Arctic Ice: Prediction That Isn't

A recent story suggests, based on computer modeling, that the decline of arctic sea ice may be a good deal further in the future than various people predicted, fifty  or sixty years instead of five or ten. It struck me because I had some posts a while back pointing out that a NASA/JPL web page was misrepresenting the facts on the subject, claiming a continued decline in the face of a (perhaps temporary) reversal. 

The other thing that struck me about the post was that the author did not understand the nature of computer modeling and how you test it:

Accuracy of future predictions was checked by running simulations of the late 20^th century. The Model replicated the events of the past well enough to suggest that its forecasts of possible futures are realistic.

How are such models created? By fitting to past data. Having used that data in constructing the model, it is no longer available to test it.  Someone is said to have claimed that with ten parameters he could fit the skyline of New York. Assuming he did it, it does not follow that by keeping the same regression coefficients while increasing the range of the parameters he could predict the skyline of the rest of the country.

In order to test the predictions of a model you need to do it against actual predictions—information that didn't go into building the model.

What the Tea Party Gets Wrong

Suppose the government wants, for some reason, to subsidize biofuels. There are at least three different ways to do it:

1. For every gallon of biofuel produced, the government will pay the producer a one dollar subsidy.

2. For every gallon of biofuel produced, the government  gives the producer a one dollar tax credit.

3. Make a regulatory rule that forces people to use more biofuels, such as a requirement that gasoline can only be sold if combined with at least ten percent biofuel.

The first two differ only in labeling. They have the same effect on the federal budget. They provide the same amount of subsidy. They are both, in fact if not in form, federal expenditures. The only difference is that the second is an expenditure masquerading as a tax cut.

A lot of expenditures pretend to be tax cuts. If you look at the CBO figures for federal income tax in 2007, you discover that, on average, the bottom 40% of the income distribution not only does not pay federal income tax, it is paid federal income tax—a federal welfare program in the form of tax credits. And since these are 2007 figures, it is Bush's federal welfare program, not Obama's.

As best I can tell by news stories dealing with recent budget controversies, this simple point has somehow been missed by the Tea Party Republicans. They insist on taking such expenditures at face value, as tax cuts rather than expenditures, hence oppose their elimination. They have thus fallen for the very simplest scam operated by their opponents in both parties, and by doing so come out against instead of for cutting federal expenditures.

The third alternative is also an expenditure, also a subsidy, should also be opposed by those who wish to reduce the role of government in the economy. But at least it uses a less obvious device, goes to more trouble to hide what it is, than number 2.

Capital Gains Taxation: Contra Landsburg

Steve Landsburg has a recent post on capital gains taxation in which he makes one odd but arguably legitimate point while missing two other and, I think, more important ones. The result is that he gets the wrong answer to the question of how capital gains ought to be taxed. 

His central claim is that the tax rate on capital gains ought to be zero—more precisely, that if it is zero, taxpayers with capital gains will end up paying as taxes the same proportion of their income as those with other forms of income. It's not as screwy a claim as it seems at first; here is the argument, in my words not his. It assumes a 50% income tax, no capital gains tax.

Taxpayer A earns $1000, pays $500 in taxes, has $500 left to spend on her own consumption—half as much as she would have in a world without taxes.

Taxpayer B earns $1000, pays $500 in taxes, uses the remaining $500 to buy an asset which then appreciates at 5%/year for the next twenty years; for simplicity we ignore compounding. B then sells the asset for $1000, which he can spend on himself. 

In a world without taxes, B would have had $1000 to invest and so would have ended up with $2000. Hence A and B have both paid the same tax rate—50%. The only difference is that A chose to take her (taxed) income in the form of $500 of consumption in (say) 1990, B his (taxed) income in the form of $1000 of consumption in 2010. Each ended up with half the consumption he would have had in a world without taxes, so is really being taxed at 50% on all income. Hence, Steve argues, the fair capital gains tax rate, the rate that treats capital gains like other income, is zero.

What is being described here as capital gains looks an awful lot like interest. B could, after all, have deposited his $500 in a bank at 5% instead of buying an asset that appreciated at 5%. From an economic point of view, interest is what people are paid to postpone their consumption, making resources available for other people to use productively. It's not obvious why interest should not count as income and be taxable as such.

