Social Security – Will Wilkinson

The "Annually Appropriated/Authorized Until Revised" Spending Distinction

July 18, 2008 Will Wilkinson

My colleague John Samples, a distinguished political scientist and scholar of American politics, writes to me to pithily explain the discretionary/non-discretionary distinction:

Discretionary spending goes through the annual appropriations process. Such spending has to be approved each year, in form at least. Non-discretionary spending has permanent appropriations which are determined by demographics in tandem with specified entitlements. Non-discretionary spending can be changed but only by reopening the authorization for the program. Revising an authorization is quite difficult to do politically. This is not a principled distinction. It is just a political difference. Presumably constituencies of non-discretionary spending are more powerful than those supporting discretionary spending.

So the real distinction has to do with the political ease of exercising discretion, and not the legal possibility of exercising discretion.

What, then, would be a descriptively accurate way of making the distinction? Suppose a responsible journalist wanted to convey the facts as they actually stand rather than reinforce the political choice to make the distinction in this confusing way? Spending that is “annually appropriated” versus spending that is “authorized until revised”? That would be accurate, wouldn’t it? (Please come up with something better!) But “authorized until revised” leads you to think immediately of the possibility of revising it. “Non-discretionary” or “mandatory” leads you to think that that the spending is somehow legislatively off-limits, which is of course false and so misleading.

This is one of those cases where I think language must have political effects. So here’s one for you, George Lakoff. The labeling of Social Security and Medicare spending as “non-discretionary” encourages the sense that the programs are part of what Cass Sunstein calls “the second Bill of Rights” — that they are underwritten by some kind of legal guarantee that goes beyond the will of the current congressional majority, when in fact reopening authorization, though politically difficult, could in principle happen at any time. Strangely enough, I think the “non-discretionary” frame is partly what led Social Security reformers to attempt to motivate reform as necessary to avert a “crisis.” It’s non-discretionary! Like a barrelling train with the brakes out heading for a cliff! Here is out plan to use massive dirigibles to lift the train aloft! However, by successfully rebutting the “crisis” meme by pointing out how easily the benefit formulas can changed–the brakes work fine, thank you–defenders of the status quo loosened the sense of guarantee crucial to the social insurance frame. So what we saw were Democrats simultaneously arguing that we can just change the formulae whenever, so no worries, while at the same time fretting over how certain changes, like progressive indexing, would dislodge the sacred myth of universal social insurance.

By the way, I just glanced at Jason Furman’s short paper about progressive indexing. I really like progressive indexing.

I also like mandatory retirement accounts for paternalistic reasons that are also sort of libertarian. Means-tested benefits for old people is a better idea than our stupid current system, but would encourage too little retirement savings. Why? Old people are so politically powerful that these benefits will be too high to make saving rational. So forcing people to transfer their own money to their future selves prevents them from later forcing others to transfer them money when old.

Safety Nets, Growth, and Liberation from Family

March 27, 2007 Will Wilkinson

In his by-request post on safety nets, Tyler writes:

Most of all, the welfare state liberates the productive and the creative from their sometimes burdensome family ties. The welfare state is the Randian’s secret dream, and that is what clinches the case for a government safety net.

I don’t think I understand. How many productive and creative people take advantage of government assistance programs for the poor, or are liberated by them? It would seem that genuinely productive, creative people would need them least.

Probably Tyler means that Social Security and Medicare allow the productive and creative to foist their poor parents off on the state. True. But increasing incomes also allow the productive and creative to foist their not-so-poor parents off on “assisted living” facilities. Wealthier parents don’t need their kids’ money, and wealthier kids can afford to have somebody else worry about their parents. How much has the deadweight loss of our actually existing, almost entirely middle class to middle class “social insurance” tranfer system decreased economic growth over the last half-century? It is not obvious that the history of our real system compares favorably to even a slightly higher-growth counterfactual in terms of the kind of liberation from burdensome family ties Tyler is talking about.

If you’re a Sen-type positive liberty advocate like Tyler, and don’t so much care about the coercion implicit in transfers, your problem with safety nets ought to be the potentially psychologically debilitating effects of transfers on the recipients with respect to a sense of control, self-efficacy, motivation, etc. I do not doubt for a second that many, many people have been genuinely helped by public assistance. I do, however, have some doubt that the overall effect has been positive relative to some of the potentially feasible alternatives.

