Biofuel economics and the fallacy of composition

A few years back, when corn was trading for less that $2/bushel, ethanol started to make sense to a lot of people, and it wasn't just politics. I remember meeting young policy wonks in bars in Washington, DC and they would enthusiastically tell me how things had changed, that corn was cheap, that they could get more ethanol from each bushel of corn, that it was a good substitute for MTBE to boost octane and make the fuel burn cleaner, and oil prices were expected to rise. It really did make sense, they told me, and there was support on all sides. All of this was true. They were sincere.

What a long strange trip its been since then.

I was skeptical from the beginning for the following reason: the stylized facts told me the demand for corn was very steep. I was pretty sure corn supply was steep, too. That meant the price of corn would go up a lot if ethanol production increased to any significant fraction of the fuel market. The problem was that these non-economist policy types were taking the price of corn as fixed. They didn't consider how much the price would rise once ethanol was scaled up by a significant amount.

This is what some refer to as the fallacy of composition. Each farmer and each ethanol plant takes prices as given. Individually each is too small to influence price. Collectively, however, they can have a huge influence on price. What is true for a small part of the market is not true for the whole market.

This is basic economics. It's just supply and demand. Part of the problem was we didn't really have good clear estimates of supply and demand elasticities, which made this difficult to explain in a concrete way. That is what motivated my study with Wolfram Schlenker. I guess we were a little late to the table.

Now focus is turning toward other kinds of biofuels, like cellulosic fuels derived from jatropha, switchgrass, miscanthus, sorghum, algae, and waste wood products. The key economic question is how these biofuels will scale up. It's not enough to show that they may make sense at current prices, because if these fuel sources are used on a large scale, prices will go up. The key economic question concerns what activities these alternative biofuels would displace if done on a massive scale. What do the supply and demand curves look like for competing uses of biofuel resources?

These alternative biofuels may hold promise. But I still worry that people are not asking this key economic question. So, if you're doing work on biofuels, please keep in mind that prices for raw biofuel materials will change and that the key economic question concerns how much prices will change. If you don't account for these price changes, you're committing the fallacy of composition, just like those policy wonks singing the virtues of ethanol back when corn was less than $2/bushel.

Ethanol featured on the Lehrer News Hour

I saw this a day late on the DVR.

It's nice to see high-level exposure to the issue.

But I can't believe they didn't mention ethanol's influence on food prices.

Tim Haab likes cash for clunkers

Env-Econ directs us to cash for clunkers.

NEW YORK (CNNMoney.com) -- With auto sales at crisis levels, Washington is trying to figure out how to get Americans buying cars again.

Several ideas are on the table, but two of them are really making the industry pay attention. One plan is to make new car costs tax deductible. The other is to give rebates to Americans with old cars so they can better afford to buy new ones, a program otherwise known as "cash for clunkers."

....

The second plan is more politely known as "fleet modernization." It combines economic as well as environmental goals in one package.

Under a bill introduced by Sen. Dianne Feinstein, D.-Calif., owners of older cars would get vouchers worth thousands of dollars toward the purchase of newer, more fuel-efficient vehicle. For the customer to get that cash, the car dealer would have to certify that the trade-in was getting scrapped and not resold. The car's vehicle identification number (VIN) would be tracked to make sure it never shows up on a vehicle registration again.

Crushing the old car has two benefits. First, it ensures that the consumer's purchase of a more efficient vehicle actually has a net environmental benefit. Second, it prevents a glut of used cars on the market, which would reduce trade-in values for new car buyers, which would cut into the sales incentive effect.

In addition to boosting car sales, the plan could save an estimated 80,000 barrels of oil a day, a spokesman for Feinstein's office said. Because it holds appeal for both environmentalists and the auto industry, the cash-for-clunkers idea seems to be generating the most discussion.

Tim Haab likes the idea. Steven Levitt disagreed with the idea back in August 2008. Levitt is worried about buying clunkers that aren't being driven very much. Levitt apparently has never lived in California. But August seems like a lifetime ago in the auto industry and economy as whole.

Anyway, there has been research on this topic. I think the seminal paper is by Alberini, Harrington, and McConnell. Levitt should have checked it out. The benefits in terms of pollution reduction aren't bad, and I don't think earlier studies counted carbon. Today this seems like a policy that would have a double dividend of a high multiplier effect in a sagging economy plus pollution reduction. I think I'd like this idea even if it were more than $1300 per car.