No doubt some of what shows up as capital gains is in fact implicit interest, but I don't think that is the natural way of looking at most of it. I think it is more accurately viewed as the return from a particular form of skilled labor. A speculator/investor spends time and effort figuring out what firms and what assets are going to increase in value and investing in them, improving the allocation of capital, nudging markets a little closer to efficiency, and being rewarded, assuming he does a good job, with income above and beyond the normal return on capital. I do not see why the income from that form of labor is a less suitable subject for taxation than income from digging ditches. And since the return is not a fixed proportion of the amount invested, as in Steve's case, but a function of the time and effort spent investing it, Steve's argument for implicit taxation does not apply.

There is, however, a serious problem with treating capital gains as ordinary income—much of it, perhaps most of it, is not income at all.

To see why, imagine that you buy a house for $100,000 and sell it, twenty years later, for $200,000. Over those twenty years, not only has the price of housing gone up, the price of practically everything has gone up, with the result that two dollars at the end of the period will buy about what one dollar bought at the beginning. Measured in nominal terms, by counting dollar bills, you have made a capital gain of $100,000 and will be taxed on it. Measured in real terms, by what those dollar bills will buy, you have made a capital gain of zero. Prices in our society trend up over time, so measured capital gains overstate—for long periods greatly overstate—actual capital gains.

Combine these two arguments and you have a simple conclusion. Capital gains ought to be indexed—measured in real rather than nominal terms, purchasing power not number of dollars. That done, they ought to be taxed as ordinary income.

The Syrian Bodycount Puzzle

"a crackdown that, by the count of some human rights groups, has killed more than 2,000 people. Hama, the victim of one of the bloodiest moments in modern Middle East history, is a national symbol of violent repression. The military crushed an Islamist revolt there in 1982, killing at least 10,000 people. The assault in Hama last week, in which more than 200 people were killed, inflamed sentiments across the country" 

             (recent news story)

Syria has been in revolt for months. News stories describe attacks by tanks and troops on nonviolent protesters. At least one story on the recent attack on Hama made it sound as though the city had been bombed out and destroyed. 

Checking Wikipedia, the population of Hama is 696,863. 200 people killed comes to fewer than one in 3000.

Which raises an obvious puzzle. The figures on bodycounts are coming from the opposition, which one would not expect to minimize them. Yet they seem surprisingly small, given the scale of the violence.

Which makes me wonder if, at this stage of things, a lot of what is happening is bluff—security forces killing a handful of demonstrators in order to shut down demonstrations, while refraining from the sort of large scale violence of which they are surely capable for fear of setting off a level of conflict with which they cannot adequately deal.

---

"Syrian troops stormed the port city of Latakia and sprayed it with gunfire on Saturday, killing at least two people" (Aug 13 news story)

Probability of U.S. Default

"The United States can pay any debt it has because we can always print money to do that. So there is zero probability of default" said Greenspan on NBC's Meet the Press.

 The first sentence is true. The second is not.

Suppose that avoiding default requires money creation on a scale that would set off an inflation rate of fifty or a hundred percent a year. The U.S. could--but would it? That would be a de facto default although not a de jure one, since creditors would be being paid back in inflated dollars. And it would have all the added costs of inflation.

Would Greenspan advise doing it? Unlikely.

Reality Based Community?

I was struck by a recent post to a NYT blog on the subject of "What Happened to Obama." The author is identified as a psychology professor. His thesis is that Obama should have told a story to the American people  that made sense of what happened. The story he should have told is the standard left of center view of who did what and was at fault. No mention of the fact that the collapse started in an industry dominated by two giant firms, both created by the federal government, both (along with private firm) long pressured by politicians of both parties to make it easier for people to buy houses with borrowed money.

Three things struck me about the piece. The first was that depended on the author's opinions about subjects in which he had no expertise—in at least one case, in which his factual belief was strikingly false ("a deficit that didn’t exist until George W. Bush gave nearly $2 trillion in tax breaks largely to the wealthiest Americans and squandered $1 trillion in two wars." I suppose he could mean that there was at least one post-war year in which the budget was in surplus—but that would require him to care about facts, not stories.) He simply took for granted, as almost everyone on his side does, the 1960's Keynesian story that deficits reduce enemployment—despite the fact that the predictions the administration made based on that theory have not turned out to be true.

The second thing was the confidence with which he wrote. He has no doubt that his opinions on subjects in which he has no professional expertise are correct—presumably because they fit his political views and those of the people he knows. He probably does not even know that there are professionals in the field, including ones with Nobel prizes, who are skeptical of the economic theories he takes for granted.

