Tuesday, May 5, 2015

Was There Ever Really a Big Tradeoff Between Equality And Efficiency?

Probably not.

In the last couple of days there was a big meeting at the Brookings Institution to celebrate the 40th anniversary of the publication of Arthur Okun's influential 1975 book, Equality and Efficiency: The Big Tradeoff, an idea that went into the textbooks and became a VSP de rigueur truism for quite some time.  The conference was asking its relevance today, which it looks like some questioned.  Many big cheeses were participating, representing a range of views from Mankiw on the right to Heather Boushey on the left, with Larry Summers making the introductory speech, and Janet Yellen in the audience.  But the real  question is: Was it ever at all true to begin with?

Mark Thoma provides some quotations from Summers's opening address.  He clearly seems to think that the relationship has gone off the rails.  The quoted material talks a lot about rising inequality and increasing problems in the financial sector, with the undertow of slower growth and his bugaboo of secular stagnation.  Nothing I saw there looked unreasonable.  But, again, I am wondering if anybody asked the deeper question in a location where the late Okun was long highly revered, and I think still  is.  I never met the man, but plenty there think he was just wonderful, and maybe he was.  But I think he was never right about this Big Tradeoff, or not very right.

OK, so back in the day we had the Big Poster Boy of communist states with lots of equality, think Maoist China, but not a lot of either growth or high income.  Yes, there was certainly an argument there: a society with essentially no incentives to make capital investments or technological innovations, with those incentives maybe leading to greater inequality (and indeed the fall  of communism did lead to higher inequality, with China in particular growing rapidly since, an outlier to support Okun's argument, even as many other former socialist states did  not do all that well).   So, he had some things going for him.  But all along there was this problem of Latin America and parts of Africa, with the highest levels of  inequality in the world and pretty pathetic growth and income records.  This led to an eventual  modification of the truism, usually voiced as indeed going against this deep truth, the idea that the relationship between at least per capita income and inequality being an inverted U-shape.  It also increasingly was noted that the source of inequality was important, with it coming from being "earned" rather than inherited or through corruption making a big difference.

That the relationship was not  all that it was cracked up to be dates back at least to the turn of the century, even among those who ideologically might be inclined to support it.  Thus, in 2000 Robert Barro published a paper in the Journal  of Economic Growth, Inequality in a panel of countries.  He summarized that there is "little overall relation between income inequality and rates of growth and investment."  Barro also noted the likely inverted U-shape  relationship between per capita income and inequality.

In the previous year, a long paper by Aghion et al in the JEL argued that it looked increasingly like if anything the relationship might be the other way around, with more equal  naitons growing more  rapidly and programs to increase equality being associated with accelerated growth.  They posed the Philippines and South Korea, equal in per capita  income in 1960, but with the Philippines having twice the quartile income ratio of South Korea.  We know what their subsequent growth records have been.

I would push this forward to note the experience of Latin America, one of the poster boys for high inequality and poor growth performance.  During the Great Recession this was arguably the best performing region in the world in terms of growth compared to its several decades past.  Curiously, it was the only major region of  the world where incomes were becoming more equal, if still more equal than in the rest  of the world.  This does not prove anything, but then that China fits Okun's story does not prove it, and that nation increasingly looks like a big outlier, quite aside from its decelerating growth rate.

Update: I read the speech by Larry Summerslarrysummers.com/2015/05/04/okuns-equality-and-efficiency at the event. He praise Okun as brilliant, of whom Paul Samuelson basically said that he never said anything incorrect, someone not only insightful in economic policy, but also of philosophical depth.  His only comment on how things were back then was that the income distribution did not change.  He did not directly comment on the basic issue of whether or  not there really was an equality/inefficiency big tradeoff then.  He does now think that rising inequality has been linked to slowing growth. 

Second Update:  Mark Thoma links for 5/6/15 to Heather Boushey's talk at the Brookings conference, and after long discussing how inequality has soared since Okun's day, she finally comes to the same conclusion I do, that Okun was basically wrong and that his legacy is part of our problem today.  There  never was a Big Tradeoff, just a Big Myth about there being one (latter is my terminology).

Barkley Rosser


Peter T said...

I am struck by this phrase: "think Maoist China, but not a lot of either growth or high income". No high income certainly. But growth? Do not near-universal literacy, basic health care, huge amounts of infrastructure, the imposition of peace all count as growth? My Chinese acquaintances certainly thought so.

Second point. Why would you even want to think about growth and inequality divorced from the character of production and the structure of the economy? To do so has more the character of a philosophical musing than an empirical inquiry.

Cory said...

The link has "2005" in it when it should be "2015"

rosserjb@jmu.edu said...

Thanks, Cory. You were right. Works now.

Peter T.,

Yes, there was some growth. There was also a massive famine during the Great Leap Forward that probably killed more than any other in world history.

rosserjb@jmu.edu said...


