the environment and the goals of key actors have changed. Because
both the context and goals associated with retrenchment are distinctive,
whether theories of welfare state expansion offer insights into the retrench-
ment process must be considered an open question. 1 will argue that some
factors heretofore critical to processes of program enactment and expansion,
such as the role of organized labor, are of declining importance. Others,
such as the design of political institutions, are of continuing significance but
have different consequences because of the distinctive character of retrench-
ment politics. Finally, some previously peripheral factors move to center
stage. Among these, the policy feedback from previous political choices are
most important.
 me vuineraduny ol m0dalvidual prograras. 11e resmpicrico 01 Many CICUMICHLS 01 UE
ORGANIZED LABOR, PROGRAM CONSTITUENCIES,
welfare state suggests that although unions and left-of-center parties may play
AND THE WELFARE STATE
a key role in welfare state development, programs may be sustainable even
Because retrenchment politics generally involves efforts to prevent the formation where that support weakens.* In short, the Thatcher and Reagan records of mixed
of active opposition, the organizational capacities of welfare state supporters and limited success raise serious questions about the applicability of the domi-
matter. In this respect, retrenchment resembles welfare state expansion, with the nant paradigm for studying welfare state development to the study of retrench-
interests of important societal actors playing a part in determining political out- ment.
comes. In critical ways, however, the role of such actors in retrenchment rep- The challenge to the power-resources approach advanced here needs to be
resents a substantial change from the welfare state politics of the past. carefully qualified. This analysis compares two unabashedly conservative ad-
In particular, the impact of organized labor has decreased considerably. This ministrations; both inherited what Esping-Andersen would call liberal welfare
is a dramatic occurrence, for labor movements have long been seen as central state regimes. In part, this selection of cases bypasses key issues raised by
to welfare state development. During the past decade the power-resources per- power-resources analysts. In a comparison of, say, the United States and Swe-
spective, which attributes cross-national variations in social provision largely to den, the strength of left parties and labor movements might well have emerged
differences in the distribution of political resources among classes, has been the as more important.* In addition, just as the application of arguments derived
leading approach in comparative politics to explaining patterns of welfare state from studies of expansion to problems of retrenchment is a questionable practice,
expansion.” Class-based struggle over social provision may be intense, it is ar- the reverse warning should also apply: The current analysis may shed little light
gued, because social programs have a significant impact on the bargaining po- on earlier chapters in the history of the welfare state.
sition of workers and employers in the marketplace. Many social programs limit Caveats aside, this analysis suggests limitations in the power-resources ap-
the economic vulnerability of wage earners and increase worker solidarity. Ac- proach. Even where power resources are held relatively constant, important dif-
cording to power-resources theorists, encompassing, centralized unions, strong ferences in outcomes remain to be explained, especially when one examines
parties of the left, and weak or fragmented conservative parties all contribute to individual programs. The power-resources argument, which focuses on system-
the expansion of these programs. level variables, is unlikely to account for such variations. And given that
Given the consequences of many social-policy initiatives, the argument that Thatcher and Reagan each faced severely weakened unions and left-of-center
their development turns on the balance of organizational power between rep- parties, a power-resources perspective would imply much more dramatic
resentatives of labor and capital has great plausibility. Indeed, the power- changes than have actually occurred.
resources approach has had considerable success in accounting for cross-national Esping-Andersen has argued that ''a theory that seeks to explain welfare-
variations in social provision during the three decades following World War Il. state growth should also be able to understand its retrenchment or decline.””*
Furthermore, the replacement of crude social expenditure data with more fine- On the contrary, power-resources arguments have limited relevance because wel-
grained distinctions among patterns of social provision has greatly improved the fare states are now mature and retrenchment is not simply the mirror image of
explanatory power of the model.* welfare state expansion. In a context where public social provision is just emerg-
Nevertheless, a power-resources perspective cannot explain patterns of re- ing, the existence of broad organizations pushing a social-policy agenda is likely
trenchment in the United States and Great Britain. Reagan and Thatcher's re- to be crucial. However, the unpopularity of program cutbacks will give politi-
cords reveal two key features: Their overall impact has been modest and their cians pause even where unions and left-of-center parties are weak. Equally im-
success has varied widely across programs. Á power-resources perspective portant, maturing social programs develop new bases of organized support that
would predict neither of these outcomes. In both countries, the political and have substantial autonomy from the labor movement. This shifting base of sup-
economic resources of the left have diminished considerably. Rates of unioni- port may have consequences for the dynamics of policy development, but it is
zation have plummeted; left-of-center parties have been weakened. Although clear that the weakening of the labor movement does not translate automatically
power-resource arguments suggest that this shift should have sharply altered the into a commensurate weakening of the welfare state.
character of the welfare state, this has not been the case. There have been Analysis of the contemporary welfare state's supporters must shift from or-
changes, but most programs remain largely intact. ganized labor to the more varied constituencies of individual programs. Interest
Furthermore, the variation among program outcomes in each country has groups linked to particular social policies are now prominent political actors.
been extensive. This intrasystem variation suggests important limits to the ex- As 1 will argue in detail later in this section, the rise of interest groups is one
planatory power of systemic variables, including power resources. Án across- of the clearest examples of how policy feedback from previous political choices
the-board diminution in labor strength cannot account for major differences in can influence contemporary political struggles. Interest groups did not build the
 WEHdIC state, OU Ud WEHNdIE Side COMUIDUICO MUSHtly to tE GCVCIODMMENE Ol
“NEW INSTITUTIONALISM' AND 1HE FUL11 IS
an “'interest-group society.”” As Jack Walker noted in his detailed investigation 7, Lo,
CON
SOCIAL POLICY 4
of interest groups in the United States, '“the steady expansion of the federal
government figures as one of the major causes of the recent growth of new Patterns of governance matter. The **New Institutionalist”* resurgence in polit-
organizational devices for linking citizens with their government.*”” ical science reflects a renewed appreciation of how relatively stable, routinized
By the time a politics of austerity began to emerge in the mid-1970s, most arrangements structure political behavior.” The political institutions of different
social programs in both Britain and the United States were connected to exten- countries vary along crucial dimensions, such as the rules of electoral compe-
sive networks of organized social support. The recipients of various benefits — tition, the relationship between the legislative and executive branches, the role
pensioners, the disabled, and health-care consumers — were the most prominent of the courts, and the place of subnational governments in politics. These insti-
of these. There were also, however, a range of public-interest organizations tutions establish the rules of the game for political struggles, shaping group
seeking to protect the position of the unorganized. Finally, the providers of identities and their coalitional choices, enhancing the bargaining power of some
public services had a large stake in sustaining levels of social expenditure. In groups while devaluing that of others. Political institutions also affect the ad-
this respect, organized labor (public employee unions) continued to be of sig- ministrative and financial capacities of states. Furthermore, institutions influence
nificance. Their interests, however, were now linked primarily to the employ- the ability of policymakers to achieve the degree of insulation from social pres-
ment-generating effects of specific public programs rather than to the broad sures that may allow relatively autonomous initiatives, building on (or reacting
consequences of generous public provision for the bargaining position of work- against) actions of their predecessors.
ers, and their power was exerted more through individual unions than through New Institutionalist research has raised significant questions about the dy-
broad union confederations. namics of welfare state development. New Institutionalists have noted that the
The analysis offered in Chapter | suggests that these groups will continue to power-resources approach has little success in accounting for pre-World War II
play an important role in welfare state politics. A strong and effectively mobi- social-policy developments. The role of Social Democratic parties and union
lized opposition is likely to be able to face down even determined retrenchment pressure in the formation of early welfare states was limited. New Institution-
advocates. This effectiveness, however, is open to challenge. Interest-group alists have stressed that political institutions must be considered consequential
strength depends not simply on formal properties like the size of membership, structures. Power resources, these authors note, are themselves partly the result
but on the ability of the group's leadership to convince policymakers that the of institutional variables.
membership can actually be **delivered.”” This is especially true in the case of This approach to comparative social policy has been developed primarily by
the mass organizations that support social policies. Unlike some groups (e.g., Theda Skocpol and her collaborators.'” Their research agenda has centered on
business associations), their power depends less on financial resources (over explaining ** American exceptionalism”' — the belated and halfhearted develop-
which leaders may have direct control) than on the ability to offer or withhold ment of social-welfare policies in the United States. Just as a focus on Sweden
electoral support. Influence on policy depends on whether leaders can actually was central to the development of the power-resources model, concentration on
mobilize members to reward or punish policymakers for particular courses of the United States has underscored the importance of political institutions.
action.* The American state is distinctive.'' Its formal institutions create a marked
It is here that the interplay between the strategic options of retrenchment dispersal of political authority (both within the central government and between
advocates and the mobilizing potential of welfare state constituencies becomes central and local authorities) that often allows well-placed minorities to veto
relevant. The constituencies of welfare state organizations must be activated to unwanted policy change. The administrative capacities of the federal government
be politically effective. This is not an automatic process. Interest-group leaders are relatively feeble, impeding the design and implementation of extensive pol-
must persuade policymakers that their membership is so concerned about an icies. The representation of interests, in significant part because of institutional
issue that the wrong decision will have political consequences. The efforts of structures, has not taken the form of strong ideological parties and comprehen-
retrenchment advocates, as 1 have argued, are directed precisely toward blocking sive '*peak associations.”? Instead, parties have been weak, operating more as
this activation. Under the right circumstances, strategies of obfuscation, division, intermittent electoral/patronage machines than as promoters of coherent policy
and compensation can be highly effective. So one must look at other elements alternatives.'? Institutional decentralization also encouraged the emergence of
in the political system, such as the structures of political institutions and previous highly fragmented interest groups, while the parceling of political authority al-
policies, to judge the prospects for effective interest-group challenges to re- lowed many of these groups to gain significant access to policymakers. These
trenchment initiatives. characteristics of state structure, Skocpol and her colleagues have persuasively
Figure 2.1. Patterns of institutional design political authority. This is a result not only of its parliamentary system, but of
a first-past-the-post, single-member district electoral system, which limits the
Horizontal integration
potential for weak coalition government (Israel's system of proportional repre-
(concentration of authority within nat'l gov*t) sentation represents a contrasting case). In the United States that authority is
widely dispersed. These differences were in full force during the period under
Low High examination. The Thatcher government always had a substantial parliamentary
Vertical integration Low United States Canada majority. Reagan, on the other hand, operated in a context of '*divided govern-
(concentration of authority at
ment.'” He had to contend with a Democrat-controlled House of Representatives
nat"! level) High Israel Britain
throughout his two terms; in 1986, Republicans lost control of the Senate as
well.
argued, must be incorporated into an understanding of social-policy development A high degree of vertical integration is often considered a prerequisite for
in the United States. strong government.!'* Many have argued that because the concentration of po-
My analysis of retrenchment draws heavily on this school of thought. How- litical authority lowers the number of effective veto points, governments oper-
ever, 1 must stress that we are talking about a family of arguments. Some in- ating in parliamentary systems will have a much greater capacity to pursue
stitutional variables are much more relevant to retrenchment politics than others. radical policy change. So long as the governing party or parties has a majority,
The following discussion considers how institutional variables have been linked legislation can be passed even over heated opposition. By contrast, the American
to patterns of welfare state development, then evaluates their role in retrench- system of checks and balances can lead to deadlock and inaction, especially
ment politics. It is necessary to sort out the possible roles of two factors: the when a majority of Congress is loyal to a different political party than is the
structure of formal institutions and the policy-making capacities of government president. Party lines in the legislature are much more fluid in the United States,
authorities. A third factor, the role of policy feedback, requires a separate and and even if the president is of the same party as a majority of members of
extended analysis. Congress, he may have limited leverage to pass his proposals. There are, more-
over, many points in the legislative process at which opponents can effectively
Formal institutions block policy changes. Furthermore, legislators are more likely than in a parlia-
mentary system to be held individually accountable for their actions.'*
A central feature of New Institutionalism has been a renewed appreciation for Indeed, there is considerable evidence that a low level of horizontal integra-
the ways in which the structures of formal political institutions influence social tion probably slows the expansion of sociál programs.'* The multiplication of
processes. Analysts working from a variety of perspectives have argued that veto points has made it easier for determined minorities to thwart reform. In the
institutional structures profoundly influence the viability of alternative political United States, for example, even though clear majorities of the electorate have
strategies. Institutions steer individual choices in particular directions and thus often favored some form of national health insurance, and although Democrats
help to shape political outcomes.
controlled both branches, these political resources were not sufficient to over-
In discussing the impact of formal institutions on welfare state politics, two
come the institutional advantages of determined opponents.'*
distinct dimensions of institutional design need to be distinguished. One dimen-
As a basic account of the impact of horizontal integration, this is persuasive.
sion is horizontal integration: the extent to which power within the national In general, major policy change is likely to be easier to achieve in Britain than
government is concentrated or dispersed (e.g., parliamentary vs. separation-of-
in the United States. Whether vertically integrated political institutions facilitate
powers systems). The second is the level of vertical integration: the degree that
retrenchment, however, is less clear. Although parliamentary systems concen-
power is concentrated nationally or devolved to more local government author-
trate authority, they also concentrate accountability; the former tendency facil-
ities (e.g., federal vs. centralized systems).
itates retrenchment, but the latter impedes it.
