Sorry, Folks: One Way or the Other, You'll Never Be Able to Completely Count on Retirement
Social Security. Corporate Pensions. 401ks. They're all risky, and that risk isn't going away.
For over 100 years, the Studebaker Corporation made vehicles, from wagons for the Gold Rush to tracked vehicles for World War II. And of course, the cars that we mostly know them for: streamlined, a little stodgy, and very much of their era.
That's because their era ended, in 1963, when the company closed its South Bend plant. That wasn't a surprise; sales had been suffering for years, due to high costs and concerns over reliability. What was a surprise was the state of its pension fund, which was disastrously underfunded. No, wait--hold that rant about greedy bosses looting their pension plans unbeknownst to the innocent, victimized workers. The UAW, which represented Studebaker's employers (some of the highest paid in the auto industry, by the way), had not only allowed the company to stretch out its payments into the fund, but had arguably actually encouraged it, because the alternative was lower wages.
Nonetheless, workers were devastated. A deal was worked out whereby those in or very close to retirement got some of their pensions, those over 40 got a fraction of the actuarial value, and everyone else got nothing. Unions pressed Congress for some sort of government solution. And 10 years later, we got ERISA, which regulates private pension plans, and also insures them, through the Pension Benefit Guarantee Corporation. The idea is that sound regulation and government-backed insurance would prevent pension risk, in much the way that regulation plus insurance had stabilized the banking industry. For a while, it seemed to be working. But then the risk began popping up in other places.
We shouldn't have been surprised--however much financial planners may boast of reducing risk, it is rare for financial risk to actually disappear. What we do is move it around, maybe transform it a bit. But the risk always remains. Unfortunately, the transformation is often good enough for us to believe--temporarily--that the risk has gone away. That is the recurring theme in the history of retirement planning--and, I'm afraid, this article. There is no system that gets rid of retirement risk. It just shifts who bears the risk. And don't think that there's some easy political fix, because the same political incentives that have stymied sound regulation of whatever system you want to fix, will also shape whatever new system you want to create.
To see what I mean, let's return to the regulation of defined benefit pension plans, the shiny new system that was supposed to prevent the sort of catastrophe that had happened to Studebaker workers. ERISA mandated that companies had to keep their pensions funded at all times--if a company had a shortfall, they had to make it up immediately, and no, we don't care if the union said it was all right. But this, it turned out, created a new problem. The assets in pension funds tended to fall precipitously during recessions. So, of course, did company profits. So the law demanded that companies put millions of dollars into pension plans just when they were least able. Pushing companies into bankruptcy wouldn't do anyone any good: the PBGC would have to make up the shortfall, the workers would get less (because the PBGC makes them take a haircut), and of course, the corporate shareholders would lose their whole investment.
Companies could have dealt with this problem by overfunding during boom years. Companies and governments in 1999 should not have been assuming that growth would continue at 8% (or higher!) forever. As granny used to say, the wheel goes round and round, and sooner or later, the fly on top will be the fly on the bottom. A sound pension should plan for the time on the bottom, not extrapolate from the moment on top.
But even if they wanted to do this (and I'm not sure that many did), there were structural reasons that they couldn't. Instead of requiring them to behave sensibly, the government required them not to.
Overstuffed pension plans were often an attractive target for LBO operators, who would "unlock" the cash (and pay it out to bondholders or themselves). Worse, they became a target of the IRS. An overfunded pension fund can, in slightly unscrupulous hands, be used as a tool for tax management: stuff the funds in during very profitable years, take them out later when you want them. The IRS takes a dim view of such maneuvers, and therefore essentially forces employers to stop contributing to overfunded defined benefit pension plans, or add new benefits. This is why, in 2001, companies and governments found themselves with such massive holes in their pension plans.
