Welfare as Poverty Insurance

Most of the arguments I've heard in favor of "the welfare state" (i.e. a government guarantee of a basic standard of living, a.k.a the social safety net) appeal to compassion, or justice. This works for people who are already sympathetic to such appeals (such as myself), but it really does no good for those who don't put much stock in loose and subjective ideas like justice and compassion to begin with. Fortunately, this paragraph from University of Arizona Professor Lane Kenworthy's blog suggests an alternate—and, I think, fruitful—approach:

As people get richer, they tend to be willing to buy more insurance and more services. We observe this both among individuals and among countries. Some insurance and services are provided at good quality and price by private markets. Others less so. That’s the underlying reason why nations have tended to expand social policy as they grow wealthier.

Once we start thinking of welfare as insurance rather than charity, it opens a whole new series of possibilities. Most importantly, it gives us a way of addressing opponents of welfare programs (mostly conservative economists and their students), on their own terms, by providing an account of welfare programs that appeals not to any sense of morality or altruism, but rather rational choice, fully consistent with the idea of individuals maximizing their well-being.

Even though I myself don't fully believe in the conception of rational choice as self-interested utility maximization (for the same reasons that Amartya Sen doesn't either), I'll make an attempt here to defend the welfare state on this premise. This short post, of course, cannot hope to cover all the relevant issues. But I'll try to make a good sketch of what such an argument might look like.The exposition of the argument (below) makes four claims that collectively show how government-provisioning of poverty insurance is both consistent with rational choice and certain principles of liberty.

Claim #1: It is sensible to think of welfare policies as poverty insurance.

Birth is a kind of lottery. By random luck, and through no fault of our own, we end up either in a poor household, or a well-off household. Every one of us, when we are born, stands an equal chance of landing in a poor household. Therefore, we all stand to gain from a welfare program that would make the poverty that we would potentially face less dire.

Even after we are born, we face a non-trivial chance of becoming poor in the future. Upward income mobility is a blessing for poor people, but downward income mobility is a significant risk for the non-poor:

Most policy-makers view poverty as akin to a stagnant pool which may be slowly drained as each individual or family trapped there is helped to climb out. What this imagery and the policies it suggests disregard, however, is the fact that of those who are poor in any particular year, most were not poor in the preceding year. In other words, at any given time, many of the poor are newly poor. According to the University of Michigan‘s Survey Research Center study, Five Thousand American Families, the “persistently poor” (defined as those who were poor during all five years studied) amount to only 9 percent of the total poverty population. [link]

If it's not us that falls into poverty, then it is perhaps our children who will. Most often sons do not end up in the same income bracket that their fathers belonged to:

This table (source) shows intergenerational mobility between income quintiles in the U.S. Class 1 falls is the bottom quintile (the 0 to 20th percentile) and Class 5 is the top (the 80 to 100th percentile). So even if you're born into the top of the distribution, you have a 9.5% chance of ending up at the bottom. That's probably much higher than many of the other risks that we insure against. (Lest one underestimate what it means to land in the bottom quintile: according to this U.S. Census Bureau report, people in the percentile made $17,984 in 2003.)

So everyone, by the lottery of birth, faces a risk of facing poverty. Above this, people who are born into non-poor families face the risk of falling into poverty; or those born into poverty who manage to exit it face a non-trivial risk of falling back into poverty. The more severe the poverty, the more anxious people are to mitigate these risks; hence the need for insurance.

Claim #2: Government can provide poverty insurance more effectively than the private sector.

In a private poverty insurance market, I imagine that parents would buy policies that would cover themselves and their children. The premiums would probably be determined by a number of factors, such as general macroeconomic conditions, and also the policy holder's likelihood of falling into poverty. Rival plans would compete on, perhaps, how many generations they cover, or how generous the benefits are.

