You are currently browsing the tag archive for the ‘econometrics’ tag.
It seems like discussion of Piketty’s Capital has run its course and much of the commentary has moved on (though not necessarily from the broader topic) so now is as good time as any to peer back and reflect on how the debate around the book ended (if such a thing can be summarized). From my own vantage point, the debate about the book (not necessarily the discussion) stalled out around a single question, so I will do my best to restate and clarify that question so as to focus where more evidence and argument is needed, should this be a conversation anyone wishes to resume. None of this is new, exactly, but it’s worth recanting given the importance of the question and the stakes surrounding it.
Around 1800 AD, living standards in some countries began to rise substantially, and over the past 200 years, that rise (as measured in GDP per capita) has been on the order of a factor of 50. This generally seems to correlate with other indicators of increased living standards to a degree that, with some exceptions (such as thinly-populated resource-rich countries) it is generally, though not universally, accepted practice to use GDP per capita as a good-enough shorthand for broad living standards. Whatever the case, exactly how and why this increase transpired is still a matter of debate, in no small measure because most people would find it desirable to replicate the phenomenon in those areas that have not yet experienced it. Indeed, some countries that did not begin experiencing the phenomenon in its initial emergence have experienced it since, leaving, essentially, three groups of countries – those who have experienced it, those who have not, and those in transition.
Piketty’s book, while not exclusively, overwhelmingly is focused on the first kind of country. A compelling portion of his narrative is documenting that transformation, yet the broader focus of the book is on what has transpired since that transformation was consolidated in the era following the Second World War. There are two key factors to be documented. The first is that the countries that have fully experienced this transformation are themselves not ‘complete’ in this regard – average living standards (recent economic troubles excepted) continue to rise and are generally, though not universally, expected to continue to rise in the absence of extreme calamity on the scale of global catastrophic climate change. The second is the change in the distribution of income – since a moment of ‘peak equality’ in roughly 1970, most of the countries Piketty analyzes have seen a sharp increase in inequality, the specific degree of which dependent on method of measurement but whose general contours is not really disputed. This, Piketty and many other believes, poses a problem for these countries that is not alleviable solely by continuing increases in average living standards or aggregate wealth and income growth.
Piketty devotes a lot of space to developing a simple model of how the aggregate quantity and distribution of capital can drive income inequality. This remarkably simple model requires only three input variables – the growth rate of the economy, the average return to capital, and the savings rate (perhaps better phrased as the rate of capital formation relative to national income) – to generate a long term prediction of two key ratios: the ratio of capital to income, and the capital share of national income. From there, wealth inequality can be used directly to compute a floor on income inequality – for example, if 1% of the population owns 50% of the national wealth and the capital share of income is 30%, then that 1% captures, at a minimum, 15% of national income.
And here we arrive at the crux of the debate. Piketty’s model implicitly assumes a certain exogeneity between those three input variables and the two ratios they converge towards, ie, that they are not inherently correlated with each other. This exogeneity poses a fragility in Piketty’s model and a challenge to mainstream economic theory. The fragility is that, if they are strongly correlated (in the direction such correlation is expected), and especially if there is iterative feedback between them over time, then Piketty’s model no longer produces outcomes in which wealth inequality drives income inequality. The key example here is the average return to capital; were it to fall in proportion to the rise of total capital accumulation, then the capital share of national income would be invariant to the quantity of capital, and thus largely undermine the mechanism by which present wealth inequality drives future income inequality. Furthermore, were this anticipatable decline in the in return to capital to drive a decline in savings, the capital/national income ratio would converge at a substantially smaller value than that projected by extrapolating from the initial period. This further depresses the likelihood of ever-increasing wealth-driven income inequality.
