And what, you ask, do I think the real debate is (or should be)?
I'll have some thoughts on that later. It's probably not what you think I'm going to say.
Social Security benefits are progressive--benefitting lower income workers with higher replacement rates. The system is widely touted as Insurance against poverty in Old Age or in the case where a spouse or dependents Survive the main breadwinner. OASI=Old Age and Survivor's Insurance.
However, benefits are at least marginally tied to your lifetime wage profile--like a defined contribution plan or a forced savings program. So what is it? Insurance or forced savings? Brad DeLong and Mark Thoma emphasize the insurance aspect. Actually, they take particular objection to those who call it welfare. Most assuredly, Social Security is not welfare, so we've got no argument there. Angry Bear comes closer to a direct comparison of the insurance vs. savings question. He says,
...all of the privatization plans I've seen discussed would replace Social Security with a defined contribution style system in which the more a person makes, and the luckier that person is, the more that person will have at retirement, and vice-versa. The insurance function of pooling and attenuating risk is totally removed; or, more accurately, reversed.
So I think it is useful to think of a Social Security program like ours as having an insurance component to prevent poverty and indexed benefits as a political appeasement to those who are less likely to need the insurance.
Where do you draw that line? What is the appropriate level of insurance? Can you have private accounts (add-on or carve-out) and maintain the insurance aspect? How do we get from where we are to where we want to be?
This is the discussion that we aren't having, but should have. It doesn't matter where you stand on trust funds or private accounts to discuss the appropriate level of social insurance. If we could get a handle on that question, then we could make progress on the rest of it. The consequences of this question run deep. If you believe that the insurance portion of the system should be enhanced above what it is now, then private accounts are a non-starter. If you want to allow people to self-insure more than they do now, then you would likely think that private accounts are a vehicle for them to choose their level of risk. This is a normative question. This is a matter of your view on how much risk is acceptable in a person's lifetime consumption profile, how replacement rates should vary with lifetime income, etc.
Trust funds are a distraction when you look at the problem this way. All these posts about the trust fund are arguing over minutia--how real are the assets, how shortsighted the credit market is for not recognizing that the trust fund has been raided, and so on. The bonds are real. The promise to pay benefits will be kept (subject to modifications by Congress, but the modifications are not likely to be severe). The trust fund is being used by both sides as a rhetorical device to advance their point of view by quibbling over the value of a bond. These matters, while interesting, are nothing compared to a frank discussion of the role of Social Security as insurance and the appropriateness of having the government run a pension program whether it's with private accounts or not. This discussion would not take place in a vacuum, but would necessarily take place in the context of the overall tax and spending environment at the macro (government finance) and micro (overall tax progressivity and the effect on households) levels.
Debating the progressive price indexing of the Pozen plan is a start. Brad DeLong's piece in Slate is a good opening salvo. In it, he implicitly acknowledges the political reasons for less draconian cuts.
Pozen's proposal caps the maximum Social Security retirement benefit at roughly $22,500 dollars a year (adjusted for inflation). Bush's private-accounts plan—which would allow people to contribute 4 percent of their wages—makes retirees repay the taxes they diverted into private accounts out of their standard Social Security benefit. Medicare premiums are already deducted from your Social Security check. Deduct the claw-back for the private-accounts diversion as well, and by late in this century the odds are that—at least for the upper middle class—the standard Social Security check would be zero. Social Security would no longer be a universal program: It would be a program in which the half of America that is richer and more powerful and more likely to vote sees large chunks of its money going in and nothing coming out. (emphasis mine)
Appeasement for those less likely to need insurance might be a political necessity to keep the program viable. This has always been a standard argument against means testing.
I say again, this is the right discussion to have.
I, for one, would approach such a discussion with an open mind.
I agree and have talked about this quite a bit in the post you linked and in others on this same issue. This is a normative issue, a choice society must make, it is not something economics has the answer to. We can advise on efficiency once the choice is made, but we cannot make the choice itself.
I think about someone who worked hard their entire life and gave to this society - volunteered in schools, and so on, worked hard everyday to support a family, saved stocks and bonds, but unexpected health costs (or a hurricane hits the house, etc., some uncontrollable shock to savings) wipes out the asset accumulation. The person lives to be 95 and has done everything right in life but had bad luck and is now too old to find work even if they were capable. What type of life does society owe such a person? What if it is you or your mom or dad? My view is fairly generous in this regard, but it is ultimately a normative decision and something to be resolved through the political process. You can argue anywhere from an absolute bare subsistence level to a fairly generous level of support, but there is no correct answer as to what is fair that we will all agree upon other than that a floor exists. And I suspect there are a few who would even argue against a minimal level of support.