Now, why couldn't such a market work? In short, for many of the same reasons that completely private health insurance doesn't work (or more accurately, unpalatable). Prof. Kenneth Arrow already showed in the 1960s why the market for medical care has certain characteristics that make for "the laissez-faire solution for medicine [to be] intolerable." A market for poverty insurance would suffer from many of the same defects. Specifically:

1. Moral Hazard. As Arrow notes, "What is desired in the case of insurance is that the event against which insurance is taken be out of the control of the individual." But this is clearly not the case with poverty insurance. An individual who takes a poverty insurance policy can afford to be less careful about staying out of poverty, perhaps to the point of abuse. Insurance companies, of course, will write contracts that stipulate that they can deny coverage in the case of recklessness, but it's very difficult for the insurance company to make a reliable decision about the degree of culpability for someone who has fallen into poverty. As with health insurance, the lack of reliable procedures in a poverty insurance business will lead to mistakes: some people deserving of insurance will be denied, and some who should be denied will receive the benefits.

2. Adverse Selection. The people who are most likely to get covered are the ones who need it least. In the same way health insurance companies screen out people with unhealthy lifestyles, or with a predisposition for disease, poverty insurance companies will likely vet out people who seem likely to fall into poverty (being an artist will be the new preexisting condition).

3. Other-regarding preferences. Just as we are concerned about the health of others, we also care about their living standards. People prefer a healthier society to a less healthy one, and a less poor society to a more poor one. In economics, we say that health and poverty are externalities: that is, an individual's private good health or good fortune is not just a benefit to himself, but also seen by others as a benefit to themselves. In the face of positive externalities, the market will provide a level of insurance below the optimal amount (and significantly so, since our preferences for the well-being of others are generally quite strong).

Of course, showing that a certain market has defects does not necessarily imply a need for government provisioning. The electronics market, for example, also contains asymmetry of information, a market defect, but there's no one is calling for government intervention there.

The key difference is the severity of the consequences. In the electronics market, the worst that could happen is to leave the store dissatisfied. Severe poverty, however, can lead to death—or, at the least, it inhibits an individual's ability to receive the nutrition and education necessary to reach her potential, which results in a loss not just to herself but to all of society. That is why failings in the poverty insurance market would take on a certain moral urgency, and would be much more harshly criticized than failings in other markets.

Government, unlike private companies, is in a position to address most of the problems caused by the three aforementioned defects of a poverty insurance market. By offering automatic universal coverage, for example, the government can circumvent the adverse selection problem altogether. Universal coverage would also provide a level of welfare provisioning more consistent with our other-regarding preferences.

Admittedly, universal coverage does not solve all problems. Specifically, universal coverage by itself does not change the moral hazard of insuring another person for risks he has some control over. This is part of a major source of the frustration over welfare that lead to the overhaul of welfare in the U.S. in 1994. Welfare reform was meant to address the perception that welfare made people lazy, or went to undeserving candidates (such as so-called "welfare queens"). These are legitimate concerns, and they require that the government, like any private insurance company, craft its policy carefully. But then we enter the debate of what kind of poverty insurance we should have, and we're no longer discussing whether or not we should have it in the first place.

Claim #3: The majority of people in a sufficiently affluent society will rationally choose to vote for some level of poverty insurance.

Whether or not people would freely contribute to a poverty insurance program is not the relevant kind of rational choice in this case. After all, the main risk of poverty comes from the lottery of birth, so it doesn't make sense to collect payment for insurance after the main risk event has already happened. A welfare program needs the money of rich people in order to work, but since the rich have avoided the main risk (that of being born poor), they would significantly undervalue the benefit of poverty insurance, and hence underpay its true value.

In order for people to truly gauge the full risks of poverty, we would need to ask them how much they would be willing to contribute before their position in the income distribution is decided. This pre-birth state is similar to Rawls's idea of the "original position," except in this case I allow information about the structure of the society, such as the income distribution, the degree of income mobility, and the like. From there, then, just as in Rawls's original position, people in this pre-birth state would try to advance their own best interest, but in good-faith, knowing full well that whatever position that is adopted there would be binding once they find out their place in the income distribution.

