Do Taxes Matter?

Elsewhere, I’ve written about the long-run negative consequences of taxes, and the possibility that the top of the Laffer Curve is not so far off. Karl Smith takes exception to these arguments, when looking at this graph:

The supply of working-age people into the labor market seems to be largely driven by changes in monetary policy, rather than changes in taxes. For Karl, this seems to imply that the choice to enter or exit the workforce is largely pre-determined, and responsiveness to to taxes must be pretty low (he has another recent post that makes the same argument; though it is only shows up in my Google Reader).

I’m pretty skeptical of this line of analysis. Rather than eyeball aggregate data like this, I’d rather look at evidence that spans different countries over time, carefully calibrated analyses of specific tax cuts, or modeling based on economic data. All three methods seem to suggest that the responsiveness of people to taxes is pretty high. Even Saez, who recently won a MacArthur genius award and is generally regarded as being on the left on these issues, has a recent paper that argues the impact of taxes can be sizable, especially for those that itemize deductions and higher-income folks (his paper also suggests that the impact of taxes is taking place through channels other than simple labor supply). I think this is all evidence we should take seriously, regardless of what aggregate data suggest.

But none of this directly answers his chart. Well, here are a few more charts:

This one shows the average federal tax/GDP ratio of the American economy going back a few decades. Across a wide variety of tax regimes that have drastically changed average and marginal tax rates, the average tax/GDP ratio has stayed within a fairly tight band. It seems to be driven far more by changes in the business cycle (especially the stock market) than changes in statutory tax rates.

I wouldn’t subscribe to this interpretation entirely: but one plausible reading is that the long-run Laffer curve maxes out for a pretty low value. Successive tax cuts from the high rates of the ‘50s have been accompanied by economic expansion. Tax hikes from the ‘80s have been accompanied by economic slowdown. A tax/GDP ratio of 18% is, if not optimal, close to a maximum of what can be obtained without seriously affecting work incentives. European countries get higher tax/GDP ratios primarily by lowering GDP, not through higher tax revenue.

Here’s another graph in support of that interpretation:

Somewhat surprisingly, America gets about the same tax income per capita as Western European countries with far higher tax rates. That’s because America’s lower tax rates balance out Europe’s lower income almost exactly. I don’t like generalizing about “Europe”, which contains a variety of different systems, including the low-tax, high economic freedom Switzerland. But a simple reading of this graph would reinforce the conclusion that higher tax rates primarily lower economic output, rather than raising tax revenue.

Back to Karl's original graph — note that labor force participation rates are lower than Europe, at least normally. This was not the case several decades ago, which suggests that “culture” is not the reason. The culprit, depending on the country, seems to be a combination of rather high rates of taxation, high rates of people on disability, and the disincentives of social security provision.

I’ll add finally that this debate plays into whether or not we have a sustainable fiscal future. If you look at this CBO graph:

You’ll see that they project rapidly rising entitlement spending, driven primarily by Medicare and to a lesser extent by Social Security. You’ll also see that they project a rapidly rising tax/GDP ratio under the "baseline scenario", which will increase tax/GDP ratios far above historical norms (though this scenario would call for a full repeal of all of the Bush tax cuts, and so is not probably the path we will take). If the CBO is right, we can easily achieve massive increases in tax revenue without seriously affecting work incentives. If this story above is right (and, again, I’m not sure I’d agree with it entirely), a sharp rise in taxes/GDP would not be accompanied by large rises in tax revenue, but rather by stagnant growth in GDP.

The point here isn’t that we can raise tons of tax revenue by cutting taxes — I think the opposite is true. Also, I suspect labor force participation is pretty fixed for a large chunk of the population. But I think the aggregate data cut both ways, and the best evidence we have suggests that taxes have pretty bad effects.

To make a petty partisan point, I’ll also add that though the Republican plan for dealing with deficits in their Pledge has gotten a lot of flack, I don’t quite see how the Democrats are targeting this issue either. After slamming Bush for a decade for his irresponsible tax cuts, they now plan to extend almost all of them — except for those that may actually grow the economy, as the CBO suggests. Meanwhile, they are rolling out new expansions of healthcare entitlements and plan to hold the line on Medicaid, state and local spending, and Social Security. Conflict efforts in Afghanistan will ensure that large military cuts are not forthcoming either. How exactly is this supposed to be affordable? I really hope Karl is right on this one.

Rajan, the GSEs, and the Housing Bubble

Krugman and Rajan have had another round of combating articles going around.
I want to focus on one bit: the role of the government in creating the housing boom. This story is an important part of his book, Fault Lines, and I’ve argued before that this association is a little dodgy.

