Why WEP?

Filed Under (Retirement Policy, Uncategorized) by Jeffrey Brown on Aug 23, 2010

One of the most despised provisions of the Social Security regulations is known as the WEP – an acronym for the “Windfall Elimination Provision.”  This provision is poorly named, poorly designed, and poorly understood.  But that does not mean it should be eliminated.  While the Social Security Administration does a truly horrible job of communicating it, the WEP (or something like it) has a legitimate reason for existing.

What is the WEP?  It is a provision in the law that alters the way Social Security benefits are calculated for individuals who work for state and local employers who do not participate in the U.S. Social Security system.  For example, the earnings of employees of public universities and public schools in Illinois – who participate in Illinois SURS and Illinois TRS – are not covered by Social Security. 

Illinois is not alone.  Approximately one fourth of all public employees in the U.S. do not pay Social Security taxes on the earnings from their government job according to the U.S. Government Accountability Office (GAO).  This includes approximately 5.25 million state and local workers, as well as approximately 1 million federal employees hired before 1984. 

However, many of these public employees – including the author of this blog – will still qualify for Social Security benefits, either as a result of switching between covered and uncovered employment at some point in their career or because they simultaneously work two or more jobs that span both covered and uncovered employment.  For example, a teacher in the State of Illinois may spend his summers working in covered employment.  Alternatively, a professor may spend part of her career working at a private university covered by Social Security, and part of her career working for a state university that is not covered. 

If Social Security benefits were calculated as a simple “linear” function of lifetime earnings, this would not present any problems.  If you earned 50% of your lifetime income in Social Security, you would just get 50% of the benefit that you would have earned had all your earnings been covered.  The only thing Social Security would need to know is how much you paid into Social Security.  Whether you have other “uncovered” earnings would be irrelevant.

But Social Security does not have a “linear” benefit formula.  Rather, it is explicitly designed to offer a higher ratio of benefits-to-taxes-paid for low income workers than it offers to higher income workers.  It is designed this way in an attempt to redistribute income from the rich to the poor.

And therein lies the problem.  If Social Security only observes part of a person’s total earnings (e.g., they know someone’s earnings from a summer job, but not their university salary), then they might mistakenly classify this person as a low-income individual, even though they might be a high income individual who just had a small part of their earnings covered by Social Security.  As a result, blindly applying the same benefit formula to this person gives them a benefit that is too high relative to other individuals who have the same total lifetime earnings!  In essence, we would be paying too much to people who only worked a small part of their career under Social Security.      

In order to adjust for this, the Windfall Elimination Provision (WEP) was enacted as part of the 1983 Social Security Amendments.  This provision is meant to downward-adjust the Social Security benefits of affected workers in order to eliminate the “windfall” (a poor choice of words, I am the first to admit!) that arises when, for example, an individual with high lifetime earnings (based on both covered and uncovered earnings) would appear as if he or she were a low earner when evaluated solely based on covered earnings. 

It is easiest to see the problem that would be created if there were no WEP provision in place through an example.  Consider the three individuals shown in the table below.  “Larry” is a very low income worker who works his entire life under Social Security, with an average lifetime monthly earnings of only $500 per month.  Using the 2008 benefit formula, Larry would have a full benefit $450, or 90% of his pre-retirement income.  “Mo” is a higher income worker with all of his earnings covered under Social Security, thus having an average monthly income while working of $5,000.  Under the benefit rules, Mo would have a full benefit of $1891.34, or a 38% of their working life income.  Thus far, this example simply illustrates the “redistributive” nature of the benefit formula, as Larry receives a higher replacement rate than does Mo, owing to the fact that Larry has lower lifetime earnings.