What stance towards welfare would participants in this pre-birth position adopt? Unlike Rawls, I don't think that there will be unanimous consent, at the very least because people face varying levels of aversion towards risk. Since we don't have unanimity, it makes sense then to take a vote. The result of the vote will fall according to the preferences of the median voter.

Of course the choice of the median voter will depend on the factors specific to the country: the level of income mobility and the severity of the poverty, as noted above, but also the national income the degree to which the culture values equality.

In the developed world, I believe that the median voter will almost certainly opt for a level of poverty insurance greater than zero, for the very reasons that Prof. Kenworthy pointed out at the beginning of the discussion: as people grow richer, they become more willing to spend their money on insurance (that is, they would be willing to forgo a higher GDP in exchange for a welfare program, or they would be willing to pay the taxes necessary to fund the program if they were wealthy). In fact, it is probably even rational to do so: since individuals face diminishing returns to wealth and income, they would likely face a higher expected utility, given the right conditions, from shoring up their income in the event they end up poor rather than betting on making even more money if they end up rich.

Claim #4: Collective decision making does not violate the liberty of dissenters.

There will likely be some participants in the pre-birth position who will opt for no poverty insurance. But since the majority carries, and the majority will most likely opt for some non-zero amount, these dissenters will be forced to contribute to a program that they do not want. How is their liberty not violated? (Note that dissenting to contribute in the pre-birth position is radically different then dissenting to contribute in real life.)

This question is similar to a problem that Ronald Dworkin has posed in his article "The Majoritarian Premise and Constitutionalism" (In Philosophy and Democracy: An Anthology, edited by Thomas Christiano, pgs. 241-257): In a democracy we like to think of ourselves as a self-governing people. But how can that make any sense if you're on the losing side of the vote? That is, how can one reasonably think of himself as self-governing if he has to follow the majority will, which may not be his own?

The answer, Dworkin says, is that it depends on what view of collective action you take.

Now, government is required because there is a need for collective action. But Dworkin notes that there are two views of collective action: one, which he calls statistical, is where the group action is just some agglomeration of individual action with no sense of acting as a group. The example he gives is the stock market; when the "stock market" falls 20 points, there's no concerted effort behind it. The second kind of collective action is communal, where the group action cannot be understood as anything but as a group. We cannot make sense of a member of an orchestra, for example, except by considering her role in terms of the whole piece.

Dworkin thinks this distinction is very important. The only way for a person on the losing side of a vote to still think of himself as self-governing is if he adopts a communal view of the collective action of voting. He needs to think of himself as a genuine member of the political community, such that collective act of the group is also my act as well, in the same way that the victory or defeat of a sports team is the victory or defeat of each player, even if their individual contribution did not make a difference one way or another.

Now, to be a genuine member of a democracy, what Dworkin calls a "moral member," certain conditions must be satisfied. People need to be a part of the decision-making process (i.e. they should have a vote, a say); they need to have a stake in the decision (i.e. they need to consider themselves as a respected part of the group, and other people should also see them as a respected part of the group); and they need to be morally independent of the decision (i.e. they should be allowed certain private space to make personal decisions). Dworkin explains these concepts in more detail in the article. The upshot, though, is that once one is a moral member, one is then, in fact, self-governing. There is no meaningful loss of liberty when a moral member ends up on the losing side of a fair and inclusive vote on a problem that requires collective action (and poverty insurance is one of those problems).

The pre-birth position described in section three is a position of perfect equality, set up in such a way that it already establishes all the conditions necessary to make all parties a moral members of the political body. Therefore, even though some individuals in the pre-birth position would perhaps vote for zero poverty insurance, when the majority passes a non-zero level of insurance, they cannot claim that they were unjustly or unfairly treated.