Friday, February 10, 2012

Thoughts on Income Inequality - Part I

Ever since I wrote this post on President Obama’s increasingly strident language about the unfairness of reported increases in US income inequality, the issue has only become more prevalent in our Oped pages, political blogs and journals. The fervor of this newfound fascination is perhaps not surprising given the popular backlash against mind-numbing Wall Street compensation during these times of high unemployment and stagnating wages.

I have earnestly sopped up much of the recent writing and some of the scholarship on the issue, even wading into the pages Mother Jones (here is a good example). And while many of these authors bring up reasonable and concerning points, I find that they all suffer from the same fatal flaw of logic: they assume that the simple act of decreasing income inequality will, by itself, help the neediest among us. I do not necessarily subscribe to this view, and will outline common flaws in the typical analysis below.

First, let me make a stipulation: in any given society, there is a meaningful correlation between freedom and wealth. If you disagree with this stipulation, then you can stop reading right now – you will disagree with my conclusions. However, in order to disagree with it, you’d have to disagree with quite a lot of scholarship and some very simple, but quite compelling analysis. For instance, the chart at right from economist Scott Sumner is a linear correlation between the Heritage Freedom Index and GDP per capita for all the countries of the world with GDP per capita greater than $23,000. The picture is quite demonstrative: it suggests that the more economically free a country, the higher its GDP per capita.

Now, I’ll acknowledge that there are numerous problems with such an analysis. Historical circumstance, lack of natural resources and other considerations may depress a certain country’s GDP per capita in ways that would not be captured by a Freedom Index. Conversely, countries like Norway, with its vast oil wealth, and Switzerland, as the banker to much of the world, skew the analysis in another direction. Therefore, I only introduce this chart as reasonably compelling evidence that, on average, within certain boundaries, the more free a country, the more chance it will have larger GDP per capita. Finally, I hope it goes without saying that although economic freedom may be a necessary precondition for wealth, it certainly is not sufficient.

Next let me stipulate that the tools and policies which a country like the United States might employ to reduce income inequality would, in all cases, serve to reduce economic freedom. The most common solutions to the income inequality dilemma are increasingly progressive taxation schemes, transfer payments such as food stamps and unemployment insurance, and, in some cases, increased regulation to reign in the worst perceived cases of wealth generation (think of the call for regulation of compensation during the financial crisis and the more recent calls for hedge fund taxation reform and registration). Each of these things might be effective in reducing income inequality, but they would undoubtedly also place a country lower on the Economic Freedom Index. When someone finds a way around this, please let me know – we’ve solved the problem!

One implication of my initial stipulation is that there is an unintended consequence of the adoption of these policies: the whole pot gets smaller. If freedom means wealth, and reducing income inequality requires reducing freedom, then reducing inequality requires reducing wealth. Liberals, including many economists, tend to ignore this point in their editorializing; sometimes it’s simply due to ignorance, sometimes I’m convinced it’s disingenuousness.

Now, these policies may be implemented with endless variety, each having a different effect on GDP. In economic terms, the negative multipliers of various policies are different, and economists could fine tune a set of recommended policies to minimize the effect on the country’s ability to provide for itself. Unfortunately, I suspect that the Policies which might have the greatest effect on income inequality might also be the ones which depress GDP the most.

But let’s revisit my initial thesis: that liberals’ distaste for income inequality, qua income inequality, is misguided. Let’s look at the current income distribution in the United States. In its various reports on gross income, the Commerce Department divides Americans into quintiles. The chart to the right shows the mean gross income for households in each quintile, with the poorest 20% of households bringing home $11,000 per annum and the richest 20% bringing home $169,000. Indeed, the rich have a lot more income than the poor.

Two comments are in order before we proceed. First, by organizing the analysis by quintiles, I am really talking about the vast majority of Americans. Conversely, when one sees statistics about income inequality bandied about in the liberal press, the conversation usually compares the top five percent, the top one percent, or sometimes even the top one-tenth of one percent of earners to the rest of America. While the numbers might look incredibly compelling when viewed this way, the numbers of the truly, truly wealthy are so small that it is of little use to economists or legislators in crafting policy. There are not enough billionaires to make a real difference on the deficit, for instance, solely by taxing them.

