This NYT article about using behavioral nudges to get people to save more is fine, so far as it goes. But it’s another example of policies which go through an outrageous amount of complexity to encourage people to do the right thing, instead of doing the right thing for them.
There are a lot of potholes on the way to a happy retirement. First is to save enough. Second is to save in appropriately diversified vehicles (e.g., index funds). Third is to choose a mix of assets that’s appropriate for your particular stage of life — e.g., more stocks when you’re young, and more bonds when you near retirement. Fourth is to not outlive your savings; the way to do this would be to buy an annuity, but the annuity market (as I understand it) is not as well developed as the rest of the retirement-savings industry.
So then we (as a society) compile all of this knowledge about financial best practices, then try to convince people to use it, then outsource preparations for retirement to employers, then encourage employers to nudge their employees into doing the right thing. Not only is this ineffective; it’s infuriating and exhausting.
Until fairly recently, we tried another outsourcing approach: encourage employers to provide pensions (“defined-benefit retirement plans,” in the jargon, as opposed to “defined contribution” plans like 401(k)s). Pensions required employers to set aside money today for their employees’ retirement, but this of course presents a problem: what if the employer goes out of business before paying out those retirement benefits? To address this possibility, we built a regulatory infrastructure to try to regulate these pensions. If all else failed, we created a government agency as a backstop.
Somehow the straightforward approach of just providing strong retirement benefits to all Americans, then paying for those benefits with progressive taxes, hasn’t yet taken hold. In the U.S., “inflation-protected retirement annuity with survivor benefits, paid for through taxes” is pronounced “Social Security.” Social Security is expected to be part of a three-legged stool, the other two legs of which are pensions (nowadays IRAs and 401(k)s) and private savings. Those other two legs are increasingly weak: the overall national private saving rate has been declining for decades, and briefly went negative in 2005. That savings rate already includes 401(k) and IRA contributions, so one of the legs of that stool is practically nonexistent. As for the 401(k) leg: only 66% of private employers even offer retirement benefits, and only about 3/4 of the people who have access to them use them. I can’t find statistics on how much people contribute to such plans, but the amount must be less than the low overall savings rate. [1]
In this light, behavioral workplace nudges to get people to save more seem like a last-gasp rearguard effort. If we, as a society, believe that saving for retirement is important, why don’t we, as a society, reflect this belief in our policies? The standard nudgey answer is “libertarian paternalism”: set appropriate defaults on retirement savings, and allow people to override those defaults if they wish. It seems pretty clear that libertarian paternalism doesn’t work. Libertarian paternalism seems like nothing so much as resignation in the face of a hostile political climate; it resembles the early Obama administration, hoping against all evidence that the GOP would take half a loaf. “Conservatives and liberals disagree on a lot of things,” I imagine libertarian paternalists saying, “but surely they’ll agree on market-friendly solutions like behavioral nudges.”
Even if these nudges worked, Social Security would still be better. First, there are enormous economies of scale from administering Social Security centrally rather than outsourcing pension management to millions of employers. Second, those employers can’t be expected to be any good at choosing retirement options for their employees. My employer makes software; it doesn’t make retirement funds. Why would we expect my employer to be any good at offering 401(k)s? Why would it want to offer 401(k)s? It wants to make software and it wants to lure talented employees with high salaries; all else is noise.
Social Security isn’t perfect. It’s not progressively taxed, for one thing: it’s a flat tax, and you don’t contribute on any dollars you earn above $118,500. It’s regressive on what’s paid in, but progressive in what’s paid out: higher earners can expect to get back a smaller fraction of what they paid into Social Security than lower earners do. And it only pays out $1,335 per month on average, which still amounts to 39% of elderly people’s income. So it needs to be more generous and more progressive. But it’s a start, and the infrastructure is already in place. Removing the taxable maximum would be a hard-fought battle, but would be comparatively easy to implement once we’d made the decision to do so.
I like to imagine a thought experiment. Back in 1970, I imagine someone told my parents (who weren’t yet parents) that in 50 years they would be nearly four times as wealthy as they were back then. My imaginary interlocutor would then ask my parents what to do with that windfall. They might have felt perfectly well off with the income they were earning back then, so the thought of quadrupling it might have seemed outrageous. Maybe they should set some of it aside for retirement? Set aside even more of it and allow themselves to retire early? How about setting aside some of it to provide the world’s best universities, for free? There’d be a long list of choices they could make. They could choose to buy a larger house, though maybe they’d look around at the house they have and think, “No, this is a fine size house; we’ll choose to spend our money on other things” (unlike Americans as a whole, whose homes are 42% larger at the median than they were in 1973).
Like Odysseus tied to the mast, every decision that my parents made in their mid-twenties would bind them in their working years and on into retirement, and the binding would be handled through the tax code. Would they lament the money that never made it to their wallets, but rather was spent on social goods? Rather than after-tax paychecks that were on their way to quadrupling, they’d get free university educations for their children, and they’d have enough money to retire comfortably. A lot of the arms races that we fight with our neighbors would never have been fought: rather than build a larger house simply because that’s what everyone around us does, and because we must keep up with the Joneses, the tax code would disarm everyone at once. To the extent that we build larger homes out of an arms-race mentality, rather than because any individual person wants a larger home, this multilateral disarmament would help everyone.
Of course this story isn’t complete. For one thing, real per-capita disposable personal income doesn’t capture the story of rising income inequality over the last 30-40 years; the story of American growth hasn’t been a story of rising tides lifting all boats. A more accurate measure might be median household income, which has barely budged since the early 80s. This may, in fact, strengthen my story. Which do we prefer, as a society: quadrupling our income, but putting most of that extra income into a few hands, or sharing it more broadly and investing in our future through strong public universities, public-health spending, and basic research that helps everyone?
The story of increasing wealth may also be missing a key component: health care. Maybe our parents would have loved to have set aside money to build for the future, but they didn’t have that choice: much of what would have gone into their pockets as increased wages went, instead, into health-care costs borne by their employers. First, I have my doubts that this was actually a problem: the real disposable personal income number already factors in the health-care CPI. And second, the fact of rising health-care costs in the U.S. results, at least in part, from our fragmented system of care. That is, rising health-care costs have been very much a social choice. Imagine again that our parents, bound like Ulysses, were asked in the early 70s to make a choice: organize medical spending in a deliberate way (à la the VA medical system, or à la Medicare), or continue with the chaotic and spectacularly inefficient way we’ve organized it.
Barney Frank is supposed to have said that “government…is simply the name we give to the things we choose to do together.” Let’s get back to thinking of things we can do better together, as a society. Social Security is a good start, but it’s just a start.
[1] – I don’t know whether home equity is counted in the NIPA definition of savings. Even if it is, I’m a little confused about how owners’ equity is counted. How could equity be cut in half and then return to trend? In any case, equity per capita only hit a peak of about $45,000, so it alone won’t rescue Americans’ savings rates. And we can’t compare savings rates to equity amounts; that’s comparing a flow to a stock. The relevant comparison would be either the rate of change in equity to the rate of change in savings, or the total amount of savings to the total amount of equity. I can’t figure out a quick way to do the former; the latter doesn’t seem particularly valuable.
Steve Reads 