Sorry, but Steve Roth’s critique of my Inequality and Growth paper for CAP is misleading and inaccurate. Moreover, he had to work pretty hard to miss my points linking inequality to macro-instability and thus persistent periods of weak growth.
Not only does Roth somehow miss my emphasis on what I think are critically important linkages between inequality and growth, ones I repeat below. What’s particularly unfortunate about his critique is that while he purports to solve the hard problem that I laid out in the piece–correlations over the long-term between inequality and macroeconomic growth are hard to find in the US data–he fails to do so.
As I stressed in my piece, this is a key missing piece of evidence, and while I don’t see how he missed the other connections I did make in my piece, I read his critique with the expectation that he had found evidence tying inequality directly to slower growth through the consumption channel. He does not.
Instead, Roth just throws up a few FRED charts that merely obfuscate, arguing that they reveal evidence of a “massive, three-decade secular decline in spending relative to wealth over 35 years — the very same period over which we’ve seen massive growth in wealth inequality.”
That’s arguing by adjectives, not evidence. If all those “massives” are operative, it should be straightforward for him to show us convincing evidence linking these dynamics lead to slower growth of GDP or more pointedly, aggregate consumption. That’s the question I tackled and I encourage him to find what I could not. That would be an important contribution.
(BTW, if someone can think of a way to examine this through a state panel, versus the national data that I and Roth use, that increase in observations and variance would surely help in the search for the signal. The IMF cross-country work certainly points in this direction, but again, the challenge is to find this relationship in the US data, give the post 1970s increase in inequality.)
But—and here’s where he works hard not to see the more nuanced linkages I try to make between inequality and growth—the absence of this macro evidence in the aggregate variables tying growth to inequality does not at all mean inequality is unconnected to negative growth outcomes.
My key point, which again, is stated clearly throughout—in the intro, I signal this as “perhaps the most interesting finding in the report”—is that inequality contributes to credit bubbles which has obviously and negative impacts on growth (this is not an original finding; I’m building off of the work I cite by others, featuring in particular an important and under-appreciated paper in this space by Cynamon and Fazzarri).
I then devote a dominant section to this destructive chain of events. From the paper:
Source: my CAP paper.
I’m not saying this is the only way inequality hurts growth. Throughout the piece I present the logic of the basic argument regarding marginal consumption propensities—the idea that concentrated wealth among those with relative low propensities to consume should show up as weaker aggregate real consumption (per capita), post mid-1970s, when inequality started rising. I don’t find this in the data, and present arguments about why not, most notably wealth effects (again, note the linkage to the bubble sequencing above).
BTW, I especially suspect this effect is operative–it certainly should be–in the current recovery, given rising inequality, weak macroeconomic growth, and little in the way of offsetting wealth effects for the broad middle class. To be clear, that wasn’t the question of my CAP piece, which was a long-run analysis, looking for this relationship over the sweep of inequality’s rise since the late 1970s.
For whatever reason, Roth misses all that, but that’s not important, especially as he seems more interested in complaining that a progressive economist failed to confirm a progressive point.
What matters here is that if someone has convincing evidence of the relative propensity point, separate and apart from the bubble/bust dynamics that I emphasize at length (and Cynamon and Fazzari identify), let’s see it. It would advance this important debate and I guarantee you, I’d be among the first to highlight it. The fact that neither I nor Roth show it doesn’t mean it’s not out there. The logic is sound. But it will take more digging to make the empirical case.