The Distribution of National Income (part 3)

February 23, 2017

Previous | Next

I admit that my last two posts have been pretty heavy on the facts and figures. But now we can use the conclusions to shed some light on the political polarization of the country.

Two main conclusions of Piketty, Saez, and Zucman’s analysis stand out. First, the distribution of pre-tax income is now more uneven than at any time since the 1920s. The top tenth of the population is getting almost half the national income, while the entire bottom half of the population is getting only one-eighth of it. Second, taxation and government spending are only mildly progressive and redistributive. Redistribution reduces the top tenth’s share from 47% to 39%, while increasing the bottom half’s share from 12.5% to 19.4%.

The politics of redistribution

To start assessing the political implications of these conclusions, let’s do a mental experiment. Imagine that each of the broad income groups described in the report took a position on government taxes and spending based solely on their narrow economic self-interest. We would expect people in the top tenth of the distribution to oppose the government’s redistributive role, since they pay more of its costs and qualify for fewer of its benefits. The lower-half of the population should be more supportive, since they receive more in benefits than they pay in taxes.

However, the political stance of the remaining two-fifths–those with incomes in the upper half but not in the top tenth–is likely to be more ambivalent. Their pre- and post-tax shares of national income are about the same (40.5% vs. 41.6%). What they receive in benefits offsets what they pay in taxes. Bear in mind that post-tax income in this analysis includes all forms of government benefits–monetary transfers, in-kind transfers, and general spending for the public good. If they focus on the benefits, they may support government spending; but if they focus on the costs, they may support tax cuts. (Or they can support a lot of both, and put up with deficits and more national debt.)

Since the major political parties disagree so much on taxes and spending, we would expect higher-income people to prefer the Republican Party and lower-income people to prefer the Democratic Party. This is true up to a point. Income is a fair predictor of party affiliation and voting, and the effect of income on voting has actually increased as the gap between rich and poor has widened. Gelman, Kenworthy and Su reported, “For the nation as a whole…there is a broad similarity between the trends in income inequality and the rich-poor gap in partisan voting. Each declined after the 1940s and then rose beginning in the 1970s or 1980s” (Social Science Quarterly, December 2010).

Gallup surveys have found that Democrats are much more likely than Republicans to believe that the present distribution of wealth is unfair, and that higher-income groups should pay more taxes.

The role of beliefs

Narrow self-interest is not the only basis on which people vote, however, even on questions of economics. Beliefs about how the economy works or should work are important, as well as beliefs about the impact of public policy on the general prosperity. Politics in a democracy is partly a struggle for the hearts and minds of the people, especially the hearts and minds of the middle class. They may align themselves with either the rich or the poor, depending on whose interests they think best represent the general good.

The upper-tenth have a disproportionate share of the money, but only a minority of the votes. To have their way politically–and they’ve been doing a pretty good job of that lately–they need good arguments against high taxes on the rich and high spending for the less fortunate.

One of those arguments is the appeal to meritocracy. Higher-income people can defend the very unequal pre-tax income distribution as a reflection of people’s real contribution to society. The successful deserve what they get; the unsuccessful deserve less; and the trouble with redistribution is that it punishes achievement and rewards failure. A related argument is that the rich are the job creators who use their incomes and wealth to invest in economic growth for the benefit of all.

Support for these views is widespread. Gallup has reported that when Americans are given a choice between taking steps “to distribute wealth more evenly” or “to improve overall economic conditions and the jobs situation,” people of all political affiliations and income levels prefer the latter by a wide margin.

That helps explain the working-class conservatism reported, for example, by J. D. Vance in Hillbilly Elegy. Although many low-income whites have more to gain from government spending than they have to lose from taxation, they cling to an ideology of self-reliance and hostility to government “handouts”. Reliance on government carries with it a stigma that I see as partly a racial stigma. Slavery, segregation and discrimination impeded black achievement and fostered government dependence, contributing to a stereotype of black laziness. Whites could maintain their sense of superiority by dissociating themselves from such dependency. That meant dissociating themselves from Big Government and liberal politics, especially after the Democratic Party embraced the civil rights movement in the 1960s.