Holes that still haven't been filled, because the companies that have these pensions are mostly older industrial firms at substantial risk from low cost competition, at home and abroad. No one wants to risk the whole show by forcing them to top up their pensions too fast. At that, they're in better shape the public worker pension funds, which have historically been regulated with much laxer standards for
The PBGC is now itself underfunded. Why? For the same reason that it hasn't been able to get companies to top up their pensions. In theory, the PBGC should work like a normal insurance company: riskier companies should pay higher premiums. In practice, who are your riskier companies? The ones on the verge of bankruptcy, with large, unfunded pension obligations. If PBGC makes them pay the actual, actuarial cost of their shortfall, that might be what pushes them over the edge--at which point, that unfunded liability now belongs to the PBGC.
A private insurer might simply walk away and refuse to insure the risk, but the PBGC can't. So what do they do? They charge lower premiums than they should. They are abetted in this by Congress, because virtually every district contains one shaky company with a Sisyphean gap between the assets and liabilities in its pension fund.
Public sector pension funds are accruing even worse disasters: the accounting standards are much laxer, politicians have a history of giving "free" gifts to their supporters in the public sector unions, and legislators, unlike corporate directors, can be pretty sure that no one's going to prosecute if they just vote not to make the pension contribution this year. So we came up with another solution: IRAs, 401(k)s and similar "defined contribution" plans. These plans definitionally can't be underfunded, because there's no promise to pay. You get whatever is in the fund when you plan to retire.
It sounded great before we did it. Now that the first generation to use these plans is hitting retirement, it turns out that there's a problem: most people didn't save enough. Or they saved in risky assets that underperformed the market. Or maybe they sensibly bought index funds, but they didn't realize that on January 2nd, 2013, the S&P 500 would close at basically the same level as it had in March of 2000.
All of which is to say that we didn't get rid of the risk: we transferred it. Companies used to bear all the investment risk; now people do. And those people aren't necessarily well-positioned to handle it.
Josh Barro argues that we should therefore give up on the idea of investing for retirement, and help people rely more on Social Security. I'm going to excerpt more generously than I usually do, because it's an interesting argument, and I want to present it fairly:
It doesn’t make sense to finance retirement in such a risky way. Retirement savings exist disproportionately for the benefit of people with low or moderate means and a relatively low tolerance for risk. If retirement assets were invested safely, they would not be expected to grow faster than the economy as a whole.
So 401(k) and traditional pensions are both just efforts to finance retirement on the cheap by taking on excessive risk. The problem created by risk manifests itself in different ways with the different vehicles.
With 401(k)s and IRAs, retirees tend to invest in equity-heavy portfolios and then get hammered if they have bad luck in the stock market. This leaves retirees with far less wealth than they expected, needing to work longer or significantly lower their standard of living. Even worse, recessions that reduce the value of retirement assets will also tend to hit wage income and home equity.
The risk situation with pensions is more complicated because pension payouts to retirees are guaranteed regardless of the plan's performance. But just because investment risk is taken off the retiree does not mean that it goes away. In the case of a corporate pension, investment risk is shifted to the employer; for public employee pensions, the risk goes to the government and taxpayers.
Laws over the last four decades have made private-sector pensions more secure, but in doing so such laws have also increased the apparent costs of pensions. Private employers must calculate such costs on the basis of the risk borne by the employee; they cannot write down the cost because they expect high returns on risky investments.
That is, private pensions no longer rely on the premise that retirement can be made cheaper through investment in assets that grow faster than GDP. But such a free lunch was what made the plans attractive for employers in the first place, and as employers have faced the plans' real costs, they have increasingly eliminated them.
In the public sector, the free lunch lives on in the financial statements of pension funds. Governments fund their pensions based on an expected rate of return on a risky portfolio of assets, most commonly between 7.5 and 8 percent a year, far above the roughly 5 percent growth path we might expect for nominal GDP.
When fund assets underperform, taxpayers must make up the difference. In essence, governments are writing insurance policies to their employees that pay out when the market does badly and collect when it does well. This is not a “win some, lose some” situation, because a government’s overall fiscal fortunes are closely correlated to the investments pension funds make. The windfalls will come when they are least needed and the excess costs when they are least affordable. In the past few years, required pension contributions have spiked at the same time that state and local governments faced other fiscal distress.