Just a brief follow up - I agree entirely and have noted before that the question of whether a person views the current system as just insurance or insurance plus savings depends upon their (normative) view of the level of support that ought to be insured, the minimum acceptable level of societal support. Much of the argument you hear is needless because it is really an argument over this difference and there is no right answer. So until we collectively decide through the political process what that ought to be it is diffiuclt to move forward.
I will note that the average payment is around $1,000 for men and $850 for women (very, very roughly), but I don't know much about how the payments are distributed around the mean. But at that mean, if that is the sole source of income as it is for a substantial number of the elderly, then I do not consider it overly generous - but that is a personal opinion...
Yes - I think you are (both) spot on. The explicit question of redistribution is at the heart of this, but also sensitive and hard to tackle, so the political discussion gets encumbered with all sorts of baggage.
But be careful about saying "redistribution", rjw. Redistribution conjures up notions of welfare, which we want to avoid.
If you could opt out, there would be adverse selection. The wealthier would want to put all their money into their 401(k)s. So it has to be designed as an optimal insurance system when viewed from behind the "veil of ignorance" or from an "original position"--two Rawlsian ideas that I think apply here.
But as Mark points out, some might even argue against minimal support. The liberal vs. libertarian debate (or Rawls vs. Nozick) continues.
Mark,
I think we have to specify carefully what we are insuring against. To take your examples:
Unexpected health costs: National catastrophic health insurance (not necessarily national health care) could address this. (topic for another post)
Hurricane: Homeowner's insurance. Disaster aid.
Let's just say that most idiosyncratic shocks can be insured by things other than Social Security--whether the insurance is public or private.
Uncontrollable shock to savings: Such as? Would too much insurance of this type lead to moral hazard?
These aren't criticisms of your point as much as they are a request that we be very specific about what we are insuring against. The uncontrollable shock to savings is probably the most applicable event, but I would be cautious about explicitly providing complete insurance here.
By the same token, one problem I have with private accounts is that if the government offered a guaranteed minimum return (I would never suggest this, but it has been mentioned from time to time) then people are going to hold too much of their portfolio in risky assets (stocks) for too long (up to retirement). Moral hazard.
So what do you do? Should we mandate how the private accounts are diversified? That's not very libertarian! Do we let people sink or swim? That's not very liberal!
No matter how you slice it, it comes down to how much insurance do we want, how do we implement it, and how do we address adverse selection and moral hazard problems that get worse as the generosity of the insurance increases.
I admit having trouble seeing the "generosity" of the insurance.
I bought my house for $X. For whatever reason, the "value" of the house is now $Y, Y>X.
My house insurance (and the premium paid) has gone up to cover Y replacement cost.
Similarly, if I make SteadyMax-- currently $90K--I would receive current benefits of, what, about $30K/year ($2500/month) on retirement.
That looks like about 1/3 of the gross earnings, with rather little in reduced expenses (e.g., no commuting to work, but increased medical expenses).
It's difficult to see the "moral hazard" argument being viable when the standard of living from SocSec alone is explicitly not sufficient to maintain the current lifestyle--the rest of the premium having gone for Disability, Survivor and other benefits.
Ken,
If you're referring to my last sentence in my last comment, that sentence refers to various avenues to reform, not to the current system.
I'm saying that we (the voters, through the political system) can choose the level of insurance we want--more or less generous.
If you make the system marginally more generous, you would at least marginally increase the moral hazard.
But yes, someone making $90K in the current system probably doesn't experience much moral hazard.
So one point was about changing the system--choosing the optimal level of insurance. The other point is about being clear on what we are insuring.
I was going from your last sentence and juxtaposing it with your "appeasement" comment in relation to Dr. De Long's sentence.
Adverse selection doesn't apply if Barro is correct that changes to the composition of SocSec will result in rebalancing. (There is a high likelihood that shifting the SocSec TF into some combination with equities would reduce the ROI of one's portfolio, but that's a side discussion.)