Progressives need to change the national conversation about economic inequality, so that it is no longer about industrious job creators at the top, undeserving slackers at the bottom, and families in the middle who should be grateful to the rich for whatever wages they are offered. They need to challenge the dubious assumption that private wealth is always invested for the public good, while government spending is nothing but a drag on the economy. Considering our low rate of economic growth, our lagging productivity, and our wage stagnation, it isn’t obvious that concentrating more and more financial capital at the top has been such a winning strategy. Meanwhile, we cannot seem to find the money to make vital investments in our human capital, so that young people can get educations without accumulating a mountain of debt. People should not have to apologize for getting help to develop their human potential, especially when that enhances their capacity to contribute to society.

Voters shouldn’t have to choose between policies that create jobs and those that alleviate inequality. In a properly functioning democracy, they ought to go hand in hand, as they did during the postwar economic boom.

Progressive beliefs have the potential to spread to all class levels, just as conservative beliefs have. Already there are many higher-income individuals, such as Warren Buffet and George Soros, who advocate for more egalitarian policies.

Trump: populist or plutocrat?

Where does President Trump fit into the politics of redistribution? As a billionaire, he stands near the top of the economic pyramid. Like many other rich men, he sees his success as a sign of his superior merit, no matter what Trump University students or other detractors say in their lawsuits. Indeed, he declares himself to be uniquely suited to save the US economy.

Trump has filled his cabinet mainly with other rich folks who are not noted for their egalitarian views. Mother Jones reported that his cabinet selections have an average net worth of $357 million. The richest 1% of American households have an average net worth of only (did I say “only”?) $18.7 million.

Why is Trump so popular? I think primarily because he presents himself as the ultimate job creator, who will boost economic growth by bringing back lost American jobs. He will use the unorthodox strategy of getting other countries to give us more favorable terms of trade, so that our manufacturing industries prosper, presumably at someone else’s expense. All Americans will benefit, especially downwardly mobile workers, when he puts America first and makes America great again.

We are supposed to be so impressed by these promises that we overlook his tendency to favor the privileged over the rest of society. Strip away his economic nationalism, and what’s left is the usual Republican tax breaks for the rich and benefit cuts for the poor. We don’t have the detailed plans yet, but all indications point to a tax reform bill that will give the biggest reductions to the top brackets, and an Obamacare replacement that will make health insurance less affordable for the poor. Although Trump appealed to enough Democrats and independents to eke out an electoral college victory, the core of his support is among  Republicans.

The Trump administration has a real potential to exacerbate income inequality and political polarization. Maybe he can grow the economic pie so much that people don’t care how unequally it is divided, but I wouldn’t bet on it.

 


The Distribution of National Income (part 2)

February 21, 2017

Previous | Next

We have been looking at a report on the distribution of national income from the Washington Center for Equitable Growth, authored by Thomas Piketty, Emmanuel Saez and Gabriel Zucman. What makes it special is its attempt to account for all forms of income, not just those most often reported in surveys and tax returns. Based on this more complete accounting, the authors conclude that between 1980 and 2014, the top tenth of the adult population increased their share of the pre-tax national income from 34.2% to 47.0%. The share going to the next two-fifths of the population declined from 45.9% to 40.5%, and the share going to the bottom half of the population declined from 19.9% to 12.5%. During this period, economic growth was sluggish compared to the postwar era (1946-1980), but average real income more than doubled for the top tenth, while remaining essentially unchanged for the bottom half.

These figures are only for pre-tax income, however. They leave open the question of what role taxes and government spending play in the distribution of national income. Does post-tax income tell a different story?

Post-tax income

By considering the distribution of the entire national income, the report challenges the way we normally think about after-tax income. In our everyday experience, it’s what’s left after the taxes are taken out. That makes it always less than gross income. But in the national income accounting, total post-tax income and pre-tax income are the same! That’s because the national income does not go down just because some of it is taxed. The tax dollars are spent directly or indirectly on someone’s behalf, and so they can be counted as somebody’s income. Post-tax income is not a reduction in national income, but just a redistribution of national income.

The calculation of post-tax income from pre-tax income requires two steps: the subtraction of taxes paid, and the addition of government benefits received. Taxes include all levels (federal, state, local) and all types (income, sales, payroll, property). Government benefits include both monetary transfers (earned income tax credit, cash assistance payments, food stamps) and in-kind transfers (mainly health benefits through Medicare and Medicaid). Some cash income is already included in pretax income, such as Social Security payments.