At first glance, one might expect political institutions to account for different
Recall that retrenchment is generally unpopular. Because governmental power
policy outcomes in the United States and Great Britain. On both these dimen-
is more centralized in parliamentary systems, accountability is more centralized
sions, as Figure 2.1 indicates, the United States and Britain represent near-polar
as well. Governments can act to prevent groups from suffering losses, and the
extremes. Britain is a highly concentrated system, with only limited powers left
public knows this. Individual legislators in parliamentary systems are not im-
to local governments. In the United States (and by way of comparison, in Can-
mune from blame for party positions; they are in fact much more susceptible to
ada's parliamentary system), subnational political units are often critical political
swings in party electoral support than their counterparts in the United States.
actors. Within the national government, the British system radically concentrates
This may, in turn, make governments in parliamentary systems even more re-
luctant to undertake unpopular actions.'? However, this fear of negative political relatively low **accountability effects” of national political institutions to ob-
consequences will be felt not (as in the United States) through open political scure their responsibility for unpopular actions.
conflict, but through behind-the-scenes pressure within the governing party. When analysis turns from programmatic retrenchment to systemic retrench-
Furthermore, although the centralization of parliamentary systems increases ment, the impact of horizontal institutional integration remains limited. At first
the govemment's accountability, it also decreases the accountability of the op- glance, this is puzzling. Many of the policy changes that might further systemic
position. Opposition parties in parliamentary systems cannot hope to enact their retrenchment are not as unpopular as welfare state cutbacks. When systemic
own policy preferences while out of power; their only hope is to topple the retrenchment could be produced without electoral costs, the much more cen-
current government by publicizing its misdeeds. Governments in parliamentary tralized institutional structure of the British state should have given the Thatcher
systems, anticipating the high political cost of retrenchment actions, may forgo government a tremendous advantage. That systemic retrenchment did not pro-
opportunities that concentrated power would have allowed them to undertake. ceed very far in Britain reflects Thatcher's unwillingness to pursue policy op-
Thus, the theoretical case for believing that vertically integrated institutions tions that the institutional setting probably would have permitted. As 1 will argue
favor retrenchment is weak. We are left with the empirical question of whether in Chapter 7, this reflects the conflict between pursuing reforms that would
concentration-of-power effects outweigh accountability effects. The evidence to constrain the welfare state and other aspects of the Conservative government's
be presented demonstrates that the Thatcher government was as reluctant as the agenda.
Reagan administration to pursue widely unpopular policies. Nevertheless, if con- In the United States, institutional fragmentation generally posed a consider-
centrated authority does not appear to give an administration an overwhelming able barrier to systemic retrenchment efforts. Only when that fragmentation
advantage in achieving programmatic retrenchment, it does help to structure the could be turned into a political asset — that is, when lowering accountability was
choices available to retrenchment advocates. Reagan and Thatcher had to adjust important — was systemic retrenchment encouraged. There is, however, one cru-
their strategies to fit the contours of different institutional terrains. cial example of this: the Reagan administration's shrewd use of institutional
For example, different institutional settings led the two administrations to opportunities to enlist high deficits to the cause of constraining social spending.
emphasize varying strategies of obfuscation. In the United States, retrenchment This will be discussed in Chapter 6.
advocates often focused on the *'traceability”” link in voters' causal chains, What about the second dimension of institutional design, the degree of ver-
shifting or obscuring blame for unpopular policies. This was an unrealistic strat- tical integration? The evidence is reasonably clear that federalism constrains
egy in Britain. Indeed, the government's most prominent attempt to use this welfare state growth.'* The most important consequence of decentralized insti-
technique, the requirement that local governments rely on the highly unpopular tutions is the creation of ““fiscal competition”? among jurisdictions. Local gov-
“*poll tax”” to raise revenues, backfired. The concentration of accountability in ernments find it difficult to pursue redistributive policies for fear that high taxes
a parliamentary system meant that blame fell on the Thatcher government rather will lead business and wealthy individuals to move out while attracting low-
than on local authorities. income groups who would benefit from generous social programs.'” Centralized
In Britain, those seeking retrenchment used their much greater control over authorities can make such policies uniform throughout a country, limiting the
the policy-making process to devise retrenchment strategies that did not raise prospects that capital and labor mobility will pose such a dramatic threat to
problems of accountability. In other words, they worked hard to structure re- social provision.
forms in such a way that vocal opposition would not develop. For example, the This argument probably can be applied equally well to retrenchment politics.
British Conservatives? solid parliamentary majority allowed them to legislate All other things being equal, redistributive social policies transferred to local
reform in incremental stages where that tactic would help minimize opposition. authorities would be vulnerable to a downward cycle of jurisdictional compe-
In general, the concentration of political authority meant that the Thatcher gov- tition. The caveat is that where policy-making authority is shared between local
ernment was well placed to implement whatever strategies stood to lower op- and national officials, the need to produce a common agreement may represent
position. The concentration of accountability, on the other hand, meant that a barrier to change.” The Thatcher government responded to this problem by
where such strategies were unavailable the government generally had to retreat. further centralizing authority over social policy. In other words, the government
With less complete control over the policy-making process, retrenchment preferred to maximize its control over policy-making rather than exploit the
advocates in the United States tried to fashion strategies that minimized the need advantages of decentralization.
to force multiple policy changes through the numerous hurdles of the political In the United States, federal institutions created opportunities for certain re-
system. In cases where program structures required positive legislative or ad- trenchment strategies, but also posed occasional obstacles for reform. Pushing
ministrative action to continue, the Reagan administration was able to turn this authority down to states and localities was likely to put pressure on spending
institutional feature to its advantage. Retrenchment supporters also exploited the levels, and helped to shift blame away from Washington for program cutbacks.
 However, the desire to avoid both of these effects also meant that state and local of what Heclo calls ““collective puzzlement on society's behalf”” — bureaucrats
authorities often mobilized to prevent such changes. The intergovernmental are likely to be important.” In his investigation of Swedish and British welfare
lobby thus became an important source of protection for social programs in the state development, Heclo placed the autonomous actions of civil servants near
United States. The outcome of these clashes often turned on the structure of the center of the policy-making process.
particular programs. When reforms threatened key state and local interests, or A focus on administrative capacities might lead one to expect important dif-
when Washington was heavily reliant on local officials for implementation, de- ferences between the process of social-policy change in the United States and
centralizing reforms were hard to carry through. in Britain. Britain is generally considered to be a relatively **strong”” state,
If this discussion yields few broad conclusions, it reflects the fact that the equipped with a well-trained, prestigious, and reasonably autonomous civil serv-
role of formal institutions in welfare state politics is complex. Arguments sug- ice. By contrast, civil servants in the United States have less status. At the higher
gesting that one arrangement or another provides sweeping advantages for those ranks, movement in and out of public service is common. The administrative
seeking expansion or retrenchment are suspect. Particular arrangements are capacities of federal agencies, watched suspiciously by Congress as well as by
likely to be seen as mixed blessings for policymakers, enhancing certain capa- the president, are widely regarded as limited.
bilities while creating problems on other fronts.*' In both countries, institutional Again, however, one must ask whether factors important in determining pat-
arrangements served to steer policymakers in particular directions because they terms of welfare state expansion are equally pertinent to patterns of contraction.
made some retrenchment strategies appear more attractive than others. In fact, there is reason to doubt that administrative structures will be central to
the politics of retrenchment. **Can we administer it?”” is a fundamental question
when one is discussing new or greatly expanded public initiatives; but for re-
Governmental capacities
trenchment advocates, the primary goal is to dismantle existing efforts rather
Governments vary in the degree to which they possess the resources needed to than create new ones. Closing offices, curtailing services, and cutting benefits
implement strategies and policies. As Skocpol has argued, beyond **sheer sov- do not require formidable administrative capacities. Retrenchment initiatives in-
ereign integrity and .... control of a given territory?” the crucial elements of state volving efforts to promote privatization may be more complicated to administer.
capacity, the *“universal sinews of state power,'” are the presence of **loyal and Even here, however, the reliance on private-sector mechanisms eases the burden
skilled officials and plentiful financial resources.'”"*? For any given policy, de- on public bureaucracies. The following chapters in fact reveal very little evi-
cision makers must consider not only political constraints but administrative and dence that important retrenchment efforts have been scuttled because govern-
financial ones. ments lacked the administrative capacity to carry them through.
Even if policies can be enacted, they may be impossible to implement. A Nor have civil servants been crucial political actors. My examination of the
number of studies have demonstrated how the existence or lack of specific ad- Reagan and Thatcher records suggests that bureaucratic activity has played only
ministrative capacities has influenced policy development. In an influential ar- a marginal role in the politics of retrenchment. Politicians have been the prin-
ticle, Weir and Skocpol argued that previous experience with public-works cipal decision makers. Their success has depended on an ability to shape strat-
programs was an important factor in the Swedish government's ability to launch egies to minimize opposition. Bureaucrats have played some part in identifying
Keynesian policies in the 19305. Governments without these administrative ca- these strategies, but the choice of goals and policies, and the factors determining
pacities were likely to turn to more traditional policy options.?* the likelihood that particular strategies will succeed, have all been outside their
Bureaucratic capacities also matter because bureaucrats themselves are polit- control. In short, neither governmental administrative capacities nor autonomous
ically relevant. Those occupying key positions within government agencies have bureaucratic activity shaped ultimate outcomes in fundamental ways.
expertise and command significant institutional resources. They are often given Because governments need money to finance major social programs, a gov-
responsibility for devising solutions to pressing problems. Frequently, bureau- ernment's revenue-generating capacity is likely to have an impact on the welfare
crats have the strategic advantage of possessing both a longer time horizon and state. Although historical institutionalists have acknowledged this fact, they have
more focused policy goals than other political actors. Administrators are some- not developed systematic arguments about the links between financial capacities
times energetic policy entrepreneurs, devoting careers to the construction of and program enactment or expansion. There is, however, significant evidence
political coalitions that can further their policy ambitions.?* that the structure of taxation has an impact on how much money governments
Bureaucratic influence is likely to be especially prominent in laying out al- can raise, and hence on how much is available for social programs. The nature
ternatives for government action: Those within key agencies can credibly claim of this link between tax structure and financial capacities fits well with the
to know what the government can and cannot do, what will work and what will general approach to policy design advanced in Chapter 1. The less visible the
not. To the extent that policy evolves through a process of problem solving — tax system (e.g., high reliance on indirect and payroll taxes rather than on in-
 an y 22... muss seras a BUYEIIIIcI can generate without provoking a
Table 2.1. Institutional variables and welfare state politics
taxpayer backlash.?*
Although administrative capacities become relatively unimportant in retrench-
Formal institutions State capacities
ment politics, government financial capacities continue to play an important role.
That role needs to be carefully spelled out, however. For retrenchment advo- Horizontal Vertical
integration integration Administrative Financial
cates, the government's revenue-generating capacity is a two-edged sword. As
I have already noted, programmatic retrenchment is often facilitated if the gov- Expansion Positive Positive Positive Positive
ernment is in a position to compensate at least some of its potential opponents. Retrenchment Mixed-P Mixed/Negative-P Negligible Positive-P
The healthier the government's financial position, the easier it is to buy one's Mixed-S Negative-S Negative-S
way out of political troubles. The following chapters provide examples in which
the Reagan administration's retrenchment efforts would have benefited from a P = Programmatic
bit more financial room to maneuver. However, this is likely to be more than S = Systemic
offset by the negative consequences for retrenchment of robust financing. Al-
though improving the prospects for programmatic retrenchment, healthy finances topic because it is so central to retrenchment politics. One simply cannot make
impede systemic retrenchment. Increasing the government's revenue-generating sense of the contemporary politics of the welfare state without considering how
capacity makes it harder to create an overall climate of austerity to constrain the consequences of-preexisting policies structure struggles over social-policy
social provision. reform.
