Tuesday, June 30, 2009

Not the Fed's fault (at least its monetary policy)

Brad Delong is making sense:
What I do know is that the way the issue is usually posed is wrong. People claim that Greenspan’s Fed “aggressively pushed interest rates below a natural level.” But what is the natural level? In the 1920’s, Swedish economist Knut Wicksell defined it as the interest rate at which, economy-wide, desired investment equals desired savings, implying no upward pressure on consumer prices, resource prices, or wages as aggregate demand outruns supply, and no downward pressure on these prices as supply exceeds demand.

On Wicksell’s definition — the best, and, in fact, the only definition I know of — the market interest rate was, if anything, above the natural interest rate in the early 2000’s: the threat was deflation, not accelerating inflation. The natural interest rate was low because, as the Fed’s current chairman Ben Bernanke explained at the time, the world had a global savings glut (or, rather, a global investment deficiency).
Exactly. Monetary theory explains how if the Fed keeps interest rates "unnaturally" low, we'll experience inflation, a general increase in the price level. It does not explain how unnaturally low interest rates can produce massive relative price distortions and mispricing of risk. In fact, as I've discussed, the "global savings glut" is not enough to explain these either, although at least it's a little more coherent—to my knowledge, blaming the Fed for "unnaturally" low rates that didn't produce any inflation is a simple contradiction in terms.

At times like these, I wonder whether I'm just confused, and I have to admit that it's a distinct possibility. Yet I simply cannot understand how anyone can blame the housing bubble and financial disaster on the Fed's monetary policy. A post from David Beckworth (via another post by Delong) that purports to establish that rates were unnaturally low, for instance, seems to assume that the "natural" real interest rate is endogenously determined within the US—by American preferences and macroeconomic variables—even when the effects of global capital on the interest rate have been discussed to exhaustion. And again, this is hardly relevant, because there was no general price inflation, and I know of no conception of "natural rate" that means "whatever rate interacts with confused and irrational investors' psychology to prevent a bubble in real asset prices."

Now we're at the main point: you can argue that the Fed should have raised rates artificially high as a contractionary policy to burst the housing bubble before it became too dangerous. But this is not a very "natural" policy on the Fed's part, and it's not clear that it is even a good approach. I enjoyed this comment from Robert Waldmann in the comments section of Delong's post:
I definitely go with 3 mistakes not 4 as no way am I ever going to be drier than Alan Greenspan. The point, I think, is that there were much better ways to control the bubble than to impose high unemployment. In particular Greenspan had much finer tools.

The FED has regulatory authority which it didn't use. Even a Randian lunatic who is opposed to regulation in principle could have done things with lower cost per benefit than drive up unemployment.

For example he could have said "there is a housing bubble. It will burst. Only a total fool would count on an endless rapid increase in the relative price of housing.

Or "There is irrational exuberance in the housing market"

Or people securitizing and trading mortgages are going to regret it (a prediction or a threat doesn't matter he's Alan Greenspan).
Implementing contractionary monetary policy and creating unnecessary unemployment is hardly some obvious measure that the Fed missed out of incompetence. It is a speculative, indirect intervention into the market, even if it seems wise in retrospect.

Although many blame-the-Fed commentators seem to think they're grounded in monetary theory, they actually seem to be relying on ad-hoc theorizing about the nature of asset price bubbles. Since we don't have a very good understanding of the dynamics behind bubbles, they could be right or wrong, and without more solid quantitative footing, it's impossible to do any real cost-benefit analysis weighing the costs of unemployment against possible abatement of the housing bubble. There's a lot of false certainty here.

Sunday, June 28, 2009

Gas prices in 2008: not just speculation

Amid a generally unhinged rant in Rolling Stone against Goldman Sachs, Matt Taibbi restates one of the most clearly wrong economic fallacies around: that the spike in gasoline prices in 2007 and 2008 was the result of speculation, untethered to the reality of supply and demand.

To see why this is wrong, we need to think for a minute about what supply and demand mean. As we all learn to exhaustion in Econ 101, the "market price" is the price at which the quantity suppliers are willing to provide equals the quantity consumers are willing to buy. If the price of a good is higher than the one determined by current supply and demand, we have a surplus, as illustrated by this nice graphic cribbed from Wikimedia Commons:


In other words, if "speculators" were driving the price of oil far above the equilibrium price determined by supply and demand, by definition there must have been a surplus in the market, and a substantial increase in the amount of oil sitting around in storage.

If this happened, no one seemed to notice. Take this chart of OECD commercial inventories from the Energy Information Administration:

As we can see, the inventory levels before 2008 weren't any higher than usual, and the runup through mid-2008 was only slightly above the typical seasonal pattern (represented by the light blue colored band), despite the fact that oil skyrocketed above $120 by the beginning of May. Had speculation truly been responsible for the bulk of the oil price increase, and prices far above the level determined by supply and demand, we would have seen a substantial runup in inventories—remember that this isn't complicated theorizing on my part, but the very definition of supply and demand! While I'm open to the possibility that the extreme highs in the oil spot market, like the peak of $145 per barrel in July, were mostly speculative, it's clear that the majority of the price increase was the simple consequence of market forces. The only large increase in inventories we see in this diagram comes as the global recession sets in and demand plummets—after the price spike had already ended.

This is a very simple story. I've written about it extensively, and so has Paul Krugman (a slightly better credentialed economist than me!). It's remarkable that so many writers still fall for such a transparent fallacy.

The ideas man

Kevin Carey on Thomas Friedman's latest:
In a canny act of preemptive self-parody, Tom Friedman begins today's column as follows:

"I was at a conference in St. Petersburg, Russia, a few weeks ago and interviewed Craig Barrett, the former chairman of Intel, about how America should get out of its current economic crisis. His first proposal was this: Any American kid who wants to get a driver’s license has to finish high school. No diploma — no license. Hey, why would we want to put a kid who can barely add, read or write behind the wheel of a car?"

One can imagine Friedman's thoughts as he filed this beauty: "Suck on that, Matt Taibbi!" Friedman may not have invented the place-drop / name-drop / facile idea three-step, but he's certainly perfected it.

Internet space is free yet I'm still resentful of the resources about to be wasted pointing out how ridiculous this is. But okay: Many high school drop-outs live in cities where you don't need a car to get around. As for those who don't--let's say you drop out of high school because your high school is terrible or you get pregnant or there's a family emergency or you're 16 and prone to foolish choices. A couple of years go by and you realize you need that diploma. How do you go back to school if you can't drive a car to get there? Or get to your job and feed your family in the meantime? Friedman and Venter seem not to realize that a sizeable majority American teens don't attend Rydell High School. "Don't drop out, Kenickie--you won't be able to take your hot rod to the drive-in!" Plus, you don't actually need a high school education to be a good driver--for pity's sake, Dale Earnhardt was a drop-out...
The apparent lack of any quality filter in Friedman's brain remains astonishing.

Saturday, June 27, 2009

Two schools, two approaches

For anyone who thinks that the SAT is the primary culprit keeping disadvantaged students out of elite universities, let me provide an illuminating comparison:

School #1: Bates College led the "SAT optional" movement by dropping its test requirement all the way back in 1984, and has been lauded by critics of standardized tests ever since.

School #2: The California Institute of Technology places such importance on tests that the middle 50% of its first-year students on the SAT math section was 770-800, and the average SAT II math score (like all averages, weighted downward by outliers on the low end) is frequently above 780.

Now, which of these schools has a higher percentage of its students receiving Pell Grants -- a rough proxy for the percentage of disadvantaged students? It turns out that this isn't even close: Caltech has 15.3% of students receiving Pell Grants, while Bates has only 8.7%. Somehow, the school that stresses standardized tests more than any other in America manages to be a lot more economically diverse than the hip liberal arts college at the vanguard of the anti-test movement.

You may complain that these are just two data points, and you're right: this hardly provides conclusive proof. It's possible that Bates would be even more massively skewed toward affluent students without its test-optional policy, and there are plenty of other social and economic factors at work here. But if you think that standardized tests are the main barriers at selective colleges to disadvantaged students, this ought to at least make you question that belief. I suspect that other factors—namely high tuition, inadequate financial aid, the need to have carefully padded extracurricular "resumes" to gain admission, and the vastly greater resources available to rich students at private or elite public high schools—conspire to make a much bigger difference than standardized tests.