The trickiest type of government benefit to account for is “collective consumption expenditures.” This is government spending on behalf of society in general. One might apportion it equally, on the assumption that each citizen gets the same benefit from it. But the researchers distribute it in proportion to other income, reasoning that higher-income people usually get more benefits from general public spending. For example, wealthier people are more likely to live in communities where the taxes support higher spending per student in the public schools. They are also more likely to be shareholders who benefit from the profits earned by defense contractors. The authors acknowledge that “our treatment of public goods could easily be improved as we learn more about who benefits from them.”

What if the government spends more than it receives in tax revenue? Then the deficit has to be allocated to individuals too, as a kind of negative benefit. Otherwise, total benefits received would be larger than total taxes paid, making post-tax income larger than total national income, upsetting the logic of the entire analysis.

The distribution of taxes and benefits

In general, the distribution of taxes and benefits is mildly progressive, but not markedly so. With all forms of taxation considered, higher incomes are a little more heavily taxed. The effective tax rates are 33.9% for the top tenth of adults, 28.6% for the next two-fifths, and 24.4% for the bottom half. The effective tax rate for the adult population as a whole is 30.5%.

Each group’s share of all taxes paid depends on how much income they have to begin with, as well as the rate at which it is taxed. In 2014, the top tenth got 47.0% of the pre-tax income and paid 52.2% of the taxes (hardly an unreasonable burden in my humble opinion). The next two-fifths got 40.5% of the income and paid 38% of the taxes. The bottom half of the population got 12.5% of the income and paid 10% of the taxes.

On the government benefits side, the top tenth got the smallest share–26.0%–which is lower than their share of income and taxes, but still much higher than their share of population. Although they didn’t qualify for means-tested assistance programs like Medicaid and food stamps, they got a lot of the general benefits of government spending. The next two-fifths, however, got the largest share–41.6%–roughly proportional to their share of the population. What they pay in taxes they get back in benefits such as good schools. The lower half of the population got 32.6% of the benefits, which is much more than their tax burden but much less than their 50% share of the population.

The redistribution of national income

The result of government taxation and spending is that a modest portion of national income is redistributed, primarily from the top tenth of the population to the bottom half.  A simple comparison of pre- and post-tax income shows this clearly.

Because the top tenth paid more in taxes than they received in benefits, their post-tax share of national income was 8 percentage points lower than their pre-tax share in 2014 (39.0% vs. 47.0%).

For the next two-fifths of the population, pre- and post-tax income came out about the same. They started out with 40.5% of the pre-tax income, paid 38% of the taxes, got 41.6% of the government benefits, and wound up with 41.6% of the after-tax national income. All the figures are roughly proportional to their 40% population size, so this group didn’t win or lose much from income redistribution.

The bottom half of the population gained more in benefits than they paid in taxes, so their post-tax share of national income was 6.9 points greater than their pre-tax share (19.4% vs. 12.5%). That difference consists mainly of non-cash benefits. That’s because their meager pretax incomes–averaging $16,200–were taxed at 24.4%, and that more than offset any cash benefits they received. The net benefits they got were primarily from health insurance programs.

To summarize, in 2014 the US transferred 8% of the national income by taxing the top tenth of the population, with 7 points of that going to the bottom half and 1 point to the other two-fifths. The transfer reduced the top tenth’s sizeable after-tax income by 17%. But the transferred income loomed much larger in the lives of the people at the bottom who received it in one form or another. Since they had so much less to begin with, it boosted their income by 54%. In dollar terms, it meant an increase in average income from $16,200 to $25,000, a significant improvement, but still leaving them far behind everyone else.

Redistribution and the trend toward inequality

Has the redistribution of income through taxes and government spending helped to offset the trend toward greater inequality? One would expect that as the rich got richer, they would be forced into higher tax brackets, increasing the tax revenue available for redistribution. One might also expect that as incomes at the bottom stagnated, political pressure would build to increase spending to augment them.