The third and most interesting was the focus on "story." As he put it, "in similar circumstances, Franklin D. Roosevelt offered Americans a promise to use the power of his office to make their lives better and to keep trying until he got it right."

It apparently did not occur to him that reality matters—that if you give the patient the wrong medicine he may die, even if you have a good story about why it is the right medicine. It apparently did not occur to him that the outcome of the policies FDR followed was the longest and worst depression in U.S. history. Which might have had something to do with the relation between FDR's story and the reality it claimed to describe.

Hence the title of this post. The blogger in question apparently does believe what some unnamed Bush official is asserted to have claimed—that one makes one's own reality. If you only have a good enough story ...  .

The Cost of Healthy Eating: Incompetence or Fraud

According to a recent article, "A new analysis shows healthy eating can really run up a grocery bill, making it tough for Americans on tight budgets to meet nutritional guidelines. The study estimates that getting the average American to the recommended target of just one nutrient, potassium, would cost an additional $380 each year." 

Anyone who believes that should Google for "potassium supplement"—priced at $9 for a 120 potassium iodide tabs of 32.5 mg each from one source for $9, 100 caplets of 99 mg of potassium gluconate for $6.87 from another, and about ten cents a pill—with calcium and magnesium thrown in for free—from a third.

The trick is quite simple. The article pretends to be about what healthy eating costs. It is actually about what people who eat healthily spend. Higher income correlates with better education, so people who spend more also, on average, spend better, nutritionally speaking. That is no evidence that good nutrition costs more—and, as a comparison between the price of spareribs and the price of pork and beans or fruit salad would demonstrate, it often does not. Precisely the same analysis could be used to show that people who spend more on rent eat better too.

It is possible, although not likely, that an author could be sufficiently clueless to make the argument and believe it. But not this author. Reading the article it is pretty obvious what axe is being ground.

And I have difficulty believing in an author who thinks that if only the prices of apricots and raisins were sufficiently subsidized, people who currently prefer Happy Meals would switch to fruit salads instead.

Obama and Taxes: If the Truth is Too Complicated ...

President Obama's rhetoric in favor of tax increases is heavy on claims that the rich fail to pay their share, light on facts. Most of it is put in terms of claims about the federal income tax. Which is strange, since the federal income tax is paid almost entirely by high income taxpayers.

For details, see the figures provided by the Congressional Budget Office. For 2007, the latest year they cover, the bottom 60% of the income distribution paid about 1% of federal individual income taxes, the top 20% paid about 86% of the total and the top 1% almost 40%—the opposite of what Obama's rhetoric implies.

These figures are, however, misleading in two different ways. The federal individual tax is heavily weighted towards high income taxpayers, but it only produces about 40% of federal tax revenue. Payroll taxes produce about the same amount and, because they are paid only up to a maximum, the top 20% of the income distribution pays a lower share of its income in that form than the quintile below it, making the combined effect considerably less graduated than the figures on the income tax alone suggest.

The second problem should be obvious to an economist but seems to be invisible to almost everyone else. It's natural to think of the size of the check I write to the IRS on April 15th as the measure of the cost to me of the income tax—natural but wrong. 

To see why, suppose that the demand for high wage workers—top surgeons, lawyers, executives—is very inelastic. High taxes on such workers reduce the number of suitably talented people willing to enter the field, with the result that customers in need of their services bid up their wages. In the limiting case of perfectly inelastic demand, the result is to transfer all of the cost of taxation from those who hand over the money to the consumers of their services.

The same issue arises for other taxpayers as well. Payroll taxes take the form of a tax "paid by" the employee and another tax "paid by" the employer, but both are actually a tax on the same transaction—hiring labor. If an employer owes a thousand dollars in wages to an employee, it is of no economic significance whether the fraction that goes to payroll taxes is taken out by the employer before he hands over the money or by the employee after.

Who bears the actual burden of taxes depends on elasticity of supply and demand. If the labor supply is elastic, if the number of people willing to work is not very sensitive to wages, the burden ends up on the employees. If the only way of getting more workers is to pay them more and employers are willing to pay whatever it takes, on the other hand, wages rise by about the amount of the tax and the burden is born by the consumers of what the workers produce. 