Second, please note that these figures relate to mean gross income, not including transfer payments. In order to be truly fair, any analysis of income inequality should really be done on a median basis, net, after tax, including transfer payments. Those numbers are hard to come by, however (and they don’t look as compelling to liberals!), so most people use mean gross income. I’ll stick to the traditional method and remind readers of any misleading points. [For instance, if one divided the sum of welfare payments, Medicaid, food stamps, and child subsidies by the number of households in the bottom quintile and added the result to their mean income, I suspect Figure 2 would look somewhat different. It’s why conservatives often point out that the average household living under the poverty line owns multiple cars, even more television sets, and the most significant health issue for impoverished American children is obesity].

Let’s assume that a liberal administration implements a tax policy with a goal of both raising revenue and reducing income inequality. Presumably the intent of any such Policy would be to change the income distribution from the figure above to look more like the green line in this chart to the right. (Let’s ignore the fact that we are analyzing income on a gross, not net basis!).

But we’ve already established that these policies might also likely reduce the amount of GDP per capita. As such, we should ask ourselves a question: is it possible that rather than the new targeted distribution, we might actually get one which looks more like the one here to the right? This new income distribution also decreases income inequality, but everyone is worse off. Importantly, even the poorest people in the nation are worse off than before.

I would like to think that even the most die-hard liberal would concede that this is a suboptimal result, but I am not so sure. The prevalence of recent opinion pieces which conjure up happiness studies and other evidence suggesting that income inequality itself is a cause of mental distress may suggest that liberals are intent on providing cover for this dirty little secret. The line goes like this: well, yeah, maybe nobody in Europe has as much money as Americans at any level, but, hey, they’re all a lot happier than we are! This line of thinking has so many faults that it will require a separate post to deal with – here I’ll move onto more pressing topics.

Given the possibility that any particular policy might affect all points on the income distribution curve, it seems to me that each policy ought to be analyzed not solely on its capacity to flatten the curve, but also its likelihood of affecting income generally (either positively or negatively). Politicians should be required to tell us what they think the optimized income distribution curve ought to look like: how flat it should be and how much net, adjusted income should be claimed at the lower quintiles. It seems to me that the very least important consideration should be the steepness of the curve at the far right of the distribution, and yet this is the very point over which so many people are expelling so much hot air!

Unfortunately, economics being the dismal science, it’s not so easy to determine the true effect of a given policy on the shape of future income distribution curves. I suspect Paul Krugman would give a very different answer than Art Laffer if President Obama asked each of them to analyze his pending tax proposals. Nevertheless, this should not stop politicians from being required to tell us what they think the curve ought to look like and how we’re going to get there. Then, we can have a national debate comparing the merits of Laffer’s model against Krugman’s as a means to analyze the proposals.

Doubly unfortunately, this is not how politics works. I have been dismayed at the breathtaking disingenuousness of the current rhetoric surrounding Obama’s proposed “Buffet Rule” – the fuzzy notion that millionaires ought not pay a lower effective tax rate than secretaries. In the thousands of words that Obama has spent on this issue, he’s never once admitted that there is a difference between the taxation of labor and the taxation of capital, and the reason that Warren Buffett pays a very low effective tax rate is because virtually 100% of his income is from dividends and capital gains rather than salary. Somehow, it doesn’t matter anymore that we got here (i.e. Buffett paying 15% effective tax rate) for a reason. That reason is because people argued, correctly or not, that low tax rates on dividends and capital gains would be good for all of us, including those at the bottom of the income distribution. Instead of revisiting that proposition, for some reason, we’ve skipped the analysis and started appealing to Obama’s sense of “fairness.” Obama’s fairness may end up costing all of us – even those at the bottom of the income distribution – a whole lot.

In fairness to Obama, I do not know the answer. My gut tells me that low tax rates on dividends and capital gains are good for poor people, but I have not personally analyzed the issue yet to my satisfaction (another Polartics post coming!). But we should not let liberals jam policy down our throats without a clear answer from them on this issue. We should call out any politician, blogger or editorial page writer who favorably invokes the Buffet Rule, but does not tell us why they think raising Buffett’s effective tax rate won’t affect everyone negatively, including poor people. That even goes for Warren Buffett!

That’s it for now, but if you want more good thoughts on the issue, here is a splendid article (from the reliably unreliable Economist) on “fairness.”

And for those of you who suspect this article may be self-serving, unfortunately, I have not achieved that level of success yet where the majority of one’s income is comprised of dividends and capital gains; my effective tax rate last year was 30.1%.