Figure 9-1: International inequality, 1820–2013
Sources: International inequality: OECD Clio Infra Project, Moatsos et al. 2014; data are for market household income across countries. Population-weighted international inequality: Milanović 2012; data for 2012 and 2013 provided by Branko Milanović, personal communication.
Still, an international Gini treats all the Chinese as if they earned the same amount, all the Americans as if they earned the American average, and so on, and as a result it underestimates inequality across the human race. A global Gini, in which every person counts the same, regardless of country, is harder to calculate, because it requires mixing the incomes from disparate countries into a single bowl, but two estimates are shown in figure 9-2. The lines float at different heights because they were calibrated in dollars adjusted for purchasing parity in different years, but their slopes trace out a kind of Kuznets curve: after the Industrial Revolution, global inequality rose steadily until around 1980, then started to fall. The international and global Gini curves show that despite the anxiety about rising inequality within Western countries, inequality in the world is declining. That’s a circuitous way to state the progress, though: what’s significant about the decline in inequality is that it’s a decline in poverty.
Figure 9-2: Global inequality, 1820–2011
Source: Milanović 2016, fig. 3.1. The left-hand curve shows 1990 international dollars of disposable income per capita; the right-hand curve shows 2005 international dollars, and combines household surveys of per capita disposable income and consumption.
The version of inequality that has generated the recent alarm is the inequality within developed countries like the United States and the United Kingdom. The long view of these countries is shown in figure 9-3. Until recently, both countries traveled a Kuznets arc. Inequality rose during the Industrial Revolution and then began to fall, first gradually in the late 19th century, then steeply in the middle decades of the 20th. But then, starting around 1980, inequality bounced into a decidedly un-Kuznetsian rise. Let’s examine each segment in turn.
Figure 9-3: Inequality, UK and US, 1688–2013
Source: Milanović 2016, fig. 2.1, disposable income per capita.
The rise and fall in inequality in the 19th century reflects Kuznets’s expanding economy, which gradually pulls more people into urban, skilled, and thus higher-paying occupations. But the 20th-century plunge—which has been called the Great Leveling or the Great Compression—had more sudden causes. The plunge overlaps the two world wars, and that is no coincidence: major wars often level the income distribution.27 Wars destroy wealth-generating capital, inflate away the assets of creditors, and induce the rich to put up with higher taxes, which the government redistributes into the paychecks of soldiers and munition workers, in turn increasing the demand for labor in the rest of the economy.
Wars are just one kind of catastrophe that can generate equality by the logic of Igor and Boris. The historian Walter Scheidel identifies “Four Horsemen of Leveling”: mass-mobilization warfare, transformative revolution, state collapse, and lethal pandemics. In addition to obliterating wealth (and, in the communist revolutions, the people who owned it), the four horsemen reduce inequality by killing large numbers of workers, driving up the wages of those who survive. Scheidel concludes, “All of us who prize greater economic equality would do well to remember that with the rarest of exceptions it was only ever brought forth in sorrow. Be careful what you wish for.”28
Scheidel’s warning applies to the long run of history. But modernity has brought a more benign way to reduce inequality. As we have seen, a market economy is the best poverty-reduction program we know of for an entire country. It is ill-equipped, however, to provide for individuals within that country who have nothing to exchange: the young, the old, the sick, the unlucky, and others whose skills and labor are not valuable enough to others for them to earn a decent living in return. (Another way of putting it is that a market economy maximizes the average, but we also care about the variance and the range.) As the circle of sympathy in a country expands to encompass the poor (and as people want to insure themselves should they ever become poor), they increasingly allocate a portion of their pooled resources—that is, government funds—to alleviating that poverty. Those resources have to come from somewhere. They may come from a corporate or sales tax, or a sovereign wealth fund, but in most countries they largely come from a graduated income tax, in which richer citizens pay at a higher rate because they don’t feel the loss as sharply. The net result is “redistribution,” but that is something of a misnomer, because the goal is to raise the bottom, not lower the top, even if in practice the top is lowered.
Those who condemn modern capitalist societies for callousness toward the poor are probably unaware of how little the pre-capitalist societies of the past spent on poor relief. It’s not just that they had less to spend in absolute terms; they spent a smaller proportion of their wealth. A much smaller proportion: from the Renaissance through the early 20th century, European countries spent an average of 1.5 percent of their GDP on poor relief, education, and other social transfers. In many countries and periods, they spent nothing at all.29
In another example of progress, sometimes called the Egalitarian Revolution, modern societies now devote a substantial chunk of their wealth to health, education, pensions, and income support.30 Figure 9-4 shows that social spending took off in the middle decades of the 20th century (in the United States, with the New Deal in the 1930s; in other developed countries, with the rise of the welfare state after World War II). Social spending now takes up a median of 22 percent of their GDP.31
Figure 9-4: Social spending, OECD countries, 1880–2016
Source: Our World in Data, Ortiz-Ospina & Roser 2016b, based on data from Lindert 2004 and OECD 1985, 2014, 2017. The Organisation for Economic Co-operation and Development includes thirty-five democratic states with market economies.
