ECONOMISTS have two contradictory impulses when analysing policy. The first is to ignore its effects on the distribution of income or wealth, and argue that policymakers should shoot for efficiency and worry about redistribution later. This instinct often surfaces in discussion of policies like carbon taxes. These increase efficiency by raising the price of pollution, a harmful activity that, without intervention, comes too cheap for the polluter. But carbon taxes tend to hit poorer people, who spend a higher fraction of their income on fuel, the hardest. We should introduce the tax anyway, economists say, and re-distribute by other means.
This first impulse is loosely based on the “fundamental welfare theorems”, which say that whatever the initial distribution of wealth, trade will lead to an efficient outcome. The best way to redistribute is not to interfere with trade, but to divvy up “endowments”—peoples’ initial wealth. On this view, inequality can be dealt with in isolation, after figuring out how best to lubricate markets.
The second, opposing impulse, which is on display in many DC think-tanks, is to examine forensically the distributional consequences of every policy in isolation. This is becoming more common, partly because computing advances have made simulation models, which estimate how much a proposal will change the income of each social group, easier to run. As a result, few policy analyses omit to mention the “distributional benefits” or “distributional costs” of whatever is being proposed. In Britain, new policies come with an “equality impact assessment” which scrutinise their impact on specific groups.
A good example is student debt. Many young Americans feel it is an acutely stressful burden, which holds back their ability to save to, say, buy a house. Yet when politicians like Hillary Clinton propose measures to reduce student debt, economists are quick to point out that doing so would not, on the whole, help the poor. Graduates earn more than the population at large; those with the most student debt tend to be law-school graduates with lucrative careers ahead of them. Similarly, economists love to point out that higher minimum-wages do not benefit the poorest households most, because they relate to hourly pay rather than overall household income. Such polices are, economists say, “poorly targeted”.
Which impulse is preferable? The first seems wrong-headed, because it is not possible in reality to redistribute endowments. Inequality often results from trade over time—some people do well in the market, others less well—and, as a result, redistribution must be an ongoing project. When pressed on their preferred measure of redistribution, economists will often cite policies like the earned-income tax-credit, which tops-up the wages of poor workers. Yet this does not look at all like the efficient, lump-sum redistribution of economic models. Because it gets less generous as income rises, the EITC discourages low-income workers from earning more (marginal tax rates for low earners, accounting for benefit-withdrawal, can exceed 80%). The scheme is wasteful: the IRS estimates that 24% of EITC payments are improper. And other workers must pay taxes to fund EITC payments. Those higher tax rates distort the economy, too.
The second instinct, though, does not seem much better. Take student-debt again. When the British government raised tuition fees in 2010, it was so keen to make its reforms “progressive” that it introduced a complex system whereby higher-earners pay higher interest-rates. Yet it seems bizarre to redistribute within one group—graduates—rather than in society more broadly. There is no convincing reason for making high-earning graduates, as opposed to high-earners in general, subsidise education.
Furthermore, policy sometimes rightly helps higher-income folk. One example is childcare. At a conference on tax policy I attended last year, one participant pointed out that the child tax-credit, which provides money to parents with incomes up to about $95,000, does not have the “distributional advantages” of the EITC. But presumably there is value to society in helping more than just the poorest parents and children. Similarly, policies to subsidise child-care—often supported by the left—help all single-earner households become multiple-earner households, including those where the single-earner already has a high income. That is no bad thing if it helps achieve other goals, such as increasing female participation in the workforce.
The lesson is that policymaking must be holistic; proposals cannot be judged fully in isolation. Sometimes, regressive policies and universal benefits are welcome, even if an unequal income distribution is not. At the same time, it is unrealistic to expect all corrective redistribution to happen through one idealised channel like the EITC, because concentrating redistributive firepower in one area can cause collateral damage to incentives. As usual, economists will do best if they cast their most unworldly assumptions—of whatever variety—aside.