This dual impact of government finances is important because of a major
aspect of the two records reviewed in this book: The Reagan and Thatcher
POLICY FEEDBACK AND POLITICAL CHANGE
governments diverged more radically on tax policy than on any other major
domestic issue.?” Whereas Thatcher increased the British government's revenue- I have suggested that a major reason for viewing retrenchment politics as a
generating capacity, the Reagan administration curtailed that capacity in the distinctive enterprise is that the political context has changed considerably since
United States. This difference in state capacities in turn conditioned retrench- the period of welfare state development. Perhaps the most important aspect of
ment politics. Financial flexibility allowed the Thatcher government to pursue this change has been the emergence of extensive patterns of government inter-
compensation-based strategies of programmatic retrenchment impractical in the vention in social and economic life, that is to say, the emergence of welfare
United States. The Reagan administration, however, could use the deficit issue states themselves. In advanced industrial democracies, anywhere from 30 per-
as a powerful tool for constraining overall domestic spending, whereas Thatcher cent to 60 percent of gross national product is filtered through government pro-
could not. grams. Large-scale policies are a critical feature of the contemporary political
The argument advanced here is clearly informed by key elements of New environment. The arrival of big government makes an analysis of policy feed-
Institutionalist theory. However, institutional variables often play a different role back — again, the ways in which previous policy choices influence present po-
in the politics of retrenchment than the one analysts have identified in the politics litical processes — an integral part of any investigation of social-policy change.”
of welfare state expansion. The changes reflect the distinctive characteristics of Given the development of these massive public systems of resource extraction
retrenchment, such as the new concern with minimizing accountability and and deployment, it is hardly surprising that, as E. E. Schattschneider argued
blame for unpopular activities. Table 2.1 contrasts the treatment of institutional more than a half century ago, “new policies create a new politics.””?” However,
variables in studies of welfare state expansion with the arguments 1 have made political scientists were slow to incorporate Schattschneider's insight into their
about the politics of retrenchment. Many of the processes stressed in studies of models of politics. Traditionally, researchers treated policy as the result of po-
earlier periods turn out to be relatively unimportant. Whereas the concentration litical forces (the dependent variable), but rarely as the cause of those forces
of political authority clearly has a positive affect on social-policy expansion, its (the independent variable). In the past decade or so, this has ceased to be true.
impact on retrenchment appears to be more mixed; and although bureaucratic Scholars working on a range of empirical issues have begun to emphasize that
capacity and autonomy influenced welfare state development, they have limited “policies produce politics.””
implications for the politics of retrenchment. Financial capacities continue to The massive twentieth-century expansion of the public sector has clearly
matter, although in ways that pose dilemmas for retrenchment advocates. contributed to this new orientation. Increasing government activity made it
One important component of New Institutionalist analysis remains to be dis- harder to deny that public policies were not only the result of but important
cussed: the policy feedback from previous political choices. 1 have deferred this contributors to the political process, often dramatically reshaping social, eco-
nomic, and political conditions. Intellectual developments have also fostered this provides incentives for individuals to join particular groups (e.g., by banning or
shift in research. The **post-behavioralist”” emphasis on the structural constraints harassing alternative organizations). In a compelling essay on the development
facing individual actors has led scholars working from a variety of perspectives of the Swedish labor movement, Bo Rothstein has demonstrated that policy
to begin identifying the ways in which formal and informal rules of the game designs that gave unions authority over unemployment funds provided a crucial
in political and social life influence political behavior. To date, most analysis impetus to the development of powerful labor confederations.** Union admin-
has centered on formal governmental institutions and political organizations. istration of these funds gave workers a strong selective incentive to join. Roth-
However, major public policies also constitute important rules, influencing the stein's analysis indicates that union membership rose rapidly and stabilized at
allocation of economic and political resources, modifying the costs and benefits higher levels in countries that adopted such plans.
associated with alternative political strategies, and consequently altering ensuing Policies may also strengthen particular groups by increasing their access to
political development. An examination of the political consequences of policy decision makers. Analyses of European corporatism have often stressed the con-
structures is a logical extension of New Institutionalist arguments. scious efforts of policymakers to implement reforms that strengthen links be-
Research on policy feedback has stressed two arguments: that policy struc- tween government elites and key interest groups. Students of regulatory
tures create resources and incentives that influence the formation and activity of “capture”? have noted a similar phenomenon.** Interventionist government pol-
social groups, and that policies affect processes of *“social leaming'” among icies often have the paradoxical effect of making the success of particular pol-
major political actors. In this section 1 review these claims and suggest two icies dependent upon-group-controlled resources (e.g., information or skilled
additional kinds of policy feedback. In the following section, I demonstrate the personnel). This dependence in turn enhances the ability of groups to turn their
importance of incorporating these arguments into the analysis of welfare state preferences into government policy.
politics. There is frequently a strong case for standing the pluralist claim that interest
If interest groups shape policies, policies also shape interest groups. The groups drive policy on its head. Groups are profoundly influenced by structures
organizational structure and political goals of groups may change in response of public policy. These structures may give organizations a reason to exist,
to the nature of the programs that they confront and hope to sustain or modify. facilitate or impede efforts to overcome collective-action problems, or provide
Policies provide both incentives and resources that may facilitate or inhibit the access to considerable political resources. Feedback effects on interest groups
formation or expansion of particular groups. Such incentives stem primarily need to be incorporated into our analyses of policy-making.
from the major social consequences of specific government actions. Public pol- A second approach to policy feedback starts not from how policies provide
icies often create **spoils”” that provide a strong motivation for beneficiaries to resources and incentives for interest groups, but from how policies provide in-
mobilize in favor of programmatic maintenance or expansion.*” Policy designs formation that helps individuals navigate the social world's complexities. Some
can also create niches for political entrepreneurs, who may take advantage of scholars have stressed the importance of learning effects in policy-making, con-
these incentives to help **latent groups”? overcome collective action problems.*' centrating on the efforts of politicians to understand the consequences of their
The history of the now-formidable American Association of Retired People own actions.** Political-learning arguments focus on those at or near the center
(AARP) illustrates this feedback process. The inadequacy of health-care benefits of the policy-making process, and emphasize problems of bounded rationality
for the elderly provided the AARP with a niche for activity. The sale of health and uncertainty. Implicitly or explicitly, these analysts build on work in decision
insurance prior to the enactment of Medicare, and of **Medigap”” policies since making and organizational theory, which emphasizes the variety of techniques
then, has provided a strong *“selective incentive” for individuals to join AARP, used to cope with limited cognitive capacities.** Heclo summarized this per-
promoting the development of an elderly lobby that is unmatched in other coun- spective in an early but still-influential formulation:
tries.*?
Not only do public policies create incentives for interest-group activities, they Politics finds its sources not only in power but also in uncertainty — men collectively
wondering what to do. Finding feasible courses of action includes, but is more than,
may also provide resources that make such activities easier. The political influ-
locating which way the vectors of political pressure are pushing. Governments not only
ence of groups varies dramatically; some are central to the development of
“power”' (or whatever the verb form of that approach might be); they also puzzle. Policy-
policy whereas others are ineffectual, forced to accept gains and losses deter-
making is a form of collective puzzlement on society's behalf; it entails both deciding
mined by others. Public policies can clearly feed back into politics in this respect and knowing.”
too. Policies can have an effect on the resources of groups, and on the ability
of groups to bring those resources to bear on decision makers. The depiction of policy change as a learning process is sometimes presented
Sometimes government policies create interest-group resources in a straight- in sweeping terms. Heclo, for example, talks of both social leaming and political
forward sense, as when legislation gives funding to favored organizations or learning, and identifies a number of sources of such effects. Prominent among
uem, nowever, 1s the impact of previously adopted public policies, and it is commitments, in turn, may vastly increase the disruption caused by new policies,
these policy-learning effects that are relevant here. Important political actors effectively locking in previous decisions.
may become aware of problems as a result of their experiences with past initia- That this type of policy feedback has not been widely utilized in investiga-
tives. The setting of a new agenda and the design of alternative responses may tions of policy development is a significant oversight. There is good reason to
build on (perceived) past successes or may reflect lessons learmed from past believe that these lock-in effects can be of considerable significance. To date,
mistakes. economic historians interested in the development of technology have conducted
Policy learning has emerged as an important line of argument for those in- most of the work on such effects. 1 will begin by summarizing this research and
terested in how policy feedback affects the cognitive processes of political ac- then suggest how it can be extended to the study of policy feedback.*
tors. Heclo”s study of the Swedish and British welfare states remains the most Economic historians using the assumptions of neoclassical economics have
sustained effort to develop this idea. **Once implemented,”” he argues, *“a tech- recently demonstrated that under certain conditions the development of tech-
nique such as social insurance has tended to be readopted, to be considered the nology will not proceed toward the most economically efficient alternatives. The
“natural” policy response for other types of income risk.””* Overwhelmed by QWERTY typewriter keyboard is a classic example. Although more efficient
the complexity of the problems they confront, decision makers lean heavily on alternatives to the QWERTY key layout existed when it was developed, they
preexisting policy frameworks, adjusting only at the margins to accommodate were not feasible for use until better machines allowed typists to go faster with-
distinctive features of new situations. out jamming keys. However, by the time such machines were introduced, the
Although policy-learning arguments have been applied with considerable QWERTY configuration was so well established that alternatives could not gain
force in a variety of settings, their explanatory power and scope of application a foothold in the industry, and the slower standard was effectively locked in.”
remain open to question.*” So far it has proved difficult to demonstrate that Under what conditions are such outcomes likely? Brian Arthur has identified
policy learning has a significant impact on actors” political behavior other than the following factors:
simply contributing to their accounts of their actions. Heavy reliance on case
studies has also made it hard to establish how often policy learning is important. Large set-up or fixed costs. If initial costs are a high proportion of total ex-
If governments both **power”” and **puzzle,”” when should we expect to see penses, there are likely to be increasing returns to further investment in a given
one type of process or the other? Learning processes seem likely to be more technology, providing individuals with a strong incentive to identify and stick
important when policy-making remains insulated from broader political con- with a single option.
flicts. Policy learning is also likely to play a different role at different stages of
the policy-making process. Learning effects will be most apparent in the iden- Learning effects. Large learning effects, which may lower product costs or im-
tification of particular policy alternatives, since this is when detailed knowledge prove their use as prevalence increases, provide an additional source of increas-
is most crucial.*” It is less clear that policy learning is central to the formation ing returns.
of government agendas or to the final choices between alternative policies.
Despite limitations, these arguments about policy feedback are broadly per-
Co-ordination effects. In many cases, the advantages an individual derives from
suasive. Nevertheless, they miss a level of political consequences that stem from
a particular activity depend on the action of others. These effects may encourage
previous policy choices. Policy feedback not only affects the resources of qr- coordination with others in adopting a single option.
ganized interests and the mind-sets of political elites; it also creates incentives
“and provides information for individual members of the electorate. Indeed, the
Adaptive expectations. 1f it is important for individuals to **pick the right horse””
effects of policies on mass publics are tremendously important. Unless this feed-
— because options that fail to win broad acceptance will have drawbacks later
back induces overtly political action, however, it is unlikely to attract the atten-
on — individual expectations about usage patterns may become self-fulfilling.**
tion of political scientists.