A naif's view of macroeconomics, macroeconomists and the marcoeconomic situation

I'm an economist but not a macroeconomist. But like many non-macro economists I'm deeply fascinated and concerned by what's going on. If only to document my current thinking, mostly derived from others' thinking, so that I might return to it one day and see how foolish I was, here are eight thoughts.

1) It’s really hard to be sympathetic to Real-Business Cycle (RBC) theories in the current environment. But current events only make especially salient what has been clear for a long time: the evidence aligns very well with Keynesian and New-Keynesian (NK) theories and not so well with RBC. This needs to be acknowledged more widely by those doing work in the area. And the burden is on this area of RBC macroeconomics to prove its ongoing relevance.

I should mention that I am sympathetic to these theories in the following way: RBC theories have pushed economists look more deeply into the underlying rigidities that make the world look more like Keynes and less like RBC theories, or neoclassical theory in general. It’s has been important and useful to have perfect-markets rational-expectations baselines against which to compare the real world.

At this point, however, beating this dead horse of RBC appears increasingly fruitless. Yes folks, economics is about theory and evidence.

More promising areas would seem to be toward refining and pairing down New Keynesian theories, and especially (in my naïf view) the role of risk aversion (more broadly defined than EU) in business cycles. This connects with the informal notion of “confidence” or lack thereof. In Keynes lack-of-confidence is just an outward shift in the demand for money. We need a more refined view and a more direct link to the risk premium. And psychology may be big part of it.

2) Conservative economists seem more aligned with RBC and Liberal economists with NK. This pattern is not universal. Milton Friedman was a Keynesian at heart and more conservatives need to embrace this fact. Greg Mankiw, Martin Feldstein, and Ken Rogoff are all conservatives and are all Keynesians at heart. I'm sure there are examples that go the other way as well. This is not and should not be a political issue.

3) I believe the old Keynes-Friedman monetarist vs. fiscal policy debate needs to be revived, if shifted a bit to suit the times. I believe it is an open question whether monetary policy of a non-traditional variety might be more effective at reviving the U.S. economy than fiscal stimulus. Will there be another Friedman of tomorrow saying that the reason we had this recession, and perhaps a long and deep one, is because the Fed, though aggressive, was nowhere near aggressive enough? If we’re willing to spend trillions in debt for bailouts and stimulus, why not print a trillion or two of cash instead? Would that have been more effective? Why didn’t they do it? Or maybe we’ll be saying why did it take them so long to do it?

My feeling is that this particular debate has been too subdued. Perhaps for political reasons. My feeling is that left-leaning economists don’t want to talk about it because they want fiscal stimulus, and they want that stimulus because they believe the government should be bigger anyway, and this is a way to achieve that alternative goal. To some extent, they have been clear in stating this intention. In principle if not in practice, I think policy goals should be kept separate. That is, if you want stimulus because you want a bigger government, then say that. Don't use bogus arguments that tax cuts can't work or that other kinds of monetary policy can't work. And don't ignore these alternatives either.

My feeling is that the right doesn’t want to talk about it because their inner-Austrian hearts still have sympathy for the gold-standard and they don’t like the idea of the Fed printing tons of money. Some don’t realize that this is what Friedman probably would have prescribed. Some don’t want others to realize that this is what Friedman would have prescribed, because it is so interventionist and they are non-interventionists at heart.

I suspect righties also don’t want to talk about it because they like talking about tax cuts for stimulus because they think this will help shrink the size of government (“starve the beast”). This is the right’s unstated agenda, like "growing the beast" is the sometimes unstated objective of the left. But the right is not clear about this objective at all. It is really strange to watch the intellectual gymnastics here because these are often the same guys arguing that fiscal stimulus is ineffective, Ricardian equivalence holds, and Ricardian equivalence says tax cuts don’t stimulate the economy.

Both right and left don’t want to talk about printing lots of money to buy risky and long-term assets because it makes them nervous. I can understand this. But I think that’s why they should be talking about it.

Serious economic scholars should put aside their personal agendas and start thinking hard about the tradeoffs between aggressive, novel monetary policy verses fiscal policy in terms of pure stimulative capacity and effect.

4) As far as fiscal policy goes, tax cuts do have the benefit of being quick to implement. But handing out cash will not provide much stimulus (it will just be saved). Handing out cash tied to new spending makes a lot more sense. This could work, but it needs to be well designed.