Something doesn't seem right

Jim Manzi responds to Conor Clarke's post on the costs of climate change:
...But Conor goes on to argue that the costs that Waxman-Markey is expected to impose on American consumers by 2050 – about $1,100 per household per year, or a little less than 1% of total consumption – are pretty trivial, because we should expect to be so much richer by then. (I’ll note in passing that, as per my posts on this, there are very good reasons to believe that the EPA cost estimate is low, and also that costs are also virtually certain to rise between 2050 and roughly 2100 when we would expect to start getting some offsetting benefits.)

He then shows a chart making the point, basically, that 1% is a small fraction of 100%. But of course, this cuts both ways. We hear constantly about the existential threat posed by global warming – Cities underwater! Drought! Famine! Think about his graphic. The expected benefits don’t even outweigh these costs. That ought to make you stop and think.
This does make me stop and think, but not for the same reasons as Manzi. Instead, I wonder: if there is a nonnegligible chance of famine, widespread droughts, and coastal cities falling underwater, how the hell does spending 1% of GDP outweigh the costs of climate change?

This isn't reason to be immediately dismissive of Manzi's analysis, of course. We're often very bad at comparing large costs and benefits. Large numbers all seem to blur together, and we can't really intuit the difference between a risk of 0.1% and 1%. But in this case, our intuition captures an important truth. The only reason Manzi manages to arrive at such low cost estimates is that his modeling framework makes a host of questionable assumptions, all of which dramatically push down the costs of global warming:

1. His sole welfare criterion is per capita world economic consumption. This ignores distributional considerations: if, in an extreme case, the consumption of the poorest nations in the world was completely wiped out, the world would be only "slightly poorer" according to this line of analysis. It also presumes that the environment only influences welfare through its effect on economic consumption.

2. He calculates the present value of future costs from climate change using high discount rates that make losses beyond 100 years practically meaningless, and dramatically lessen the importance of the future in general.

3. He dismisses the role that extreme uncertainty plays in our understanding of climate change, on the grounds that this is an untenable application of the Precautionary Principle.

Regarding the third point, remember that we know that the North Pole was once subtropical. We're not sure how this happened, but our best guess is that a smaller climactic change triggered a massive wave of amplifying feedback, as carbon dioxide and other greenhouse gases were released into the atmosphere and caused more and more warming. What reason do we have to think that this won't happen again, especially when we're attempting an unprecedented natural experiment where the atmospheric concentration of greenhouse gases will skyrocket in a geologically instantaneous timeframe? Sure, this risk is difficult to quantify and model, which is why geological feedbacks aren't included in the IPCC estimates that Manzi likes to use, but that's hardly a reason to pretend it doesn't exist.

What use is the irrational voter?

Bryan Caplan, economist of Myth of the Rational Voter fame, notes that "retrospective voting" isn't a cure-all for voters' economic ignorance:
Voters literally know less than zero about economic policy - we would have better policies if they just voted randomly. But people who believe in "retrospective voting models" often retort that voters' policy incompetence doesn't matter. They don't have to understand economics; they can simply reward politicians who deliver the goods, and punish those who don't.

Sounds good. But does it work? Tim Harford reports on some interesting research showing that voters actually reward politicians more for prosperity that they didn't cause. In other words, voters know less than zero about causal responsibility...
I think we need to be careful when interpreting results like these. It's true that voters don't have a good understanding of economic policy, and over short time scales attribute the responsibility for economic performance in a completely irrational way. Although they understand policy lags enough to give politicians a small grace period (as Obama is experiencing now), they tend to blame whoever is "in charge" for the current state of the economy, even when there is little legitimate reason to do so.

But what does this mean? Should we abandon democracy on economic matters, because the voters are clearly too stupid to offer any useful input? Here I think it's critical to distinguish between time scales. Voters' failure to correctly attribute responsibility in the short-term doesn't obviate their long-term usefulness as a check on our governing institutions. For instance, I'm sure that a council of our most prominent economics bloggers, given dictatorial authority to manage fiscal and economic matters, would do a better job over the next few years than our bungling Congress. You could include a range of political opinion: I'll bet that if the council consisted of Tyler Cowen, Brad Delong, Paul Krugman, and Greg Mankiw, it would accomplish more and enact wiser policy than our current political leaders, even amid large gaps in ideology.

I don't, however, favor a permanent switch to dictatorship by unelected economists, and I don't think anyone does. I trust smart academic economists to make far better decisions than the median voter, but I can't be sure that they won't make some serious mistakes over the long run, and long-term democratic accountability is the only viable mechanism we have to correct this. While voters may be economically clueless, they are capable of basic observations like "we have 20% unemployment -- let's try something different" and "the economy hasn't grown for 15 years," and their retrospective voting will in the long term keep policy roughly on track. Yes, they will also produce a lot of noise, and Caplan has done political economy a valuable service by noting that this noise is in fact negative in many ways. But I shudder to think how bad the Great Depression would have been if lassiez-faire economic orthodoxy had the power to override the popular will indefinitely, and the New Deal was never even allowed to leave the ground.

The real question is how our system will implement democratic accountability. There is a tradeoff between giving elite policymaking bodies more "slack" and making them responsive to the popular will. We've created some institutions, like the Federal Reserve and the Supreme Court, that tend to function well despite (and perhaps because of) their removal from immediate democratic control. It's quite possible that we're getting this tradeoff wrong, and that we still lean too far toward the "popular will." This is a difficult question of institutional design. I just think it's important to understand this tradeoff, and that all too often we opportunistically align ourselves with one extreme (voters have the right to decide) or another (voters have no idea what's good for them) depending on the issue at hand.

Wednesday, June 24, 2009

Too big to fail?

I'm skeptical of proposals that we eliminate systemic risk (and dampen moral hazard) by preventing firms from becoming "too big to fail." First of all, I agree with Paul Krugman when he notes that there isn't some convenient size beneath which banks don't pose a threat to the system:
I’m a big advocate of much strengthened financial regulation. One argument I don’t buy, however, is that we should try to shrink financial institutions down to the point where nobody is too big to fail. Basically, it’s just not possible.

The point is that finance is deeply interconnected, so that even a moderately large player can take down the system if it implodes. Remember, it was Lehman — not Citi or B of A — that brought the world to the brink.

And as far as I know, there never was a time when policymakers could have viewed the collapse of a major money center bank with equanimity.
I also think that the "shrink the banks" philosophy rests on some shaky assumptions about the nature of financial crises. The idea is that crises begin when the failure of a large firm leads to chaos in the system. In retrospect, this almost always seems like the correct story: as long as we have a core of large, important banks, any destructive financial panic is likely to consume at least one of them, and ex post facto we can finger the first such failure as the point at which the system really started to fall apart. This hardly means that we can avoid panics by writing large banks out of existence.

To claim that a system where risk is distributed among many small banks will demand less government intervention to prevent crises, I think you need to make a strong argument for the benefits of diversification. Replacing one large institution with twenty small ones does very little if the twenty new banks fail in exactly the same way. You need to assume that this is unlikely to happen—that the individual risks from each smaller bank will aggregate to a significantly less risky whole.

I think this is a bad assumption. Recall that the most spectacular mispricing of risk in our current crisis arose in exactly the same way—statistical magic transformed baskets of extraordinarily risky securities into assets "riskless" enough to be rated AAA. This worked for a while, but ultimately the logic of crises took hold. In moments of systemic danger, you can't count on risks aggregating away. All the correlations run to one, and it's no safer to have a crowd of small banks than a few large ones.

Tuesday, June 23, 2009

Free giveaway!

After beating the drum against Waxman-Markey's free giveaway of a sizable fraction of its carbon allowances, I think it's time to clarify some economic basics. Most commentators who oppose free allowances assume that lawmakers are acting in bad faith, rewarding influential industries with thinly disguised handouts. No doubt this is an important part of the story. But I also think that many policymakers genuinely don't understand the issue: they really think that "transitional assistance" to firms will protect American jobs and production.

This is, of course, wrong. It's made dangerous, however, by the fact that it seems so plausible. Won't American companies retain more jobs if they aren't burdened by having to pay for permits?

The flaw here is simple. In the sense that's relevant to profit-maximizing companies, everyone pays for emissions permits. Firms that are forced to buy permits at auction or on the open market pay in the obvious, explicit way. But firms that receive free allocations also pay, implicitly. Each permit they use is one that they can't sell, and the marginal cost of carbon emissions remains exactly the same. Whether we auction the permits or hand them out makes no difference: the profit-maximizing course of action will not change, and firms will hire or fire the same number of workers.