Two feedback processes should be distinguished. First, policies may encour- The existence of lock-in effects in the development of technology is now
age individuals to adapt in ways that lock in a particular path of policy de- generally accepted, but one can legitimately ask whether this excursion into
velopment. By ''lock in”? 1 mean that they bring about the policy-induced economic history has any relevance to the current discussion. Douglass North
emergence of elaborate social and economic networks that greatly increase the argues persuasively that it does. The factors Arthur identifies as contributing to
cost of adopting once-possible alternatives and inhibit exit from a current policy technological lock-in — increasing returns and high fixed costs, learning effects,
path. Major policy initiatives have major social consequences. Individuals make coordination effects, and adaptive expectations — are often characteristics of
important commitments in response to certain types of government action. These institutions. Consequently, one could anticipate the same kind of historical proc-
ess, a process of path dependence, in which initial choices of institutional design the absence of conflict. In Peter Bachrach and Morton Baratz's terms, lock-in
had long-term implications for economic and political performance. leads to *“*non-decisions.””** Another problem is that comparative analysis may
This argument can easily be applied to public policies as well. North defines be required to study policy lock-ins. An analyst needs a comparative case in
institutions broadly as **the rules of the game in a society or, more formally, which lock-in has not occurred to identify the political effects of policy feed-
. « the humanly devised constraints that shape human interaction.””* This def- back.
inition would seem to encompass public policies as well as what we con- Nevertheless, there is good reason to believe that instances of policy lock-in
ventionally recognize as institutions, since policies clearly do establish rules and are widespread. Many public policies create or extend patterns of complex social
create constraints that shape behavior. The specific example North uses to il- interdependence in which microeconomic models of isolated, independent in-
lustrate his argument about path dependence is instructive. The Northwest Or- dividuals smoothly and optimally adapting to new conditions do not apply.*?
dinance was a quasi-constitutional initiative, outlining basic rules for “*the The characteristics identified by economic historians, modified to incorporate
governance and settlement of the vast area of land in the West and... a frame- aspects of social as well as technological complexity, provide an excellent start-
work by which the territories would be integrated into the new nation.” In ing point for this research. Lock-in effects are likely to be important when public
this respect, it resembled a formal institution. However, the Northwest Ordi- policies encourage individuals to make significant investments that are not easily
nance was also **a specific legislative enactment”” — that is, a public policy. reversed, or when individuals face strong incentives to coordinate their activities
By choosing such a legalistic, foundational initiative — one that created such with other social actors and adopt prevailing or anticipated standards. Policies
straightforward rules of the game — North's example obscures the broad appli- that involve high levels of interdependence, and in which intervention stretches
cation of his argument to policy feedback. Policies may create incentives that over long periods, are particularly likely sites for lock-in effects.
encourage the emergence of elaborate social and economic networks, greatly Policy feedback may also have a second major effect on large segments of
increasing the cost of adopting once-possible alternatives and inhibiting exit the public. Just as policies provide information that may produce political learn-
from a current policy path. Individuals make important commitments in response ing among policymakers, policy-induced cues may influence a private indivi-
to certain types of government action. These commitments, in turn, may vastly dual's awareness of government activity. In common with policy-leaming
increase the disruption caused by new policies, effectively locking in previous arguments (which have been applied primarily to bureaucrats), these arguments
decisions. focus on the signals that policies send to political actors. Policy initiatives may
Like more formal institutions, public policies operating in a context of com- send signals that influence individuals” perceptions of their own interests, of
plex social interdependence will often generate increasing returns and high fixed whether their representatives are protecting those interests, of who their allies
costs, leaming effects, coordination effects, and adaptive expectations. For ex- might be, and of what political strategies are promising.
ample, housing and transportation policies in the United States after World War Besides broadening the range of actors considered, this approach has a sig-
II encouraged massive investments in particular spatial patterns of work, con- nificant additional advantage over a focus on policy learning. Not only does it
sumption, and residence. Once in place, these patterns sharply constrained the acknowledge that all policy-making takes place in a context of information con-
alternatives available to policymakers on issues ranging from energy policy to straints, but it recognizes that the distribution of this information is often highly
school desegregation.** unequal. The emphasis of these arguments is on how information asymmetries
Many of the individual commitments that locked in suburbanization were create space for the strategic manipulation of policy design. Knowledge is indeed
literally cast in concrete, but this need not have been the case, nor need it be power, and the fact that policy structures can influence the role and availability
in the future. Policies may encourage individuals to develop particular skills, of information makes this an important and contested aspect of policy devel-
purchase definite varieties of goods, or devote time and money to distinct kinds opment. To rephrase Heclo, '*powering'” and “*puzzling'” are often part of the
of organizations. All these decisions generate sunk costs. That is to say, they same process; power can be utilized to facilitate or impede actors” efforts to
constitute investments that generate strong commitments. In many contexts, pol- understand the consequences of public policies.
icies may push individual behavior onto paths that are hard to reverse. Recall Arnold's concept of causal chains, reviewed in Chapter 1. The critical
Political scientists have been slow to build an examination of lock-in proc- point is that the ability of voters to reconstruct causal chains can vary indepen-
esses into their models of policy development.* One reason for this lack of dently of a policy's actual impact, and that this variation may be a product of
attention is that policy feedback of this kind has a tendency to depoliticize issues. policy design. Specific features determine a policy”s informational content, in-
By accelerating the momentum behind one policy path, it renders previously fluencing these determinants of the electorate”s reaction. Policies that distribute
viable alternatives inaccessible. The result is often not conflict over the forgone benefits widely and intermittently are likely to be less visible than those that
alternative (which political scientists would generally be quick to identify), but distribute benefits to a concentrated group and in a single package. Whether
 those affected are part of a network (e.g., geographical or occupational) allowing
POLICY FEEDBACK AND WELFARE STATE POLITICS
communication with others affected (what Arnold has called **proximity””) is
another important factor. Homeowners living near the same toxic dump and Arguments about policy feedback are essentially arguments about the conse-
dairy farmers engaged in a common profession are each likely to be part of quences of big government. As policy decisions have had increasingly pervasive
networks that facilitate communication and therefore improve the chances that effects on economic and social life, their impact on political processes has ex-
they will become aware of decisions or actions that affect them. Recipients of panded. Given that the welfare state is at the very heart of big government, it
disability payments who have their benefits cut are unlikely to be similarly should come as no surprise that studies of welfare state development have gen-
linked, since they may be scattered throughout the country and lack organiza- erated some of the most persuasive arguments about the role of policy feedback.
tional linkages. a Heclo's analysis has already been discussed. His investigation of social policy
Because a crucial factor in linking outcomes to policy is the length of the in Britain and Sweden puts great weight on incremental policy-learning proc-
causal chain, the more stages and uncertainties that lie between a policy's en- esses, in which past policies have provided the intellectual models for policy-
actment and its perceived result, the less likely it is to provoke a popular re- makers. Skocpol and her colleagues have also stressed policy-learning processes,
sponse. Policymakers have a significant degree of control over this. They may though with a twist: Orloff and Skocpol's account of the impact of Civil War
choose interventions that create causal chains of varying lengths. Ideally, they pensions in the United States emphasizes negative reactions to preexisting pol-
seek to design programs for which the benefits involve short causal chains and icies, and the ““learning”” took place mainly among social groups rather than
the costs involve long ones. Time lags, for example, add greatly to the length among bureaucrats and politicians.? Perceived by key middle-class reformers
and complexity of causal chains, so policymakers favor policy designs that ac- as a scandalous example of patronage politics, Civil War pensions were avoided
centuate immediate benefits and delay or camouflage costs. as a blueprint for incremental extensions of government activity. On the con-
Traceability — the linking of government action to specific decision makers trary, important political actors drew negative lessons that served as a check on
— may also depend on policy design.*” Policies can often be structured to the emergence of significant federal social expenditures in the pre-New Deal
heighten or obscure the role of decision makers. As Weaver has argued, index- era. Negative learning thus had a major effect on the policy agendas and political
ation mechanisms that put particular policies on **automatic”” have proven at- strategies of prominent middle-class reform groups.
tractive precisely because they reduce the traceability of outcomes to particular In other essays on welfare state development, New Institutionalists have given
decision makers.*' To take another recent example, the intricate legislative his- weight to the role of policy feedback in shaping interest-group structures. Many
tory surrounding government regulation of the savings-and-loan industry made observers have suggested that heavy reliance on means-testing in the United
it practically impossible for even the most incensed taxpayer to know which States, for example, encouraged a sharp bifurcation between groups interested
politicians to hold accountable. in middle-class entitlement programs and those concerned about the position of
Given significant evidence that the information in policies is important for the poor. Esping-Andersen has applied the inverse argument to Sweden and
mass publics, these feedback effects deserve careful attention. Certainly politi- Norway, where universalist policies produced a more solidaristic interest-group
cians” ability to raise or lower the profile of their actions for different constit- politics.**
uencies would seem to give them an important political resource. This potential The great expansion of the modern welfare state provides reason to suspect
underscores the point that although policies indeed allocate substantial resources that consequences of previous policy choices will be even more important for
and create powerful material incentives, some of their most important effects the politics of retrenchment. Indeed, 1 will argue that policy feedback was a
may be cognitive. The massive scope of public policies assures that they play crucial determinant of retrenchment results in Great Britain and the United
a significant role in our efforts to understand and act in an enormously complex States. In both countries, previous policy choices generated resources and in-
political world. centives that helped structure the development of relevant interest groups. For
The central claim of policy-feedback arguments is that policies themselves example, the fragmentation and underdevelopment of British pension policy
must be seen as politically consequential structures. The rise of active govern- contributed to a similarly fragmented and underdeveloped structure of pensioner-
ment leaves little room for doubt about this. Nonetheless, that policy-feedback interest representation. These groups, in contrast to the powerful elderly lobby
arguments are now widely applied in divergent national contexts and across a in the United States, were easily outmaneuvered by the Thatcher government in
variety of issue areas drives home the growing importance of this concept to the struggle over pension retrenchment.
the study of comparative politics. In a wide range of circumstances and in nu- Lock-in effects have been apparent as well. Where public social provision
merous ways, policies restructure politics. The immediate question, however, is generates extensive networks of commitments, retrenchment advocates find ex-
the relevance of these factors for the study of welfare state dynamics. isting policies hard to reverse. The Social Security system in the United States,
tU UL MUiIILUIJDILI 212 SM ENA E A a O >
Table 2.2. Policy feedback and welfare state politics
feedback. Retrenchment advocates in the United States found that their existing
commitments sharply circumscribed options for radical reform. Social Security?s
Feedback type Expansion Retrenchment
complex financial arrangements involved a series of implicit promises stretching
decades into the future. Huge public expenditures were essentially precommit- Interest-group activity Substantial impact (creation Substantial impact
ted. Furthermore, these promises had shaped the retirement strategies of most of spoils, niches for
entrepreneurs, selective
Americans. The Reagan administration found itself locked in by the dense net-
incentives)
work of commitments produced by previous policy choices.
“Lock-in”” effects Negligible during Substantial impact in specific
Welfare state programs do indeed distribute material resources and generate
formative period sectors
incentives that help structure retrenchment politics. Just as important, however,
Policy leaming Considerable evidence of Negligible impact
is the impact of policy design on political actors” attempts to make sense of the both positive and negative
social world. The structures of existing policies may influence the availability learning effects
of information and hence the prospects for actually mobilizing potential political Information effects Unknown Substantial impact (creation
resources. The information content of existing policy designs becomes crucial of opportunities for
in retrenchment politics because of the centrality of obfuscation strategies. As retrenchment advocates to
already noted, if policymakers are attempting to pursue unpopular courses of pursue obfuscation
techniques)
action, they will do their best to camouflage their activities; the options open to
them in this respect have already been discussed.
Policy feedback can play a crucial role in formulating retrenchment strategies.
Illustrations of this point are discussed in detail in the following chapters. Pro- to perceived limitations in specific aspects of the welfare state. Nor were their
gram structures, which establish eligibility requirements, benefit rules, and pat- agendas shaped fundamentally by the ways in which previous policymakers had
terns of service provision, provide key levers that policymakers will try to framed discussions of social policy. Instead, both Reagan and Thatcher offered
manipulate in ways that weaken prospects for mobilized opposition. Whether a virtually complete rejection of all but the most residual social policies. Al-—
policy structures produce or inhibit what John Kingdon has termed '*focusing though both the political strategies adopted and the degree of success achieved
events”? — dramatic, attention-generating occurrences — can be important for varied from program to program— and usually fell well short of such ambitious
retrenchment politics.** In some cases, focusing events may give retrenchment goals — the reformers” agenda in each case was radical retrenchment. Lessons
advocates a political advantage by forcing reform onto the agenda despite op- learned from the specific features of past policies played very little part in the
position from program supporters. As 1 argue in Chapter 3, the specific design formation of these two governments” programs.
of Social Security in the United States has produced intermittent '*trust-fund In emphasizing the role of policy feedback, this analysis follows a growing
crises.'” Although lock-in effects prevented radical reform in the United States, trend among students of policy development. Nevertheless, much dispute re-
these dramatic focusing events helped set the agenda for pension reform, cre- mains over how such feedback conditions change. 1 have suggested that it is
ating opportunities for moderate cutbacks in benefits that might have been po- possible for policy feedback to influence the prospects for policy change in four
litically impossible otherwise. The absence of such events in a particular arena ways, summarized in Table 2.2. It bears repeating not only that a wide range
may be important as well, especially when retrenchment advocates are trying to of propositions can be placed under the rubric **policy feedback,'” but that
lower the visibility of their efforts. Chapters 4 and 5, which examine housing different propositions imply quite distinctive images of the political process.
and income-support policies, show how both governments have identified pro- Policy feedback might matter because it determines how policy analysts and —
gram features permitting the introduction of automatic retrenchment mechanisms bureaucrats learn from and modify past commitments. Feedback might also be
that generate little public attention. important because it influences how social actors organize collectively, and af-
1f there is considerable evidence of the role of these various feedback effects fects their ability to mobilize resources in defense of their interests.
in retrenchment politics, policy-learning arguments appear to be less applicable. Table 2.2 also indicates which kinds of policy feedback are most important
Whatever relevance policy leaming may have in other contexts, its role in the for retrenchment politics. Past research has often emphasized the ways in which
formation of the agendas of retrenchment advocates has been minimal. Reform reformers — especially bureaucrats — develop new policies in response to the
initiatives did not percolate up from agencies dissatisfied with the workings of perceived failure of old ones. The current analysis, by contrast, downplays this
current programs. Neither administration pushed reform agendas that responded “policy learning”” aspect in favor of other consequences of policy structure.