The point about tax revenue has some truth to it. Between 1980 and 2014, the top tenth increased their share of pre-tax national income from 34.2% to 47.0%, but some of that gain was offset by taxes. Still, their after-tax share of national income went from 29.5% to 39.0%. The increase in post-tax income was about three-quarters of the increase in pre-tax income. In other words, they got to keep three-fourths of their gains.

For the other nine-tenths of the population, tax offsets worked to reduce losses instead of gains. For the bottom half, the decline in their share of post-tax income was 85% as large as the decline in their share of pre-tax income. For the two-fifths of the population in between, the decline in their share of post-tax income was only  59% as large as the decline in their share of pre-tax income. To put it another way, the government absorbed 15% of the losses for the bottom half and 41% of the losses for the two-fifths in the upper middle of the distribution.

What the country did not do in those years was increase the overall rate of taxation or make the tax rates more progressive. The average tax rate considering all taxes went down slightly from 30.8% to 30.5%. Moreover, the effective rate of taxation went down for the upper half of the population (due mainly to income tax cuts), but went up for the lower half (due mainly to increases in payroll taxes). That’s why the government absorbed more losses for the upper-middle class than for the bottom half. Redistribution from top to bottom could still go up a little, because the rich had more money that could be taxed. But non-progressive tax policies left most of the increase in inequality untouched.

As for the second point, about political pressure to increase spending on the poor, that was outweighed by pressure to cut tax rates for the middle and upper classes. Between 1980 and 2014, the percentage of national income going to finance government benefits for the bottom half remained stuck around 10%, while benefits for the upper half remained around 20%. The upper middle class played a crucial political role here. With their own share of the national income shrinking, a majority of them sided with the rich in supporting low taxes, rather than with the poor in supporting policies to reduce inequality. I will have more to say about the political implications of the income distribution in my next post.

Continued


The Distribution of National Income

February 20, 2017

Previous | Next

The Washington Center for Equitable Growth has issued a new, very informative report on income inequality. Its authors, Thomas Piketty, Emmanuel Saez, and Gabriel Zucman, are trying to improve on the way economists have measured inequality in the past:

One major problem is the disconnect between maceoeconomics and the study of economic inequality. Macroeconomics relies on national accounts data to study the growth of national income while the study of inequality relies on individual or household income, survey and tax data. Ideally all three sets of data should be consistent, but they are not. The total flow of income reported by households in survey or tax data adds up to barely 60 percent of the national income recorded in the national accounts, with this gap increasing over the past several decades.

Why is there such a discrepancy between the national income accounting and the personal reporting? The main reason is that when people report their income on a survey or a tax return, they are thinking of income actually received in cash. But some forms of national income accrue to individuals whether they see cash from them or not. Employers contribute to workers’ pension plans or subsidize their health insurance. Corporations make money on behalf of shareholders that they retain for investment rather than distribute as dividends. This report aims to apportion the entire national income among individuals. It tries to account for all forms of compensation for workers and all returns on capital assets, whether taken in cash or not.

For purposes of analysis and discussion, the researchers divided the US population into three broad groups, the top tenth, the next two-fifths, and the bottom half. The unit of analysis was the adult individual 20 or older. Most of the analysis split marital income equally between spouses, for example assigning each of them $40,000 if one earned $50,000 and the other $30,000. That makes sense if couples are sharing their purchasing power. The authors also did a separate analysis of gender inequality using individual earnings. There they found that overall, men had 1.75 times as much work income as women, without controlling for hours worked or types of jobs. That ratio has been falling steadily since the 1960s, when it was over 3.00.

Pre-tax income

To appreciate the degree of income inequality the researchers found, consider the familiar analogy of dividing a pie. Imagine that you bake a large pie for a party of ten, dividing it into ten equal slices. But the first guest to dig in takes five slices! The next four guests take one slice each, leaving only one slice to be divided among the remaining five diners. In percentage terms, one-tenth of the people got 50% of the pie, the next two-fifths got 40%, and the remaining half got only 10%.

The real numbers for 2014 (the last year reported) are not far from that. The top tenth got 47.0% of the national income; the next two-fifths got 40.5%, and the bottom half got 12.5%. The average (mean) income for the groups was $304,000 per person for the top 10%, $65,400 for the next 40%, and $16,200 for the bottom 50%. (If some of the numbers sound large, remember that income is being defined very inclusively.)