Social Security Primary Insurance Amount If No WEP Adjustment Applied


Average earnings covered by SS

Average earnings not covered by SS

Average total earnings

Benefit if SS formula applied to covered earnings

Benefit as % of income if no WEP adjustment




















Now consider Curly, a public employee.  Curly’s total lifetime earnings of $5000 are identical to Mo’s.  Had all of Curly’s earnings been covered by Social Security, Curly would have the same 38%replacement rate as Mo.  However, only 1/10th of Curly’s earnings were in employment covered by Social Security; the rest were in non-covered public employment.  If Social Security applied the standard benefit formula to Curly’s covered earnings without any WEP adjustment, Curly would receive a monthly benefit of $450, equivalent to Larry.  This provides Curly with a ratio of benefits to (covered) earnings of 90%, which is substantially more generous than the 38% ratio provided to Mo, even though Mo and Curly have identical lifetime earnings.  To use the language of the provision designed to address this issue, Curly would receive a “windfall.”  The WEP adjustment is designed to calculate Curly’s benefits differently, so that they end up looking more like Mo’s, since they both have similar lifetime incomes.    

In short, because Social Security is a redistributive program, there is a real need for something like the WEP.  Most people affected by it, however, hate it.  And who can blame them given that SSA does a terrible job of explaining it?  In essence, instead of telling a retiree that “your benefit will be $800,” SSA tells them “your benefit would be $1100, but because of the WEP, it is only $800.”  But for the individual in question, the $1100 benefit is a red herring.  In no way, shape or form is the $1100 benefit a relevant amount to start with.  So SSA’s poor communication and negative framing raises a lot of hackles unnecessarily.  As a result, thousands of letters are written to elected officials every year demanding that it be repealed.  And, every year, bills are introduced in Congress to eliminate it.  And every year, those bills fail as they should.

This is not to say that the WEP is perfect.  Far from it.  I have written more extensively elsewhere that the WEP calculation may be close to correct on average, but it is almost certainly wrong for each individual.  Sadly, it hits lower income individuals harder than it should, and does not hit most high income individuals hard enough.  There is a “right” way to calculate the WEP, but implementing it requires that SSA have a full history of both covered and uncovered earnings, but they did not collect the uncovered earnings in a systematic way until the early 1980s.  As such, we probably have to wait another 10 years before they can implement the fix.  In the meantime, SSA could do themselves and a lot of elected officials a huge favor by taking the time to adequately educate affected individuals on the rationale for this program.

Who Bears the Burden of Energy Policy?

Filed Under (Environmental Policy) by Don Fullerton on Sep 4, 2009

Economists have tools to analyze the distributional effects of income taxes, payroll taxes, property taxes, and corporate income taxes.  Some existing research even looks at distributional effects of environmental or energy taxes used to help control pollution or energy consumption.  Yet most pollution policy does not involve taxation at all!  Instead, we use permits or command and control regulations such as technology standards, quotas, and quantity constraints.  Existing studies of energy policy are mostly about effects on economic efficiency, addressing questions such as: how to measure the costs of reducing pollution or energy use, how to measure benefits of that pollution abatement, what is the optimal amount of protection, and what is the most cost-effective way to achieve it.

Yet environmental mandates do impose costs, and an important question is who bears those costs.  Moreover, those restrictions provide benefits of environmental protection, so who gets those benefits?  Full analysis of environmental policy could address all the same questions as in tax analysis.  Perhaps it could use the same tools to address distributional effects – not of taxes, but of these other policies that are used to protect the environment.

Thinking about the distributional effects of environmental policy is interesting and difficult.  For example, a standard tax analysis would point out some complex implications of an excise tax: not only does it affect the relative price of the taxed commodity, and thus consumers according to how they use income, but it also impacts factors intensively used in the production of that commodity, and thus individuals according to the sources of their income.  Yet an environmental mandate can have those effects and more!  Consider a simple requirement that electric generating companies cut a particular pollutant to less than some maximum quota.  This type of mandate is a common policy choice, and it has at least the following six distributional effects.

First, it raises the cost of production like a tax, so it may raise the equilibrium price of output and affect consumers according to spending on electricity.