Zombie Reforms, Zombie Arguments

February 9, 2006 Will Wilkinson

In the Washington Post’s account of the resuscitation of Social Security reform in the president’s budget proposal, Allan Sloan writes, of progressive indexing:

This means that although progressive indexing is an attractive idea from a social-justice point of view, it would reduce Social Security’s political support by making it seem more like welfare than an earned benefit.

Matt Y. characterizes Social Security reform as a zombie that won’t (really truly) die. Sure. But how about Sloan’s undead item of liberal faith about political support, which keeps rearing its dessicated head despite the lack of any good argument.

As far as I can tell, there is little to no evidence that converting social security to a means-tested benefits program would reduce political support for it. It is true that former Social Security administrator Wilbur Cohen’s assertion that a program for the poor will be a poor program is repeated endlessly. But truth is not established by repetition. Unemployment and disability insurance, unlike Social Security retirement benefits, kick in only in conditions of necessity. Nevertheless, or perhaps due to that fact, they are very politically popular, well-funded, and face no apparent political threat of reduction.

Indeed, there is compelling evidence that means-tested retirement benefits would be too generous creating a perverse incentive for workers to save too little in order to qualify for a beefy means-tested benefit.

Xavier Sala-i-Martin and Casey Mulligan have written a fascinating account of the political economy of “gerontocracy” or rule by the old. Their key point is that during their working years, workers have fractured political interests. Teachers plump for teacher’s unions. Manufacturers lobby for price-supports. Investors fight for lower capital gains taxes. Etc. Many of these interests cancel each other out. In any case, there is no unified front. However, when people retire from their particular occupations, they leave behind a narrow sectional interest and move into a much broader interest group, retirees, who have highly unified interests, and, moreover, very low opportunity costs to political participation.

The demographic problem of social security is precisely that a very large portion of the population is soon to make the transition into this group. There is every reason to believe that retirees would lobby for big retirement benefits, and there would be no other political interest as large and unified to keep them from getting them. One of the chief economic reasons for mandatory retirement savings accounts is to guard against this much more likely contingency. (More likely than too little political support, that is.) If workers are required to save for retirement in protected accounts, they will not be able to prematurely consume their savings in order to qualify for predictably over-generous means-tested retirement benefits.

I agree that progressive indexing is attractive in terms of distribution. But since there is little reason to believe that a more means sensitive program will lack political support, why not just go all the way to a progressively designed means-tested program for the poor elderly complemented by mandatory personal retirement accounts to buffer against moral hazard?

This is, I agree, not a particularly libertarian proposal. But I think it is the best feasible option from almost every perspective with an interest in feasibility. For the life of me, I still can’t figure out how a liberal could possibly prefer the status quo over a cushy retirement safety net plus mandatory accounts.

My prediction is that nominal liberals actually won’t be able to hold out that long against the overwhelming liberal sensibleness of this, and so if Republicans can’t get the job done, Democrats will soon enough.

Dominoes vs. The Great Leap Forward

August 3, 2005 Will Wilkinson

If you missed Nathan Smith’s TCS article on the ongoing saga of Social Security reform, do check it out.

Bush’s plan for carve-out private accounts would have amounted, institutionally, to a sort of Great Leap Forward. DeMint’s plan will set in motion incremental changes which may be compared to knocking down a row of dominoes. The first domino to fall is the payroll tax surplus in the Trust Fund. The second domino will be the excessive scheduled benefits that drive the program into long-term bankruptcy. The third domino will be the restriction of personal retirement accounts (initially created as lockboxes to stop the raid on the Trust Fund) to T-bonds. When that falls, all Bush’s Social Security reform goals will have been accomplished, and we’ll have a system of forced savings and private accounts.

Suppose that the DeMint plan passes and personal accounts are created from the surplus, then fast-forward two years. Now every working person under 55 — well over 100 million Americans — will own a personal retirement account consisting of US Treasury bonds. Since everyone and his brother knows that Social Security can’t pay promised benefits in the long run, most young people will see these accounts as their sole source of real retirement security. But they’ll also realize that the personal accounts are too small to underwrite a comfortable retirement. Moreover, they will learn that new money will cease being deposited in their accounts after about 2018, when the Baby Boomers’ retirement puts an end to the surpluses.

At this point, there will be pressure from younger voters to increase the size of their personal retirement accounts. If, up until now, the Social Security program has consisted of one-sided class warfare, with the old fighting against the young and the young not defending themselves, personal accounts will clarify younger generations’ stake in the fight.