5) On the other hand, spending could have a bigger multiplier but it could take longer—too late, some say, to be effective. The way things are going, this may not be such a bad thing as we are likely to be well below full employment for at least several years. Still, quicker is better, all else the same.

(4) and (5) are to say that fiscal stimulus is not just a question of tax cuts or spending. Like most things, the devil is in the details, and that has not been acknowledged as widely as it should be, probably due to sublime objectives to ‘grow-the-beast’ or ‘starve-the-beast’. Other non-Keynesian reasons for growing or shrinking the government should be put aside; or if, they are part of the rationale, make sure it is spelled out. No sublime motives, please.

6) We should worry more about unintended consequences of greater prospective debt. Economists worried about this a lot a few years ago. Now the problem has become much, much bigger. The worry is that foreign investors will get worried about the size of that debt and pull their money out fast. If they do, all the options we currently have would go by the wayside and we would be in really bad shape—that means high interest rates, runaway inflation, and the Fed would have to tighten and the government would have to reign in spending at the worst possible time.

I think this is unlikely. But the downside is really ugly. More should be worrying about it and writing about it. I’m guessing this is a small deal given foreign debt is all dollar denominated. This essential fact, in addition to the US being so large, is what makes such a crisis seemingly implausible. If people flee the dollar and dollar-priced assets like treasuries, the problem will be strongly self-correcting—inflation will make our debt much smaller, making our debt situation look much better, which will attract investment back. It will also increase demand for exports. This thinking makes me lean more in the direction of more ambitious monetary policy. A good bit of inflation seems like just what the doctor should order. It’s a good way of spreading the pain while simultaneously bringing back full employment. I may be way off-base here. I’d really like to read more from the guys who are smarter and know more than I do.

7) Regarding bank bailouts: It seems these have been and will continue to be necessary. This does not need to be done in a way that incites future moral hazards. The media jargon “bailout” is thus deeply unfortunate, as have been some of the particulars about the way this has been done so far.

Bailouts should not be a happy time for those getting bailed out. All shareholders should be wiped out. In my view this should be standard operating procedure on all future bailouts and the Treasury should unequivocally state this SOP now.

There are reasonable, non-moral-hazard-inducing, and tax-payer friendly ways of recapitalizing the banks. Several have suggested that in addition to shareholders being wiped out, institutional (non-FDIC insured) debt should be partly or mostly wiped out in exchange for shares in the new, recapitalized banks. This approach may recapitalize all banks without any taxpayer cash infusion. Firing all current CEOs without golden handshakes should also be part of the deal. It may or may not involve a short period of nationalization. This makes a lot of sense to me.

Call it a government-managed bankruptcy.

8) Finally, I’m rather ashamed at the low level of debate. I’m ashamed of my discipline of economics when past and prospective Nobel-Prize winners say and write things that are so plainly false and so plainly politically motivated. I’m ashamed that our field is so populated with ideologues wedded so firmly to moral philosophies and beliefs that it dramatically infects their professional work and objectivity on matters of pure positive economics. This whole episode makes it clear to me that our discipline is not yet the science that I would like it to be and that (hopefully) I strive for it to be. We are in the Dark Ages of economics and I hope current events ultimately bring about its renaissance.

An interesting use of data: Obama votes and cotton production circa 1860

This semester I am teaching the second semester of "core" microeconomic theory to a crop of 25 PhD students. The class is a notorious tough slog--a lot of material, a lot of math, and way more abstraction than students see in undergraduate training. This is stuff all Ph.D.s need to know. But getting so much of it over a short period of time can sometimes stunt natural creative tendencies.

So to help keep students' minds open, I'm adopting a set of assignments that George Akerlof gave when he taught macroeconomics at Berkeley. The first assignment is to find an interesting use of data (preferably from one of the best economics journals) and summarize it in 400 words, plus maybe one figure or one table. We will then meet in small groups and discuss them.

In class I gave a couple examples from the academic literature that interested me. Here is one I just stumbled upon on the internet. The black dots indicate cotton production circa 1860. The colored counties show the 2008 Presidential election results, more blue is more votes for Obama.



The map sure is striking. A little tickle in the back of my mind is thinking of all those papers with demographic variables on the right-hand-side of regressions and thinking they really are not so exogenous. (Well, actually, I've always thought that.) A form of Tiebout sorting? Strong historisis or path-dependency? Some of both, probably. I'm especially curious about the exception in southern Tennessee/Alabama border. What do you think?