Aren't I overstating my case? Not really. Under even the most imaginative scenario, the benefits to actual laborers (rather than shareholders) from this policy are minimal. Perhaps the cozy financial situation of firms that receive free allocations will make it more difficult for them to lay off workers, from a purely public-relations standpoint. But when soft influences like public image conflict with hard profit maximization, this economist's bet is on the latter; and regardless, the vast majority of any job cuts will come through attrition rather than immediate layoffs.

Most of the time, policymaking is hard—there's a legitimate tradeoff at the heart of almost every policy dispute. This is not one of those times. Handing out allowances is a simple giveaway to firms and their shareholders, and it needlessly makes Waxman-Markey less progressive.

(You can find a extensive post on the topic here.)

Even worse than it looks

Conor Clarke provides some nice graphs of the distributional impact of Waxman-Markey, according to the most recent CBO analysis.

One of the unfortunate limitations of the analysis is that it groups taxpayers into quintiles. This gives us a rough sense of where the benefits and burdens fall, but it misses some very important variation within each quintile. Consider the estimated net cost per household:

Lowest quintile: -$40
Second quintile: $40
Middle quintile: $235
Fourth quintile: $340
Highest quintile: $245

As I noted earlier, the bizarre cost curve—which rises rapidly, and then falls for the richest quintile—is entirely due to profits from free carbon allocations, which go mostly to wealthy shareholders. This is where talking about the "highest quintile" becomes misleading. I suspect that the majority of households in the top quintile experience a higher net cost than the average fourth quintile household. When we average the net cost over the entire group, however, windfall gains from assets that are most often held by the extreme rich—the top 1%—are attributed to the "highest quintile" in general. This hides the likely reality that the "winners" from a policy of allowance givaways are an even narrower, wealthier subset of the population.

Monday, June 22, 2009

Carbon policy will make a difference

Opponents of unilateral US action on global warming usually claim that without the cooperation of China, India, and other major nations, we can accomplish "almost nothing," and it's wasteful to even try. This is bad analysis.

First, policies that accomplish "almost nothing" on a worldwide basis can still be very beneficial. The idea that a cut of 5% in world CO2 emissions is so insignificant that we might as well pretend it's zero is a classic logical error. You can't use your subjective judgment about what constitutes "insignificance" to arbitrarily set benefits at zero before even comparing them to costs! As we emit more carbon, the consequences for the climate become more difficult to predict, and the last 5% of emissions are the most dangerous. On a worldwide basis, they will do a great deal of harm.

You might claim that carbon emissions will simply shift overseas, but this goes much too far. Not all carbon-intensive production will move, and an enormous fraction of our emissions come from electric power generation, transportation, and other functions that can't be readily outsourced. An aggressive emissions policy by the United States will lead to real cuts in worldwide carbon output, and if you acknowledge the tremendous worldwide costs of climate change, you have to admit that there's potential for global benefit even from unreciprocated American effort. If you dispute this on technical cost/benefit grounds, fine, but you need to make this more explicit than a handwaving dismissal of "small" progress.

Admittedly, these are global benefits we're discussing. If you assume that there is zero chance of multilateral bargaining on climate change, and that the United States will be completely alone in whatever it does, the American economy itself will be a net loser. But even putting aside the absurdity of the assumptions, does this really mean that we should do nothing?

Let's try a thought experiment. Say that households in the United States emit a large amount of deadly poison, which kills tens of thousands annually around the world. Other countries emit poison too, but the United States is responsible for a vastly disproportionate amount. Even if it doesn't act as part of an international coalition, and it will suffer a net economic loss from changing policy alone, isn't the US morally obligated to address this? As long it's the single most culpable nation in the deaths of countless people, I think it is. When poison coming from the US is killing thousands elsewhere, it's patently unethical to interpret policy through the narrow prism of American benefit. I cannot imagine any consistent moral philosophy suggesting otherwise.

Of course, "poison" here is a philosophical stand-in for climate change. I plead guilty to overdramatizing my example, but I don't think that it makes any difference to the core issue: whether it's right for the United States to disproportionately harm the rest of the world as long as it's in the nation's interests to do so. In fact, thinking about a case as simple as "poison" allows us to clarify our moral intuition, because it sets aside complications—like the temporal displacement of cause and effect, and the fact that costs lie on a probability distribution rather than at a single point—that have no bearing on the basic moral question yet tend to confuse us.

And remember that the US can still benefit! There is no basis for assuming that concerted global action on climate change is impossible. Yes, China and India will probably be reluctant to take costly action when the US still emits far more carbon per capita. (Frankly, this is reasonable: why should your underdeveloped nation make cuts when a far richer country emits so much more?) But this won't last forever. If we make cuts now, our per capita emissions will be much closer to theirs in just a few decades. At that point, the need for dramatic changes will be even clearer, and since there will not be such obvious "winners" and "losers,", a fair and Pareto-improving policy will be accessible. If we do nothing, however, we risk a perpetual cycle of blame and inaction, as large developing countries refuse to make sacrifices when the US is still polluting so much more.

It's easy to dismiss carbon policy because it "won't make a difference." But I have never seen anyone making this argument appropriately deal with the economic, moral, and political issues I've discussed here.

Sunday, June 21, 2009

Mental accounting and the F-22

Whenever I criticize overpaying for fighter planes, I'm told that I should be more considerate of the fighter pilots' lives. As the saying goes, "you can't put a price on life," and it's wrong to put fighter pilots in greater danger because we want to save on equipment.

But of course, we do put a price on life in almost every decision we make. Drive an additional block to save a little on gas this morning? You just implicitly put an upper bound on the price of your life! After all, you decided that your savings on gas outweighed the slight additional risk of dying from a crash. In fact, you also put a limit on the value of random strangers' lives, because your savings apparently trumped the risk that you would kill someone else along the way.

How much would the marginal improvement in survivability provided by the fancier F-22 cost per life saved? Given the high cost of fighters, the relatively low death rate, the fact that the F-22 isn't a decisive improvement over the alternatives, and the market returns from making a large capital investment somewhere else, I suspect it's at least half a billion dollars. Here's the key question: could we invest that $500 million somewhere else and save many more lives? Of course. There are all kinds of medical interventions, safety measures, and public health programs that could do far better. In fact, I suspect that you could save many more lives in other branches of the military—say, the Marines—for that amount. Why, then, should a sliver of the Air Force receive such lavish protection?

I think that this is a classic case of mental accounting, an irrational tendency first documented by Richard Thaler. Most people will drive a few blocks to save $5 on a $10 lunch, but the majority won't do the same to save $5 on a $1000 television. We naturally judge costs relative to other expenditures we make at the same time or place, even when it makes far more sense to look at them in absolute terms. Hence our attitude about fighter jets. While it would seem exorbitant to spend $10 million apiece on super-high-tech uniforms for every Marine (even if it saved a few lives), it seems perfectly reasonable to spend far more to achieve the same benefits for pilots, simply because we're already spending so much on planes.

The only reasonable retort, I think, is that because we invest in such incredibly expensive pieces of equipment, we should have the best pilots possible, and higher risk makes that attracting good candidates more difficult. But there are cheaper and more effective ways to attract them: given the stakes involved, I wouldn't mind bumping up elite pilots' salaries to $400,000 a year. It certainly makes a lot more sense than spending billions in quixotic pursuit of "safety," when we're far less generous in every other walk of life.

Free handouts

Conor Clarke, Ryan Avent, and others have been blogging about the new CBO analysis of Waxman-Markey, which estimates the economy-wide cost of the bill at $175 per household. Buried in the last page of the document is a very interesting table outlining the distributional effects of the bill—for instance, the average net costs in each income quintile:

Lowest quintile: -$40
Second quintile: $40
Middle quintile: $235
Fourth quintile: $340
Highest quintile: $245

As is evident, the bill starts out progressive, but ends up costing more for the fourth quintile than the highest. This is entirely due to a specific category: "Allocation to Businesses and Net Income to Domestic Offset Producers." Handing free carbon allowances to businesses is equivalent to delivering giant wads of cash to their shareholders, and since the highest quintile holds a disproportionate amount of equity wealth in the US, it receives most of the benefits:

Lowest quintile: $65
Second quintile: $90
Middle quintile: $140
Fourth quintile: $230
Highest quintile: $885

I wish Waxman-Markey opponents who are rightly appalled at these handouts would attack the handouts directly, rather than obscure the issue and attack the idea of cap-and-trade in general...