Nice account of what perhaps should have been the strategy all along. I don’t have a good independent sense of what DeMint’s odds are. Probably not great, but better than than is being reported.

The Case for Carve Outs

July 6, 2005 Will Wilkinson

My brilliant colleague Jagadeesh Gokhale explains why the notion that personal accounts and Social Security solvency are unrelated is a canard.

Here’s the core of the argument:

The difference between the two projections of future benefit levels funded out of present law payroll taxes — higher ones under “add-on” personal accounts versus lower ones under a “status-quo” hike in payroll taxes — constitutes the basic case for “carve-out” personal accounts. How come? If “add-on” accounts to pay for benefits that are promised but unpayable under present law effectively increases saving and investment and preserves work incentives, the (lower) level of payable benefits under a “status quo” payroll tax hike could be financed with a less than 12.4 percent payroll tax rate under the “add-on” policy. That implies room for a “carve out” — that is, for diverting a part of present law payroll taxes into personal accounts.

How large would be the size of a feasible carve out? Would it ultimately completely do away with the need for “add-on” contributions? These are difficult questions to answer. Two considerations suggest, however, that the scope for carve-outs could be large. First, several studies report that payroll taxes add significantly to marginal tax rates — especially for households’ secondary earners — and that labor supply is quite sensitive to higher taxes. Noteworthy here is a recent study by economics Nobel laureate Edward Prescott that attributes the significant decline in European labor supply relative to the United States since the 1970s to higher European social insurance taxes.

Second, loss in annual output because of the savings-reducing impact of the current Social Security system’s pay-as-you-go financing structure is estimated to be of the same size as total current outlays on Social Security. That is, were the existing system based entirely on “add-on” personal accounts, the gain in annual output due to higher saving and capital formation would have been about as large as total current outlays on Social Security.

It’s a complicated argument, but nobody ever said getting it right is easy.

New Cato Social Security Choice Paper

June 28, 2005 Will Wilkinson

I have a new paper out today in Cato Social Security Choice series. It’s called “Noble Lies, Liberal Purposes, and Personal Retirement Accounts.” If you’ve been following the Social Security posts on the blog, lots of the paper will seem familiar to you, but there’s a lot of new stuff in the paper that I encourage you to check out.

Here’s the executive summary:

Opponents of President Bush’s proposal to make individually owned personal retirement accounts a part of the Social Security program routinely charge that it is motivated by ideological animosity toward the values Social Security is supposed to embody, such as equality and social cohesion. However, a frank look at the Social Security status quo reveals that the program is very poorly designed to realize liberal ideals. Social Security has a barely progressive overall structure, if it is progressive at all. The huge volume of transfers inherent in the system accomplishes very little income redistribution within generational cohorts. Furthermore, it works to the disadvantage of current workers, who will receive a smaller “return” on their payroll taxes than do current retirees. The terms of the imaginary “compact between the generations” are manifestly unfair.

What is worse is that the Social Security status quo embodies a government-perpetuated deception designed to generate its own political support by misleading voters into believing that their payroll taxes entitle them to later benefits. The architects of Social Security created a structure and accompanying rhetoric that were specifically intended to encourage the false belief that the system provides a kind of insurance, similar to private insurance based in contract and property, and therefore involves a binding entitlement to benefits.

However, there is no justification for this deception on contemporary liberal grounds. The persistent intentional misrepresentation— the “noble lie” — embedded in the structure and language of the Social Security system is in fact antithetical to the ideals of transparent government, open democratic deliberation, and equality among citizens — ideals at the core of contemporary liberal thought.

A system of personal retirement accounts plus a means-tested safety net would serve the “social insurance” function better than the Social Security status quo according to liberal standards. Contrary to critics of reform, personal retirement accounts would materially enhance equality and social cohesion by more fully integrating workers into the market, providing everyone with a stake in its growth, closing the gap between the investing and noninvesting classes, and making more salient the mutuality of interests in a market society.

The paper is here.

If You Like Social Insurance So Much, Then How Come You're Against It?

June 13, 2005 Will Wilkinson

This is from a now very old Barry Schwartz column about why people are too stupid to manage their own finances, especially if their finances involve a Social Security personal retirement account, in the NYT. Old, but so bad I can’t help talking about it.