Second, it may reduce production like a tax, reduce returns in that industry, and place burdens on workers or investors.

Third, a quota is likely to generate scarcity rents.  For simplicity, suppose pollution has a fixed relation to output, so the only “abatement technology” is to reduce output.  Then a restriction on the quantity of pollution is essentially a restriction on output.  Normally firms want to restrict output but are thwarted by antitrust policy.  Yet in this case, environmental policy requires firms to restrict output.  It allows firms to raise price, and so they make profits, or rents, from the artificial scarcity of production.  Just as tradable permit systems hand out valuable permits, the non-tradable quota also provides scarcity rents – to those given the restricted “rights” to pollute.

Fourth, if it cleans up the air, this policy provides benefits that may accrue to some individuals more than others.  The “incidence” of these costs and benefits usually refers to their distribution across groups ranked from rich to poor, but analysts and policy-makers may also be interested in the distribution of costs or benefits across groups defined by age, ethnicity, region, or between urban, rural, and suburban households.

Fifth, regardless of a neighborhood’s air quality improvement, many individuals could be greatly affected through capitalization effects, especially through land and house prices.  Suppose this pollution restriction improves air quality everywhere, but in some locations more than others.  If the policy is permanent, then anybody who owns land in the most-improved locations experience capital gains that could equal the present value of all future willingness to pay for cleaner air in that neighborhood. Similar capitalization effects provide windfall gains and losses to those who own corporate stock: capital losses on stockholdings in the company that must pay more for environmental technology, and capital gains on stockholdings in companies that sell a substitute product.

Capitalization effects are pernicious.  A large capital gain may be experienced by absentee landlords, because they can charge higher rents in future years.  Certain renters with cleaner air might be worse off, if their rent increases by more than their willingness to pay for that improvement.  Moreover, the gains may not even accrue to those who breathe the cleaner air!  If households move into the cleaner area after the policy change, then they must pay more for the privilege.  The entire capital gain goes to those who happen to own property at the time of the change, even if they sell it at the higher price and move out before the air improves.  Similarly, new stockholders in the burdened company may be “paying” for abatement technology in name only, with the entire present value of the burden felt by those who did own the stock at the time of enactment, even if they sell that stock before the policy is implemented.

Sixth, strong distributional effects are felt during the transition.  If workers are laid off by the impacted firm, their burden is not just the lower wage they might have to accept at another firm.  It includes the very sharp pain of disruption, retraining, and months or years of unemployment between jobs.  These effects are analogous to capitalization effects, if the worker has large investment in particular skills – human capital that is specific to this industry.  If the industry shrinks, those workers suffer a significant loss in the value of that human capital.  They must also move their families, acquire new training, and start back at the bottom of the firm hierarchy, with significant psychological costs.

The challenge here is that many of these effects of environmental policy are likely to be regressive.  Consider the six categories just listed.  First, it likely raises the price of products that intensively use fossil fuels, such as electricity and transportation.  Expenditures on these products make up a high fraction of low income budgets.  Second, if abatement technologies are capital-intensive, then any mandate to abate pollution likely induces firms to use new capital as a substitute for polluting inputs.  If so, then capital is in more demand relative to labor, depressing the relative wage (which may also impact low-income households).  Third, pollution permits handed out to firms bestow scarcity rents on well-off individuals who own those firms.  Fourth, low-income individuals may place more value on food and shelter than on incremental improvements in environmental quality.  If high-income individuals get the most benefit of pollution abatement, then this effect is regressive as well.  Fifth, low-income renters miss out on house price capitalization of air quality benefits.  Well-off landlords may reap those gains.  Sixth, transition effects are hard to analyze, but could well impact the economy in ways that hurt the unemployed, those already at some disadvantage relative to the rest of us.

That is a potentially incredible list of effects that might all hurt the poor more than the rich.  The challenge for those of us who want to claim to do policy-relevant research, then, is to determine whether these fears are valid, and whether anything can be done about them – other than to forego environmental improvements!