Too much emphasis on averages

The Quick and the Ed links to a study by the American Institutes for Research that translates state-by-state math scores in the US on the NAEP into scores on the internationally benchmarked TIMSS test. The moral? Our best-performing states, like Massachusetts and Minnesota, have test scores right up there with the world's best, and indeed better than all nations outside of East Asia. Take a look:

International Grades for Countries in 2007 Mathematics, Grade 8
  1. Taiwan: 598
  2. South Korea: 597
  3. Singapore: 593
  4. Hong Kong: 572
  5. Japan: 570
  6. Massachusetts: 547
  7. Minnesota: 532
  8. North Dakota: 530
  9. Vermont: 529
  10. 13 other American states...
  11. Hungary: 517
  12. England: 513
This analysis isn't perfect, of course. We don't have actual state-by-state TIMSS scores, and we can't accurately predict what they might be. But the study makes the comparison in the most sensible way possible given our limited data, using equivalent national samples on the NAEP and TIMSS. A state whose NAEP score is 0.2 standard deviations above the national student mean is mapped to a TIMSS score 0.2 standard deviations above the national mean. I suspect that this is quite accurate as a reflection of the performance of America's states on an international basis.

At this point, however, I think that the obvious question is missed: why do we always look at averages, anyway? They provide a useful summary statistic, but they don't tell us much about what's going on in specific parts of the distribution. Take, for instance, this list of the cutoff PSAT scores for "National Merit" status, which are set at the 99th percentile in every state:

1. District of Columbia: 221
1. Massachusetts: 221
3. Maryland: 220
4. New Jersey: 219
5. Hawaii: 218
5. Connecticut: 218
5. Virginia: 218...

47. West Virginia: 202
47. South Dakota: 202
49. Arkansas: 201
49. Wyoming: 201
51. Mississippi: 199

The test is scored out of 240, and it's clear that the differences between the states at the top and bottom are enormous. A student at the 99th percentile in Mississippi would probably be at the 96th percentile (if that) in Massachusetts, implying that the number of students at the "top end" differs by a factor of four between states. These gaps are larger than anything you'd suspect by looking at averages, and although there are similarities between the two state rankings, there are also a lot of differences. (Most notably the District of Columbia, an extreme underperformer in average performance that rockets to first place when you look at only the top slice of students.)

You can plausibly argue that this is more the result of economics and demographics than real differences in the quality of schools. And certainly a high 99th percentile doesn't excuse the District of Columbia's school system, which does an abysmal job for so many of its students. But I think there's a strong case that the performance of the top 1% is a lot more important than average performance for long-term economic growth. The defining technological innovations of our time come from people at the extreme right end of the ability distribution, not the "average." And since it's clear that the two metrics can diverge so markedly, we should be careful to examine how our policies affect the best students, rather than just the average ones.

Saturday, June 20, 2009

What's wrong with low real interest rates?

After my last post questioning the "savings glut" explanation of the financial crisis, I think it's time to ask a broader question: why do so many people think that low real interest rates lead to speculative buildup and ultimate collapse? There is no fundamental reason why low interest rates should cause massive mispricing of assets or risk. As my earlier post mentioned and Bryan Caplan emphasizes, most attempts to blame the Fed and its monetary policy commit the Banana Fallacy:
Remember the Fable of the Banana Subsidy? Government subsidizes bananas. People buy a ton of bananas and store them on their roofs. The bananas weigh so much that their roofs collapse. Then people say, "It's the government's fault!" - which is true in a sense, but also misleading.

I see the same ambiguity in a Forbes interview with fellow GMU prof Todd Zywicki...

If you pay attention, you'll notice that Todd's making a Banana Subsidy argument. The government cut interest rates, and then... banks started offering loans to people who wouldn't be able to pay them back - and borrowers accepted.

Maybe Todd's story is right. If he is, the Fed made a terrible mistake. Unfortunately, for Todd's story to work, we also have to admit that business and consumers are so clueless that they're habitually on the edge of disaster.
What charitable interpretations of this story are out there? I can think of a few:
  1. If low interest rates are liable to undergo a sudden runup, previously economical investments may go bad. But this raises a question: if a sudden increase in short term interest rates is so likely, why isn't it priced into long term interest rates at the beginning? For this explanation to work, the change has to be in large part unexpected, or at least uncertain. As an example, even if all investors admit that a reversal of the current global "savings glut" will take place sometime in the next 50 years, the exact time is unknown, and anticipation of this change plays only a small role in current interest rates, despite the fact that it may cause serious dislocations whenever it happens. (I'm not sure I believe this story, but it's at least vaguely plausible. I should repeat, however, that it has very little to do with what happened in our current financial crisis.)

  2. Downward movement in the real interest rate allows new classes of consumers to borrow economically, for new purposes (i.e. housing). Since these marginal consumers have little established borrowing history, it is difficult to evaluate their risk. Still, even if lenders appropriately internalize this uncertainty, they may think that it is profitable to engage a new base of consumers. The financial troubles come when they are wrong in the aggregate. (Their errors can be highly correlated.)

  3. Elaborating on #2, as a decline in real interest rates makes homebuying at high prices more economical for a new set of consumers, housing prices are driven up. This process does not happen overnight, however, and limited knowledge about the new consumer base means that no one knows when it will stop. Prices rise in anticipation of future increases. Misreading speculative runups in price as further evidence of strong fundamentals, investors continue to pour money into the market until the ultimate crash. This is a classic bubble. (The extent to which it can be a rational phenomenon, however, is severely limited.)
Possibility #1 suggests that low interest rates themselves can be risky, assuming they're based on an unstable financial arrangements. Possibility #2 explains how low interest rates can lead to a higher chance of massive risk mispricing, and possibility #3 helps to explain how they can trigger a "bubble." Note, however, that all these explanations depend on more than low interest rates. The real problem is the change in interest rates, which acts as a threat in case #1 and as a trigger for mispricing in #2 and #3.

Needless to say, it's important to know whether low interest rates themselves lead to asset bubbles and instability, or whether rapid changes in interest rates are to blame. The policy implications are very different, and I want to hear better explanations from neo-Austrians who blame the Fed for private banks' massive failure to price risk.

Thursday, June 18, 2009

Costs multiply

A post from last month at Climate Progress points us to a new MIT study in the Journal of Climate, which makes a frightening median projection of 5.2 degrees Celsius (9.4 degrees Fahrenheit) warming by 2100, along with a 5% chance of warming greater than 7.4 degrees Celsius (13.3 degrees Fahrenheit). The research program's website makes a critical point in its explanation of the results:
There is no single revision that is responsible for this change. In our more recent global model simulatations, the ocean heat-uptake is slower than previously estimated, the ocean uptake of carbon is weaker, feedbacks from the land system as temperature rises are stronger, cumulative emissions of greenhouse gases over the century are higher, and offsetting cooling from aerosol emissions is lower. No one of these effects is very strong on its own, and even adding each separately together would not fully explain the higher temperatures. Rather than interacting additively, these different affects appear to interact multiplicatively, with feedbacks among the contributing factors, leading to the surprisingly large increase in the chance of much higher temperatures.
By interacting multiplicatively, even small errors in several parts of a climate model can combine to produce massive swings in the final result. And it doesn't stop there: these uncertainties in climate modeling then interact multiplicatively with uncertainties in economic modeling, producing even larger swings in the estimated cost of global warming. As I explained in a post on the topic a little less than a year ago, the results can be overwhelming:
When considering the implications of the three concerns I have raised, one question is particularly important to keep in mind. What happens if the concerns are simultaneously valid—if the Nordhaus/Manzi damage estimates are biased down by more than one of the weaknesses I mention? To a first approximation, they multiply. This means that if the failure to analyze extreme uncertainty, the high discount rates, and the crude worldwide aggregation of damages present in Nordhaus's model each lower the optimum carbon tax by a factor of two, a better estimate would have a starting carbon tax approximately eight times as high. Since Nordhaus's "optimal policy ramp" starts with a tax of $7.40 per ton of carbon dioxide, such a recalculation would bring us to almost $60 per ton, which is close to what many carbon tax advocates suggest. More interestingly, if each of the above failures causes a miscalculation by a factor of three, the total estimate may be off by twenty-seven. This brings us to a very high carbon tax of $200.
Unfortunately, these costs are all too plausible. At the very minimum, I think that the failure to appreciate the uncertainty in our models deflates our cost estimates by a factor of four, unreasonably high discount rates lower them by a factor of two, and using aggregate world GDP rather than country-by-country modeling also lowers them by a factor of two. This already means that we're off by approximately a factor of 16—so while economist William Nordhaus's models of climate change suggest an optimal carbon tax of $7.40 per ton, a more appropriate level is at least $120.