So we're talking "moral hazard" primarily in that either (1) SocSec discourages savings by lower-paid workers [since we appear to agree that the moral hazard on High Earners and SteadyMax people is marginal at most] or (2) SocSec encourages people to make economically-adverse decisions by lowering their risk.
To the extent that (2) encourages entrepreneurial activity (e.g., a 35-year-old decides to take a few years to try to build her own business, knowing that she will be able to take a job in four or five years and not jeopardize her SocSec benefits), I'm inclined to believe that the benefits of that option indicate that no changes should be made.
So the question is if (1) is likely to produce a disincentive to work among the low- and medium-earners--and whether that is necessarily a bad thing if so.
We can probably agree that the most valuable aspects of SocSec that are not available in the market (and, arguable and I'd be willing to make the argument, cannot be provided on a cost-effective basis except by government) are the Disability/Survivor benefits.
Using Cato's assumptions that's about 20% (2.4/12.4 being close enough to 2.5/12.5 for discussion purposes) of the presumed cost. If we remove that "premium" from consideration--no one gets a refund on auto or homeowner's insurance just because they never make a claim--the ROI on SocSec becomes reasonable, but not excess.
Which is a very long-winded way of saying that I don't see any reason to change the system so long as it is (1) only generating 30-35% of pre-retirement income at best, (2) providing marginal but detectable entrepreneurial incentives, (3) providing insurance that is most cost effective if centralized.
To the extent that SocSec indirectly subsidizes Long-Term Disability insurance (i.e., does so to more of an extent than would be directly justified by the lower payout), there is certainly a skewed benefit in the system. But the benefit should be skewed only to the extent that slices of the market reduce the ability of the insuring firm to manage their risk by comparison with centralised insurance--exactly the inefficiencies that lead to profit potential in the first place.
The adverse selection I'm talking about is where only the high risk people want insurance. As far as the political appeasement goes, I think this definitely applies to what DeLong is talking about. If the Pozen plan goes through, the wealthy will want out and DeLong's comment seems to suggest (and I'm inclined to agree) that 50 or 75 years from now, they could potentially use their political power to do so. Then adverse selection kicks in with force. Only the high risk people are looking for the insurance. Bad outcome.
Your statement of the moral hazard is pretty much what I had in mind. However, I have been leaving DI out of the equation. In my previous posts, I have only been talking about the returns to OASI. The return to survivors is in there in an important way though. If you look at the return to women it is quite high as they are often the surviving spouses who paid in less to the system themselves. Many men actually get negative returns, but I would think that the household's returns are approximately an average of the two.
I don't think we're that far apart on the concept of what Social Security does or should do. I'm just trying to engage on the subject of how much of it is insurance and whether it is well designed as an insurance system. I haven't totally made up my mind yet on that. I do think that it would be unlikely if the present system is really optimal (and that depends on your social welfare function anyway). But the issue is still pension vs. insurance and what the government should focus on providing. In principle, I'd try to shore up the insurance and diminish the government's involvement in providing a pension. But that might only be possible in a perfect "blackboard" world where DeLong's concern about the Pozen plan can be ascribed to "simply politics."
But thanks for your comments. This is the discussion I want to have.
Admittedly it's not adverse selection in the traditional sense. But I've adapted the concept to apply to Social Security as insurance. The point is not so much about asymmetric information, but that the low risk individuals find the insurance too expensive.
Hi William:
Great post, and good comments thread.
I am still not at all clear what the insurance aspect of social security, since "aging" is not a risk, it is an inevitabiility, and the other risks notes (hurricanes, medical bills) aren't age specific at all.
I think that people should always be insured against things like hurricanes and medical catastophes, and if such a calamity impoverishes someone, then they should be helped because they are poor. Age does not come into it.
The only pure "old age" risk is the risk that you might live too long, but this can simply be dealt with using annuities. Again, upon retirement, the retiree would put all his savings into an annuity and have that be his "pension".
So what risk are we left with -- the risk that the retiree has not saved enough before he retired? I don't see how this is a risk. The retiree 1) chose when to retire and 2) saved throughout his life. This is not a risk, this is a bad choice.
I am supportive of (and understand) forced savings so ppl don't make the bad choice I've detailed above. But I cannot see the insurance aspect so social security. I don't understand why insurance against unexpected poverty should be limited to old people.
-winterspeak
winterspeak,
Good question. I will answer it tonight or tomorrow. (Tonight if I need a break from grading exams--tomorrow if I don't.)