Income Inequality: Screw U Kuznets

Travis Fast

Some other social science blogs set me to thinking about income inequality.

What I want to look at today are two rather less obvious but I think more interesting questions that arise when we move away from static comparisons between countries and rather focus on comparisons drawn over time. One of my central gripes with the Varieties of Capitalism literature, which is by now moribund due to a preoccupation with demonstrating the superiority of the putative Scandinavian model, is that it fails to attempt to account for changes over time when drawing its comparisons between countries.

For example, take a hypothetical situation in which three countries are compared across a range of qualitative vectors. At t-1 it may be that there exists n degree variance between the countries in question. Suppose we repeat the same analysis for two further points in time and record a similar degree of variance. Do we then conclude that there still exists a significant degree of divergence among the three model cases?

Perhaps, but that depends on what question we are trying to answer. For example, it may be that all three of the models in question have registered increasing degrees of income inequality yet at the same time the degree of difference between the models has remained the same or similar. Such a result would suggest that there is common underlying force driving income inequality in all three models despite the formal persistence of institutional variation. Static comparisons force us to focus on the variation because that is all we have to compare. If we were instead to compare trends over time we may be forced to account for some other variable along side of institutional variation which is of a higher order in causal determination.

Below I have drawn a simple graph of US, Swedish, British, and Canadian GINI scores over roughly the same points in time.


When I see a plot like this I am immediately suspicious of any purely technical explanation such as a skill biased technical change. The Canadian case might fit this description but the US and the UK look very suspicious. And I suspect changes in the political economy of macroeconomic regulation from good old fashion union busting resulting in the break-down of collective bargaining along side tight money, privatization, regressive tax shifts, in the context of increasing international competition—in short politically driven restructuring—provide a more promising description of the forces responsible for the rising GINI and their relative timings than any technical explanation rooted in some natural process of technological change which privileges highly skilled workers. Indeed, in Sweden despite the increasing levels of income inequality there still remains a relatively high degree of wage compression (that is, there is a narrower spread in the initial distribution of income). But that said all four trends are up with Canada appearing to have come a little late to the rising income inequality party. And that too seems odd given the trend in the US. If this were some shift in the mix and composition of some aggregate production function why was Canada so out of step with the US? If anything Canada has been more dominated in the past by services than the US. So yes institutional variation (including juridical regulation) matters but there seems nonetheless to be some kind of broader force at work here, which appears to be engineered.

But then Smith was out to prove that god was an engineer and Marx perhaps was right: god is us. Now for Kuznets. Here is the riddle: why does the Canadian income inequality index look like an inverted Kuznets curve when plotted for the last 35 years? I thought the richer we got the more egalitarian income distribution was suppose to be. Give that man a faux Nobel Prize. I wait for a REAL economist to explain it away.


Source data can be found here:

3 thoughts on “Income Inequality: Screw U Kuznets

  1. Stephen thanks for stopping by and taking up my gracious offer:

    Kuznets’ curve was constructed not by looking at a single country and plotting the evolution of income inequality and development over time but rather by creating a plot of countries “stage of economic development” and distribution of income shares. The inverted U then is not constructed by plotting a single country over time but by using a static snapshot of many countries at different stages of development and their corresponding distributions of income shares.

    Implicit here is the argument that as countries get richer the distribution of income becomes more egalitarian not less. That in fact held for Canada until the late eighties. Despite the fact Canada is much richer today then during the late eighties and early nineties income inequality has been increasing. But the really interesting thing is that this compression of income inequality and its subsequent breakdown is better explained via recourse to politically led restructuring of those institutions which effect income distribution. Recall that in Kuznets classic “Modern Economic Growth: Rate, Structure and Spread” he includes a whole array of political variables that are necessary for economic development which simultaneously push in the direction of increasing income compression. Sung to: “You can’t have one without the other.”

    Alas this was the hope of that time or what I might call the secondary Keynesian conceit: namely, that whether from socialist or capitalist forms of planning, the conditions for economic growth pointed in the direction of increasing income equality. In this sense Kuznets was put to good use suggesting that the choice between socialism and capitalism was a false choice because all roads led to an egalitarian Rome.

    Also recall that it was on this fundamental axiom of equality that first Hayek, later Friedman, and finally supply siders crossed the Rubicon when they launched their attack by insisting that inequality was a necessary by-product of economic growth.

    To be fair Kuznets, the political economist, needs to be distinguished from the sterilized Kuznets qua the “REAL” U curve economist. From the former we can learn much; from the latter all we can do is shoot lame duck covering laws in a barrel. I should also say that from actually reading Kuznets one could stay well within the spirit his work and derive a plausible theory of the causes of increasing income inequality in rich countries since the mid 1970s.