Is Ethanol Worth 100 Million Starving People?

Via Mark Thoma:

Jeffrey Sachs reports that there are 100 million more chronically hungry people in the world today as compared to two years ago. A key reason for the increase, and plausibly the main reason for the increase, is U.S. ethanol subsidies, which have diverted food grains to fuel production and increased prices of staple grains worldwide. Wolfram Schlenker and I estimate the price increase to be 22-67 percent for corn, soybeans, wheat and rice. It's hard to translate price increases into hungry people, but 100 million sounds plausible to me.

All told, there are about a billion chronically hungry people worldwide.

Sachs says poor countries might be able to produce a lot more food on their own. They just need more money to pay for fertilizer and modern high-yielding seed varieties. One would think higher prices would spur the adoption of these techniques. But they need financing, which is difficult for almost everyone these days.

Maybe addressing this problem would be a good place for the new Obama administration to "extend opportunity to every willing heart — not out of charity, but because it is the surest route to our common good. "

In light of Obama's support for ethanol subsidies, I would hope that he would at least be willing to compensate for the harm this does to the most disadvantaged people on the planet.

Do we need more enviromacroeconomists?

There’s an ongoing spat between Joe Romm’s Climate Progress and Tim Habb’s and John Whitehead’s excellent and well-visited Env-Econ blog. The link here connects the spit balls.

Joe Romm started it, saying things like “economists can‘t walk an chew gum at the same time” and "economists don't understand climate science." The first broad swipe was actually a retort to Rob Stavins who argued that environmental policy and macro stimulus policy were different things that needed different policy tools; the second a retort to Robert Mendehlson, who, well, can say some provocative and controversial things about climate change. Obviously those guys are speaking for themselves, not the profession.

Anyway. I don't care much about the cat fight but I do think there might be something interesting and relevant beneath the surface of Stavins' point and Romm's attack.

John Whitehead and Tim Habb jumped to Stavins’ and profession’s defense. Here's Tim trying to clarify the point:

Joe [Romm] summarizes Rob [Stavins]'s argument as follows:

In short, whatever we do to address climate must not attempt to create jobs. And whatever we do to create jobs should make no effort whatsoever to get off our self-destructively unsustainable economic path. That would not be a Pareto optimum, I guess.

Seriously, Dr. Stavins, just because you haven’t figured out how to walk and chew gum at the same time, doesn’t mean nobody else can.

WRONG. At the risk of putting my words in Rob's mouth (and I have no intention of doing so because that would significantly decrease Rob's intelligence), the point is not that we shouldn't try to meet both goals--good dinner and good shower--but rather, the policy of addressing both at once is STUPID. Determining the best way to make dinner in the shower will result in a bad shower and bad dinner (both outcomes will be suboptimal). There are better ways to reach both goals.

Targeting climate change under the mask of stimulus policy may result in suboptimal stimulus, and targeting stimulus under the mask of climate policy may result in suboptimal climate outcomes.

The critics of Stavin's argument seem to be arguing that the only acceptable way to stimulate the economy is by addressing climate change. The irony is that their own recommendations may very well be counter to their ultimate goal. Good intentions--unintented consequences. That (I think) was Stavin's point. A soggy dinner indeed.

Certainly Rob Stavins doesn’t speak for all economists. I don't understand or appreciate Joe Romm's tone any more than Tim and John do.

But is Rob Stavins (and John and Tim) right about keeping the policy goals separate? For most kinds of market failures I think he would be. This is a classic result in micro: optimal policy dictates that you need as many instruments as you have policy failures. Let’s take an example from environmental economists who are accustomed to externalities. Say you have two of them, pollution and traffic congestion. The best policy taxes pollution at its marginal social cost and driving at it’s marginal congestion cost. One might try to deal with both using a single tax on gas, but that’s an imperfect solution to the problem—it wouldn’t do much to entice people to spread out their commuting times in order to reduce traffic congestion.

But in this case, I wonder: macro and micro are different animals. So different, in fact, that many and perhaps most non-marco economists don’t understand it. For example, here is Eugene Fama, widely believed to be on the short list for the Nobel Prize, mistaking an accounting identity for a behavioral macro theory. And here is Tim Habb applauding the non-bailout of Lehman Bros., a policy decision that pushed the financial crisis and whole macro-economy over the brink. (Macroeconomists were wary, and for good reason, it turned out). These guys are great in their respective micro fields. But they seem way out of their element when it comes to macro.