The mistaken idea that we should finance retirement with especially risky investments has led to major errors with both types of retirement vehicles. With individual accounts, people undersave in part because of the expectation that small equity investments upfront will produce big retirement income due to high returns. With government pensions, politicians overpromise on benefits because part of the cost is hidden.
Social Security is also based on a bet about future economic performance, but it’s a much more reasonable bet. Forget the trust fund -- Social Security is based on a bet that the payroll tax base and annual benefit payouts grow at approximately the same pace.
In practice, that’s not quite happening: As the share of the working-age population shrinks, the growth of benefits is outpacing the growth of taxable wages. But funding gaps in Social Security open up much more slowly than in pension funds and 401(k)s. The roughly 16 percent funding gap we must close in Social Security over the next 75 years is easily manageable compared with shortfalls in public employee pension plans and the woeful insufficiency of 401(k) balances: just $42,000 for the typical middle-income retiree.
Oddly, one of the top virtues of Social Security is that it is unfunded. When there is a pool of money sitting around to finance retirement liabilities, there is always a temptation to chase yield, because if you can achieve a high return, you don’t need to set aside as many assets.
But Social Security isn’t any ordinary unfunded liability. In most cases, an unfunded liability is an even bigger invitation to mischief than a funded one. Although pension funding rules allow politicians to understate the cost of benefits, not prefunding at all can allow them to treat the cost of providing benefits as zero. This is how state and local governments have found themselves under more than a trillion dollars of health-care liabilities to retirees, far more than they ever realized they were promising.
Social Security stays unfunded and sustainable because of two of its other features: It is universal, and it has a dedicated tax source. If an unfunded benefit grows to have a surprisingly large cost, but it only covers a small share of the population, the government is likely to end up simply passing that cost along to taxpayers. That’s hard to do if almost everyone is a participant. Social Security’s universality gives politicians an incentive not to overpromise and individuals an incentive not to overdemand.
Although Social Security could theoretically be financed through general revenue (and was, in part, during the payroll tax holiday in 2011 and 2012), there is strong political pressure to ensure that Federal Insurance Contributions Act tax revenue and benefits paid stay in line over the long term. This means that any major benefit increase would have to be accompanied by a FICA tax increase, ensuring that benefits do not rise above a level that is fiscally and politically sustainable.
There are a lot of minor objections to this. For example, the 16% average shortfall is an artifact of the accounting process. Right now the difference between payroll tax revenue and Social Security's expenses (including checks) is about 10%. It goes to about 25% in 2033. Since pension accounting of any sort weights near years more heavily than later years, we get an "average" shortfall of 16%. But since the Social Security Administration "saves" by lending the money to the US government, which then spends it, this is interesting only to accounting nerds. In 2033, we will have to find 25% of the programs' revenue somewhere. That may be a small funding gap compared to the more freewheeling state pension systems, but it will not be small to the US taxpayer, or to the beneficiaries.
But these are minor sideshows compared to the two big ones. The first is that Josh is muddling together two problems: investment risk, and savings risk. Investment risk is the risk that your savings will underperform; savings risk is the risk that you'll save too little.
Barro writes as if the main problem with pension and 401(k) underfunding is that people are dependent on volatile assets that grow faster than GDP, and are therefore unpredictible. But most of the problem in 401(k)s and public sector pension plans has nothing to do with this undeniable truth. Most of the problem stems from the fact that companies and individuals simply didn't save enough. Yes, their undersaving was abetted by rosy projections, but after covering both the state pension crisis and personal finance, I am pretty skeptical that the rosy projections caused the undersaving. Rather, in most cases the rosy projections were chosen in order to enable the undersaving, in the same way you enable yourself to eat that delicious dessert right now by assuming you'll eat nothing but cocktail onions tomorrow.
If most people were putting 15-20% of their income into savings, they would be fine. But most people weren't. They were putting in 4%. Or 0%. To this day, when I tell people that they need to save 15% of their income, I am greeted with the same sort of disbelief that might greet a suggestion that they move to Tanzania and become big game hunters. Most of these people do not even have anything so firm as a rosy projection in their head; they just think they're supposed to be able to start tucking away a modest portion of their income, say 5%, sometime in their forties, and then retire comfortably.
