Bubble Land

There's been a lot of talk about bubbles in commodity prices, including food commodities like corn, wheat and soybeans.  I've been critical of this, as have a few others who I admire.  Now there's a lot of talk about farmland values and I've been asked a couple times whether I think there's a farmland bubble.

My answer:  I don't know.  It's hard to get enough data on farmland prices, the nature of the transactions, and whether we have reckless lending and leverage that would be the telltale signs of excessive speculation.  It's not just that there are so few farm transactions.  It's also that much of the data we do have is made utterly useless due to USDA's antiquated definition of a farm (more on this some other time).

Perhaps most importantly, farmland is a durable good that cannot easily be reproduced, and buying and selling it involves huge transactions costs.  These features probably make farmland susceptible to a bubble, far more so than nondurable and easily tradeable commodities like corn.  Possibly even more than houses.  But they also make its value potentially sensitive to fundamental forces, like low interest rats and high commodity prices.  Calling a bubble is tough business; if it were easy, bubbles wouldn't exist.

Before getting into land pricing fundamentals, I'm going to try to clarify why price bubbles in consumable non-durable commodities are so rare.  The trick here is to follow the path of how a speculative expectations could drive the market, and then check to see how powerful are the natural corrective forces keep bad expectations from becoming self-fulfilling.  

The usual story is that Wall Street brokerage houses and various annuities are now investing in commodity indexes.  This basically means a lot of outside money is buying commodity futures that promise to buy so much of a commodity delivered on a particular day and location, and rolling them over to the next contract just before they expire.  This is an easy way to invest in commodities without ever really holding the stuff physically.  Indexes do this with lots of commodities and different delivery dates simultaneously.

Now, suppose all this money were to come flooding into commodity futures and the other side of the market---those selling commodity futures---did not immediately adjust their activities, and so futures prices began to rise.  At this point, people working close to the ground in these markets, especially those with access to storage facilities, would see higher futures prices and might be inclined to sell at those  prices rather than today's.  To be sure they had enough product to sell at that higher future price, they would begin to hold more inventories.  But if those in the product market started holding more inventories, they would then be keeping product off the market, thereby driving up current prices as well.

Now, as current prices increased, the incentive to store product and capitalize on higher future prices would be diminished.  And equilibrium will settle out for some basis, which is the difference between the current spot price and the futures price, where inventory holdings would make sense given the cost of storage.

One thing to note here is that spot prices and futures prices do generally move up and down together, due to just the phenomenon described above.  While the basis does change over time, it changes far less in absolute terms than does the overall price level.

So how could speculation continue to drive up prices?  We would need to have continuing cycles like the one described above, where futures prices would be driven up by bullish market expectations, which would drive up futures prices, and commodities would continue to be withheld from the market and stored. If bullish expectations for higher future demand or lower future supply did not materialize, even greater inventories would have to be accumulated to maintain high prices.

Now, if all these expectations were false or misguided, it would be tempting for those storing all the corn, soybeans or oil to dump their inventories.  Furthermore, the people closest to the ground and seeing where the supply and demand come from, would be the ones to make this call.  They would have everything to gain, and little to lose by selling short.  And if outside investors could do the basic math and find that all the price increases were simply coming from higher inventory holdings, a few might quickly become bearish.  After all, if inventories are transparent to the market, this wouldn't be very difficult thing to discern. Some more would simply need to be more willing to sell futures than buy them.   

Finally, keep in mind that worldwide inventories for any of the consumable commodities rarely exceeds more than a small fraction of one year's consumption.  In a commodity equilibrium, these inventory holdings essentially comprise a relatively short-run gamble (or insurance) for the day when a big positive demand shock or negative supply shock causes prices to spike, giving occasional big rewards to storers (and speculators), while ultimately serving to calm price volatility.

Contrast this story with what happened in houses or the technology stock market bubble in the late 1990s. Gold might be a similar case, different from other commodities because it is durable and not consumable (we don't burn it or eat it).  In these cases the time horizon of the investors is many decades.  The gamble was that dividends in the form of rents would rise somewhat faster rate going forward.  Or that interest rates trending lower, raising the present value of the future stream of dividends by a lot.  Another story---credible at the time---was that the magic of modern finance (insurance, options, derivatives, etc.) and modern Federal Reserve policy was to reduce the cost of bearing risk, thereby limiting risk premiums.  Add these fundamentals---which can, in principle, cause huge change in the value of long-lived assets---to a general euphoria and unguarded optimism and you have the makings of a bubble.  To really get things going, add nearly unlimited borrowing capacity, shady lenders, and an ability to highly leverage ones positions.  Then add middle men who gain on the upside but do not lose on the downside of the gambles they've made and can hide their true positions, sometimes even from themselves, in a blizzard of options and credit default swaps.

Thus, I think the biggest difference between consumable commodity prices and long-lived assets are the time horizon of the investment.  They also are thickly traded markets with low trading costs.  It's a lot easier to dump a position in commodities than it is to dump a house.  Stocks also have low trading costs and that is a key reason whey stock-market bubbles are likely rarer than home-price bubbles.  The third issue is leverage: the ability to buy a house or buy stocks using other people's money.  I haven't seen this kind of thing in consumable commodity markets, at least not on the same scale.

Now things may be different for farmland values.  Here we have a market that shares many characteristics of houses.  We also have very low and declining interest rates that would make real assets like farmland seem especially attractive.  In principle, very high prices might be justified given the right combination of expectations going forward.  So, to understand whether these expectations are reasoned decisions or frothy hopes of blind speculators, I think we need to know something more about those who are buying farmland.  Are they highly leveraged?  Who's lending, and on what terms?

The last time I had access to decent data (when working at USDA) it looked like few farmers were highly leveraged.  Most of the major operators leased their land on annual contracts at very reasonable rates.  Their crops were mostly insured (at subsidized rates).  Most of the land that farmers did own appeared to be owned free and clear.  Major debts were for operating costs. And the landowners leasing the land out were a widely mixed variety of former farmers, their families and outside investors, who likely held land as a small part of a much larger and more diverse portfolio. Owners were not highly leveraged.

Now, it's been a few years since I've been able to look at that kind of data and perhaps the fundamentals have changed.  In housing the fundamentals changed fast between around 2004 and 2006. I don't have access to the mirco data on farm balance sheets, the nature of the farmland mortgage lending market, or even good farmland price data to get a good sense of where things stand right now.

Also if you've followed this blog you'll know I'm reasonably bullish on food commodity prices.  A couple years of record crop harvests could bring prices down a fair amount.  But I'm not especially optimistic about that. And demand growth could eat record harvests real fast.  So maybe high farmland values are worth it.