Book Review: The Great Risk Shift (part 2)

This is the second part of my review of Yale political science professor Jacob S. Hacker's The Great Risk Shift. The first part can be found here.

Hacker sees the Social Security system as a clear "win" in insulating people from risk. He describes it in glowing language (p41):

The intelligence of insurance became genius when insurance principles were coupled with the power of the state to require participation …

The "require" part bothers me tremendously. I want the freedom to choose the the level of risk I undertake. I am hostile to mandatory programs. Their common fault is that they actually harm some people. For some specific examples, see my sixteen reasons why people should be allowed to opt out of Social Security. To force people to participate in a program that they know harms them is absolutely wrong and should not be tolerated. To praise it is horrifying.

Imagine the ridiculousness of a national mandatory weight-loss campaign, with the stated goal of improving everyone's health, but that had no provision for the underweight to opt out. These people would be harmed by the program. It may sound silly to suggest that such a stupid program could be created, but Social Security has exactly this structure, in forcing everyone to save for retirement even if a person rationally expects to die before they could collect benefits. (This is the first of my sixteen reasons.)

Mandatory participation is a flaw, not a feature. For a solemn underscoring of this point, you should learn how Social Security affected the Amish. The story is horrifying to read, and the resulting public outcry actually created the (much too weak) freedom to opt out of Social Security on religious grounds.

I was troubled by a passage in the book (p48-50) describing the phenomenon of moral hazard in health insurance. The economic arguments are stated briefly and reasonably, but followed by this:

Of course, this critique … was directly at odds with the conception of insurance that had emerged out of the New Deal. Social insurance … was supposed to provide subsidized coverage to high-risk groups and those who couldn't easily purchase commercial policies. Social insurance was supposed to protect beneficiaries from the intrusive and stigmatizing interventions into private conduct that had been so characteristic of assistance policies in the past.

This amounts to an admission that "social insurance" is a misleading name. Hacker wants forced charity in the name of insurance. He spills no ink combating the moral hazard arguments, and I can only suppose from this that he accepts them as true but thinks they're unimportant. A pity.

There are more misleading statistics in a passage about unemployment (p73):

The decline in labor force participation has greatly masked both the severity of the recent 2001 recession and the anemic "jobless" nature of the recovery that followed. This is because most of those who have left the labor force, or opted not to enter it in the first place, are not "actively seeking work" and hence are not counted among the formally unemployed. There is good evidence, however, that many of these potential workers would be in the labor force were the opportunities for them better. In 2005, according to an analysis by Katharine Bradbury of the Federal Reserve Bank of Boston, the total labor force "shortfall" — compared with similar points in the business cycle in the past — was as high as 5.1 million men and women. This amount would raise the official unemployment rate to 8.7 percent, a level not seen since the steep recession of the early 1980s.

Firstly, it is a truism to say that "many of these potential workers would be in the labor force were the opportunities for them better."

Secondly, Hacker's comparison of the official unemployment rate in the early 1980s to the modern official unemployment rate plus the labor force "shortfall" is invalid. Unlike measures are incommensurable. To have a valid basis for comparison you must either include the "shortfall" in both time periods or exclude it in both. (Here's a chart of the unemployment rate from 1970 through 2006.)

I wonder about cherry-picked dates on p122:

Over the fifteen years between 1983 and 1998, the typical family approaching retirement saw the wealth earmarked for its retirement decline, not rise. The stock market skyrocketed, 401(k)s exploded, but the typical family saw its retirement wealth drop.

Of course, 1983 was a year of major reforms to Social Security. These reforms were supposed to permanently save the system (heh). The reforms included benefit cuts, so I'm curious to know whether the early endpoint included the projected Social Security payments the government knew it wouldn't be able to make. It would be misleading to include the unaffordable projected benefits! Alas, there is no footnote in this section, so we'll just have to wonder.

There are more funny numbers on the same page:

We are told that the "average" American has tens of thousands of dollars socked away in a 401(k), but in fact roughly three-quarters of account holders have less than the widely cited average of $47,000. The median among account-holders — which is a better measure of what's typical — have around $13,000.

A better measure of what's important is the median among account-holders nearing retirement. The balances of thirty-somethings certainly pull down the numbers if they're included; they haven't reached their peak earning, saving, and compounding years yet! Further, keep in mind that 401(k)s have only existed since 1978. They haven't been available for the whole working lives of people nearing retirement today, so have had limited time to grow.

A few pages later (p124-125) is a revealing example of myopic thinking. Hacker disapproves of people who take lump-sum distributions from their defined-contribution benefit plans without rolling the funds into another tax-advantaged retirement account. "They do not roll over the funds despite the fact that they must pay income taxes on all their benefits, as well as a penalty of 10 percent if they are younger than fifty-five." Hacker notes that in many cases people are using their retirement savings as a buffer against economic insecurity (relocation, caring for a family member) and concludes that this behavior "leaves families more vulnerable to one of the greatest risks of all: retiring without adequate income."

And what of the fates of people in these situations who have Hacker's favored defined-benefit pension plans? He says nothing, so I will. The pension is inflexible and does not buffer them from economic insecurity during their working years. What would people do? Rack up credit card debt? Go bankrupt? The consequences could be horrible, but surely Hacker could console them with the news that at least their sacred retirement income will be intact! Sure, they might object that the well-being of their family in the present day is more important than their retirement lifestyle. But Hacker believes that people would be better off trusting his judgement than their own, and believes it so strongly that he would use the government to compel it against their will.

Another curious item is on p133 in a discussion of Social Security individual accounts:

… half of Americans who set up private accounts would fail to do better than they would under Social Security even if they used the conservative investment strategy (50 percent bonds / 50 percent stocks) suggested by the President's Commission to Strengthen Social Security.

My head is spinning. "Even if"?!? How about "because"? Holding half your funds in bonds is not a growth strategy. An allocation like that may make sense for someone nearing retirement but not for the several decades leading up to it.

I appreciated the following comment on the failed Clinton health care reform on p148:

One memo on Medicare by a top architect of the Clinton plan declared that "Medicare's entire history should be a lesson on how not to structure a national health program," ignoring that Medicare was the only national program the United States had and one that was overwhelmingly popular.

When a quasi-socialist (what else could a "top architect" of the Clinton plan be?) who shares the same policy goals as Hacker declares that Medicare is horrible, and Hacker meets it with the equivalent of "but it's popular!", I do not know whether to laugh or to scream. What the hell does popularity have to do with anything? Isn't it remarkable that someone ideologically predisposed to favor Medicare believes that its structure is flawed? This is another instance where I sorely miss a footnote; I would have liked to understand the concerns of the "top architect" — coming from a "Clinton top architect" they're probably politically bulletproof!

As I read the book I made approximately one hundred distinct notes on passages that struck me as biased, misleading, or just plain wrong. I simply do not have the endurance to present them all.