This shift in emphasis reflects the distinctive qualities of retrenchment politics: TH
widespread public provision, extensive networks of organized supporters, and
the unpopularity of cutbacks. In this environment, achieving policy change is The politics of programmatic
likely to be less a matter of learning than one of identifying strategies to weaken
or outflank political opponents. retrenchment
Previous policies matter not because they have transmitted particular lessons
to public officials. Welfare state retrenchment is far too controversial and im-
mediate for mass publics to generate such an insulated process of bureaucrat-
driven policy change. Instead, previous policies matter because they help shape
the distribution of political resources (including information) and provide the
raw material from which retrenchment advocates must try to design successful
strategies. In this sense, the current analysis tries to bridge some of the gap
between New Institutionalist arguments about policy change and those that place
greater emphasis on social groups. Social forces are important, because advo-
cates of retrenchment are unlikely to succeed in the face of substantial political
opposition. Nevertheless, institutional factors — including the structure of formal
institutions, but especially the consequences of previous policy initiatives — are
central in determining whether this political opposition actually emerges. As the
following chapters explore in detail, retrenchment advocates were able to suc-
cessfully pursue strategies of obfuscation, division, and compensation only
where institutional structures and existing policy designs were favorable.
 3
Retrenchment in a core sector:
old-age pensions

This chapter contrasts the initiatives of the Thatcher government and Reagan
administration in a core area of social policy: old-age pensions. As might be
expected, attempting to cut these popular social programs was politically dan-
gerous; each government experienced both severe setbacks and occasional suc-
cess. In the end, however, the Thatcher government implemented far-reaching
and probably irreversible reforms in pension provision. This outcome is espe-
cially striking because it defies the conventional wisdom that middle-class en-
titlements are inviolable. Furthermore, it stands in stark contrast to the repeated.
and politically costly failures of Thatcher's efforts to reform the other pinnacle
of British social provision for the middle class, the National Health Service. In
the United States, Social Security emerged from the Reagan years essentially
intact. Although the Social Security amendments of 1983 produced some sig-
nificant reductions in future pension benefits, a number of glaring setbacks over-
shadowed this single and limited political success. In contrast to Britain, reform
in the United States modestly scaled back the existing pension program rather
than refashioning policy in line with conservative preferences.
A satisfactory account of these events must explain both the marked diver-
gence in final outcomes and the patterns of success and failure in each country.
Why did some initiatives fail when others did not? Why did the ultimate form
and scope of retrenchment differ substantially in the two cases? My answer to
these questions stresses the crucial role of preexisting pension structures — the
feedback effects of previous policy choices. In Great
Britain, the fragmented
nature of state provision left potential opponents of reform divided and weak.
Specific features of existing programs, including indexation mechanisms and the
strong linkages between public and private pension systems, provided the
Thatcher government with opportunities to obscure the extent of public-sector
retrenchment. Thatcher did not always get her way, but she controlled the po-
litical agenda and ultimately engineered a major transfer of responsibility for
retirement provision to the private sector.
In the United States, on the other hand, a single, mature program of public
provision dominated the field of old-age security. The system's scope generated
PP A O . de .
a strong and coherent base of political support while creating extensive long-
term financial commitments. Both of these conditions diminished prospects for

53
major policy change. Not only did the design of Social Security prohibit radical sions.* Although the system was somewhat more generous to low-income work-
reform, it also dictated the circumstances under which limited retrenchment ers, both contributions and benefits were to be tied to previous earnings. Since
would prove possible. Cutbacks occurred only in the wake of financial imbal- its inception, middle-class groups have had a strong stake in Social Security.
ances in the pension system itself. The scope and structure of Social Security Broad political support helped sustain a long and uncontroversial period of ex-
produced trust-fund-driven politics. Despite attempts to seize the initiative, the pansion unique in the history of American social-policy development.
Reagan administration's role was essentially reactive. Efforts to produce re- By contrast, Britain adopted a system of flat-rate pensions in 1908.* The
trenchment outside the context of trust-fund difficulties invariably resulted in system became contributory in 1925, but serious discussion of adding eamings-
politically costly defeats. related provision did not begin until the late 1950s. Because the scope of any
The evidence presented here, 1 will argue, strongly supports the claim that such expansion would have a major impact on the room for private provision,
standard explanations for patterns of welfare state development such as national debate between the Conservative and Labour parties was intense. Successive
political culture and the power resources of liberal or social democratic oppo- governments proposed legislation implementing earnings-related provision in
nents cannot account for this important divergence between the experiences of 1958 (Conservative), 1969 (Labour), and 1972 (Conservative), but each plan
these two conservative administrations. The structure of formal institutions was was scrapped when the next election brought the opposition party to office.*
of greater significance, but made less difference than one might expect. Instead, Finally, in 1975 — four decades after the U.S. Congress passed the Social Se-
policy feedback was crucial. Previous pension-policy choices provided the raw curity Act — a new Labour government was able to forge an all-party consensus
material from which retrenchment advocates had to try to design successful on a new plan, the State Earnings-Related Pension Scheme (SERPS).
strategies. Existing policies influenced the setting of political agendas and the These distinctive paths of national-policy development yielded quite different
prospects for demobilizing opponents of pension cutbacks. Furthermore, by cre- public-pension programs by the end of the 1970s. The systems differed in in-
ating extensive patterns of commitments, existing policies in the United States clusiveness, maturity, treatment of private-sector options, financing, and index-
created lock-in effects that greatly increased the cost of pursuing major reforms. ation mechanisms. Although some of these differences might appear to be
matters of detail, each was to have important consequences during the struggles
over pension reform in the 1980s.
THE STRUCTURE OF PENSION POLICIES IN BRITAIN
AND THE UNITED STATES
Inclusiveness
The growth of public pensions has been at the heart of the development of the
modern welfare state. Like all advanced industrial countries, Britain and the For the overwhelming majority of Americans, public-retirement provision meant
United States greatly expanded state pensions in the past half century. Despite a single program, Social Security.* Regardless of age or income (except for
this basic similarity, the two systems differed in crucial respects, including the those well below the poverty line), Americans expected to turn in their retire-
scope and maturity of public schemes, the role of earnings-related benefits, the ment to the same source of public benefits. In Britain, on the other hand, public
treatment of private pensions, and the structure of financing. Surprisingly, it was provision was much more fragmented. The Basic Pension was indeed universal,
the United States — generally and with good reason categorized as a '*“welfare but its benefits were far lower than those available through Social Security in
state laggard'” — that created the_more extensive and resilient public-pension the United States.” As a result, a significantly larger proportion of the elderly
sector. AAA 7 were dependent on means-tested support.* SERPS added significantly to the
Indeed, although the United States and Britain are often grouped together as confusion. The system was of no relevance to the already retired, who had not
““liberal”” welfare states, they represent near opposites in pension development. had time to build up significant entitlements; it would be of great importance
As John Myles has noted, the postwar period was marked by a gradual conver- to younger age cohorts, however. These groups were further divided, however,
gence of pension systems in advanced industrial societies.? Everywhere, gov- between those directly covered by SERPS provisions and the significant pro-
ernments moved to combine adequate minimum benefits with provisions that portion of participants who had opted to **contract out”? into a private scheme
partly reflected workers' pre-retirement earnings. Efforts to transform pensions (see later in this chapter). In short, whereas in the United States virtually eve-
from a safety net into a “*retirement wage”” required a move toward earnings- ryone perceived a stake in Social Security, the impact of pension reform would
related benefits. Significantly, the United States adopted such an approach from vary significantly in Britain, depending on a citizen”s age, income, and status
the outset, whereas Britain did not successfully incorporate earnings-related ben- in SERPS.
efits into its pension system until 1975. Furthermore, the existence of an extensive and universal pension system in
The Social Security Act of 1935 created a system of earnings-related pen- the United States created a highly visible set of government '*spoils”” that helped
yper mus mv unIzaLIvn UL a sirong pensioner lobby.” The American Association practice, however, the financial mechanisms in the two cases operated quite
of Retired Persons (AARP) is one of the most powerful interest groups in Wash- differently. In Britain, the existence of a Treasury contribution insured that the
ington. In the much more fragmented British policy environment nothing similar system would be in balance regardless of economic developments. In the United
emerged.'” States, however, funding was based completely on payroll taxes. 1f projected
payroll-tax revenues were insufficient to meet scheduled benefits, the Social
Security system was said to be out of balance, triggering calls for corrective
Maturity
measures to prevent ““insolvency.””
Although SERPS had been introduced only a few years before, by 1980 Social Conservatives (including, in this regard, Roosevelt himself at the time of the
Security was rapidly approaching **maturity.”* A mature system is one in which Social Security Act) had always maintained that a strict reliance on contributions
the retired population has paid contributions for an entire working life, and is would impose a discipline on the system that general-revenue funding would
consequently entitled to full benefits. The transition to maturity takes many years remove. When poor economic performance caused a trust-fund crisis in the mid-
— for SERPS, it was anticipated to occur some time after 2030. What makes 1970s, conservatives clung tenaciously to the principle of exclusive reliance on
maturation important is the **pay-as-you-go”” financing common to both the payroll taxes. Congress rejected a Carter administration proposal that would
British and American systems. Rather than building up savings as a private plan have solved the imbalance with an injection of general revenues, choosing to
would, the public sector pays benefits out of the current working population's rely instead on sizable increases in the payroll tax.'* Thus, in contrast to the
payroll taxes. Public pensions constitute an intergenerational contract in which British system, the American pension program continued to be vulnerable to
each generation relies on the one following to pay its benefits. The maturity of any projected shortfall in payroll-tax revenue.
the pay-as-you-go system in the United States means that extensive payroll taxes
are already committed to the current generation of retirees. Although this had
Indexation
become true of the smaller Basic Pension scheme in Britain by 1980, it was not
yet the case for SERPS. Public-pension schemes usually contain some mechanism of dynamization to
insure that benefit levels are adjusted to take into account the significant in-
creases in prices and earnings that occur both in the course of a worker's career
Treatment of private provision
and during retirement. Although these key economic variables change slowly,
In the United States, private pensions are designed to supplement the Social long-term shifts are likely to be dramatic: A pension system that made no ad-
Security system. When Social Security was enacted there was some discussion justments for inflation, for example, would in time play an increasingly marginal
of providing an **opt-out'” clause for those with private pensions. However, role in retirement provision. Both the British and American systems provided
because private schemes then offered only meager pensions to a tiny fraction for automatic annual adjustments in pension benefits. In Britain, the Basic Pen-
of the elderly, the bill's drafters rejected this proposal.'' In Britain, by the time sion was adjusted in line with increases in wages or prices, whichever was
an eamings-related provision finally reached the political agenda, private pen- higher, The provision for wage-based increases was designed to allow pension-
sions were a much more plausible alternative. The first Conservative plan of ers to share in economic growth, maintaining the relative positions of workers
1958 introduced the concept of contracting out, meaning that participation in and retirees. In the United States, Social Security provided for annual cost-of-
the public earnings-related plan would be compulsory only for those without living adjustments (COLASs), compensating retirees for inflation.
employer-sponsored private pensions. This arrangement was continued follow- However, Social Security also possessed a second important dimension of
ing the complex interparty negotiations over SERPS. The new legislation dynamization that the Basic Pension scheme in Britain lacked. Because changes
actually included generous public support for the private option. Thus, whereas in real earmnings over a worker's lifetime are likely to be substantial, some ad-
in the United States private pensions supplemented public earnings-related ben- Justment must be made in benefit calculations or a worker's early years of earn-
efits, the British arrangement has been aptly described as **a structure of sub- ings will dramatically lower his or her pension entitlements. Thus a retiree's
sidized competition.””'? initial Social Security benefits are based on lifetime earnings, adjusted for econ-
omy-wide earnings growth during his or her working years. The same is true
for SERPS. Unlike flat-rate schemes, earnings-related pension systems have a
Financing
very significant degree of benefit dynamization built into them.