In the current case there do seem to be synergies between macro and the environmental policy. A good place for the government to spend is on public goods, the realm where environmental economics lives. Of course, it’s always a good idea for government to spend on public goods that pass a basic cost-benefit test. But now we’re in an unusual time when there are additional benefits to government spending, all else the same. This means we should work down the list, toward projects with some public benefits that might not exceed cost in normal times but are worthwhile today when we are at less-than-full employment and monetary policy seems to have its brick wall (unconvential monetary policy, like the Fed monetizing long-term debt nonwithstanding). This, it seems to me, is the way we maximize bang-for-the-buck in government spending.

There are, of course, other considerations, like the size of the multiplier. Projects with more public goodness may or may not have multipliers larger than other spending possibilities. I’m no expert on multipliers—I just know what I read in macro textbooks—that multipliers increase with the marginal propensity to consume. That means multipliers are higher if money is given to poorer people who tend to save very little.

All this suggests an interesting set of tradeoffs when considering stimulus policy, and that public goodness, including environmental public goods, are surely an important consideration. In this case, I think, the market failures cannot be entirely separated. To consider these tradeoffs well we need someone who is both a good environmental economist and a good macroeconomist. To my knowledge, that is a rare animal.

Maybe we need more enviromacroeconomists.

Update: I have not done justice to the nuanced view in Rob Stavins' original quote. In the original quote Rob said "that is an illustration of the fact that it is not always best to try to address two challenges with what in the policy world we call a single policy instrument.” He was far from absolute. Tim Habb was a little less nuanced in defending Stavins' from Romm's attack. I probably should have quoted Stavins, not Habb in my post. And like I said, I agree with Stavins: usually we don't want to address two challenges in a single instrument. I found the debate interesting only because I think the current macro situation may be an exception.

Commodity price puzzles

There are many puzzles about the behavior of commodity prices. And things seems to be getting stranger by the day. Mind you, these puzzles aren't new, they just seem more extreme than in the past.

A few of them:

(1) Non-renewable resource prices don't trend up at the rate of interest, or at all, as Hotelling's rule predicts (But they sort of look like asset prices in that they are near random walks).

(2) Agricultural commodity prices display "too much" autocorrelation. Storage should make corn prices behave more like oil prices, but not exactly. But agricultural commodities display about as much autocorrelation as oil does. (See work by Deaton and Laroque)

(3) Firms and/or individuals store commodities even when the futures price is less than the spot price ("backwardation" in commodity jargon). This seems like throwing money away. The academic literature calls it a "convenience yield". I'm guessing it could be a risk premium of some kind--if storage pays off, it tends to do so when the storer really needs the cash, kind of like insurance. In any case, it's weird--cash seems a lot more convenient than a pile of grains expected to lose value every day.

(4) At times, and especially recently, futures prices do not converge to spot prices. This is making it more difficult for farmers to hedge their price risk. Look up Scott Irwin at the University of Illinois for more on this.

(5) And now we have "Super Contango," the extreme opposite of backwardation. Here's Kevin Drum:

Normally, it costs more to buy a barrel of oil for delivery six months from now than it does to buy a barrel of oil today. After all, if you're not going to take delivery of my oil until July, then I'm going to want the spot price of the oil plus the cost of storing it plus the cost of having to wait for my money. So maybe a barrel of oil that costs you $38 today will cost you $41 for delivery six months from now.

But instead of $41, what if the July price is $53? Then anyone who wants to can make a guaranteed killing. Accept the contract, buy a tankerful of oil, store it for six months, and then deliver it. Even after the cost of storage and the interest on the loan you took out to buy the oil, you'll make a quick and easy twelve bucks per barrel profit....

My gut instinct tells me the root cause or causes of these puzzles are in some way related to broader financial puzzles, like the large and highly variable risk premium on stocks. That we observe these strange things becoming stranger in a time of broad financial problems, credit constraints, and a huge risk premium, seems unlikely to be a coincidence.

Discount rates and greenhouse gas emissions

How much should we spend to curb greenhouse gas emissions?

A big part of this question concerns the discount rate: the rate used to balance future expected benefits against today's costs of reducing emissions. Common thinking among some economists is that since you can put money in the stock market and expect a decent return over the long run, benefits in the distant future should be worth a lot less in today's dollars. This thinking suggests we shouldn't spend much on reducing greenhouse gases emissions.

That first-cut thinking by economists gets the economics wrong.