And I'm not exaggerating when I say "at a minimum": these are incredibly conservative estimates for the gap between analyses like Nordhaus's and more complete models that throw away the false certainties and simplifications. If you have any doubt about the effect of "fat-tailed" uncertainty on the costs of climate change, you should read Martin Weitzman's brilliant analysis of the problem. Meanwhile, the effects of even small changes in the discount rate are mathematically inevitable: the difference between damages 100 years from now discounted by 2% and discounted by 3.1% is already a factor of three.

The moral? Don't let artificially low, "rigorous" estimates of the costs of climate change lull you into complacency. As I've said before, if superficially valid estimates defy all reasonable intuition, you ought to see whether there's a problem with the underlying analysis. When global warming threatens to move average temperatures more than our transition out of the Ice Age, melting the ice caps and disrupting climates throughout the world, there's probably something wrong with any calculation that the optimal policy is to nudge up the price of coal power by less than a cent per kilowatt-hour.

Wednesday, June 17, 2009

Beware selection bias

Robin Hanson isn't going to take it anymore:
Listening to my son’s high school graduation ceremony last night, I was struck by how completely implausible were many speaker claims, such as:
  1. Never let anyone tell you there is something you can’t do.
  2. You’ll have setbacks, but never let them discourage you.
  3. If I can succeed, so can you.
  4. We’ll always treasure our memories of high school.
  5. We students are so thankful to have such a friendly principal.
I was embarrassed to be associated with such transparent falsehoods, but apparently I’m in a minority. What obvious lies have you heard at commencement, and why do you think such lies were told?
Indeed. Whenever you hear successful people offering you banalities like "if you just follow your dreams and believe in yourself, you'll go far," you should be very skeptical. Perhaps in their case it worked out, but to extrapolate from their experience is to fall victim to the simplest form of selection bias. Luck is an important component of success, and it's only natural that successful people will have been disproportionately lucky, having achieved their goals against often overwhelming odds.

I like to keep this principle in mind when thinking about how I might become successful (not that I know enough to have an inkling about what "success" might be). Generally speaking, it's a bad idea to rely on others' path to success as a model for your own. You can assimilate their useful lessons, of course, but all too often their "big break" will have been the result of dumb luck, and it's tempting to sit back and wait for such bursts of fortune to come to you. Not everyone is lucky, and you should make plans based on the actual distribution of luck, rather than the skewed one you see among prominent people.

Maybe this sounds hopelessly pessimistic, but that's not how I see it. I'm not against chasing your dreams; I just think that you should chase them in the most effective way, and that means being aware that your idols aren't a very representative sample.

Savings glut: redux

Official friend of the blog Matt Zeitlin links to my post questioning the "savings glut" interpretation of the financial crisis, and writes:
...Most economic theory seems to suggest that saving is generally very good, and moreover, a lot of economists and policy types are always recommending that we tax consumption so as to encourage saving. If that’s true, then Matt Rognlie is certainly right to say, what’s wrong here, the global savings glut theory of the crisis or lots of economic theory? And how should this affect our views about saving?

I don’t think this two views are very much in conflict. For one, the global saving glut types are saying that problem isn’t too much saving, but that some countries (America) save far too little and other countries (China) save far too much. So, all those questions about tax policy towards saving and consumption don’t seem all that relevant when discussing a global savings glut.
The question, then, is why foreign saving is more dangerous than domestic saving. There is one very good explanation: artificially high savings in other countries, and the concommitant deficits in the United States, create unsustainable imbalances that threaten to suddenly unwind, taking world financial stability with them. This is a very serious concern—it's been discussed by leading economists for years—and it still poses a threat. But it has very little to do with what happened over the last few years. We did not experience a sudden run on the dollar, and foreign creditors did not withdraw their investments en masse.

If we want to establish that excessive global savings, more than a high level of domestic savings, is likely to lead to the kind of financial crisis we actually experienced, we need a different argument. In this vein, commenter Tord Steiro points us to a paper on vox by Daniel Gros, which proposes:
Another way to look at the same phenomenon is to note that the increased demand for US government debt by emerging economy central banks led to lower yields, thus forcing those savers in the OECD countries which would normally have held government assets to frantically “search for returns”. But this was a search for yield on safe (and liquid) assets. The AAA tranches on securitised US mortgages (and other debt) seemed to provide the safety plus a “yield pick up” without any risk, at least in the sense that the securities were rated AAA.
As long as US house prices kept on increasing and unemployment remained low, actual delinquencies remained low and there seemed to be no reason for market participants to question the high ratings of these securities, even though the incentive for the ratings agencies to provide favourable ratings were well known. AAA-rated residential mortgage-backed securities thus provided an important source of liquidity by their widespread use as collateral.
This seems like a reasonable interpretation—massive buying by central banks pushed down the yields on truly safe, liquid assets (Treasuries), creating demand for more "safe" assets with higher yields. But it's not completely convincing. There will always be a demand for "safe" assets that provide higher yields—after all, if you believe that they're truly safe (or can pass them off as such to investors), they're a free lunch! No complicated theory of international finance is necessary to explain this. The real failure is in the markets and rating mechanisms that created the illusion of safety for profoundly risky investments.

Including some excerpts from a speech on the topic by Ben Bernanke, Matt adds:
But the way Bernanke tells the story is that this global savings glut was real and actually led to a decrease in long term interest rates...

And while Bernanke, who was giving this talk in September 2007, doesn’t connect the glut to the crisis, plenty of other people have done so and are perfectly reasonable. This is not to say that savings and low interest rates always means asset bubbles and subsequent crashes, but just it’s worth watching out for when high savings and one part of the world leads to massive overinvestment in another.
Sure, and insofar as low real interest rates encouraged the bubble in housing, we can blame the the global "savings glut." But this still begs my original question: if the main problem is that interest rates were too low, how should this change our views on optimal taxation, where the supposed advantage of consumption taxes is that they encourage savings and thus lower long-term real interest rates?

Our problem, I think, is that we lack a robust theory of asset price bubbles. We're reduced to making shaky inferences about complicated patterns of causation, and often we focus on factors whose role isn't at all clear. How important, exactly, are low interest rates to the formation of bubbles? It's a subtle issue, and I really don't know.

This all reminds me of a brilliant post by Tyler Cowen on a related topic, the role of the Fed's low interest rates in causing the housing boom. He wrote:
Keep in mind that no entrepreneur looks at price signals exclusively, rather they interpret prices in the context of the real economy and other bits of knowledge. Was it so hard for investors to say to themselves?: "I see that one price (short-term rates) has changed in favor of greater housing investment. But other parts of my brain tell me that real estate prices won't go up forever, levered positions are dangerous, and that I should be cautious."

Let's say that the government subsidized the price of bananas, you bought so many bananas, put them on your roof, and then the roof collapsed. Is that government failure or market failure? The price was distorted, but I still say this is mostly market failure. No one made you put so many bananas on your roof.
Exactly. We probe too deeply for "root causes" (the price of bananas) when the true, policy-relevant issue is already in front of us: how to stop idiots from putting too many bananas on the roof.

"Monetizing" the debt: a clarification in terms

The buzz in some circles is that we'll have to "monetize" the federal debt, printing money to satisfy our obligations. I want to put aside for a moment the question of whether this will actually happen and clarify some vocabulary, because monetization is not the right term for the majority of the reduction in our debt burden that would result from an inflationary monetary policy.

When debt is monetized, it's converted into currency. The Federal Reserve creates $100 billion in new monetary base, uses it to buy Treasuries, and thus reduces the amount of government debt held by the public. There's another way, however, that printing money can ease the debt burden, and it's almost certainly more significant.

Aside from a small percentage (~10%) issued in inflation-protected securities called TIPS, most of our national debt is held in nominal bonds, which pay a fixed dollar amount that doesn't depend on inflation. If the price level unexpectedly increases by 100%, therefore, our real debt will be cut in half—adjusting for inflation, our bonds will be worth half what they once were, and be half as difficult to pay back.

Why will this have a larger effect than printing money? Subtracting excess reserves—which are kept out of the broader money supply by the Fed's current policy of paying interest on reserves, and will be unwound as the economy begins to recover—the monetary base in the US is a little less than $1 trillion. Accordingly, if we doubled the monetary base, we would be able to monetize a $1 trillion of debt. This is no small amount, but it only makes a dent in the federal debt held by the public, currently a little over $7 trillion. If left out in the economy, however, the new monetary base would cause prices to double, cutting the debt held in nominal bonds in half. The arithmetic here is simple: $3.5 billion is a lot more than $1 trillion, and the main effect is from inflating away the value of our debt, not from monetization specifically.