Technically, public-pension systems in both countries possessed separate finan- There were, then, differences in the programmatic structures of retirement
cial accounts that were supposed to maintain something resembling balance. In provision in Britain and the United States. Although many of these may seem
 matters of technical detail, they proved highly consequential when conservative fruition in the Social Security Act of 1986, was an overhaul of earnings-related
critics of public social provision were elected. Ronald Reagan inherited a uni- pension provision.
versal, cohesive, and mature public-pension system, financed by employee and The Thatcher government's plan for limiting the Basic Pension provides an
employer contributions to an earmarked trust fund, and providing eamings- excellent example of what can be called '*implicit privatization.”” The govern-
related benefits that workers could supplement through private arrangements. ment sought to freeze the scope of public provision. In a growing economy, this
Margaret Thatcher confronted a system financed by a combination of payroll meant that expansion would be channeled into the private sector, leading to a
taxes and general government revenues, with much more limited, flat-rate ben- gradual but substantial shift in the balance between the two sectors. The (first)
efits. A new earnings-related scheme augmented this system but did not yet pay 1980 Social Security Act changed the basis for uprating (indexing) the Basic
significant pensions, and it competed directly with private alternatives. These Pension. From 1973 to 1980, benefits had been uprated in line with higher of
different program structures created distinctive opportunities and constraints for prices or earnings, which in practice generally meant eamings. Essentially, this
those seeking fundamental change in pension provision. upratings formula meant that pensioners would share in the benefits of economic
growth, and the relative role of state pensions would remain stable. The 1980
act, however, provided for upratings only in line with prices.
THATCHER, REAGAN, AND PENSION POLICY Though the act generated limited political controversy, it had substantial
long-term implications. Because earnings generally grow faster than prices,
Both Thatcher and Reagan were strongly committed to reducing government
uprating pensions in line with prices meant that economic growth would grad-
spending and transferring public-sector activity to the private sector wherever
ually diminish the relative role of state pensions. Even in the short run, the real
possible. Because pensions account for such a large share of social spending,
value of pensions has lagged far behind earnings. By 1988, the new upratings
they became an obvious target for politicians interested in scaling back the
formula meant that a married pensioner received £65.90 rather than £79.90 per
welfare state. Nevertheless, if the size of expenditures attracted budget cutters”
week, a reduction of almost 20 percent. The change lowered expenditures by 4
attention, it also both reflected and enhanced the programs” political strength.
billion pounds a year.'* If this continued, the long-term effect would be dramatic.
Pensions provided large benefits to a population widely regarded as deserving.
As the government actuary pointed out, in forty years the Basic Pension for a
Perceptions that benefits had been earned by a lifetime of contributions greatly
single pensioner would, assuming 2 percent real annual earnings growth, pro-
strengthened the general sense that pensions constituted an untouchable entitle-
duce a replacement rate (the pension benefit as a percentage of previous earn-
ment in both Britain and the United States.
ings) of roughly half that provided in 1980.'*
Behind the superficial similarity of these two middle-class entitlement pro-
This change substantially reduced expenditure growth at low political cost.
grams lay important differences, however. Pension policy in the United States
Much smaller policy shifts often generated massive public reactions, but for a
had developed through a long series of largely incremental and consensual steps,
number of reasons opposition in this case was muted. For opponents, one prob-
and events would demonstrate that Social Security was firmly embedded in
lem was the difficulty of communicating the argument that earnings-based in-
American political life. British policy had developed fitfully, and it was not until
dexation was the appropriate policy. As long as the Thatcher government could
the mid-1970s that Britain truly embarked toward a mature pension system. This
claim that it was maintaining the Basic Pension's real value (and it did so
was forty years after the United States, and it would prove to be too late.
repeatedly), it was difficult for program supporters to portray the new upratings
formula as a major assault on state provision. Nevertheless, as Thatcher herself
said in a 1986 interview, **If you actually hold [public expenditure] against a
The Thatcher government and state pension provision
background of growth you have got what you want.”*'* The decremental nature
British pension policy must have seemed a dubious target for reform. Because of the policy change further diminished opposition. There was no highly visible
existing benefits were ungenerous by international standards, cutbacks were hard restriction of benefits. Instead, a single reform produced a stream of automatic
to justify. Although the new SERPS scheme did promise eventual expansion, it annual adjustments in benefits. Each adjustment was relatively small, but the
was the product of an agreement for which the Conservatives had repeatedly cumulative impact was substantial. Finally, the weakness of the pensioner lobby
expressed support. Nonetheless, a government committed to reducing state in Britain undermined prospects for forceful opposition. Given the complexity
spending was bound to look for cuts. Despite emerging rather slowly, Conser- and decremental nature of the reform, a sustained effort to explain its impact
vative efforts to change pension provision eventually produced major reforms. and mobilize those who stood to lose was required. The lack of anything equiv-
Thatcher's government followed two strategies. Initially, it moved to stop the alent to the AARP meant that in the absence of highly visible cutbacks, a mo-
growth of the basic state pension. The second track of policy, which reached bilization of the elderly was unlikely.
 If the new upratings formula was a major political victory, it nevertheless quired to make minimum mandatory contributions to either an employer's oc-
provided an incomplete solution. The government realized that SERPS promised cupational scheme or to an employee's **personal pension.””
to substantially increase public spending. Even so, a direct attack on SERPS These changes met the government's major ambitions for pension policy.
was politically risky. The program had all the features that were supposed to First, SERPS would be eliminated. Admittedly, the process would be slow, given
make it untouchable. Benefits were universal, the link between contributions and that the desire to limit political opposition required a long transition. Reductions
entitlements was relatively tight, and the recipients commanded widespread pub- in anticipated SERPS expenditures were not expected until 2002. Nevertheless,
lic support. Furthermore, changes in policy would have required an embarrassing by 2033 the plan would save an estimated 75 percent of the enormous projected
public reversal. The Conservatives had publicly backed the enactment of SERPS expenditures.** Second, and equally important, the introduction of per-
SERPS, and the Thatcher government had repeatedly promised to maintain it. sonal pensions was intended to fill many of the gaps in private provision that
During the prelude to the 1983 general election, the prime minister denied hav- previously had led to demands for state pensions.
ing any “plans to change the earnings-related component of the state pension.””'” If the government was pleased with its proposal, however, almost nobody .
If so, then plans must have developed very quickly after the election. else was. Despite a commanding majority in Parliament that all but guaranteed
The control of public expenditure was a top priority. Social-security outlays victory, the government was unprepared for the magnitude of public hostility
were rising, largely because of Britain”s staggering unemployment rate. Nev- that the green paper produced. Opposition from the traditional supporters of
ertheless, the government — and especially the prime minister and the treasury social programs was expected. The bold decision to completely abolish SERPS
— considered retrenchment within the heartland of the welfare state essential to gave the Labour Party a clear target for criticism. The Trades Union Congress
economic recovery. The new chancellor, Nigel Lawson, released a major report called the proposals a **colossal breach of faith on the Government's part.*”?*
in early 1984 expressing the need for expenditure cuts. The treasury analysis The “poverty lobby'” worried about the position of women and the low paid,
specifically identified pensions as *“the major source of future pressures”” within arguing that private alternatives would offer far less for them than SERPS., An
the social-security budget. The prime minister, in an interview with the New analysis by the highly respected Institute for Fiscal Studies (IFS) confirmed this
York Times in January of the same year, spoke of the need to confront a “*social- view. Although a single man of forty could expect transitional SERPS payments
security time bomb.””'* and his private contributions to produce a pension equal to roughly 80 percent
These comments placed a heavy burden on Norman Fowler, secretary of state of that SERPS would have provided, a pension including benefits for a spouse
for social services. Press reports suggested that Fowler had resisted treasury would be reduced to only 70 percent. The IFS added that because private plans
demands for deep cuts in 1983 by agreeing to a review of the entire social- would not duplicate SERPS's generous treatment of shorter work histories,
security budget.'” Fowler asserted that the reviews would **constitute the most working women could expect to do significantly worse.**
substantial examination of the social security system since the Beveridge Report More surprising was the: harsh criticism from usually reliable government
forty years ago.***” The prominence of pension expenditures ensured that SERPS supporters: employers and the occupational pension funds. Employers were
would receive special attention. alarmed by the cost. In effect, the green paper called for transforming a public
The government accepted submissions from interested parties until the end pay-as-you-go scheme into a private funded system. Where SERPS had antici-
of July 1984. A special cabinet subcommittee chaired by the prime minister pated that future benefits would be paid from future contributions, the green
herself reviewed Fowler's preliminary proposals in early 1985.*' According to paper called for the gradual accumulation of assets to pay retirement benefits.
a Financial Times report, the treasury pushed hard for cuts at these meetings, This produced a '*double-payment'* problem: Employees (and employers) were
seeking reductions on the order of £2-4 billion, or between 5 percent and 10 asked to continue making National Insurance contributions to pay for current
percent of social-security spending.” retirees while making mandatory contributions to private schemes to fund their
Expectations that Fowler might offer a “new Beveridge”” proved false, but own retirement.
the green paper released in June 1985 contained one undeniable bombshell: the In the green paper the government acknowledged the need for higher con-
proposal to abolish SERPS.* In a bow to political realities, the government tribution rates: **The move to additional funded pension provision will be taking
called for a lengthy transition period. Nevertheless, the plan was radical. Al- place while the cost of the pay-as-you-go state system continues unchanged. The
though those within fifteen years of retirement would stay in SERPS, men under total volume of resources being devoted to pensions will, therefore, increase.'*?”
forty and women under thirty-five would lose all SERPS benefits. Those in For both employees and employers, the reform meant higher payments for lower
between would receive partial SERPS benefits because they would have limited benefits. The Financial Times projected the increased annual cost to employers
time to make new arrangements. In place of SERPS, employees would be re- at roughly £1.5-2 billion.”
 Not only did the plan require employers and employees to make higher con- The white paper reaffirmed the earlier emphasis on expanding private pro-
tributions for lower benefits, but in the short run it promised to worsen the vision through a combination of occupational schemes and personal pensions.
government's fiscal position. The state (as an employer) would pay increased Occupational schemes were given increased flexibility. In addition, employees
contributions for its employees, while the expansion of tax-free private-pension could opt out of occupational plans, setting up personal pensions instead. In this
provision would lower government revenues. Concern for these fiscal implica- case, employers would be required to make contributions to the personal pension
tions had delayed the release of the green paper, as the treasury argued with in lieu of contributions to the occupational scheme. As an incentive to contract
Fowler over the budgetary costs of his plans. The social-security review was out, the government said it would add an additional rebate of 2 percent of
intended to lower spending, allowing Lawson to offer preelection tax cuts. In- eamings to the standard rebate for newly contracted-out employees until 1993.
stead, the chancellor faced the prospect of higher tax subsidies, which meant These major reforms were enacted in 1986.
lower revenues and less room to cut income taxes.?” Although other aspects of the government's social-security legislation met
The insurance companies who ran the occupational schemes saw their **ex- fierce parliamentary resistance, the pension reforms sailed through with limited
panded opportunities”* as a decidedly mixed blessing. The erratic earnings pat- dissent. The 1986 reform was generally regarded as a significant retreat for the
terns of many of the private sector's new clients made them expensive to include Thatcher government and further proof of the privileged status of middle-class
in plans. Private schemes would also lose some of the generous state subsidi- social benefits.** After all, the government had failed to eliminate SERPS. How-
zation for contracted-out workers that SERPS had provided. Furthermore, in- ever, this interpretation misses the significance of the government's revised pro-
surers feared that if the 4 percent mandatory contributions produced meager posals. In combination, thé reforms of the Basic Pension and SERPS represent
pensions, they might face pressure to provide more generous benefits. a dramatic change in pension policy. The new arrangements will have significant
Employers and insurers also complained about the administrative headaches repercussions for income distribution, the roles of the state and private sector
the scheme would create, in that for at least fifteen years companies would have in pension provision, and the evolution of state finances.