Moving people to higher Social Security benefits is a much better solution for the undersaving problem than for the "equity returns are volatile problem". But there are many other solutions which are just as good, like Denmark's quasi-mandatory system which forces almost everyone to put 9-15% of their income into a defined contribution plan. Average assets in these plans total almost $200,000 per person.
To be sure, even this may not work--people might run up more debt in the expectation of future pension benefits, undoing all your hard work. Arguably, that's why debt soared and savings collapsed as house prices rose. I don't think it's particularly likely, but it's worth mentioning. However, since social security can have similar, but even worse effects (as we'll discuss in a moment), it's probably not worth worrying about too much.
So back to undersaving. Even once you've accounted for undersaving, Josh is right: there is some residual investment risk in private pensions and saving, which is not present in the same way in social security. But this doesn't prove as much as he suggests, because--and at this point, I expect, you can just chant along with me--you don't get rid of the risk, you transfer or transform it.
The first thing to point out is that the general fund--which is what we're relying on to fund that 25% gap--is actually extraordinarily sensitive to changes in asset prices. You would think it shouldn't be, since after all, the general fund is where we put our income taxes. But those income taxes are highly progressive. No, no, don't shout at me--I know you think that we don't tax the wealthy enough. But that is neither here nor there; in the technical, economic sense, a "progressive" tax is one that falls much more heavily on the wealthy than the poor. Whatever you think of the overall shape of the tax code, the income and corporate taxes, which provide the bulk of revenues to the general fund, are actually much more progressive than most other countries'. In our system, the top 1% of earners pay 39% of the income tax.
Here's the result: while payroll tax collections stay pretty steady during recessions, general fund tax receipts fall by much more than GDP.
The post 2009 payroll tax numbers are affected by the tax cut the administration pushed for to stimulate the economy; without it, the payroll tax collections would have stayed pretty steady.
But what about the Bush tax cuts? You are probably thinking. Didn't they push down revenues in 2001 and beyond? Yes, they did. But the CBO has estimated that these accounted for less than half the change. Most of it came from the crash of the dotcom boom.
As of now, we are plugging the growing gap in our social security fund with tax revenues that have exactly the same problem as the assets in pension funds: they grow faster than GDP when times are good, and slower when times are bad. Now, Josh suggests that we could fix that by increasing our regressive payroll taxes, which he thinks is the natural inclination of our legislators. I'm afraid I don't see that; over the last 20 years, those legislators have made the tax code much more progressive, and both parties seem pretty firmly united in opposition to hiking taxes on the middle class.
In most recession years, this is not a problem: the government can borrow. But at some point, we may not be able to borrow, or may not be able to borrow at attractive interest rates. Certainly, the closer that debt gets to 100% of GDP, the larger this risk looms.
So now we've got another pension plan that whose "asset base" overperforms in good years, and underperforms in bad ones. What happens if the government hits the wall? Do we raise taxes in the middle of a recession? Cut benefits? Or borrow at usurious rates? All three of these things risk pushing us into financial crisis in which taxes and benefit cuts and borrowing go up even faster.
In short, we have exactly the same risk as a company: that our obligations will exceed what we can possibly pay. We've transferred it, but it hasn't gone away.
To be sure, we've also transformed it. For pension plans, this risk is idiosyncratic: every year, some pensions will fail, but most will continue to pay out. For social security, this risk is systemic: most years, there will be no failure. But when there is a failure, it will hit everyone. And in many ways, this will be much worse than individual failures, because it doesn't just shift the risk, but actually makes it worse:
1. Social security encourages behavior which undermines the actuarial soundness of social security itself. There's substantial evidence that the more generous your social security system, the fewer kids your population will have, which makes it harder to pay out generous social security benefits. We are now counting on having each retiree be supported by just two workers, which is not a very safe bet.
Social security also encourages people to retire as early as possible--the majority of retirees retire before the age of 65.
