How Interest on Reserves Affects Inflation

Bruce Bartlett sent me a new paper by Peter Ireland.

This paper uses a New Keynesian model with banks and deposits, calibrated to match the US economy, to study the macroeconomic effects of policies that pay interest on reserves. While their effects on output and inflation are small, these policies require important adjustments in the way that the monetary authority manages the supply of reserves, as liquidity effects vanish and households’ portfolio shifts increase banks’ demand for reserves when short-term interest rates rise. Money and monetary policy remain linked in the long run, however, since policy actions that change the price level must change the supply of reserves proportionately.

I found the long run neutrality of monetary policy to be quite interesting, as I’d assumed that relationship broke down with interest on reserves. The fact that IOR has little effect on inflation and growth is also interesting, but I would caution that this sort of finding needs to be interpreted with caution.

In December 2007, the Fed was trying to decide between cutting rates by 1/4 and 1/2 point. They actually cut them by 1/4 point, and the Dow promptly fell by 300 points. Because fed funds futures showed a 58% chance of a 1/4 point cut and a 42% chance of a 1/2 point cut, we can infer that the Dow would have risen about 400 points with a 1/2 point cut. (One of the few things even anti-EMH types accept is that the expected return on the Dow over any 2 hour period is roughly zero.)

Are there any models that predict that a 1/4 swing in the fed funds target would mean 700 points on the Dow? I doubt it, because most models look at these things rather mechanically. But in the real world the effect of a change in almost any monetary policy variable (fed funds rate, the base, IOR, M2, etc) depends almost entirely on how the change impacts the expected future path of policy.

I agree with Peter Ireland that a decision to pay IOR will have very little macroeconomic impact, ceteris paribus. On the other hand, ceteris is rarely paribus. It’s possible that an IOR decision might lead to changes in the expected path of monetary policy.

For instance, it might lead to fears that future QE would be less effective, as banks would have an incentive to hoard any extra cash. Or markets might have already been expecting QE (to provide liquidity during a banking crisis) and the additional step of IOR might lead them to think the QE will not immediately drive short term rates to zero.

Since the impact of a policy depends on its impact on the expected future path of policy, it is almost impossible to model these effects. They are highly contingent on the economic situation in which they occur. For instance, the December 2007 quarter point cut would normally have had little market impact, but coming on the edge of the Great Recession, its impact was greatly magnified. It made investors far more pessimistic about future Fed policy (correctly pessimistic, I might add.)

Does this mean there is no hope of ever being able to estimate the impact of policy decisions? Far from it. Louis Woodhill looked at the only three IOR decisions in the Fed’s 98 year history. In each case stocks fell very sharply around the time of the decision:

At the time of the Fed’s IOR announcement, the S&P 500 was down by a total of 12.18% from its pre-Lehman close, 15 trading days earlier. However, the day that the Fed announced IOR, the S&P 500 fell by 3.85%, and it was down by a total of 17.22% three days later.

On October 22, 2008, the Fed announced that it would increase the interest rate that it paid on reserves. The S&P 500 fell by 6.10% that day, and it was down by a total of 11.11% three days later. On November 5, 2008, the Fed announced another increase in the IOR interest rate. The S&P 500 fell by 5.27% that day, and it was down by a total of 8.60% three days later.

To play it safe it’s probably better to go with the single day returns, as the EMH suggests the effect on asset prices should be immediate. On the other hand, the concept of IOR was fairly unfamiliar to Wall Street, so arguably there might have been some delay as the program was discussed and explained. But even using the more conservative one day window, what are the odds of three drops like that occurring on the only three days in history when the IOR was raised? I’d guess no more than 1 in 10,000. Economists get published with results no more unlikely than 1 in 20, and yet I am so skeptical of statistical significance that even 1 in 10,000 seems merely suggestive to me. I think IOR might have had a significant contractionary impact, but I am not certain.

Whenever the model says one thing and the markets say another, I always go with the markets. The markets seemed to think the IOR program was a big mistake, and the QE2 program was an important step in the right direction. That’s all we know right now, and probably all we’ll ever know.

PS. After the third and final increase in IOR, the S&P500 actually fell 10% in just two days. Thus the market declined over 38% in 10 trading days–October 6, 7, 8, 9, October 22, 23, 24, 27, and November 5, 6. Think about what it means for US equities to lose 38% of their total value in 10 trading days. Coincidence? Maybe, but a pretty unlikely one. Using 2 day windows the total drop was about 24%. Even the three single day drops add up to more than a 15% decline.

And then there is Sweden, with its negative IOR and record RGDP growth. Hmmm . . .

Memo to Bernanke: Cut the IOR to 0.15%. It will give banks a bit less incentive to just sit on all the QE you’re sending their way. But it will still be high enough to prevent the MMMFs from going belly up. And you guys at the BOG can do it on your own–no pesky regional bank presidents to deal with. Even Ron Paul will approve—less subsidy to fat cat bankers. Git er done.