to run two schemes in tandem. Finally, pension providers worried that pensions The new scheme's distributional impact will depend on how adequately pri-
would once again be a '*“political football.”* Even before the green paper was vate alternatives substitute for SERPS benefits. The new arrangements will work
released, Labour Party leader Neil Kinnock announced that a Labour govern- against those with lower earnings and less stable work histories, since the new
ment would reverse any major modifications of SERPS.*” What was the advan- formulas will not provide the relatively generous compensation for these factors
tage of introducing complex new arrangements if the next government would that SERPS offered. Because women are likely to suffer from both liabilities,
render them obsolete? they are especially apt to be worse off. The new system will be particularly
Fowler's **fundamental examination”? of pensions had apparently missed damaging to middle-aged workers, who will have limited time to build up new
some important political ramifications of reform. The government was prepared entitlements. The government's own statistics suggested that benefits for those
for complaints from the opposition parties, but probably not for their vehemence. staying in SERPS would drop sharply — as much as 40 percent for widows in
Equally important, the withering criticism from the Confederation of British some age and income groups.?**
Industry and from major private insurers caught the government unprepared. The changes will sharply curtail the state's role in providing earnings-related
Members of the government's own pensions-review team, including Stewart pensions. Because public pensions have been cut, efforts to achieve higher re-
Lyon of Legal and General Assurance (Britain's largest pension fund), were tirement incomes will be increasingly channeled toward the private sector.
highly critical of the proposals. Lyon argued that the review team had never Although the public-sector earnings-related system will continue, the new rules
considered a plan to phase out SERPS.”' are structured to make the private-sector option attractive to anyone with suf-
By the fall of 1985, a rattled Thatcher government was searching for a grace- ficient time to make significant contributions. The extra 2 percent annual re-
ful way out. Faced with heavy opposition from interest groups and rising dis- bates, widely derided as a ““bribe,”” will further enhance the private sector's
comfort among backbench Conservative MPs, Fowler began to backpedal. When appeal.** By the end of 1990 — only eighteen months after personal pensions
the government's legislative proposals finally emerged in December, the pension became available — 4 million people (about 15 percent of the working popu-
provisions had been substantially revised.*” The white-paper proposals, which lation) had already contracted out and set them up. That this promises fun-
became the Social Security Act of 1986, reversed the decision to abolish SERPS. damental long-term change is signaled by the fact that three-quarters of those
Instead, the government substantially lowered benefits while continuing to en- choosing the personal-pension option were in their mid-thirties or younger.”
courage private alternatives.** As with the green-paper proposals, benefits were In the long run, SERPS expenditures will drop sharply. The changes intro-
to be phased in gradually, with the transition affecting those retiring between duced will make little difference in the short run because of transition arrange-
the years 2000 and 2010. ments. However, by 2021, SERPS expenditures are projected to drop by well
 over 50 percent compared with pre-reform estimates, from £16.4 billion to £7.1 In 1981, the simultaneous arrival of a conservative president and a financial
billion.** crisis made Social Security appear increasingly vulnerable. Despite assurances
Given the repeated shifts in British pension policy, the crucial question is that reforms adopted in 1977 had left Social Security financially sound, new
whether these reforms can be sustained. For example, a future government projections indicated looming deficits. Stagflation was the major source of dif-
would not be obligated to limit itself to adjusting Basic Pension benefits for ficulty. Although the Social Security Administration had predicted 28 percent
inflation. By adopting price indexation, however, the Thatcher government has inflation and 13 percent real-wage growth from 1978 to 1982, the actual expe-
assured that the relative position of the Basic Pension will fall in the short run. rience was 60 percent inflation and 7 percent negative real-wage growth." A
Future governments are likely to be condemned to a game of perpetual catch- new rescue package was needed.
up. During the spring of 1981 there were several small skirmishes over Social
In theory, the government's changes in eamings-related provision could also Security, but the Reagan administration's efforts centered on cutting taxes and
be reversed, but the Conservatives” 1987 electoral victory sharply increased their other, politically weaker areas of domestic spending. Growing concern about
chances of institutionalizing a new pension regime. Now that a large number of Social Security finances eventually provoked a presidential response. In the past,
people have made sizable contributions to private pensions, the '*privatization Reagan had expressed interest in moving Social Security to a voluntary basis
constituency”” will be politically powerful. Fowler certainly recognized the im- while expanding private alternatives.*' With other items high on the administra-
portance of quickly cementing his new approach, having offered substantial tion's agenda, however, no long-term reform strategy developed. Instead, Social
financial incentives (which have no clear policy justification) to those opting for Security policy was subsumed in the president's confrontation with Congress
personal pensions by 1993. over the budget.
There is some reason, then, to accept the Economist's enthusiastic claim that Social Security's widely publicized troubles seemed to offer the Reagan ad-
the shift to personal pensions **could revolutionise personal finance. It could be ministration a good opportunity to restructure the program.* However, Office
as important and irreversible as this government's other two privatising inno- of Management and Budget director David Stockman, who took the lead in
vations, the sale of council houses to their tenants and of nationalised industries formulating administration policy, was preoccupied with obtaining immediate
to individual investors.***” In the wake of four consecutive election defeats, budget cuts. One White House aide reported that Stockman hoped for *'phe-
pressures on the Labour Party to come to terms with Britain's ''new individu- nomenal'' cuts and saw Social Security as '“the best way to get a balanced
alism'' continued to mount. Efforts to reverse the expansion of personal pen- budget.””** The OMB director also believed the publicity surrounding Social
sions are unlikely. The Labour Party has come to recognize that its electoral Security's projected deficits provided a unique opening. The perception of crisis,
viability depends on maintaining support from precisely those middle-income Stockman thought, might **permit the politicians to make it look like they're
groups likely to be alienated by promises to restrict investments in personal doing something for the beneficiary population when they're doing something
pensions. Despite significant setbacks, the Thatcher government succeeded in toi.
radically reforming British pension policy. Later events would prove Stockman right: Trust-fund crises did create op-
portunities for retrenchment. However, Stockman feared that preoccupation with
Pension policy in the Reagan years the trust fund would lead to long-term changes that were of little use in his
battle to balance the overall federal budget. As Stockman told the Washington
The Reagan administration's hostility toward Social Security and the system's Post's William Greider, he had no interest in spending **a lot of political capital
financial difficulties fueled a number of heated political conflicts over pensions solving some other guy's problem in 2010. The Social Security problem is not
policy in the 1980s. Two results of these struggles are striking. First, the ad- simply one of satisfying actuaries . .. lts one of satisfying the here-and-now of
ministration's overall ability to achieve pension retrenchment was limited, es- budget requirements.””** By focusing on immediate cuts to reduce the budget
pecially when compared with Thatcher's success. Second, the results of the deficit, however, Stockman redefined the issue and precipitated a major political
individual initiatives varied widely. When struggles over Social Security high- blunder.
lighted the administration”s desire for cutbacks and involved an effort to enlist OMB designed a proposal that called for approximately $45 billion in Social
Social Security in the battle against the budget deficit, they were inevitably Security cuts. The centerpiece was a 25 percent reduction in benefits for those
unsuccessful and politically costly. On the other hand, where Social Security's choosing early retirement, scheduled to take effect January 1, 1982. The pro-
financial imbalances became the issue, there were opportunities for limited struc- vision was designed to save $20 billion, but because early retirement had be-
tural reform. come a frequently used option, it would have sharply reduced benefits for 1.4
a 94 Dunon cut in basic benefits and a three-month cost-of-living allowance benefit reforms. The biggest short-term source of revenue came from the accel-
(COLA) delay.** eration of payroll-tax increases. Because these increases were already scheduled
Reagan's top advisors hoped to keep the president one step removed from to occur in a few years, they did little to reduce the long-term deficit. More
the process, insulating him from potential criticism. At Stockman's behest and significant in the long run was the increase in payroll taxes for the self-
with Reagan's approval, Health and Human Services secretary Richard Schwei- employed. The 1983 reforms also continued a well-established practice of rais-
Ker reluctantly introduced the proposals as his own in May 1981.* The admin- ing revenues by expanding coverage to previously uncovered groups — a rather
istration"s recent string of dramatic legislative victories may have made painless approach since the newly covered workers would also be entitled to
Reagan's aides overconfident. Schweiker's initiative was quickly linked to Rea- benefits. All new employees of the federal government and nonprofit institutions
gan, and the Democrats promptly responded. Amid widespread outcry against would now be included in Social Security.
the proposals, the Senate voted 96-0 to oppose any **unfair'? or **precipitous”” The 1983 amendments included substantial modifications of benefits. First, a
cuts in Social Security. The Reagan '*honeymoon'” was over. six-month COLA delay contributed significantly to both short- and long-term
The cuts eventually enacted in 1981 were far more modest, less controversial, deficit reduction. This proposal had no effect on nominal benefits; it would take
and clearly inadequate to solve the Social Security system's financial problems. place only incrementally and almost invisibly. A second important change made
However, neither the president nor congressional Democrats wished to take the 50 percent of benefits subject to income tax if adjusted gross income. was
lead in proposing painful solutions. President Reagan's advisors had convinced $25,000 or more for single returns and $32,000 for joint returns. Because these
him to propose a COLA delay in a new round of budget cuts announced in income levels were not indexed, the change would generate large incremental
September, but this time the White House cautiously floated the proposed cuts savings. As inflation increased nominal incomes, more and more pensioners
in advance. When Republicans in Congress were unenthusiastic, the administra- would find their benefits subject to tax.*' Unable to reach an agreement on how
tion dropped the idea.** to eliminate the remaining shortfall, the commission left to Congress the choice
Reagan's new reluctance to tamper with Social Security reflected the Dem- between more reliance on taxes or an increase in the retirement age. Congress
ocrats” success at portraying the administration as a threat to the elderly. As the voted to raise the retirement age to sixty-seven in a series of small steps begin-
trust funds dwindled, the Reagan administration feared that a new initiative ning in the year 2000. Since most people already retired before sixty-five, few
would give the Democrats another opportunity to raise the '*faimess”” issue. expected the result to be later retirement. Instead, as the Democrats” represen-
Seeking a way out, Reagan announced the appointment of a commission, rep- tatives on the commission pointed out, retirees would receive significantly lower
resenting a range of interested parties, to develop a plan for eliminating the early-retirement benefits because of the reform.
projected trust-fund deficits. Conveniently, the commission would not report Although the 1983 Social Security amendments were widely seen as a com-
until after the November 1982 midterm elections. promise based largely on administration concessions, the result was a sizable
Such **blue ribbon”” commissions can serve an array of purposes, but they cut in long-term benefits. The bulk of the reduction in long-range imbalances
very rarely produce policies. Even though the administration may have seen the came from benefit decreases rather than tax increases.*? In short, the 1983 act
commission as a useful delaying tactic, House Speaker Tip O”Neill and Senate revealed that American political institutions, even at a time of divided govern-
Majority Leader Howard Baker each appointed members whose support would ment, did not prohibit significant retrenchment. A recent analysis indicated the
be essential to any successful bargain.*” To the surprise of many, key commis- magnitude of the reductions involved. Between 1985 and 2030, replacement
sion members eventually worked out a satisfactory compromise with Reagan's rates (benefits as a percentage of pre-retirement earnings) for a sixty-five-year-
aides.*” Unlike every other confrontation over Social Security in the 1980s, this old retiree are projected to fall from 63.8 percent to 51 percent for low earners
struggle actually produced significant program cuts. and from 40.9 percent to 35.8 percent for average earners.**
Social Security”s financial structure played the central role in producing this Nevertheless, the reform did not alter the fundamental structure of the Amer-
outcome. Ás conservatives had long maintained, the existence of a separate ican pension system, and suggested that the Reagan administration had given
"trust fund”* that could not be allowed to go **bankrupt”” structured the political up any radical plans for Social Security. Events after the 1984 election confirmed
debate. Faced with the specter of imminent insolvency — probably the only the difficulty of imposing Social Security cuts in the absence of a trust-fund
outcome that would cause more political damage than benefit cuts — a wide crisis. Conservative desires to limit government spending combined with con-
range of political actors were forced to embrace a compromise solution. tinued concern over the deficit to keep expenditure cuts on the political agenda.
The commission's diverse composition and Social Security”s popularity in- In turn, the recognition that major budget reform was unlikely without a con-
sured that the proposals did not call for a radical restructuring of the program. tribution from Social Security led politicians to return repeatedly to the topic of
 pension cutbacks. In each case, however, Social Security's reputation as the markets were steadying diminished the sense of crisis. In this context, House
untouchable ““third rail”? of American politics was confirmed. Republicans signaled their unwillingness to go along with Social Security cuts.