One advantage of these particular dividing points is that they clearly distinguish between one group whose share of national income is roughly proportional to its size (the two-fifths) and two groups whose share is either disproportionately large (the top tenth) or small (the bottom half).

In addition to the enormous differences in shares, the three groups differed in how much of their income they derived from returns on capital as opposed to their own labor. The top tenth got 43.0% of their income from capital, compared to 17.9% for the next two-fifths and 5.1% for the bottom half. Ironically, in a country that prides itself on its work ethic, the most meager rewards go to those who have to rely the most on their labor.

Trends in inequality

In order to study trends over time, the researchers compared two 34-year periods, 1946-1980 and 1980-2014. The first period includes the postwar economic boom. The second period begins with the year Ronald Reagan was elected president, although I don’t know how much that affected its selection as a dividing point. The authors do suggest that changes in public policy were at least partly responsible for the increase in inequality that has occurred since 1980.

The period after World War II was a time of rapid economic growth and broad-based increases in income. Pre-tax income (adjusted for inflation) increased 79% for the top tenth, 105% for the next two-fifths, and 102% for the bottom half over those 34 years. Because the increase was less for the top tenth than the other groups, the distribution became a little more egalitarian. The share of national income going to the top tenth declined from 37.2% to 34.2%.

The period since 1980 has been a time of both slower economic growth and very unevenly distributed gains. Pre-tax income increased 121% for the top tenth, 42% for the next two-fifths, and only 1% (!) for the bottom half. The rich got richer and the poor got left behind. As a result, the distribution of national income became noticeably less egalitarian. The share of the top tenth rose from 34.2% to 47.0%, but the share of the lower half dropped from 19.9% to 12.5%. That top share is similar to what rich people were getting back in the 1920s, before the Great Depression. Over the course of the past century, income inequality has gone down but then gone back up. At the highest levels of income, the return to inequality has been even more dramatic. Average income for the top 1% increased only 47% during the postwar era, lagging well behind general economic growth; but it rose 205% after 1980, far exceeding general growth. For the top 0.01%, where the average income is over $28 million, the increase has been 454%.

Although global trends such as outsourcing and automation have produced gains for capital at the expense of workers, the authors point out that not all countries have experienced the same extremes of inequality as the United States has. Although economic growth has been slower in France, the lower half of the French population has shared in the national growth as the American lower half has not. As a result, “While the bottom 50 percent of incomes were 11 percent lower in France than in the United States in 1980, they are now 16 percent higher.” America’s self-image as a unique land of opportunity is no longer secure.

Income redistribution?

Pre-tax income does not tell the whole story, however. The taxation of income provides some potential for redistribution, as those with higher incomes are taxed in order to provide some benefits to those with lower incomes. In my next post, I will discuss the report’s comparison of pre- and post-tax income to see how taxes and government benefits are distributed, and what effect they have on income inequality.

Continued


Global Inequality (part 3)

August 8, 2016

Previous | Next

This final post on Branko Milanovic’s Global Inequality will focus on the future prospects he sees for reducing economic inequality in the world. He discusses three kinds of inequality: inequality among countries, inequality within poorer countries, and inequality within richer countries, especially the United States.

Inequality among countries

Milanovic expects the median incomes of countries to continue to converge as the economies of poorer countries continue to grow faster than those of richer countries. However, he expects this convergence to be limited in several ways. Currently “it is only Asian countries that have been catching up with the rich world.” Progress has been much slower in other parts of the world, especially Africa. And although growth in China accounts for a lot of the global convergence in incomes, this may not always be the case. In a few years, China may be so far above average that further development there may increase global inequality rather than reduce it. Then continued convergence will depend on what other countries do.

Although the narrowing gap between rich and poor countries is encouraging, it is occurring so slowly that “one cannot expect global inequality to be reduced by more than one-fifteenth of its current level” over the next twenty years.

Inequality within poorer countries

Milanovic suggests that countries like China are passing through a phase of the first Kuznets wave that more developed countries experienced at least a century ago. They are experiencing a familiar pattern in which industrialization initially increases opportunities for the few, and only later for the many, as discussed in the previous post. Although the data are somewhat sketchy, Milanovic sees signs that economic inequality has peaked in China and is beginning to subside. Mass education and a reduced supply of cheap, unskilled labor would be among the reasons for the transition.