Of course, that latter point is also true of private pensions. But if IBM's pension plan encourages its salesmen to retire a couple of years early, this has basically no effect on the value of IBM's pension fund (provided that it is sensibly and legally invested in something other than IBM stock.) But if people leave the workforce at 62 instead of 66, that actually has a substantial impact on social security's finances, because that's four fewer years of someone paying into the system.
2. Idiosyncratic failures are less likely to be catastrophic. It's very bad if your pension fund collapses. But you may have a sibling whose pension fund has not collapsed, or a spouse. Your children have a spare bedroom. You have social security benefits. And unless you live in a company town, you can probably pick up a part-time job--and if you do, you can move to one. But if everyone loses half of their social security benefits, your social network is probably suffering just as badly as you are.
3. Social security shrinks the investment pool. I suppose this is actually just an example of the first point, but it seems worth mentioning on its own. Making social security more generous lowers the incentives to save, even as it depresses investment in more children. The only hope of supporting 30% of our population in retirement is for productivity to grow very fast, and the best hope for that is investment. But the government doesn't do this sort of investment very well (for all that I am a big fan of the NIH, it will not, by itself, transform the economy). So beefier benefits mean that you need more investment to make up for the falling birthrate--and instead get less of it. Or depend on foreigners for all your capital, which creates all sorts of problems in the future (for all that I am, again, a big fan of open economies and FDI).
Say it with me one more time: you don't get rid of the risk, you transfer or transform it. We transformed idiosyncratic risk into systemic risk. We reduced the volatility of most years, by increasing the "tail risk": the chance of a rare, catastrophic event that will be insanely destructive, at just the time when retirees and their families will be least able to absorb it. You can see this happening in Greece now, where all the systems that are dependent on government payments are basically paralyzed. This includes things like pharmacies, which sometimes don't have the cash to order new drugs, and oops! that's a bit hard if you're diabetic.
This is why most pension experts look to what they call the "three legged stool": some sort of basic social security system (America's is, believe it or not, much admired for having fewer perverse incentives than the systems in most countries); some sort of employer based system (defined contribution is becoming the standard everywhere, not just here) and some sort of employee savings. That's because each of these has different weaknesses and strengths.
Employer pensions offer effortless savings (just deduct it from the paycheck), moderate administrative costs, and can offer expert investment management advice. Defined benefit systems can shield employees from investment risk, but add the risk of catastrophic, unplanned-for failure that may even take the company down with it.
Government social security systems rely on a much larger asset base (the tax revenue of the whole country), and can prevent idiosyncratic investment or pension failures from being catastrophic. But they increase the tail risk that you will have your benefits cut off at the worst possible time--especially since they create perverse, system wide incentives to have fewer kids, fewer savings, and shorter work-lives.
Individual savings eliminate the perverse incentives. If the market's down, work longer while it recovers. If asset prices are underperforming, better put more money in now. But they're much more volatile than governemnt or employer pensions: if I happen to have to retire in a down market year, I'm in trouble. And they're more vulnerable to timing issues. For example, I missed a lot of prime savings years in my late 20s and early 30s because I was first, in grad school; and then unemployed; and then working for literally barely enough money to eat and pay my student loans. That means I have to save more now--a lot more.
But while we can't get rid of the risk, we can diversify it, so that we're bearing some of each kind of risk: a little less idiosyncratic risk, and a little more tail risk, not all one or all the other.
The problem is, whatever sort of risk we're not having right now looks like the most attractive one. In 1980, when corporate pensions were melting down and social security was approaching a funding shortfall, the risks of individual savings looked more enticing and everyone (including, cough, me) thought that this is how all retirement should be funded. Now that 401(k)s are in trouble, people (including Josh Barro) think Social Security or the old-style corporate defined benefit pension has a more attractive package.
But in the end, they're the same package. There's no way to take the risk out of betting on the future; by the time you can predict the future accurately, it's already the past.
We're just picking how we want to take our risk, not whether we want to take it. And if there's one thing we should have learned form the financial crisis, it's this: the minute we decide that we don't have to make that choice--that we have figured out some way to get rid of the risk altogether--is generally the moment that the universe decides to give it to us, good and hard.