Consider that we might also think of current emission reductions as insurance. We buy insurance all the time and insurance premiums typically have negative expected returns. We're willing to accept that negative expected return because insurance pays big indemnities in unusually bad circumstances when we really need the money.

Curbing greenhouse gas emissions seems a lot more like buying insurance than investing in stocks, and this pushes the discount rate down a lot.

Let's step back and spell out all the pieces.

There are three essential considerations in choosing the discount rate: (1) the benefits will accrue many years in the future; (2) the benefits, like climate change itself, are highly uncertain; (3) the discount rate itself is uncertain.

(1) means small differences in the discount rate have a huge effect on the bottom line because the discount rate interacts with time exponentially; (2) matters because risk premiums appear to be really high; (3) matters for the same reason (1) matters, as I'll explain below.

Standard theory says the discount rate (d) is the sum up of three parts:

d = eta + g + r

eta is the value of consuming something today verses consuming it tomorrow, holding all else the same. Stern set this parameter to 1% and some felt that number was too low, probably because they misinterpreted this number as the discount rate itself. I think 1% is reasonable. Those hugely worried about intergenerational equality might set eta equal to zero. In any case, its probably smallest and least important part of d.

g is economic growth, adjusted by the inverse elasticity of intertemporal substitution, if not 1 (log utility). Greater growth means lower marginal utility in the future and a higher discount rate. Given historical growth rates, log utility would put this number around 3%, higher if the elasticity of intertemporal substitution is low.

r is the risk premium. Because investments in reduced GHG emissions will pay off most in bad times, if climate change turns out to be a very bad thing rather than a mild or even beneficial thing, the risk premium is negative. This gets at the insurance/stock-market dichotomy above. Investments in assets like stocks, which pay off most in good times, have a positive risk premium. Given historical estimates of risk premiums, this could be a big negative number.

A reasonable person can probably justify anything in the range of -3 to + 5% for the sum of thee three numbers. That's a big range. And yes, it could be a negative number. How could it be negative? It's that negative risk premium, which could be huge. Subtract a big risk premium from a typical real bond rate and you're in negative territory.

Okay, now (3), what's the implication of the discount rate itself being uncertain?

Weitzman showed that this pushes the effective rate down. This follows from the simple fact that the present value of future benefits decline at a decreasing rate with the discount rate. In other words, the value function is convex. If it turns out that the discount rate is low or negative, expected returns are absolutely huge. So, we should err on the downside when we choose a
discount rate.

In sum, I have no idea how much we should spend reducing GHG emissions because I have no idea how costly reductions would be, the costs should warming occur, or how GHG affects all the possibilities. But I think a low discount rate, in the ballpark of zero, should be used when weighing current expenditures to future expected benefits. Moreover, I don't think a zero discount rate should be controversial among serious economists.

Nick Stern used low number but not as low as zero, and he got a lot of grief for it. I don't think grief was justified, except maybe he could have done a better job explaining why the number should be so low.

For more on this, read Martin Weitzman. John Quiggen has a nice writup on this too. I think I've summarized the essence of their arguments here.

How much has ethanol influenced commodity prices?

It's no news that commodity prices have been unusually variable over the last few years. Oil and gas prices have been most prominent in the press, because that's what has most affected people in this country.

In my little corner of academia it's been about agricultural commodities. Agricultural commodities have also been the big story in the poorer half of the world living on $2/day. They spend most of their money on food staples, like rice, wheat, corn, palm oil, and soybean oil. When these prices double or triple (or more) it's a pretty big deal. It makes our problems with gas price fluctuations seem trivial by comparison.

With the financial crisis and global recession bringing prices back down, the news has subsided. But before the price drop there were lots of prognostications about the underlying causes the boom in commodity prices. Four key reasons came out of the discussion:

1) Ethanol.
2) Asian demand growth.
3) Drought in Australia.
4) Slowing of yield growth.

I briefly summarized these factors here. In addition to these factors (and partly because of them) were export restrictions, particularly of rice in south Asia. These restrictions kept domestic prices lower in the countries that imposed them, but further increased world prices.

What's less clear are the relative magnitudes of these various factors.

In many ways, the first factor is most interesting, since growth in ethanol demand was driven mainly by U.S. policy: a 51 cent/gallon subsidy, a tariff on less expensive imports from Brazil, plus mandates establishing significant growth in production going forward. In the U.S., ethanol is made from corn, an important food staple (it feeds most livestock, among other things).