Many times, commentators who understand these issues perfectly well nevertheless use "monetization" as a catch-all for the total effect of inflationary monetary policy on our debt. I don't think this is wise. Confusion about monetary economics is already overwhelming, and when it's not difficult to use more precise vocabulary, we should make the extra effort.

Tuesday, June 16, 2009

Preexisting conditions

Obama wants to stop private insurers from denying coverage on the basis of prexisting conditions. This is certainly a noble goal, but I can't understand how it works in the context of a private system.

By all accounts, it's unlikely to be possible unless there's a mandate forcing everyone to become insured. Otherwise, healthy people will avoid buying coverage, knowing that they can always obtain it later if their situation changes. This tilts the risk pool further toward high-cost customers, which forces an increase in prices, which induces more people to drop out of the insurance system, which causes further price increases, which... well, you get the picture. It's virtually inevitable that without a mandate, such a policy will widen the ranks of the uninsured.

A mandate, however, doesn't strike me as a panacea. After all, insurance isn't binary: you're aren't simply "insured" or "uninsured." Much of the point of having private insurers is that the market is free to offer a variety of different packages, which can differ in the services they provide and risks they cover. But if insurers aren't allowed to deny coverage, this system breaks down. If my insurance company offers better coverage for syndrome X than the competition, everyone with syndrome X will sign up for my plan, making it impossible to continue without massive premium hikes.

The net effect will be a race to the "bottom," as defined by the government mandate. Any services beyond the core ensured by the mandate will attract a disproportionately expensive group of customers, and are likely to prove unsustainable. This doesn't mean that private insurance will disappear, but it does render many of their supposed benefits meaningless. If the mandate will effectively determine the scope of every plan's coverage, why not kill the middleman and provide public insurance?

It's not as if this is some unusual argument that I've invented—it's a classic story of adverse selection. I simply don't understand what we can do about it. Is there something I'm missing?

Monday, June 15, 2009

Rankings really can be destructive

Ryan Avent takes aim at Newsweek's obscenely bad ranking of American high schools, which evaluates schools solely on the rate at which students take AP tests:
Last week I attended the graduation ceremonies for Bell Multicultural High School, here in the District. My wife taught English there until recently, and had had many of the graduating students in her classes.

She was an AP teacher, as was every upper class English teacher, because Bell is involved in an experimental AP for all program. The idea, according to principal Maria Tukeva, is that the big failure of urban school systems is that they fail to adequately challenge the students. Set the bar appropriately high, and students will learn how capable they actually are.

Or some such bunk. One wonders why, if this is the case, teachers aren’t made to teach graduate level courses — surely the children will rise to the occasion. Hell, just pass out the latest academic journals, and watch the magic happen. If this is what one believes, then simple “AP for all” seems to reek of the soft bigotry of low expectations. Come on, woman, have some faith!

In practice, the policy is a disaster. Teachers find themselves standing in front of classrooms that contain special needs students, students who are unable to read, and recent immigrants with little to no command over English, alongside average students who wouldn’t normally take an Advancement Placement course, and a handful of kids who would. They’re given an AP curriculum and told not to fail. And they go to work...

At the end of the year, many of the students graduate and get to attend the ceremonies I recently observed... And of course, a lot of others aren’t there, because they aren’t graduating. It’s a shame, because they don’t get to hear principal Tukeva hail the students as graduates of one of the top high schools in America, according to a prominent set of rankings. They don’t get to hear the assembled guest speakers hail the principal as a true leader, an innovative thinker on the subject of education reform.
The damage done by Jay Matthews—an education writer who happens to have control of a prominent newsmagazine's ranking—is overwhelming. Certainly the idea of implementing higher standards has merit in some circumstances. But when you equate "higher standards" with the arbitrary label of "AP," the consequences should be obvious: a burst of superficial and often destructive reform, as ambitious principals decide that downloading curricula from the College Board is the solution to every educational ill.

In reality, the point of AP classes isn't only to impose some special curriculum. Like mere "honors" classes, they help schools target students' needs more effectively. Serious students with college-level skills do better when they're placed with similarly capable peers, in an environment where teachers can cater to their abilities. Immigrants who need remedial work in English, on the other hand, should have instructors that can deal with their difficulties, not ones who opine about Shakespeare to kids who can't even read the newspaper. Lumping everyone in the "AP" bucket makes this impossible, and misses one of the main advantages that advanced curricula provide in the first place.

The coolest map you've ever seen

In case anyone hasn't seen it, I thought I'd do a public service and link to an incredibly fascinating map—the Census's display of the most common ancestry in each county. Since it's too large to embed on this page in a way that does it justice, I simply encourage you to click and see.

It's awesome. Did you know that Dominican is the most common ancestry in Manhattan? That Appalachia's largest ancestry is simply "American," self-reported despite its absence as an option on the Census forms? (Nate Silver used this data to augment his election models.) That the vast national interior has German as the most common ancestry—not surprising—but Utah is mainly "English"?

If you're curious, here's the full Census brief where the map originated.

Should we smooth gas prices?

The generally very perceptive Felix Salmon makes a point that I don't think is quite right:
How to structure a gas tax? You could make it a flat X cents per gallon; alternatively (and this is essentially what a cap-and-trade system does, too) you could make it Y%, with the tax increasing with the price of gasoline.

Today, Jim Surowiecki comes up with a third option, where the tax decreases when the price of gasoline goes up...
Surowiecki makes a strong case that consumer behavior, when it comes to reducing gasoline consumption, only really changes when there’s a spike in gas prices. As a result, his proposal would seem designed to have the least possible effect on gasoline consumption, and on our dependence on oil. Sure, it’s a sensible way of raising government revenues and reducing the fiscal deficit.

Either you want to effect consumer behavior and reduce gasoline consumption — in which case you actually welcome price spikes. Or else you want to smooth out price spikes, in which case you slowly boil the frog (to use one of the stupidest metaphors ever) and keep consumption high. But you can’t have it both ways. Which is it to be, Jim?
My reading of Surowiecki (and the evidence) isn't that only sharp price changes affect gasoline consumption. A $4-a-gallon price tag will influence consumers, even if it's the product of a slow rise rather than a sharp spike. Instead, the problem Surowiecki rightly identifies is that rapid price swings cause a collapse in broader consumer confidence, damaging the economy. Their effect on consumption is so sudden that the producers don't have time to adjust, leaving us with idle capacity and structural unemployment. Needless to say, this isn't good.

Granted, price volatility reduces consumption, but the question isn't whether volatility has an effect—it's whether it is the most efficient way to accomplish our goal. There's some average tax level $X that will accomplish the same long-term reduction as a lower tax $Y implemented in a way that encourages price volatility. Assuming that they have the same effect, what's better for the rest of the economy—the higher average tax that smooths out prices, or the lower tax that leaves price spikes intact? I can't be positive without more elaborate modeling, but I'm almost sure that the higher tax is better. For the most part, avoidable uncertainty is not a efficient way to implement incentives. Indeed, Surowiecki's entire point is that gasoline dependence is odious because price changes lead to economic instability.

A related question, where I think that the issues are clearer, is the design of carbon policy. When I argue that bankable permits are an essential part of cap-and-trade, some environmentalists I know are mystified. After all, price instability will make carbon-intensive activities even less attractive. If it leads to lower emissions, why do I want to change it? Because it's an inefficient way to achieve our goals. You can implement an equally effective policy by lowering the cap slightly and allowing bankable permits, with far lesser economic and political costs. (Actually, it's not so clear that uncertainty reduces emisisons at all under cap-and-trade, because there's no incentive to push emissions below the cap. The only reason I can imagine is that consumers, fearing possible price spikes, will be more aggressive in moving away from carbon dependency, allowing future legislators to revise the cap downward. But this is pretty indirect, and the point about efficiency still holds—it's easy to forget that policies might "work" but do so in a needlessly costly way.)

Now, where does this all leave us with regard to the original proposal—Surowiecki's sliding gas tax, which decreases as the nationwide price goes up? Like I've said, all else being equal, I think that price volatility is a bad way to discourage consumption. But all else isn't equal: our power to limit volatility in the price of gasoline is limited by the reality that oil is a market, where the price much adjust to balance supply and demand. If $4 gas is necessary to reduce demand to the quantity that producers can supply, it doesn't matter what the tax is—in equilibrium, the after-tax price will be $4, period.