On the other hand, many poor countries are still in an even earlier phase of the transition, in which inequality is increasing because the benefits of economic development have yet to be experienced by large portions of their populations.

Inequality within richer countries

I found Milanovic’s views on this topic a bit confusing. On the one hand, he maintains that information technology and globalization have initiated a new Kuznets wave of economic change, in which inequality is rising again but will ultimately fall. To make that case, he needs reasons for the fall as well as the rise. Otherwise, he cannot distinguish his theory from more pessimistic assessments like Piketty’s Capital in the Twenty-first Century, which sees rising inequality as a fundamental feature of capitalism. However, when Milanovic tries to identify mechanisms by which inequality might fall, he expresses little confidence that they will work any time soon.

Milanovic identifies five “benign forces” that could theoretically reduce inequality:

  1. Political changes could result in higher and more progressive taxation. However, the global mobility of capital makes it easier for the wealthy to escape taxation. In addition, many citizens of modest means have trouble supporting higher taxes, even when that might be in their own best interest.
  2. The widening wage gap between more and less educated workers could be narrowed by improvements in the quantity and quality of education. However, Milanovic has trouble imagining that average years of education could rise above thirteen. He also thinks that improvements in the quality of education “face natural limits, given by the aptitude and interest of students to excel in whatever they choose to do.” Some would object that if better education were more widely available, more students would rise to the occasion.
  3. As the technological revolution proceeds, innovations that originally profited the few can be more widely adopted. On the other hand, the ownership of capital has become more concentrated lately, so the control of profitable innovations remains largely in the hands of a few.
  4. As wages rise in poorer countries, workers in richer countries should face less competition from foreign low-wage labor. However, it could be a long time before poor countries outside of China and a few other Asian countries experience much wage growth.
  5. Technological progress could raise the productivity of low-skill workers specifically. But this would go against the historical experience of capitalism, in which technological change normally boosts the income of the more skilled over the less skilled.

I found the last point especially troublesome, since it seems to me to undercut one of the strongest reasons for a Kuznets curve in the first place. Surely the mass-production technologies of the twentieth century helped bring many blue-collar workers into the middle class by boosting their productivity, raising their wages, and making former luxuries like automobiles more affordable. If we are looking for mechanisms for reducing inequality in the new wave of change, shouldn’t we be looking for a new productivity revolution along the lines suggested by Rifkin’s The Zero Marginal Cost Society or Paul Mason’s Postcapitalism? Milanovic  doesn’t anticipate anything that radical, but maybe the falling inequality phase of the alleged Kuznets curve won’t work without some fairly dramatic change. Ironically, Milanovic begins his chapter on future inequality by criticizing previous attempts at prediction for assuming too much continuity from the present to the future.

The United States: A “perfect storm of rising inequality”?

Milanovic is especially pessimistic about reducing inequality in the United States. He provides five reasons he expects the rise in inequality to continue:

  1. The share of national income going to capital rather than labor will remain high, especially since businesses find it economical to replace labor with machinery.
  2. The income from capital will remain highly concentrated.
  3. The people with the highest incomes will also be the ones who can save and invest the most, so the same people will be getting most of the benefits from both labor income and capital income.
  4. These labor-rich and capital-rich individuals will also tend to marry each other, so that wealth and income are even more concentrated for households than they are for individuals.
  5. The rich will use their political power to support policies that protect their economic interests at the expense of those of the middle class and the poor.

Milanovic concludes:

It is hard to see where any forces might come from that could counter rising income inequality in the United States….Forces promoting offsetting policies such as more widespread education, a higher minimum wage, and more generous welfare benefits seem weak compared with the almost elemental forces that favor greater inequality.

By this time, the reader who has followed the argument from the beginning may be wondering what happened to the original idea of the Kuznets curve, with its rise and fall of inequality. Well, “the second Kuznets curve will have to repeat the behavior of the first if inequality is to decline again. But it is doubtful whether this second decline will be accomplished by the same mechanisms as those that reduced inequality in the twentieth century….” What mechanisms Milanovic does suggest are mostly political, especially changes in tax policies and improvements in public education, changes that seem unlikely in the light of his previous remarks. The reliance on political rather than economic mechanisms sounds more like Piketty than Kuznets.