Rather than building on the outcome of 1983, struggles over Social Security Although Senate Republicans continued to push for sweeping reforms, such
since 1984 have echoed Stockman's debacle of 1981. Social Security's high open opposition within the GOP ruled out any major cutbacks. With Social
profile and massive scope have attracted the attention of budget cutters. These Security excluded, agreement on broad contributions from tax increases, defense
same characteristics, however, meant that serious proposals immediately gen- spending, and other domestic programs also proved elusive. The highly touted
erated a chorus of dissent, forcing policymakers to retreat. The idea of cutting budget summit ultimately settled for a minimal, face-saving package of new
Social Security to lower the federal deficit reappeared in 1985, although this revenues and spending cuts, with Social Security left untouched. :
time the initiative came from Senate Republicans rather than the White House. Social Security”s recent history is easily summarized. Extensive conservative
Having been burned on Social Security before, President Reagan was reluctant efforts to erode public-pension provision have resulted in marginal change. Rea-
to pursue the matter. However, leading Republican senators such as Robert Dole gan administration attempts to go on the offensive against this extensive system
and Pete Domenici worried about both the economic and political consequences of retirement provision collapsed in the face of massive and unified resistance.
of continued high deficits (twenty-two Republican senators were up for reelec- Increasingly, the administration found itself on the defensive, forced to respond
tion in 1986). to a policy agenda not of its own choosing. Although some cutbacks resulted |
Through a long and tortuous process of negotiation, Dole was able to steer from trust-fund pressures, the Social Security program survived — indeed almost
a package through the Senate that included a one-year COLA freeze.** Hopes flourished — through a decade of budgetary austerity.
that the initiative would produce a bandwagon effect rapidly dissipated. Reagan
offered only lukewarm support, and Democrats refused to come on board. The
POLICY FEEDBACK AND PENSION REFORM:
proposed budget passed the Senate 50-49 on an almost straight party-line vote.
EXPLAINING OUTCOMES
Nor was the Democratically controlled House more obliging. Despite moments
when a complex package that would have included defense and Social Security Despite one limited success, the Reagan administration's effort to restructure
cuts and tax increases seemed possible, Reagan ultimately deserted the Senate state pensions was largely thwarted. The Thatcher government, by contrast,
Republican leadership. The president cut a deal with Tip O'Neill preserving overcame setbacks to introduce fundamental policy reforms. What accounts for
Social Security in return for higher defense authorizations. The embittered sen- the difference? The evidence supports the view that the structure of preexisting
ators had rediscovered the dangers of tampering with such a popular program. pension systems was crucial. Before reviewing the case for this proposition, it
Though the Reagan administration surely hoped that by now it had heard the is worth considering some alternative hypotheses.
last of proposals for Social Security cuts, dramatic events produced a renewed One possible explanation would be the Thatcher government's greater polit-
struggle only two years later. On October 19, 1987, world financial markets ical savvy in executing a politically treacherous undertaking. There is little doubt
collapsed, with the Dow Jones losing almost one quarter of its paper value in a that Stockman's poorly planned initiative in 1981 put the president on the de-
single day. With analysts arguing that the budget deficit was partly to blame, fensive, aroused opponents, and dampened any enthusiasm Reagan and his ad-
political pressures for a govermmental response mounted rapidly. Reluctantly visers might have had for radical reform. Nevertheless, the Thatcher government
accepting the need for a **budget summit,'” Reagan excluded only one possible made mistakes as well — witness the embarrassment when usually reliable allies
target for deficit reduction: Social Security. In a press conference on October roundly condemned the green-paper proposals. It was not competence that dif-
20, he announced that he was **putting everything on the table with the excep- ferentiated the two reform efforts.
tion of Social Security, with no other preconditions.””** The carefully worded Nor is a focus on the two countries” political cultures of much help. Although
statement, designed to signal the president”s willingness to consider tax in- some scholars have identified elements of public opinion as a source of cross-
creases, provided clear testimony to Social Security's special status. national differences in policy, political culture is generally used to account for
The mathematics of deficit reduction and the absence of viable alternatives less activist government in the United States.** In this case, the United States
kept pushing Social Security back into the negotiations, however. Given its has developed and sustained a more extensive public sector — exactly the op-
antipathy to social spending, the Reagan administration probably was not averse posite of what an explanation based on political culture would anticipate..
to this development — providing that all key participants were willing to share A third approach might emphasize the role of the power resources available
the blame. Early in the second week of discussions, word leaked that COLA to labor movements and their political allies in the two countries. Nevertheless,
limitations were being discussed. Even as the powerful lobbying apparatus of although the position of the left was weak in both cases during the 1980s, the
Social Security advocates moved into action, a growing sense that the financial political outcomes were dramatically different. If anything, a focus on power
 resources would suggest that reform should have been easier in the United that the British government eventually found it possible to minimize the political
States, where levels of unionization are far lower and there is no true social- costs associated with reform. Ultimately, the explanation for this divergence lies
democratic party. Whatever the role of labor-movement strength may have been in the policies each government inherited. Five differences between the two
during the era of welfare state expansion, it cannot account for the different pension systems were especially important. First, policy in the United States
outcomes of welfare state retrenchment efforts in these two cases. was built around one dominant, unifying program, whereas the fragmented Brit-
Finally and most plausibly, one could stress important differences between ish system consisted of a number of overlapping components. The greater co-
the formal institutions of the two political systems. This explanation is at least herence of Social Security in the United States created broad common interests
broadly compatible with the outcomes examined here. It is indeed far easier to among recipients, whereas distinct groups of British pensioners relied to differ-
legislate reform in Britain's parliamentary system, where the prime minister can ent extents on various forms of provision. Not surprisingly, the unifying role of
usually count on solid majority support, than in the political obstacle course Social Security made it much easier to mobilize the American elderly in op-
confronted by an American president.” Once the Social Security Act was pre- position to retrenchment proposals.
sented to Parliament in January 1986, its eventual adoption was a virtual cer- Second, the preexistence of contracting-out arrangements in the British sys-
tainty. tem facilitated the privatization process by masking the radical nature of the
Political institutions clearly did play an important part in the development of government's proposals. In the United States, allowing people to opt for private
pension systems in both countries. An institutional structure that makes conti- provision in return for reduced state entitlements would have constituted radical
nuity an easier course than change facilitated the long, incremental, and rela- reform. By contrast, in Britain dramatic change could be implemented quietly.
tively consensual expansion of Social Security. Britain's failure to follow the After failing to openly introduce an entirely new system, the Thatcher govern-
postwar international trend toward earnings-related provision was in large part ment adopted a seemingly incremental, modest revision. However, even though
attributable to highly centralized political institutions that translated frequent the basic alternatives were left in place, the attractiveness of each option was
alternations of Conservative and Labour governments into equally frequent pol- markedly changed. By recasting the rules of competition between the public and
icy reversals. private sectors, the government initiated a process that will transform the relative
Furthermore, the reality of divided government certainly presented significant roles of both in the long run.
problems for the Reagan administration. However, there are serious problems The third important difference was that Social Security”s eamings-related
with an explanation of retrenchment outcomes based primarily on the structure benefit structure offered greater protection against changes in indexation that
of formal political institutions. As I argued in Chapter 2, the theoretical case for would produce **implicit privatization.”” As noted earlier, earnings-related pen-
expecting centralized systems to be more successful is suspect. For a government sion systems must compensate for earnings growth over time, or their benefit
seeking to implement unpopular policies, the greater centralization of British structures become distorted. This characteristic builds an important element of
political institutions was a two-edged sword. Centralized government concen- benefit dynamization directly into the structure of a pension system, providing
trates power, but it concentrates accountability as well. The Thatcher govern-
ó 7 A : z a crucial protection against the gradual erosion of benefit levels. It is a protection
ment's greater institutional control must be weighed against the greater that SERPS shares, but the Basic Pension does not. The relative role of the latter
likelihood that it would be blamed for unwanted reforms. program in retirement provision has dwindled as a result.
If there are good theoretical reasons for questioning an institutional expla- The fourth major difference between the two pension systems lies in their
nation, the evidence suggests that political institutions provided onlya modest financing structures. Social Security politics were trust-fund-driven in the 1980s
advantage to the Thatcher government. Institutional design cannot account for because the program's design directs attention to the balance between payroll
the Reagan administration's single success or the Thatcher government's sig- taxes and program'outlays. In the early 1980s, trust-fund deficits created an
nificant failures. Despite the reality of divided government, under certain con- opportunity for significant retrenchment. Thus, programmatic structure also ac-
ditions the Reagan administration achieved significant cuts in public pensions. counts for the one case where pension cutbacks did occur in the United States.
Despite its institutional advantages, the British government could not act with However, the logic of the trust-fund issue led not to radical reforms but rather
impunity: Again, the green paper's demise suggests that Thatcher was hardly to efforts to “balance” the existing system's finances. In Britain, by contrast,
insulated from public pressure. Initial proposals had to be recast in a manner to infusions of general revenues have prevented trust-fund balances from becoming
make them acceptable, if not to the Labour Party, then at least to a wide range an important component of pension politics. The Thatcher government therefore
of interests outside the government. Failure.to do so has necessitated retreat, not had greater freedom to shape the political agenda.
only in the case of pensions but in other policy arenas such as health care. A final and crucial difference was that the public earnings-related pension
The difference between the situations in Britain and the United States was system in the United States was **mature,”” whereas the British one was not.
 > mps vu a pay-as-yUu-go Dasis it is very hard of program supporters to activate their political resources when existing pro-
to introduce fundamental alterations. Because private plans must pay bene- grams were threatened. E e
fits from earnings rather than from taxes, privatization means introducing a Second, in the United States, program structures had a decisive impact on
'"funded”” system. As mentioned, this presents a double-payment problem: Tax- the political agenda for Social Security. Trust-fund crises, which brought atten-
payers must continue to finance current retirees while saving for their own re- tion to the issue, were generated by the specific financing provisions of existing
tirement. policy. The status of the trust fund not only determined when reform was dis-
The double-payment problem is likely to create insurmountable political dif- cussed, but channeled reform efforts in particular directions by determining the
ficulties. The Thatcher government was well aware that the immaturity of definition of a solution.
SERPS provided a brief window of opportunity for privatization. As then-junior Finally, policy structures created lock-in effects: elaborate social and eco-
The
minister John Major noted in the House of Commons debate over pension re- nomic networks that rendered once-possible alternatives no longer feasible.
form, ““The way in which SERPS works means that every year of delay leaves lock-in effects in American pension policy were very powerful, as the Reagan
people clocking up expensive rights which must be honoured in the future.” administration discovered when confronting dense networks of commitments
In the United States, the double-payment problem was not a future prospect but produced by previous policy choices. Social Security”s complex financial ar-
an immediate reality. The financial resources needed to build a private-sector rangements involved a series of implicit promises stretching decades into the
alternative were already committed, through payroll taxes, to the current gen- future. Furthermore, these promises had shaped the retirement strategies of most
eration of elderly. Those advocating steps equivalent to Britain"s SERPS reform, Americans. With huge public expenditures essentially precommitted, the Reagan
like the Cato Institute”s Peter Ferrara, remained politically marginal figures.*” administration found its options extremely limited.
Thus, differences between the pension systems inherited from the past largely The multiple effects of policy feedback on the prospects for pension retrench-
determined the fate of efforts to formulate new policies. Although pensions ment seem clear. Reagan and Thatcher's efforts to reshape pension systems were
constituted **middle-class entitlements”” in both countries, the programmatic strongly constrained by the structure of programs already in place. Originally
structures in the two cases were quite different. These structures, rather than produced by social pressures, extensive public-pension systems now influence
differences in political culture, the power resources of the left, or the nature of political processes in fundamental ways.
formal institutions, account for the Thatcher government's far greater success.
Past research has often emphasized the influence of policies on administrative
capacities and on the way in which reformers — especially bureaucrats — develop
new policies in response to the perceived successes and failures of old ones.
However, neither of these two feedback processes appear to have been important
in this case. By the standards of contemporary governance, the administrative
demands of running a pension system are not particularly onerous. In neither
country is there any evidence that significant alternatives were ruled out because
of the absence of adequate bureaucratic capacities. Nor can a political-learning
argument explain the patterns of policy development in the two countries. Re-
form initiatives were launched by elected officials, not bureaucrats, and were
dictated by Reagan and Thatcher's general hostility to public social provision
rather than by any lessons learned from previous pension policies. Results dif-
fered not because policymakers learned different things from existing policies,
but because those policies influenced the political costs associated with reform.
The current analysis stresses three other consequences of policy structures.
First, the structures of existing policies influenced the prospects for actually
mobilizing potential political resources. Existing policy structures provided op-
portunities for the Thatcher government to lower the visibility of important
initiatives: The change in indexation methods that will gradually diminish the
role of the Basic Pension and the new SERPS rules, and which have fueled a
shift to private alternatives. The structure of British pensions limited the ability