This is complicated a little by the fact that oil is a world market, while tax policy is implemented at the national level. With this in mind, if we implement a gas tax that moves opposite to the world oil price, we will achieve a more stable price in the United States. But this will be at the expense of even more extreme moves in other markets; since America will no longer be as responsive to price signals, prices elsewhere will have to adjust more sharply than before. If this induces other major consumers to implement similar policies, it won't be long before almost all countries attempt to stabilize after-tax prices using the same approach. And at this point, the entire policy will be worthless: after-tax prices have to move to equate supply and demand, and even the best efforts of governments can't change this on a worldwide basis. Indeed, during price spikes, a broad effort to lower gas taxes is essentially a massive transfer to oil producers.

So the ultimate answer of this seemingly endless post is no: Surowiecki's sliding tax is not a good idea. But there's a lot of nuance here about the effects of instability, and even if we rightly reject his proposal, we should make sure it's not for the wrong reason.

Sunday, June 14, 2009

Flashback: the inflation worries of summer 2008

I just came across an old post of mine from 2008, where I angrily defended the core CPI against critics who claimed that it hid inflation that would soon roil the economy.

An obvious thought comes to mind: if these people actually influenced our monetary policy, and the Fed had raised rates in September 2008 in response to "worrying" CPI increases (which were really just in food and oil), where would we be now? My guess is that in such a world, 10% unemployment would seem like a rosy vision of recovery.

Nearly everyone failed to anticipate the depth of our financial crisis, or address systemic risks before they blasted their way into the popular consciousness. But compared to the idiocy of mid-2008 inflation hawks, these failures seem minor. If you thought that the summer of 2008 was a good time for the Fed to tighten the money supply, you should be permanently disqualified from offering economic commentary.

Projecting the future

If you're interested in how the world will look in 50 years, it's hard not to find the UN World Population Prospects fascinating. Among many other findings, these summary tables list the current most populous countries alongside the projected most populous in 2050:

Current:
  1. China: 1,346,197,000
  2. India: 1,198,372,00
  3. United States of America: 315,419,000
  4. Indonesia: 230,452,000
  5. Brazil: 194,481,000
  6. Pakistan: 181,507,000
  7. Bangladesh: 162,531,000
  8. Nigeria: 155,553,000
  9. Russian Federation: 141,574,000
  10. Japan: 127,593,000
2050 (Projected):
  1. India: 1,614,000,000
  2. China: 1,417,000,000
  3. United States of America: 404,000,000
  4. Pakistan: 335,000,000
  5. Nigeria: 289,000,000
  6. Indonesia: 288,000,000
  7. Bangladesh: 222,000,000
  8. Brazil: 219,000,000
  9. Ethiopia: 174,000,000
  10. Dem. Republic of the Congo: 148,000,000
The big stories here are obvious, and guaranteed no matter what methodology is used for projections: India moves ahead of China, high-fertility African nations like Nigeria, Ethiopia, and Congo experience massive growth, and so on. But for the subtler questions—how quickly China's growth will turn to decline, or whether the United States can grow even in the absence of immigration—the answer isn't really clear, and the UN projections aren't as edifying as one would hope. As Bryan Caplan points out, the "medium variant" UN projections actually rest on a single, arbitrary assumption, wherein the model assumes that fertility rates across the world will converge to 1.85.

I understand the UN demographers' predicament. Any more sophisticated model will inevitably make assumptions that (at least to some observers) appear even more arbitrary and wrongheaded, and for political purposes it's safer to go with the crude but simple approach. It's also difficult to imagine what alternative methodology could consistently be applied to the entire world and still produce reasonable results. Caplan's suggestion—that we extrapolate current trends, assuming that existing declines in fertility will continue in the near future—raises a host of potential problems. How do we extrapolate? Presumably we'll need to choose an arbitrary functional form, and unless human biology makes it possible to have negatively many children, we'll have to place some arbitrary floor on each country's fertility. To make matters more complicated, fertility in a lot of developed nations (the US, the UK) is now rising. How long will this increase last, and where will it stop? How do we "extrapolate" it? Certainly fertility in the US is not going to rise above 2.5, but how can we embed this intuition in a model? How can we predict when currently low-fertility countries will take a sudden upward turn, like the US, which has gone from 1.79 to 2.09 children per woman over the last 30 years? This is hard stuff!

On a case-by-case basis, however, I think that it's possible to make informed guesses that improve on the UN's projections. The best example is China. Currently fertility there is stagnant at about 1.77 children per woman. Perhaps it will stay at that level, or nudge upward slightly, making the UN's 1.85 estimate look reasonable. But the closest cultural analogues to what a fully developed China might look like, Hong Kong and Macao, have fertility of around one child per woman, the very lowest in the world. Taiwan is at 1.1. South Korea and Japan, the other large, developed nations in East Asia, are also near the bottom of world fertility, at 1.22 and 1.27 children per woman respectively. This seems to be a regional trend.

Meanwhile, urban China has lower fertility still, and if this holds as the nation rapidly urbanizes, it's hard to imagine that we'll see any increase in fertility. Indeed, the overwhelming weight of circumstantial evidence suggests a decline in the making. True, the fertility level has leveled off over the last decade, but this may be a response to earlier misestimation: as women of the new generation delayed having children, the total fertility statistics measured this as a "decline," and now they're playing catch-up. (This seems to be a common story worldwide.)

You could argue that the "One Child Policy" — something of a misnomer, since it doesn't prevent many women from having more than one child — has artificially deflated fertility, and women will have more children as the government eases. I'm not so sure. The fact that fertility declines were partly coercive doesn't negate their potential to set societal norms. If one or (at most) two child families become the standard in Chinese culture, it doesn't matter how this standard was originally put in place. And again, the regional comparisons are so overwhelming: developed East Asian nations have the lowest fertility in the world, bar none.

With all this in mind, I suspect that the birth rate in China will continue to decline, and by 2050 we will see a China with a population of 1.2 billion (and falling). But there's also a very good chance that I'll be wrong. Hardly anything is more difficult to predict than demographics, and few in 1970 had the slightest idea about the world 40 years hence.

Is a savings glut really our problem?

Many commentators are connecting our present financial crisis with broader global imbalances, describing the massive dollar-denominated bond holdings of Asian central banks as part of a "global savings glut" that has led to asset bubbles in the US. While there may be some truth in this analysis, I'm not sure that we have carefully considered its implications.

First of all, what's a "savings glut"? In classical growth theory, high savings are good: they lead to greater capital accumulation and higher long-term output. In more newfangled endogenous growth theory, high savings are even better, leading not only to more output but also to more output growth, which has virtually unbounded economic implications when sustained for a long enough period. Indeed, many economists support replacing income taxes with consumption taxes, either explicitly through a VAT or retail sales tax or implicitly through expansion of tax-deferred savings vehicles like IRAs, because income taxes are biased against saving.

Yet the "savings glut" concept presumes that we are actually suffering from too much saving—and not temporarily, during a recession (which almost by definition involves too much saving), but structurally. This isn't domestic saving, of course, but it's not clear what difference the origin of savings makes for the "savings glut" argument, unless we find the exact mechanism through which foreign investment specifically encourages asset price bubbles.

Does this mean that a tax system rigged against saving isn't so bad? Admittedly, the one kind of saving that our tax system doesn't discourage is investment in homes, and you could make an argument that funneling excess savings into inflated property prices is the essence of a destructive "savings glut," while "excess" savings don't do the same damage when placed in other kinds of investments. But this is both question-begging and unlikely to provide useful lessons about policy. We're not going to have the exact same kind of bubble repeat itself in the next couple decades, and if only real estate is capable of transforming savings into an economically destructive force, this entire line of reasoning has little relevance for the coming years. Essentially, the savings glut idea is only meaningful to us in the short and medium term if it's more general, but then we must reconcile its inconsistencies with our other views about capital taxation and the desirability of saving.

So here's the first question: if savings are actually too high, how does this knowledge change our understanding of optimal taxation? Are taxes that encourage consumption over savings actually beneficial? (Even with them, apparently, we're suffering from an excess of cheap money.)

Second, how do we know that in a world with lower savings, destructive asset price bubbles won't develop? Say that the current level of savings is Z, and the amount pushed into frothy, overpriced assets is X. If our savings level was Y (less than Z) instead, would X decrease by a corresponding amount, or never get off the ground in the first place? Or would the bubble assets appear so attractive that our other, productive capital assets would suffer, leaving us worse off than before? Maybe they would decline proportionally, leaving us with fewer wasted investments but also less useful capital? Answering this question requires a deep understanding of the dynamics behind bubbles, an understanding that we don't really possess.