This is also precisely the challenge to mainstream economic theory. These correlations and feedbacks are precisely what are predicted by fundamental, strongly-held ideas about economics held by economists; most centrally that investment behavior is driven by that most central economic force, supply and demand. Piketty, however, is not simply laying down an alternative model, but an empirical challenge to this challenge. The most crucial assertion made by his model – that the return to capital fails to decline in proportion to the supply of capital – is not simply a theoretical alternative but one derived from the meticulously researched and calculated estimates in his unprecedented data. As I myself pointed out in my write-up of Piketty’s book, the data show that the return to capital is sufficiently resilient to its accumulation to justify Piketty’s model. At least, that is, without controlling for any additional factors.
And here is where debate stalled, with one side asserting that theory demands these variables be tightly correlated, and the other side responding that empirics demonstrates that they are not. The problem, of course, is that macroeconometric panel empirics is extremely sensitive to model specification, to the point of being perhaps the perfect example of how any decent statistically-versed researcher with strong priors can generate the outcomes from the data they which to receive. Certainly it is more than possible to generate a superfluity of complex models demonstrating the theoretically-predicted correlations, and these models will collectively have zero persuasive power because it is trivially easily to create as many or more equally-plausible equally-complex models that demonstrate the obvious.
Why does this all matter, to the degree it’s worth recounting in such detail to the tune of a thousand words? Because it strikes directly at the heart of the most important argument for tolerating high income inequality.
There are basically three arguments in favor of tolerating high income inequality, which I will attempt to summarize as fairly as I can.
- The ‘Just Deserts’ Position: incomes reflect the inherently just outcomes of markets. Beyond a certain threshold to prevent the worst form of miseries, it is therefore a violation of justice to take from the deserving and distribute to the undeserving.
- The ‘Pink Salt’ Position: income inequality is irrelevant except to the irremediably envious, resentful, or spiteful. What matters is preserving and increasing human happiness, which is largely driven by civil liberties, non-market institutions such as family and community, and the secondary impacts of economic progress.
- The ‘Golden Egg’ Position: income inequality may be ceteris paribus bad but aggregate economic growth is extremely good to a degree that in most plausible scenarios swamps income inequality. Furthermore, income inequality and economic growth may be conjoined outcomes of our economic system and cannot be modified independently. Therefore, we should be extremely cautious about attempting to alleviate income inequality through policies that slow the rate of economic growth, as this may reduce not just aggregate utility but the utility of those benefiting directly from redistribution.
It will shock nobody to hear that I reject outright the first argument in the strongest possible terms, and the second in quite strong terms as well. Indeed, I believe that the majority of Americans, and certainly the majority of voters in developed countries, disagree with those arguments as well. It is that third argument that gives pause to many – including, to a degree, me (though that pause is still far from convincing in my own case). The average person living in a developed country today as compared to a person living in that same country in 1800 is vastly better off, and it is not impossible to imagine that the average person living in a developed country in 2100 will be vastly better off than that average person today. Impeding our shared progress in that regard could simultaneously defer developments that improve the quality of most lives while simultaneously deferring developments (like innovation in renewable energy sources and storage) that could mitigate or reverse the worst consequences of economic growth to date.
This all converges on something of an ironic surprise. In this debate, it has been the left that has been advocating, implicitly or explicitly, on behalf of the resilience of capitalism (broadly defined) and its ability to deliver human prosperity, whereas it has been the right that has claimed, implicitly or explicitly, that capitalism and the prosperity it delivers is fragile, so much so that even increasing post-market redistribution (as opposed to pre-market regulatory redistribution through minimum wages, stronger protections for unions, and abridging the current rights and privileges of lenders and shareholders) could, to use a tired aphorism, kill the goose that lays the golden eggs. This ideological positioning isn’t wholly novel, and whether it is instrumental and ephemeral or representative of something larger remains to be seen; but it is notable, and worth pondering for what it says about the state of both the contemporary mainstream left and right movements in the United States (if not beyond).