Krugman, of course, has made similar arguments too. In his response, Rajan switches his argument up a bit. He recognizes that Fannie and Freddie actually saw a decline in their activity during the worst years of the crisis. But he doubles down on his argument that the GSEs went into subprime; and he argues that they had a crucial role in “priming” the bubble.

The first claim — that the GSEs had a heavy subprime exposure — is an argument mostly made by Ed Pinto, a former credit officer for Fannie, and Calomiris from Columbia. They seem to rely on their own classifications of “subprime” . Most other sources find otherwise. For instance, the Fed claims 84% of subprime mortgages in 2006 were made by private lenders. Rajan argues that it is crucial to see not only how Fannie/Freddie were changing their overall market share during the boom, but also how they were changing their holdings of subprime. Yet it looks like they went from holding 48% of the subprime mortgages that went into the secondary market in 2004 to 24% in 2006. The quality of subprime mortgages, too, drastically worsened during that period, so Fannie/Freddie's subprimes were of far higher quality than those held by private companies during the boom.

Rather, it looks like the real problem was that private securitization drove mortgage origination. Just under three-quarters of mortgages went to securities in 2007, up from 56% in 1994. The trend was even more stark for subprime mortgages — 32% in 1994 to 93% in 2007 . I don’t want to go into this too much more because it’s fairly well-discussed elsewhere.

Rajan’s other argument is more interesting: that the GSEs did not participate in the worst of the bust, but were responsible for setting it off. Adam Ozimek also supports this story, and offers an interesting image to describe it:

Thus, a signal which traditionally could be used to hold prices in check was gone, and the only signal market participants were left with was prices themselves. It’s as if someone turned out street lights and the only way drivers could navigate is by looking at each others headlights. It’s easy to see how this could lead everyone collectively far from the roads despite behaving rationally individually given the information available to them. This uncertainty and unanchoring of fundamentals set off the herd behavior that drove prices even higher, this lured private companies in who eventually crowd Fannie and Freddie out of the market.

That is — the initial purchases by Fannie and Freddie in the pre-boom years distorted market incentives and set off a bubble that the mortgage giants did not participate in themselves. This is a new and interesting story, but I don’t think we have a good enough understanding of bubbles in general to say that Fannie/Freddie set them off this time.

Check out, for instance, Markus Brunnermeier’s account of "bubbles" in the New Palgrave Dictionary of Economics. He’s one of the best guys doing research in this field. One of the things he points out is:

[W]e do not have many convincing models that explain when and why bubbles start

That’s partially because it’s unclear how to fit bubbles into a normal economic framework in the first place. But Brunnermeier tries anyway, coming up with a few models of how to think about this.

You could, for instance, have a “rational bubble” in which all participants recognize that the bubble will crash, but continue anyway because they believe that they can re-sell to someone else. However, rationality imposes some tough constraints on bubbles like these. You have to assume that the housing bubble would grow slower than the economy; because surely a housing bubble could never get so large that we are forced to substitute away from housing entirely. That seems to rule out the housing bubble. It’s also entirely unclear how government purchases of housing would set off one of these. Surely rational investors with rational expectations would recognize that the government's intervention would taper off as a bubble grew larger.

Alternately, you can introduce some behavioral biases. Fannie/Freddie could have introduced some “noise trading” into the world, causing some investors to mistakenly believe that prices are trending higher. The GSEs, in other words, could have acted like predatory momentum investors.

Yet Milton Friedman’s arguments against bubbles remains powerful. Why would investors not respond to the addition of crazy money by betting against the bubble? Sure, most investors can’t technically bet against housing, but home owners could choose to rent, there are REITs out there, etc. Why couldn’t rational trading eliminate the mispricing induced by government spending, assuming that the government did intervene in unprecedented amounts before 2002?

As Brunnermeier suggests, these are impossible questions to answer. The government now is clearly investing heavily in housing; yet no one fears that will spark a bubble. The government’s heavy investments in Financial firms, too, seem to have failed to cause a bubble there. In general, government investments in a field "X" don't spark bubbles. Generally -- if you're a free-market economist -- you believe that this simply results in crowding out of private money, which then goes elsewhere. If Rajan's account were true, on the other hand, I would probably be far more willing to support government intervention in general. Simply by investing a little too much, it can kick-start a bubble! Yet "kick-starting" and "jump-starting" private investments through an initial government push have been key priorities over the past two years, and I have no reason to believe any of that is working. So how is housing so different?

The state of the economics is just too limited to figure out the cause of any bubble, let alone finger any one entity as the cause of the housing bubble. Rajan's always interesting to read, and I would like it if Fannie/Freddie were implicated in the Housing bubble directly, as opposed to merely being ridiculously bad uses of taxpayer money. But I'll have to side with Krugman on this one.