To summarize, Milanovic starts with a Kuznets theory emphasizing economic reasons why inequality first rises and then falls. He does broaden it by suggesting that extreme inequality generates malign forces like violent conflicts that can destroy the wealth of some and create opportunity for others. Yet he ends with a pessimism that economic forces will reduce inequality either benignly or malignly. This leaves it an open question whether what we are living through is a second Kuznets wave at all. If it isn’t, then the first Kuznets curve was a unique historical event from which we cannot generalize, and the book’s theoretical framework falls apart.

In general, I found the book’s data very informative and its interpretations thought provoking. But in the end I found its theoretical position on the central question of falling inequality too ambiguous to be convincing.


Global Inequality (part 2)

August 4, 2016

Previous | Next

The Kuznets curve

Milanovic’s interpretation of global trends in inequality relies heavily on a modified version of an idea proposed by Simon Kuznets in the 1950s. Kuznets observed a rise and fall pattern in inequality within industrializing countries. In a graph plotting inequality against per-capita national income, the pattern appears as a curve in the shape of an inverted U. This suggests that as industrialization increases national output and average income, the benefits flow first to the few, and only later to the many. Limited prosperity widens the gap between rich and poor, but then more widespread prosperity narrows it.

Consider the United States at its peak of economic inequality around the end of the nineteenth century. Industrialization has created a wealthy class of industrialists and a very small middle class of well educated workers. Meanwhile, massive numbers of workers displaced by the decline of farm jobs have poured into the cities, keeping wages low for those without capital or urban skills. Then consider the country fifty years later. Population growth has slowed along with urbanization, since urban families have fewer children than farm families. (The postwar baby boom will change that temporarily, but that is more of a response than an impediment to widespread prosperity.) Education and skill levels have risen. Assembly-line technology has boosted worker productivity, helping to justify higher wages; and perhaps requiring higher wages because businesses need consumers to buy their mass-produced products. Workers have mobilized both in the workplace and in the political sphere to fight for better wages, better working conditions, and a social safety net. A large middle class, a hallmark of general prosperity, has emerged.

Many economists once accepted the Kuznets curve as a fairly good description of trends in inequality, generalizable to many industrializing nations. However, since the 1980s, the downswing in inequality has come to an end in many rich countries. “The indubitable increase in inequality in the United States, the United Kingdom, and even in some fairly egalitarian countries like Sweden and Germany, is simply incompatible with the Kuznets hypothesis.”

Recent thinking about inequality is less confident that any long-term decline in inequality must or will occur. One popular theory says that the future of inequality depends on the outcome of a “race between education and what is known as skill-biased technological progress (that is, technological change that favors high-skilled workers).” Inequality could decline if less skilled workers can acquire skills quickly enough, but there is no guarantee that will happen. In Capital in the Twenty-First Century, Thomas Piketty is especially pessimistic. He sees rising inequality as the historical norm, because the rate of return on capital tends to exceed the general rate of income growth. The twentieth-century decline in inequality in the US and other developed countries was due to a special set of circumstances that need not be continued or repeated. Only radical political action, such as higher taxes on wealth, can counteract the economic forces favoring inequality.

Multiple Kuznets waves

Milanovic’s main contribution to this discussion is the suggestion that the Kuznets curve associated with industrialization is not the only such curve. Fluctuations in inequality have occurred in repeated waves over the past 500 years. Milanovic regards the most recent rise in inequality as just the upswing phase of a new Kuznets wave, which will eventually have its downswing phase as well. That means that the Kuznets curve remains relevant to our understanding of the present, since it does not just describe a one-time, historically unique feature of industrialization.

Milanovic see Kuznets waves even in preindustrial times, although his description of them makes me wonder whether Kuznets himself would recognize them. “Before the Industrial Revolution, when mean income was stagnant, there was no relationship between mean income level and the level of inequality. Wages and inequality were driven up or down by idiosyncratic events such as epidemics, new discoveries…, invasions, and wars.” For example, a widespread epidemic might raise wages by making labor more scarce, but the gains would be wiped out in Malthusian fashion as population growth resumed.