This brings me to the final defense of privatization: the payroll taxes you pay are your money, and you ought to be able to do what you like with your money. This, I suspect, is the real justification behind the move to privatize, and it is the worst reason of all. The payroll tax is not “your” money; it’s our money. Social Security was created as an insurance scheme, not a pension scheme. It was meant to provide a safety net, to protect the unlucky from immiseration in old age. The benefits we get are not payouts from accounts in which we have accumulated our own private stash. What we get is largely determined by what we earned, but we keep getting it even after we’ve taken out every penny we put in. And if we happen to die early, someone else reaps the benefits of our contributions.

That’s refreshingly frank: it’s not your money! OK. So, it’s “ours.” Let’s just skip over the fact that the entire Social Security system and the disinformation one regularly gets from the SSA is specifically designed to encourage the sense that there is some kind of property-like nexus of entitlement between the payroll tax and retirement benefits. (And that social insurance systems like ours are commonly called pensions systems around the world.) Perhaps Schwartz will forgive American voters for having the wrong idea here, and not realizing that they do not in fact have any right to their benefits.

Anyway, what is it that we’re doing with “our” money? Well, we’re sending over 90% of it back to the same income bracket from whence it came, that’s what! Now why would we be doing that if what we wanted to be doing was “protecting the unlucky against immiseration in old age”? (Not to sustain the illusion that our payroll taxes do in fact belong to us as individuals, for sure!) I mean, wouldn’t it be silly to pretend to “insure” people by taking money away from them (thereby increasing their exposure to risk!), and then simply replacing it later? That sure would be silly! Schwartz gestures toward the redistributive function of the program, but . . . there is almost no redistribution! And . . . it isn’t progressive!

An authentic government old-age insurance program might look like government disability insurance. You pay taxes into the program and then you get money back if and when you need it–sort of like the way actual insurance works!

Hey liberals! Since you insist on talking about social insurance, why not stop dissembling and plump for a system that is actually sort of like insurance? Why not not defend a disability insurance model of old-age insurance, where you get it only there is some actual threat of immiseration? We can fund it with a dedicated payroll tax and everything. It really will not function like a pension at all. It will be a safety net for people who need it funded by people who don’t. Isn’t this exactly what liberals should want? TPM Cafe? Left2Right? Somebody? Why don’t you love this idea? Really, I want to know.

A Rare Triumph for Liberals

May 20, 2005 Will Wilkinson

Here’s E.J. Dionne:

Indeed, the Social Security debate so far has been a rare triumph for liberals: For the first time in a long while, core liberal principles are actually winning in a public debate. The idea that Social Security is an insurance program and not an investment plan is gaining traction.

Why E.J. Dionne thinks deception is a “core liberal principle” totally eludes me. Here are two options about Social Security as insurance.

(1) It is not insurance.
(2) It is “insurance,” in some loose sense. But then so is (a) means-tested welfare or (b) mandatory personal retirement accounts plus a safety net.

If (1) is true, then Dionne’s claim is a lie. If (2) is true, then saying that Social Security is insurance does nothing to tell us why we ought to prefer the status quo over (a) or (b). As it happens, there is NO “core liberal principle” that I can think of that supports the fevered defense of America’s badly structured historic welfare programs.

Check this:

* The Social Security tax is regressive.
* The overall benefit structure accomplishes, on net, either no downward income redistribution, or a small amount of upward redistribution. (I.e., it is either close to a wash, or regressive, redistribution-wise.)
*The system is structured to disadvantage current workers over current retirees, and is thus invalid as a “compact” between generations, if we take the contract metaphor seriously.
* Because the Social Security tax hits least wealthy workers hardest, Social Security prevents many from accumulating wealth, and reinforces the divide between investing an non-investing classes.
*Social Security makes it impossible for many of the least wealthy to accumulate wealth that they can pass on to their children or grandchildren, thereby helping to perpetuate generational inequality.

It is crucial to note that whatever else it might be, Social Security is not PRIMARILY insurance, if it is insurance at all. The redistribution to the elderly poor it does manage to effect is incidental to the huge volume of transfers back and forth from within the same income bracket. (Net income-related redistribution come to less than 10% of total transfers) There’s a huge amount of deadweight loss in all this pointless churn.

A system of personal accounts plus a means-tested safety net would:

* Be more progressive in every way.
* Eliminate most of the unjust intergenerational transfers that are at the heart of the current system.
* Almost entirely close the gap between the investing and non-investing classes.
* Help the least wealthy workers accumulate inheritable wealth.
* Protect the elderly against poverty AT LEAST as well.