At the recent ASSA meetings in San Francisco, Wolfram Schlenker and I weighed in with our own analysis. Our view: we cannot get very far in ascribing relative magnitudes without having a clearer idea about the world supply and demand of food commodities. In our paper we estimate the supply and demand of calories produced from corn, soybeans, wheat, and rice combined, which comprise about 75% of the calories from which food is derived worldwide.

First some numbers to put things in perspective: ethanol now consumers about 1/3 of U.S. corn production and is mandated to grow another 50 percent or more in the next six years; U.S. corn production is about 40 of world corn production and comprises about 80 percent of world exports; the U.S. also produces nearly 40 percent of the world's soybeans, about 10 percent of the world's wheat, and some it's rice, too. All told, the U.S. is, by far, the world's largest producer and exporter of these four key staple commodities, about 23 percent of world caloric production and a larger share exports (total caloric exports are harder to quantify).

Adding it all up, U.S. corn-based ethanol uses a consumes more than 4% of world caloric production and will soon consume between 6 and 7%. Since markets are forward looking, we'll use 5% for our back-of the envelope calculations.

The figures below show the caloric breakdown by crop worldwide and the U.S. share of the total. The trend up is due mostly to technological advance--the "green revolution"--which has brought greater yields per acre.


From bottom to top, the crops are soybeans, rice, corn, and wheat.



An interesting aside: Note how fluctuations in corn production, and U.S. total caloric production, appear to mirror the world aggregate fluctuations. That's because U.S. corn production is so high and geographically concentrated in the Midwest. Most of the rest of world production is more geographically dispersed, so the fluctuations tend to average out. This means fluctuations in world caloric production come disproportionately from the U.S., as illustrated in the plot below.



Now let's turn to supply and demand. The figure below provides some powerful clues. It shows the world caloric price (the calorie-weighted average across crops) together with world caloric production and consumption. Consumption is smoother than production because storage smooths out the fluctuations. Prices fluctuate negatively with production, which is indicative of supply shocks and price movement along a downward sloping demand curve.

Here's the key thing to note: the size of the quantity fluctuations, especially consumption quantities, are quite small while price fluctuations are quite large. I see no way this can happen unless both supply and demand are extremely steep (inelastic).

We estimate the supply and demand elasticities using variations in weather, approximated by country and-crop specific yield shocks that are then aggregated to a worldwide shock. These fluctuations show an even stronger negative association with prices than the aggregate production variations shown in the graph above. The aggregate production fluctuations include expansion and contraction of land cultivated, not just weather. In other words, they include movement along both the supply curve and the demand curve.

The graph below shows our aggregated world weather-induced yield shock together with prices (in log terms). Note that the scale for shocks on the right (the shock is measured in proportion to trend production) is on a scale equal to 1/10 the scale on the right for log prices. This means small yield shocks have a big effects on prices, despite the fact the storage smooths yield shocks over multiple years.

We estimate demand by measuring the relationship between consumption changes and price changes caused by weather-induced yield shocks. Since the quantity changes are very small and the price changes are big, we get an inelastic demand curve: quantity demanded goes down about 1% for a 25% increase in price.

We estimate supply by measuring the relationship between quantity produced and price changes caused by *past* weather shocks. Because of storage, which smooths shocks over several years, past weather shocks affect both current and future prices. In the current year, farmers can't respond to the higher prices by producing more, but they can respond in future years. The result: quantity supplied goes up about 1% for a 10% change in price.

The nice thing about using weather to identify these demand and supply relationships is its near randomness. It's about as close to an experiment as economists can hope to find in naturally occurring data.

So what do these elasticity estimates imply about the effect of ethanol on the prices of food commodity staples? The 5% shift in demand for calories caused about a 34% increase in the average price of calories derived from corn, soybeans, wheat and rice. Accounting for uncertainty puts the price increase in the range of about 22% to about 67%.

R is Rising from inauspicous beginnings

I've long been a big fan and a frequent user of "R," the free and bleeding-edge statistical software and programming language.

In the past when I mentioned R people hadn't a clue what I was talking about. That has been changing at a steady but seemingly exponential rate.

So the New York Times recently had two pieces about R, here and here. And there is discussion among practicing data geeks about the future of R (bright?) and the future of SAS (not so bright?). See Andrew Gelman's excellent blog where he weighs in here, here, and here.

For the record, I use SAS too.

You can learn about and download R here.