Thursday, 22 October 2009

The Roger Farmer paradox

Roger Farmer likes unorthodox monetary policy.

At the turn of the year he proposed that central banks should buy and sell equities, targeting a stock price index as a method of controlling asset prices. My own version of this proposal was slightly different.

Now he's suggesting that they use quantitative easing as a monetary tool, independently of interest rates. The idea is that even once central banks have started to raise interest rates to control inflation, they should separately adjust their balance sheet, changing the composition of the stock of savings in the economy, to combat unemployment. If QE is a useful tool now to help raise economic output, why shouldn't it be useful later, when inflation is taking off?

Now admittedly I'm not a trained monetary economist, but I have a bad feeling about this idea. Aren't interest rate targets and QE both different manifestations of the central bank's ability to control the money supply?

Crudely speaking: if money is tight, interest rates are high. If money is loose, interest rates are low [Update: see comment from Rob and my response below]. If interest rates are at zero, money can be loosened through QE instead. The reason the central bank can influence interest rates throughout the economy is that it can create or absorb unlimited money until the market rate rises or falls to meet its target.

So let's say we are back in the realm of normal monetary policy, with interest rates at (say) 3%. Does the composition of the central bank's balance sheet matter? If the Fed holds lower-quality commercial loans instead of Treasuries, will that make a difference to the economy? I can't see that it will.

Let's say they sell $1 trillion of government bonds and buy $1 trillion of commercial loans. This affects the overall return on savings in the private economy, as the average return on the purchased commercial loans was higher than that on the government bonds which have replaced them. This lowered return on savings will slightly increase the demand for investment.

But a lower return on savings is exactly equivalent to a lower interest rate - which will lead to higher inflation than the bank's target. So to achieve its inflation target it will have to increase interest rates by reducing the money supply, and this will sterilise the change in balance sheet composition.

Farmer says:
A high fed funds rate is bad for the employment outlook because it depresses the values of corporate bonds and public and private equity. Investors move out of real assets that create jobs and into barren federal securities.
But a high fed funds rate depresses the value of Treasury bonds too! Investors don't move into "barren federal securities" because of collapsing corporate bond prices. Instead, companies issue fewer bonds because the coupons they'd have to pay makes it more expensive to borrow for investment, and some projects are no longer worth doing. Governments are less likely to be able to reduce deficits in response to expensive financing, so they will probably keep issuing bonds. And to be able to increase interest rates, the central bank has reduced the money supply; which means there simply is less money to invest, and a higher proportion of this smaller pool goes to financing public deficits.

QE, in other words, is just another tool for adjusting interest rates. It should be effective only when the normal way of changing market interest rates is ineffective, because the central bank's policy rate is zero. And even then, monetary transmission mechanisms seem to make it less effective than standard theory would predict.

It happens that I agree with Roger Farmer on several things: his approval of Tom Sargent's quote "it takes a model to beat a model", his emphasis on psychology in macroeconomics (though I haven't read his books) and that Akerlof and Shiller's Animal Spirits is a bad book. But I think he's got this one wrong.

The end of the Akerlof and Shiller review highlights his focus on the 1970s stagflation episode and the breakdown of the Phillips curve tradeoff between unemployment and inflation.

No doubt this explains his appetite to find independent tools to control inflation and unemployment; but MV = PY is a single equation, and we'll need more finely distinguished mechanisms than pure monetary tools to independently manage P and Y. There is the potential for central banks to influence investment behaviour, but I don't think buying corporate bonds - unless there is a clear market failure, which will probably not be the case with policy rates at 3% - will be an effective one.

The abstract of Farmer's new book, Expectation, Employment and Prices, says:
Central bankers throughout the world are talking now about developing a second instrument of monetary policy in addition to controlling the interest rate. This book directly addresses this issue and offers new creative monetary policy proposals and suggestions for the design of new financial institutions for the 21st century.
I'll be interested to read whether there are any more subtle - or indeed logically possible - proposals than this QE idea.


Rob said...

Crudely speaking: if money is tight, interest rates are high. If money is loose, interest rates are low.

I'm even less of a monetary economist than you are (in fact, I'm not an economist at all), but I've read the argument that this does not always hold true. Scott Sumner: Interest rates are a very misleading indicator of monetary policy. Both in the early 1930s and late 2008, falling rates disguised a tight money policy. The rates were actually falling for two reasons. Expectation of recession led to less borrowing and thus lower real interest rates. And inflation expectations also fell sharply.

As I say, I'm very much a layperson and it often seems to be the case that just as I feel I understand something relating to economics, I read another opinion arguing the opposite!

Leigh Caldwell said...

Hi Rob

You're absolutely right - I'm a regular reader of Scott's and I was thinking of him while writing this.

Therefore, my "crudely speaking" prefix should probably have been even more qualified. What I should have said is "a larger money supply correlates with lower interest rates" - other things being equal.

When Scott refers to tight and loose money, I believe he means the size of the money supply relative to what would ensure steady growth in nominal GDP. Other factors apart from interest rates can influence the size of money supply that's appropriate for the economy. That is, other things are not always equal. In the example you quote, expected recession means that interest rates fell even without the money supply increasing; and thus the correct action for the economy would have been to increase the money supply and reduce interest rates even more.

This is why the economy requires different interest rates at different times; and when the economy requires an interest rate less than zero, QE or similar policies are the only available response.

Roger E. A. Farmer said...

Thanks for your thoughtful posts. Historically central banks have targeted one price (an overnight interest rate) to hit two targets -- inflation and unemployment. According to received wisdom, the real economy moves around independently of monetary policy, and there is nothing the Central Bank can do about long-term unemployment. I disagree with received wisdom.

Long bonds and stocks are both claims to future income streams. The Fed and the Bank of England have each succesfully raised long bond prices through quantitative easing. Central Bank interventions in these markets have, in my view, also been responsible for raising the value of the stock market. Long bond prices can, and do, move independently of the overnight rate. The stock market can and does move independently of short rates. There is no reason that a Central Bank cannot independently influence overnight rates (to target inflation) and prices of stocks and long bonds (to target unemployment). The key is understanding that the natural rate hypothesis is false and that the level of real economic activity is influenced, in the long run, by policy.