Saturday, June 25, 2011

Did the Fed cause the increase in world food prices?

Long and messy post, and I started with an idea that does not seem to be completely the right one. However, it was an interesting research... So here it is:

I mentioned in May a discussion following a Telegraph article arguing that the Fed's easing policy was somewhat a cause of the Arab Spring. This was related to several discussions by Joseph Stiglitz and others that the Greenspan era of low interest rates led to high oil prices. The Lex Column in the FT yesterday argued that
 the expansion of speculative capital is a manifestation of the global increase in cheap money. Food markets are likely to stay messy as long as real interest rates stay negative.

I wanted to check quickly what could be said on commodity prices.

There were two booms in commodity prices, one starting in 07 and the other starting in 09 and crept up in June 2010. Here is the FAO food price index. Interestingly, you can note the recent plateau that has been reached for the last couple of months
FAO price index, 2002-2004=100
First or all, the usual reproach to the Fed comes from the fact that a rise in food prices is seen as bad for the developing world. Food riots were testimonies to that line. The World Bank estimates that "44 million people may have fallen into poverty in low- and middle-income countries due to the rise in food prices since June 2010". Interestingly though, new research qualifies this idea:
A striking feature of the self-reported food insecurity data is that it suggests that global food insecurity went down from 2005/06 (the pre-crisis period) to 2007/08 (the food crisis period). 
That's an interesting question, but not the one I am interested here.(I guess one driver of the qualified result is that farmers' income rises if commodity prices boom, so food insecurity might fall for part of the population)
My  problem with the argument of the recent easing in US monetary policy causing the increase in commodity prices through a weakening of the dollar is that it seems to imply that the two booms were completely different. Indeed, Fed policy actually became stricter in 2005, while the first commodity boom arose in 2007. From a recent paper by Li, Li and Yu, the Fed was extremely accommodating between 02 and 05 due to the 2001 recession, but they came back to a "normal" Taylor rule starting in 2006. 
3-month Treasury yield. Red line is the actual realization. Dark Blue line is pro-active rule, that would respond more than one-to-one with inflation.  Pink one is the accommodating rule, where the response would be less than one-to-one.
And if we look at other measures of a policy easing, it is not clear that anything drastic happened before the first commodity boom, either on the expansion of the monetary base, the 10 year treasury yield or the Fed's balance sheet:
10-Year Treasury Yield, 2001-2011

Adjusted Monetary Base, 2001-2011

Fed balance sheet, 2007Q3-2008Q4
The FT came back on this today, with a column linking the rise in the Commodity Research Board index and Quantitative easing with a nice plot going from August 2010 to Jun 2011. But if you look back, here is the CRB index:
And once again, the monetary policies in 2007 and in 2010 were quite different!

So what could have driven the two booms? First, there are structural factors. For instance, increase in yields are slowing down for various cereals and have fallen short or population growth in the last 20 years
Diets have changed towards more meat, which is more resource-intensive(you need to feed your cows with cereals, e.g. corn...)
In 2000, 56% of all the calories consumed in developing countries were provided by cereals and 20% by meat, dairy and vegetable oils. By 2050, the FAO thinks, the contribution of cereals will have dropped to 46% and that of meat, dairy and fats will have risen to 29%
According to the FT, Goldman Sachs reported in 2007 that India's meat consumption rose 40% in the previous 15 years. In Foreign Affairs, Runge and Runge argue that there are 3 structural factors:

  • The previously mentioned slowing in the rate of increases in crop yields
  • A diminution of research expenditures(e.g. global agricultural aid to developing countries for research "fell by 64% between 1980 and 2003")
  • Global food supplies have begun to fall relative to demand. 

But the steady rise in demand cannot really explain two huge non-steady peaks

So what happened? There are several candidates for the recent booms. First, the production of ethanol has crept up in the past decade. World production of ethanol was multiplied by 5 in the last 10 years.Runge and Runge argue that demand for corn used for ethanol production rose from 200 millions bushels per year in 2005 to 800 million per year between 05 and 09. Steven Rattner in the NYT today provides other numbers relative to the US production of corn(and the US is the top corn exporter, and exports more than 4 times as much as the second-ranked Argentina) and its recent diversion to ethanol:
Eating up just a tenth of the corn crop as recently as 2004, ethanol was turbocharged by legislation in 2005 and 2007
One other candidate, as Joseph Stiglitz evokes, is the rise in oil prices, since this also started to speed sharply only recently. Oil is important mostly for the price of nitrogen fertilizers, even if those mostly use natural gas(yeah, I don't know much about that, but there are some interesting graphs here). And the spike fits perfectly with the rise in food prices:

 Obviously, the main problem is that the correlation is perfect, but the proof of causation is hard: basically, we're just saying that commodity prices moved together... However, I'd take this as a convincing explanation provided one can tell me why oil prices had an impact starting in 2007 and not before, and why monetary policy is to blame for the 2007 and 2008 skyrocketing rise.

Lastly, the more interesting factors, in my opinion, are weather events. First, in 2007. In Australia, a drought led to  a wheat harvest half of forecast(and comparison to forecast is what counts, rational expectations anyone?). There were huge floods in the UK and Northern Europe while an heat wave struck Southern Europe. The NYT reported on August 7h 2007 on a report by the World Meteorological Organization, the UN weather Agency
global land surface temperatures in January and April were the warmest since such data began to be recorded in 1880, at more than one degree Celsius higher than average for those months.(...)South Asia's worst monsoon flooding in recent memory has affected 30 million people in India, Bangladesh and Nepal(...)Heavy rains also hit southern China in June(...)England and Wales this year had their wettest May and June since records began in 1766(...)Germany swung from its driest April since country-wide observations started in 1901 to its wettest May on record(...)Mozambique suffered its worst floods in six years in February(...)In May, Uruguay had its worst flooding since 1959.
For the 2010 boom, the World Bank reports that the increase in food prices has been unevenly distributed, depending on weather conditions
The transmission of higher global maize prices is varied and has depended significantly on domestic harvest conditions. Countries in Sub-Saharan Africa have benefitted from excellent maize harvests, which have led to a sharp fall in prices(...)on average, maize prices were lower in 2010 in comparison to 2009 in Uganda (52%), Rwanda (37%), Kenya (33%), Malawi (30%), Ethiopia (22%), and Tanzania (19%).(...)Several Latin American countries saw the price of maize rise dramatically in the last half of 2010 as dry weather lowered yields—the largest increases were witnessed in Brazil (56%) and Argentina (40%).
Actually, the recent FAO reports do not show strong issues with weather events in the past year. This seems to be in line with some of the differences observed on the impact of speculation during the two booms. I haven't been able to find inventory data, but to check whether speculation plays a role, one needs to look at 2 things, from what I know: whether there is an inventory build-up for commodities, and whether there are discrepancies between marketed and non-marketed commodities. Paul Krugman had a great post at the beginning of the year mentioning a build-up in inventories of cotton and iron ore in China, arguing that this was a main difference with the 2007 commodity boom where there was no hint of speculation. But:
So the case for a speculative component is a lot stronger this time around. But — and this is important — the speculation is not being driven by financialization, by all those index fund investors going long. Cotton hoarding seems to be taking place at the level of individual Chinese farmers and factories, with no indication that they’re being influenced by the futures market. And iron ore hasn’t been available for futures-market speculation: the first futures markets there came into existence just a few days ago.
For at least some commodities, then, we’re seeing a real demand boom, which may be getting reinforcement from speculative hoarding, but with this speculation taking old-fashioned forms rather than involving Wall Street. 

No comments:

Post a Comment