Compare China between 1949-79 and since. Earlier, probably most equal income distribution in the world and had real growth, but much faster after 79 with sharply rising inequality. It is Okun's case, but an outlier in terms of most of the world's experience.

Jamus said...

Barkley, I agree entirely that the so-called tradeoff is certainly increasingly questioned, especially at the cross-country level, and with the eroding evidence for a Kuznets' curve. And in fact the cross-country evidence increasingly points to positive growth effects from a reduction in (especially extreme) inequality.

But I would add one minor point: it is true in a very narrow, technical sense: when we define "efficiency" as a move toward Pareto efficiency, and we define "inequality" as a change to the original distribution of outcomes, then using a PPF it is straightforward to see that, except in a very special case, any move from within the production possibilities space to the frontier will be acconpanied by a change in the distribution of outcomes, relative to the original one (even as both increase their share, just not proportionately).

Bill H aka run75441 said...

Maybe I was missing it during my on and off tenure in China, where has Labor gained when compared to Labor in the US? As my one counterpart Chinese Engineer pointed out upon passing the acres of unoccupied structures, etc. , the latest infrastructural growth binge is all about growth and not about people or labor.

Don Coffin said...

Barkley--I have argues, ever since I first read Okun's book, that it contained an obvious problem. Equity is an objective, and efficiency is a means. If we want a more equitable distribution of income, we would like to achieve that distribution efficiently, right?

What Okun was actually arguing, I have always thought, was that there is a tradeoff between equity and measured real GDP. Somehow, maximizing measured real GDP got conflated with efficiency. (I will note that I also think the notion of a tradeoff between equity and measured real GDP is also likely not to be true.)

rosserjb@jmu.edu said...


I do not think the tradeoff holds when there is a move towards the ppf of a nation. Such a move could involve any possible array of distributional outcomes, including even some groups losing out, either very rich ones or very poor ones or even middle tier ones. All we know with a move towards a ppf is that more of all goods, or at least those depicted on the axes, is being produced. Who gets those extra goods, or even that somebody might actually be getting less of them in this process, is simply unknown and needs more details of how this move is happening.

Bill H,

I do not think that I said a thing about labor shares in what has happened in China. I have not actually looked at that. What I do know is that their Gini coefficients have risen dramatically, one of the largest increases anywhere. At the height of the Maoist period, the one measure of the Gini I ever saw had it at .16, as low a Gini for any nation I have ever seen, and by all accounts it was a highly egalitarian "Iron Rice Bowl," even if that measure undercounted in-kind perks that elites got back then. Today the Gini there is over .40, on a par if not higher with the US and Russia (which was in the mid-20s) at the end of the USSR. But, while I suspect labor's share has fallen, there is a problem in China of a share going to the state, given ongoing high levels of atate ownership, although in many firms this is parrial ownership. Labor share is not something all that easily or clearly measured there.

I think you are right that Okun wanted to see some greater equality but with it being achieved in an "efficient" manner, whatever that may be. Certainly in terms of general discussion this tradeoff has pretty much always amounted to "efficiency" meaning high aggregate growth (and high levels of income also), with it really being a growth-inequality tradeoff at the bottom line, even the word "efficiency" (suggesting at least Pareto superior, if not fully Pareto optimal) being used for that.

Sandwichman said...

It's ironic that it was Paul Samuelson who eulogized Okun as never having said a stupid thing. The equality/efficiency "tradeoff" is a classic "same yardstick" fallacy (See Samuelson, 1974, cited in my post Decisions... Decisions)

Thornton Hall said...

Talk about burying the lead! The following doesn't come until Bouchey's last paragraph:

"It’s hard not to see Okun’s legacy as the problem rather than the solution."

Why is it that economists can barely perceive that they are not analyzing the economy we live in, but rather, they have made the economy we live in by hiding moral philosophy in the sheep's clothing of policy advice?

When Okun was writing the transistor had been recently invented by a quasi-governmental monopoly, the Internet was in the process of being developed by Al Gore, and anti-biotic drugs had changed medicine thanks to a National Laboratory that came in after Merck failed.

The interesting question is this: what was wrong with the people who believed Okun in the first place?

Sandwichman said...

In fairness to Okun, though, he met the highest standards of his profession: "...exact remembrance of spells, unimpeachable performance of the rite, unswerving adhesion to the taboos and observances..."

Typo Boy said...

I also question the idea that China was egalitarian. Yes, the GDP was divided amazingly equally. But political power was extremely uneven. In the US right now, we have both kinds of inequality. And of course there is gender inequality and inequality among ethnic groups and so on. I'd be curious to see an expanded GINI coefficient that includes all those dimensions and graph it against GDP growth. Though GDP is not a measure of human welfare, I suspect such a graph would show a strong negative correlation - the higher full GINI was, the lower the rate of increase in GDP. Unfortunately I'm not sure that is empirically testable. Because I don't know if one could draw up a meaningful numeric equivalent of the GINI coefficient that included all those dimensions (including ones I have not mentioned such as GLBT and disability).