Similar problems arise if one tries to estimate the real effect of other sources of revenue, such as the corporate income tax. Corporations are legal people but not actual people, so taxes they "pay" end up reducing the consumption of customers, employees, or stockholders. Who bears how much is again a hard question.

The CBO attempts to take account of such issues in its calculations; its conclusion, again for 2007, is that the bottom 40% of the income distribution bears about 5% of the burden of federal taxes, the top 20% almost 70%. It finds the lowest quintile  bearing a tax burden of about 4% of its income, with the rate rising quintile to quintile to about 25% for the highest and almost 30% for the top 1%.

Whether the CBO estimates are correct, I do not know, and I doubt they do. Whether they imply that high income tax payers bear more or less of the burden than they should  depends on how you believe the burden should be distributed. But the CBO estimates, at least, do not support rhetoric implying that executives pay taxes at a lower rate than their secretaries.

And I haven't even mentioned state and local taxes...  .

Obama presumably understands all of this; if not, he surely has people working for him who can explain it to him. Explaining it to the public in the form of sound bites designed to support the policies he favors is a harder problem, but not an insoluble one. 

If the truth is too complicated ...

Schooling Compulsion, Incentives, and Literacy

There are at least two different ways of getting someone to learn something. You can offer to teach him something he wants to know, or you can compel him to learn something you want him to know. Unschooling uses the first approach, conventional schooling the second. One difference between the two is their effect on the incentives of teachers.

Consider the case of literacy. The ability to read is useful to almost everyone in a modern society, so one would like an educational system that does a good job of teaching it. It is widely believed that the current American system does not.

If the objective is to teach people to read, the obvious starting point is to ask what sorts of things those people would enjoy reading, since it is easier to get someone to do something he likes doing. The answer might be comic books, car magazines, science fiction, fantasy, soap opera summaries, or any of a wide variety of other sorts of written material, depending on the particular people being taught. 

As best I can tell, that is not the approach taken by conventional K-12 schooling. Instead, students are assigned to read books chosen on one of two criteria. Either they are books regarded as good literature—famous books from the past or current books that English professors approve of—or books believed to teach lessons that the people selecting the books want taught. That would include biblical literature in the past, patriotism—or acceptance of homosexuality, depending on the state—at present, and a wide range of other lessons, depending on current and local political fashion. While it is always possible that the books chosen would also be ones students enjoyed—I'm very fond of Kipling, some of whose stories might be assigned reading in English class—that is not what they would be chosen for, so the odds are not very good.

The ability to read is useful to almost everyone. Knowledge of and appreciation for great literature, even if we accept the educational establishment's definition of what qualifies, no doubt can enrich one's life, but on the evidence of what books people actually read it does not enrich the lives of a very large fraction of the population. That suggests that learning the former should probably have considerably higher priority than learning the latter.

In an educational environment where teachers can advise and persuade pupils but not compel them, it will, because the teachers who insist on telling their pupils to read books that the teacher likes and the pupil does not will shortly find their advice ignored. In an environment where teachers can tell students what books to read and, to at least some degree, punish those who fail to obey, on the other hand, there will be a strong temptation to assign the books that the teacher thinks the student ought to read, sacrificing the higher priority of literacy for the lower priority of literature—or, sometimes, propaganda. 

Which may explain why Johnny can't read.

I encountered a different version of the same logic a good many years ago in my own work. My Price Theory textbook was out of print.  I decided to rewrite it into a book targeted at the proverbial intelligent layman, the sort of book that gets read for the fun of it while teaching the reader the basics of an academic subject, in my case economics. My model, insofar as I had one, was The Selfish Gene, a book from which I learned quite a lot about evolutionary biology.

In the course of the project, it occurred to me that there was an important difference between the book I was starting with and the book I intended to end with; nobody would be forced to read the latter. It followed that if at any point the reader decided that it was not worth continuing, I would lose him. To deal with that problem I followed a deliberate policy of starting each chapter with a hook, a puzzle that would sufficiently engage the reader to persuade him to finish the chapter to find the solution. Economics is full of such puzzles; I don't know how hard it would be to do the same thing in another field.

The result, Hidden Order: The Economics of Everyday Life, has been by a sizable margin my most successful book. 

Incentives matter—including mine.

———

In the interest of honesty, I should add that some people are forced to read Hidden Order, because it is occasionally used as a textbook, even though that was not the purpose it was written for. But not, I think, very many people.