All this is independent of the fact that the “insurance” language amounts to a “noble lie” that ought to be anathema to liberals.

Consider James Buchanan’s analysis of Social Security as a “fiscal illusion,” which is an idea he credits to the obscure Italian public finance theorist Amilcare Puviani. According to Buchanan, following Puviani, those with political power will often erect a structure of public finance that is meant to cause “tax payers to think that the taxes to which they are subjected are less burdensome than they actually are” and “make beneficiaries consider the values of public goods and services to them to be larger than may actually be the case.”

Here is what Buchanan says about Social Security in particular:

Social Security Taxes. The modern American system of old-age and survivors “insurance” seems ready made for the Puviani criticism. It is apparent to almost everyone, without detailed analysis or knowledge of the system, that the effects of promoting the institutions under the “insurance” rubric, which implies actuarial independence and integrity, tends to conceal from participants the real flows of costs and benefits. Whether or not such was the deliberate intent of the founders of the system need not concern us here. The facts are that the system, as an independent trust-fund account outside of the regular budgetary procedures of the federal government, is not actuarially sound by private financial standards, and that the plan will depend for its continued existence on the Treasury’s willingness to finance currently claims made against the system. Contributors to the system finance only a relatively small share of the benefits that they receive, especially to this date (1966), and the remaining funds must be secured from current taxes collected from prospective beneficiaries. To the extent that the current contributor accepts the regular increases in his own taxes, as well as those nominally levied on his employer, under the assumption that, on balance, these are to be accumulated for support of his own retirement benefits, he will be less resistant to such increases than if he knew that such tax increases were simply required to meet current outpayments to beneficiaries. He operates under an illusion of the Puviani sort. If future claims against the system should be properly discounted, along with future taxes that are required to meet these claims, the entrant into the system would recognize that, in the net, the costs significantly exceed the benefits, both computed in present-value terms. The fact that there is no widespread resentment or resistance against entering the system supports the hypothesis that illusion is present, and is effective. Even for the employee who may recognize the actuarial bankruptcy of the present system, who is able to dispel the fiscal illusion, it may not, however, be rational to reject the scheme when he predicts that, during his own period of retirement, other prospective entrants can still be attracted by illusory claims of “insurance.” The system in this manner provides a continuing means through which income transfers can be made to the aged from the currently productive elements of the population, which can be “explained” or “rationalized” to many taxpayers on the basis of contributory schemes of retirement protection. There seems little question but that, if the same fiscal transfers were proposed openly and without attempts at illusion, there would be significantly greater political resistance. This conclusion can be attained, regardless of one’s own value position on the quite separate question as to whether such transfers should be decreased, kept the same, or increased.

Compare Buchanan’s analysis with a key idea from another of the 20th Century’s great contractarian political theorists, John Rawls:

if the basic structure relies on coercive sanctions, however rarely and scrupulously applied, the grounds of its institutions should stand up to public scrutiny. When . . . basic social arrangements and individual actions are fully justifiable, citizens can give reasons for their beliefs and conduct before one another confident that this avowed reckoning itself will strengthen and not weaken public understanding. The political order does not, it seems, depend on historically accidental or established delusions, or other mistaken beliefs resting on the deceptive appearances of institutions that mislead us as to how they work.

That’s Rawls in Political Liberalism talking about the publicity requirement for just institutions. Just rules for a free people in a liberal society must be open to them and MUST NOT be sustained by illusions that cause people to misunderstand the terms of our association, cripple our ability to govern ourselves through effective public deliberation, and ensure that we remain ignorant of the institutions that shape our expectations for our lives, and condition our preferences and characters.

By pushing the “Social Security status-quo is insurance line” Dionne is directly contributing to the violation of “core liberal principles” that ought to be even more important to welfare-liberals than the fiscally anti-egalitarian aspects of the status quo that Dionne, for some inexplicable reason, thinks are just great.

This is a “triumph for liberals”?!

Welfare iberals keep wondering aloud about what’s the matter with welfare liberalism. Dionne unwittingly points us toward a plausible hypothesis. Welfare liberals have become fetishists of the crumbling institutions of the New Deal/Great Society welfare state, which they mistake for actually existing expressions of liberal ideals. But welfare liberals don’t even understand what liberalism, as an idea, is, and so can’t recognize that their beloved programs fail according to what ought to be their own standards. All they can do to answer the what is liberalism question is to point at an amber-edged policy and say “that!”