A weekend thought: my father is the kind of guy who likes to come up with big monocausal theories to explain every little thing; he missed his calling as a columnist for a major newspaper. Anyway, last week we were chatting and he expounded on one of these theories, in this case a coherent and compelling narrative for the dramatic increase in dog ownership in recent years. The theory is unimportant (it had to do with a decline in aggregate nachas) but afterwards I decided for the heck of it to fact-check his theory. And what do you know? According to the AVMA’s pet census, dog ownership rates have declined, very slightly, from 2007 to 2012.
Now, I know why my dad thought otherwise – over the past few years, dogs have become fantastically more visible in the environments he inhabits, mainly, urban and near-suburban NYC. I am certain that, compared to 5-10 years, ago, many more dogs can be seen in public, more dog parks have emerged, and there are many more stores offering pet-related goods-and-services. But these are intertwined with substantial cultural and demographic changes, and authoritatively not driven by a change in the absolute number of dogs or dog-ownership rate.
It’s hard to prove things with data, even if you have a lot of really good data. There will always be multiple valid interpretations of the data, and even advanced statistical methods can be problematic and disputable, and hard to use to truly, conclusive prove a single interpretation. As Russ Roberts is fond of pointing out, it’s hard to name a single empirical econometric work that has conclusively resolved a dispute in the field of economics.
But what data can do is it can disprove things, often quite easily. While Scott Winship will argue to death that Piketty’s market-income data is not the best kind of data to understand changes in income inequality, but what you can’t do is proclaim or expound a theory explaining a decrease in market income inequality. This goes for a whole host of things – now that data is plentiful, accessible, available, and manipulable to a degree exponentially vaster than any before in human history, it’s become that much more harder to promote ideas contrary to data. This is the big hidden benefit to bigger, freer, better data – it may not conclusively prove things, but it can most certainly disprove them, and thereby help better hone and focus our understanding of the world.
There’s not much to add to the specific embarrassment and evisceration to which Matt Yglesias subjects this loathsome David Brooks column over the nexus of arrogance, lack of self-awareness, self-righteousness, and callousness that is his blathering on marriage. But Yglesias does touch on something broader worth elucidating.
In basic econometrics practice there is something called an “interaction term” that can be very very important. Without getting too much into jargon or technical stuff, the basic idea is that you are trying to isolate the effect that two or more different things have when they are combined from the effect each has when they are separate. The interaction term is what isolates and captures that difference. Sometimes, they show that the effect of one thing – say, the effect of marriage on income – is very different depending on the other thing – say, gender. Men might make more money if they get married, women might make less.
My guess is that life is full of these interaction terms, and that poverty is a really, really big one, and therefore the effects of almost everything are different on poor people. When rich parents get divorced, it is emotionally miserable for everyone. When poor parents get divorced, it is emotionally miserable for everyone and can entail substantial economic dislocations and suffering. But the solution isn’t to remain in an unhappy marriage for financial reasons. The solution is to ameliorate financial hardship so people don’t feel trapped in unhappy marriages.
Yesterday a friend of mine tweeted an invitation via a new service called Feastly. The invitation was to come to her home and eat a delicious, home-cooked gourmet meal in exchange for money. The service, Feastly, is set up to do exactly that – while it is still in private beta (and therefore cannot be fully-explored until one is invited in) it clearly aggregates offerings of that sort, sortable by dietary restrictions, price, attire, pet-friendliness, and other criteria. It’s a great idea, and one I wish I thought of.
On a social scale, I think as we see more services like this that directly connect buyers and sellers – think eBay, Etsy, ebook self-publishing – it will throw further into question whether statistics like GDP/GNI are useful metrics, not just of broader concepts like "standard of living," but of what they purport to measure. Every meal eaten on Feastly and not at a formal restaurant is one that involves an exchange of goods and services for money, and most of them will likely not be counted by current methods of measuring GDP. This issue predates the internet, of course, but the internet’s amazing power to match small-scale producers to buyers will accelerate this trend, as will the advent of 3-D printing.