Notice that this focus on idiosyncratic events removes two key elements of the original theory. For Kuznets, the rise and fall pattern of inequality was tied to mean income level, and the rise came before the fall. Without those elements, about all that remains is the observation that inequality fluctuates in either direction for many different reasons, which no one doubts. Calling the fluctuations “Kuznets waves” creates the impression of having a general theory of waves with some application to today’s economy. But if preindustrial fluctuations are merely idiosyncratic–or as Milanovic says, “Changes in inequality versus mean income are irregular in preindustrial societies but shift into regular cycles in industrial and postindustrial societies”–then we have only the one completed Kuznets cycle of the industrial era on which to base the claim that there are regular cycles at all. We may be at the beginning of a second regular cycle, but how would we know? The possibility that the industrial Kuznets wave is the only Kuznets wave is harder to dismiss. I don’t think that Milanovic ever entirely overcomes this difficulty, as much as I would like him to!

The industrial wave of inequality

Milanovic’s description of rising inequality during the early stages of industrialization follows Kuznets pretty closely. “As Kuznets argued, it is the structural movement, the transfer of labor from the low-income, low-inequality agricultural sector to the higher-income, higher-inequality industrial sector (and concomitantly, from rural to urban areas) that increases income inequality.”

For the later downswing in inequality, Milanovic’s contribution is to consider “benign” factors and more “malign” factors in combination. Benign reasons for more widespread prosperity include those I mentioned earlier, such as greater access to education and skills, technology-based productivity gains, and political mobilization for industrial workers.  For the malign factors, he draws on the work of Piketty, who argues that “the two world wars not only led to higher taxes but also destroyed property and reduced large fortunes.” Unlike Piketty, however, Milanovic sees the malign factors as consequences of economic developments, including inequality itself. The conditions leading to World War I included “very high income and wealth inequality, high savings of the upper classes, insufficient domestic aggregate demand, and the need of capitalists to find profitable uses for surplus savings outside their own country.” The result was nationalist rivalries, imperialism, colonialism, and ultimately war. The moral of the story: “A very high inequality eventually becomes unsustainable, but it does not go down by itself; rather, it generates processes, like wars, social strife, and revolutions, that lower it.”

In summary, Milanovic remains faithful to the main thrust of the Kuznets argument, that rising inequality eventually reaches some sort of limit, but he has a broader view of the mechanisms by which inequality then declines.

A postindustrial wave?

Milanovic believes that the emergence of a global information society is a development comparable in its significance to the Industrial Revolution. Therefore it has its own Kuznets curve of rising, then falling, inequality, but we have seen only the rising phase so far. As in the early phases of industrialization, the economic benefits have been going primarily to the few.  We can expect this to change only if we share Milanovic’s confidence that the general logic of the Kuznets wave applies: Inequality cannot increase without encountering some kind of limit.

Here is Milanovic’s description of the new inequality:

The 1980s ushered in a new (second) technological revolution, characterized by remarkable changes in information technology, globalization, and the rising importance of heterogeneous jobs in the service sector. This revolution, like the Industrial Revolution of the early nineteenth century, widened income disparities. The increase in inequality happened in part because the new technologies strongly rewarded more highly skilled labor; drove up the share of, and the return to, capital; and increasingly opened the economies of rich countries to competition from China and India…. The structure of demand, and thus of jobs, moved toward services, which in turn were staffed by less qualified and worse-paid labor. On the other hand, some service sector jobs, as in finance, were extremely highly paid. This widened wage, and ultimately income, distribution.

While in the earlier era it was the shift from agriculture to manufacturing that created wealth for some and subsistence wages for others, now it is the shift from manufacturing to services.  Not only are many service jobs too low-skill and low-productivity to pay a good wage, but service workers are hard to organize because they are widely dispersed in small work units.

What remains to be seen is whether the latest increase in inequality can be reversed by more egalitarian trends, as in the classic Kuznets curve, or if it will turn out to be irreversible. Milanovic prefers the first alternative, and I want to believe him. But when he tries to identify mechanisms that would restore greater equality today, his arguments get a bit tentative and half-hearted. This will be the subject for the last post.

Continued