  2. Travis, I have a short section on Kuznets in my new book, The Confiscation of American Prosperity. Here it is.

    Economists’ glaring lack of concern about the level of inequality owes a great deal to Simon Kuznets, a Nobel Laureate and brother of my adviser in graduate school. In his presidential address to the American Economic Association in late 1954, Kuznets laid out what came to be the prevailing view about the natural course of inequality. Seeing a trend toward economic equality by the end of the 1940s based on his analysis of data from United States, England, and Germany, Kuznets proposed that economic inequality naturally follows an inverted U-shaped curve. According to Kuznets, when poor countries first began to develop, inequality increased until the economy became more sophisticated, which then set in motion a trend toward more equality.
    Kuznets, who won his Nobel Prize for his careful analysis of economic data, was understandably modest about this suggestion, referring to his lecture “as a collection of hunches” rather than the sort of painstaking analysis for which he was justifiably famous (Kuznets 1955, p. 26). Kuznets’s followers were far less modest than Kuznets, recasting his casual observation almost as a natural law of market economies.
    The evidence that markets are a benign arrangement with a built-in tendency toward equality is not particularly convincing. A longer view of the developed market economies over the last century and a half suggests a different pattern.
    Market economies have a tendency toward inequality. Kuznets himself suggests that the tendency of the wealthy to have more savings tends to lead to an increasing concentration of wealth and income (Kuznets 1955, p. 7). As I suggested earlier, at some point inequality of income becomes so extreme it sets off an economic crisis. In the wake of such catastrophes, new institutional arrangements arise, which temporarily moderate inequality. In fact, first among the causes that offset the tendency toward inequality Kuznets lists “legislative interference” and “political” decisions (Kuznets 1955, p. 9).
    Writing at a time when the momentum of the New Deal had not entirely dissipated, Kuznets did not acknowledge that such egalitarian policies depended on political support. After a short time, the rich and powerful seem to be able to regroup and to assert their powers. As a result, the institutional support for a more equitable income distribution frays and the tendency toward inequality begins anew.
    Economists embraced Kuznets’s hunch so enthusiastically because it had such comforting implications. Coming in the midst of the Cold War, Kuznets’s hunch had an even more urgent message: unfettered markets could naturally accomplish what socialism could only promise.
    So, political leaders should just let business promote economic growth rather than even thinking about directly redistributing resources to help the poor, the message Lucas later echoed. Markets are not just the best means of creating widespread prosperity, but also a sure route to equality.
    Unfortunately, Kuznets’s idea proved to be overly optimistic. By the late 1940s, just as Kuznets was about to launch his theory, the forces that were promoting equality had already lost much of their momentum. Looking back from 1980, Jeffrey Williamson and Peter Lindert, authors of a classic book on inequality of income in the United States, speculated:
    ##By almost any yardstick, inequality has changed little since the late 1940s. If there has been any trend, it is toward slightly more inequality in pre-fisc [pre-tax] income and toward slightly less inequality in post-fisc income. This stability has been extraordinary even by twentieth-century standards. [Williamson and Lindert 1980, p. 92]
    Even the more modest appraisal of Lindert and Williamson turned out to be overly optimistic. In retrospect, instead of stability, a powerful trend toward inequality had already begun to take hold by the time they published their book. What appeared to be stability in 1980 was only a temporary pause. The right wing had already launched its renewed offensive to restore the sort of inequality that existed just before the Great Depression. As the Soviet government receded into history, pretenses of egalitarianism no longer had any political value.
    By this time, anything that threatened to inconvenience corporate balance sheets had to be vigorously attacked. Are wages too high? Then, attack labor. Reagan’s firing of the air traffic controllers union set the stage for a widespread attack on labor.
    Are regulations bothersome? Then, eliminate them. Within a few decades, the corporate sector was supreme, unburdened of a good part of the regulatory structure that had been in place at the end of the 1960s.
    The new economic climate freed business from much of its obligation to pay either decent wages or taxes. Despite following the recipe of the mainstream economic cookbook, the miraculous trickle down never materialized.

  3. Kuznets’ argument cannot be separated from some measure of growth. If he claims that rising inequalities are necessary for growth, then you should also look at growth. See if growth occurs with the inequalities.

    I agree with the analysis in general, that the model is based on cross-sectional data not time-series data, and that it was incorrectly concluded that rising inequalities eventually give rise to equalities.

Leave a Reply

Fill in your details below or click an icon to log in: Logo

You are commenting using your account. Log Out /  Change )

Google photo

You are commenting using your Google account. Log Out /  Change )

Twitter picture

You are commenting using your Twitter account. Log Out /  Change )

Facebook photo

You are commenting using your Facebook account. Log Out /  Change )

Connecting to %s