The explosion in social spending has redefined the mission of government: from warring and policing to also nurturing.32 Governments underwent this transformation for several reasons. Social spending inoculates citizens against the appeal of communism and fascism. Some of the benefits, like universal education and public health, are public goods that accrue to everyone, not just the direct beneficiaries. Many of the programs indemnify citizens against misfortunes for which they can’t or won’t insure themselves (hence the euphemism “social safety net”). And assistance to the needy assuages the modern conscience, which cannot bear the thought of the Little Match Girl freezing to death, Jean Valjean imprisoned for stealing bread to save his starving sister, or the Joads burying Grampa by the side of Route 66.
Since there’s no point in everyone sending money to the government and getting it right back (minus the bureaucracy’s cut), social spending is designed to help people who have less money, with the bill footed by people who have more money. This is the principle known as redistribution, the welfare state, social democracy, or socialism (misleadingly, because free-market capitalism is compatible with any amount of social spending). Whether or not the social spending is designed to reduce inequality, that is one of its effects, and the rise in social expenditures from the 1930s through the 1970s explains part of the decline in the Gini.
Social spending demonstrates an uncanny aspect of progress that we’ll encounter again in subsequent chapters.33 Though I am skittish about any notion of historical inevitability, cosmic forces, or mystical arcs of justice, some kinds of social change really do seem to be carried along by an inexorable tectonic force. As they proceed, certain factions oppose them hammer and tongs, but resistance turns out to be futile. Social spending is an example. The United States is famously resistant to anything smacking of redistribution. Yet it allocates 19 percent of its GDP to social services, and despite the best efforts of conservatives and libertarians the spending has continued to grow. The most recent expansions
are a prescription drug benefit introduced by George W. Bush and the eponymous health insurance plan known as Obamacare introduced by his successor.
Indeed, social spending in the United States is even higher than it appears, because many Americans are forced to pay for health, retirement, and disability benefits through their employers rather than the government. When this privately administered social spending is added to the public portion, the United States vaults from twenty-fourth into second place among the thirty-five OECD countries, just behind France.34
For all their protestations against big government and high taxes, people like social spending. Social Security has been called the third rail of American politics, because if politicians touch it they die. According to legend, an irate constituent at a town-hall meeting warned his representative, “Keep your government hands off my Medicare” (referring to the government health insurance program for seniors).35 No sooner did Obamacare pass than the Republican Party made it a sacred cause to repeal it, but each of their assaults on it after gaining control of the presidency in 2017 was beaten back by angry citizens at town-hall meetings and legislators afraid of their ire. In Canada the top two national pastimes (after hockey) are complaining about their health care system and boasting about their health care system.
Developing countries today, like developed countries a century ago, stint on social spending. Indonesia, for example, spends 2 percent of its GDP, India 2.5 percent, and China 7 percent. But as they get richer they become more munificent, a phenomenon called Wagner’s Law.36 Between 1985 and 2012 Mexico quintupled its proportion of social spending, and Brazil’s now stands at 16 percent.37 Wagner’s Law appears to be not a cautionary tale about overweening government and bureaucratic bloat but a manifestation of progress. The economist Leandro Prados de la Escosura found a strong correlation between the percentage of GDP that an OECD country allocated to social transfers as it developed between 1880 and 2000 and its score on a composite measure of prosperity, health, and education.38 And tellingly, the number of libertarian paradises in the world—developed countries without substantial social spending—is zero.39
The correlation between social spending and social well-being holds only up to a point: the curve levels off starting at around 25 percent and may even drop off at higher proportions. Social spending, like everything, has downsides. As with all insurance, it can create a “moral hazard” in which the insured slack off or take foolish risks, counting on the insurer to bail them out if they fail. And since the premiums have to cover the payouts, if the actuaries get the numbers wrong or the numbers change so that more money is taken out than put in, the system can collapse. In reality social spending is never exactly like insurance but is a combination of insurance, investment, and charity. Its success thus depends on the degree to which the citizens of a country sense they are part of one community, and that fellow feeling can be strained when the beneficiaries are disproportionately immigrants or ethnic minorities.40 These tensions are inherent to social spending and will always be politically contentious. Though there is no “correct amount,” all developed states have decided that the benefits of social transfers outweigh the costs and have settled on moderately large amounts, cushioned by their massive wealth.
