Policymakers should use the “fiscal cliff” and discussions over deficit reduction to fundamentally rethink the nature of U.S. entitlements. Current programs offer unsustainable benefit schedules not because old-age social insurance is not affordable per se, but because the type of social insurance provided in the U.S. involves graduated benefit structures designed for a different demographic environment. Huge savings could be gained by focusing social insurance programs on the core function of guarding against poverty and a lack of access to medical care in old age.
Social insurance systems generally take one of two forms: (1) a “Beveridge” floor, named after Sir William Beveridge, where the social insurance guards against poverty through a fixed benefit floor financed by general tax revenue; and (2) a “Bismarck” system (named after the German Chancellor) of graduated benefit levels where retirement benefits are a function of the amount of payroll taxes, or social insurance premiums, a worker made during his or her career.
The differences in financing lead directly to differences in benefit schedules. In the Beveridge model, the trade is general revenues – presumably progressive income taxation – for a basic benefit floor that guards against old-age poverty but requires households to save for themselves if they are to replace a significant percentage of their pre-retirement income. There is no connection between the taxes (or the households that pay the taxes) and the benefits (or the beneficiaries). The transfer payments are explicit, as the taxpaying household has no “right” to future benefits as a consequence of its tax payment, but is simply providing for the common good as is done with respect to general revenue-financed expenditures for the national defense.
Conversely, in the Bismarck model, benefits are roughly proportional to contributions to the system. In this model, the worker is essentially pre-funding his or her own retirement though “insurance premiums” that typically take the form of payroll taxes. The important point is that these contributions or premiums entitle the worker to some retirement income (or retiree health care), which means they are not taxes per se. Although the actual cash collected from young current workers may technically be used to finance a current retiree’s public pension, these transfers are implicit; the contribution’s “real” purpose is to establish the current worker’s claim to future benefits. (It is not difficult to see how this sort of logic can immediately lend itself to double-counting). These differences in structure lead to obvious differences in the political appeal of programs. If a new entitlement aims to provide a basic level of service or income in exchange for taxes on “the rich,” for example, it is likely to be a Beveridge system that’s a pure safety net rather than a wage replacement scheme. Since taxes are assigned explicitly to fund benefits, there is political pressure to keep the payments to a level slightly above the poverty rate and nothing more.
Conversely, if the program is “self-financed,” the size of the tax assessments has less political salience since they’re not really taxes, but social insurance premiums. The insurance may or may not be a good deal, but the taxes-as-premiums represent just one side of the equation. Benefits must also be considered. Casting taxes as social insurance contributions makes it more difficult politically to cut benefits because the worker feels that he or she has “paid in” to the system, whatever the actual arithmetic linking contributions and benefits.
In the U.S, Social Security is clearly a Bismarckian program designed to attract broad appeal as a pension program. It was said at the time (1939) and since that “a program for the poor is a poor program.” This means that if Social Security was designed to provide an income floor for retirees financed through general revenues, the payments would be regarded as “welfare” and political support for the program would wither.
Social Security provides a higher rate of return on the contributions of lower wage workers, but higher wages translate to higher promised benefits up to the annual contribution wage cap ($110,100 in 2012). The internal rate of return on the contributions of low-wage workers born in 1973 is 400 basis points per year greater than the return of the worker earning the annual contribution cap, but both receive something for their payroll taxes.
The progressive benefit formula and positive real return on payroll taxes among lower income workers makes talk of the “regressive” nature of payroll taxes a non sequitur. Payroll taxes were known as social insurance contributions as long as the Bismarckian consensus held. The Congressional Budget Office (CBO) previously described payroll taxes as “social insurance contributions” and still uses the phrase “social insurance” to describe payroll tax revenue in its monthly budget review (MBR). It dropped the “contributions” sometime around the turn of the century, although “social insurance contributions” was still used in 1998. Today, progressives focus relentlessly on the burden of the payroll tax, including the statistic that “82% of ‘working households’ pay more in payroll taxes than in federal income taxes.” But given the positive rates of return on these contributions, these households would be worse off in the absence of the Social Security payroll taxes.
Of course, much as progressives lament the payroll tax’s burden, they have no interest in reducing Social Security benefits. What progressives seek instead is a Beveridge funding structure for a Bismarckian benefit schedule. But history suggests that as soon as the link between benefits and contributions is severed, so too is the political support for graduated benefits. As Charles Blahous has explained, proposals to “leave social security alone” by closing its funding gap through general revenue transfers involve gambling the program’s political support in ways that are probably not well appreciated by the proponents of these policies.
In the current environment, conservatives would be best served by proposing a cap on the maximum Social Security benefit. The benefit received by a worker with the maximum lifetime earnings is currently $29,902, but this benefit is scheduled to rise to $73,854 (in 2012 dollars) by 2090, a 2.5-times increase in purchasing power. The increase in the real value of the benefit is the result of the wage-indexation of initial benefit levels. A cap of $29,902 in real benefits would cause the system to converge to a single Beveridge floor close to the year 2100, as the real growth of the wage-indexed benefits received by lower income workers would eventually hit the 2012 benefit cap. At that time, a retiree with scaled low earnings would enjoy purchasing power in retirement equivalent to the highest income benefits paid today. The system could otherwise remain unchanged, aside from changes to CPI and indexing of the normal retirement age.
It is clear from observing the current political debate that progressives have become unwilling to ask workers to finance their own government benefits, instead relying on higher taxes on “the rich” to cover outlays. This was true especially with ObamaCare, where a consumption tax or other reliable revenue system was rejected in favor of specious spending cuts and higher taxes on high earners. Conservatives should not lament this shift in rhetoric and policy preferences. The result is likely to be less political support for the current system and a flatter benefit system that provides important protections in a much more affordable way.