Now, there's a good chance that I'm just confused about these issues. But there's also a good chance, I think, that the "savings glut" concept has been pushed despite its incoherencies because it's connected to other serious economic threats, like the sustainability of trade deficits and our global balance-of-payments regime. Perhaps commentators who have just spent years worrying about a sudden flight from dollar reserves are prone to analyzing every event in the context of international finance, even when it doesn't really fit.

Is inflation inevitable?

There's recently been plenty of buzz about the possibility of inflation over the next decade. Worries abound, among them the possibility that the US government will be forced to use inflation to depreciate the real value of its enormous debt. I don't think this is a serious threat (at least for now), but that's a topic for another day. Instead, in this post I want to discuss the errors in a crude but common analysis of our inflation prospects: the assumption that because the Fed has been creating so much new money, inflation is inevitable.

Before I start, I should offer the caveat that monetary policy is incredibly difficult to understand, and every time I attempt to think about these issues I find myself pressing the limits of my comprehension. I am hardly an expert. Nevertheless, the monetary anxiety expressed by large parts of the commentariat is too loud and misleading for me to ignore, and I hope that I can contribute—if only in a small, small way—to clearer discussion on these issues.

First let's start with the "evidence" for imminent inflation, which can be summarized in a single graph:


This certainly catches the eye. A doubling of the monetary base in a single year! Unprecedented monetary expansion by the Federal Reserve! How can "printing money" in these quantities not be inflationary?

First there is the standard explanation: even if the Fed aggressively creates new money in a recession, people hoard it. The sudden, widespread desire to save money is the mechanism behind recessions in the first place. If the interest rate is close to zero and the economy is still slumping, in the short term it hardly matters how much money you print: it will just pile up. As Paul Krugman explains with some nifty graphs, massive increases in the monetary base are even consistent with deflation. Take a look at his incredibly revealing chart of Japan's lost decade:


For developed countries like America and Japan, inflation induced by aggressive monetary policy just isn't a worry during recessions.

This isn't the whole story, of course. What happens to all that money after the economy starts to grow again? To explore this question, it's useful to see where all the new money has gone:


Of the ~$930 billion spike in monetary base, ~$850 billion has not gone into hard currency at all; instead, it's piled up in US banks' reserves with the Fed. This is mainly due to the Fed's decision to start paying interest on reserves.

This brings us to a key point. In a sense, bank reserves have become a form of extremely safe, low-interest debt, issued by the Fed. They are an unusual kind of debt, with no maturity date or explicit guarantee of payback, but there is an implicit guarantee: the Fed has a mandate to preserve price stability, and it will not stand by and let hundreds of billions in new money plunge into the economy. Instead, the Fed will simply sell the assets it has acquired and take money out of the system.

There are three possible criticisms:
  1. The Fed will miscalculate. This is, in my mind, easily the best objection, and it reflects a classic qualm about the Fed's ability to use monetary policy to tweak the economy—since the Fed isn't perfect and relies on slightly out-of-date data, its presence may actually prove destabilizing. In this model, the Fed will find itself surprised by a sudden burst of dormant inflation, be forced to tighten policy in response, and thrust America into a "double-dip" recession. I'm not sure, however, that this concern is applicable today. Although there are multiple potential sources of inflation, the Fed's main target is simple: make sure that the massive amounts of cash it's injected into the banking system that currently sit as reserves doesn't all find their way into the broader money supply. The new policy of paying interest on reserves gives it a very useful tool to ensure that this doesn't happen in an unpredictable way.

  2. The Fed will not be able to unwind its positions quickly enough. This is a related objection, and it's part of Art Laffer's argument in the WSJ, where he claims:
    Alas, I doubt very much that the Fed will do what is necessary to guard against future inflation and higher interest rates. If the Fed were to reduce the monetary base by $1 trillion, it would need to sell a net $1 trillion in bonds. This would put the Fed in direct competition with Treasury's planned issuance of about $2 trillion worth of bonds over the coming 12 months. Failed auctions would become the norm and bond prices would tumble, reflecting a massive oversupply of government bonds.
    Obviously, Laffer isn't a very credible source on economic policy, but I'm really mystified by this argument. First, it's a little unclear about what is happening: it says that to reduce the monetary base by $1 trillion, the Federal Reserve would have to "sell a net $1 trillion" in bonds. This is true, but the Fed wouldn't be selling government bonds like the Treasury does; it would presumably be unwinding its less conventional positions in other kinds of debt. In a world where credit starts to loosen and the Fed has to worry about inflation in the first place, this riskier debt will be in demand, and it will be a perfect time to sell. (And as this speech by Bernanke on the Federal Reserve's balance sheet makes clear, many of the Federal Reserve's assets are in short-term liquidity provision, which will naturally unwind if the banking system becomes healthier.) To me, Laffer's critique just doesn't work on a logical basis: he's envisioning a world where it will be difficult to the Federal Reserve to successfully sell its assets along with a world where inflation threatens price stability, but doesn't seem to realize that these worlds can't exist at the same time.

  3. The Fed is insolvent. According to this critique, the Fed won't just face liquidity problems in trying to unload its assets. Rather, its asset positions are so bad that they won't be worth as much as the money the Fed created to buy them, and it will be forced to leave some of its new money out in the economy. I don't think many commentators offer this position explicitly, but it's the most charitable interpretation of the more incoherent rants against the Fed. It is, of course, pure silliness. The Fed has been a liquidity provider during a liquidity crisis, and a buyer of risky assets during a flight to safety—all while funding its operations with very cheap debt (the reserve balances held by American banks). If you can manage it, this is a very profitable business. As the Bernanke speech linked above makes clear, a large fraction of the Fed's unconventional assets are short-term loans to banks to provide liquidity. Barring another financial collapse, the Fed is not going to lose on these assets—and if we experience a return to financial disaster, inflation will not be our concern. The other, long-term assets held by the Fed as part of its "quantitative easing" strategy do not appear particularly risky either, but it really doesn't matter. Even in a bizarre world where the Fed experiences a 10% loss on all its investments, it will still have plenty of assets available to soak cash out of the economy and stop incipient inflation.
All in all, inflation should not be on our list of top economic worries. You don't even have to take my word for it; you can look at the bond market, where the gap between yields on 10-year nominal and inflation-protected Treasuries barely exceeds 2 percent. I don't think the market is efficient without error, but in this case we are wise to trust its judgment. The modern Fed is a serious organization committed to price stability. It has both the desire and the capacity to prevent inflation from threatening the American economy, and even a large increase in the monetary base doesn't make inflation inevitable.

Thursday, June 11, 2009

Why bankable permits are so important

Kevin Drum graciously responds to my last post on cap-and-trade in the comments section, and makes an important point:
Oddly, the best argument against C&T, I think, is that it doesn't combine well with other regulatory initiatives. Adopt a local reg to increase building efficiency, for example, and energy use goes down. But that just drives the permit price down, which will drive other carbon emitting activities up. C&T is good at forcing emissions to meet a target, but it also pretty much guarantees that you'll never beat the target. For some reason, though, critics don't often use this argument.
Indeed, the is the key inefficiency of cap-and-trade. Although cap-and-trade keeps carbon emissions below a certain level, it doesn't offer any additional incentive to lower total emissions beyond that level. In fact, the implicit price it assigns to carbon goes from zero to infinity at the cap amount, which is extremely difficult to justify. Of course, the constant price offered by a carbon tax isn't perfect either—the potential damages from climate change rise nonlinearly with total emissions, and the ideal policy is a sliding price that moves up (within limits) if emissions are too high. But in the absence of specific information telling us that the climate system will collapse when the amount of carbon in the atmosphere reaches exactly X, the carbon tax is a closer approximation.

Allowing investors to save and borrow permits, however, relieves an important part of this inefficiency. Since our emissions targets in the long run are very ambitious, it's almost certain that carbon permits will be valuable at some time in the future. Even if we overshoot our targets at the beginning, and market slack sends the carbon price plummeting, it will stabilize well above zero: it's better to simply buy permits and save them until they're in higher demand. This means that if the market can cheaply reduce emissions in the current year below the cap, it will—the need to reduce total carbon emissions determines the price, not the need to meet an arbitrary cap in any particular year. Bankable permits liberate the market to achieve long-term carbon reduction in the most efficient way possible.

Fortunately, the current Waxman-Markey bill includes provisions for borrowing and saving permits. This is incredibly important. We can't let misguided appeals to environmentalist populism ("don't let polluters put off their obligations" or "don't let speculators pile up a fortune") put this policy in danger, when it has the potential to improve cap-and-trade so dramatically.