* * *
Let’s complete our tour of the history of inequality by turning to the final segment in figure 9-3, the rise of inequality in wealthy nations that began around 1980. This is the development that inspired the claim that life has gotten worse for everyone but the richest. The rebound defies the Kuznets curve, in which inequality was supposed to have settled into a low equilibrium. Many explanations have been proffered for this surprise.41 Wartime restrictions on economic competition may have been sticky, outlasting World War II, but they finally dissipated, freeing the rich to get richer from their investment income and opening up an arena of dynamic economic competition with winner-take-all payoffs. The ideological shift associated with Ronald Reagan and Margaret Thatcher slowed the movement toward greater social spending financed by taxes on the rich while eroding social norms against extravagant salaries and conspicuous wealth. As more people stayed single or got divorced, and at the same time more power couples pooled two fat paychecks, the variance in income from household to household was bound to increase, even if the paychecks had stayed the same. A “second industrial revolution” driven by electronic technologies replayed the Kuznets rise by creating a demand for highly skilled professionals, who pulled away from the less educated at the same time that the jobs requiring less education were eliminated by automation. Globalization allowed workers in China, India, and elsewhere to underbid their American competitors in a worldwide labor market, and the domestic companies that failed to take advantage of these offshoring opportunities were outcompeted on price. At the same time, the intellectual output of the most successful analysts, entrepreneurs, investors, and creators was increasingly available to a gargantuan worldwide market. The Pontiac worker is laid off, while J. K. Rowling becomes a billionaire.
Milanović has combined the two inequality trends of the past thirty years—declining inequality worldwide, increasing inequality within rich countries—into a single graph which pleasingly takes the shape of an elephant (figure 9-5). This “growth incidence curve” sorts the world’s population into twenty numerical bins or quantiles, from poorest to richest, and plots how much each bin gained or lost in real income per capita between 1988 (just before the fall of the Berlin Wall) and 2008 (just before the Great Recession).
Figure 9-5: Income gains, 1988–2008
Source: Milanović 2016, fig. 1.3.
The cliché about globalization is that it creates winners and losers, and the elephant curve displays them as peaks and valleys. It reveals that the winners include most of humanity. The elephant’s bulk (its body and head), which includes about seven-tenths of the world’s population, consists of the “emerging global middle class,” mainly in Asia. Over this period they saw cumulative gains of 40 to 60 percent in their real incomes. The nostrils at the tip of the trunk consist of the world’s richest one percent, who also saw their incomes soar. The rest of the trunk tip, which includes the next 4 percent down, didn’t do badly either. Where the bend of the trunk hovers over the floor around the 85th percentile we see globalization’s “losers”: the lower middle classes of the rich world, who gained less than 10 percent. These are the focus of the new concern about inequality: the “hollowed-out middle class,” the Trump supporters, the people globalization left behind.
I couldn’t resist plotting the most recognizable elephant in Milanović’s herd, because it serves as a vivid mnemonic for the effects of globalization (and it rounds out a nice menagerie with the camel and dromedary in figure 8-3). But the curve makes the world look more unequal than it really is, for two reasons. One is that the financial crisis of 2008, which postdated the graph, had a strangely equalizing effect on the world. The Great Recession, Milanović points out, was really a recession in North Atlantic countries. The incomes of the world’s richest one percent were trimmed, but the incomes of workers elsewhere soared (in China, they doubled). Three years after the crisis we still see an elephant, but it has lowered the tip of its trunk while arching its back twice as high.42
The other elephant-distorter is a conceptual point that bedevils many discussions of inequality. Whom are we talking about when we say “the bottom fifth” or “the top one percent”? Most income distributions use what economists call anonymous data: they track statistical ranges, not actual people.43 Suppose I told you that the age of the median American declined from thirty in 1950 to twenty-eight in 1970. If your first thought is “Wow, how did that guy get two years younger?” then you have confused the two: the “median” is a rank, not a person. Readers commit the same fallacy when they read that “the top one percent in 2008” had incomes that were 50 percent higher than “the top one percent in 1988” and conclude that a bunch of rich people got half again richer. Peop
le move in and out of income brackets, shuffling the order, so we’re not necessarily talking about the same individuals. The same is true for “the bottom fifth” and every other statistical bin.
Nonanonymous or longitudinal data, which track people over time, are unavailable in most countries, so Milanović did the next best thing and tracked individual quantiles in particular countries, so that, say, poor Indians in 1988 were no longer being compared with poor Ghanaians in 2008.44 He still got an elephantoid, but with a much higher tail and haunches, because the poorer classes of so many countries rose out of extreme poverty. The pattern remains—globalization helped the lower and middle classes of poor countries, and the upper class of rich countries, much more than it helped the lower middle class of rich countries—but the differences are less extreme.
* * *
Now that we have run through the history of inequality and seen the forces that push it around, we can evaluate the claim that the growing inequality of the past three decades means that the world is getting worse—that only the rich have prospered, while everyone else is stagnating or suffering. The rich certainly have prospered more than anyone else, perhaps more than they should have, but the claim about everyone else is not accurate, for a number of reasons.
Most obviously, it’s false for the world as a whole: the majority of the human race has become much better off. The two-humped camel has become a one-humped dromedary; the elephant has a body the size of, well, an elephant; extreme poverty has plummeted and may disappear; and both international and global inequality coefficients are in decline. Now, it’s true that the world’s poor have gotten richer in part at the expense of the American lower middle class, and if I were an American politician I would not publicly say that the tradeoff was worth it. But as citizens of the world considering humanity as a whole, we have to say that the tradeoff is worth it.