Why the Fed can't drive

One of Milton Friedman’s favorite analogies about monetary policy was driving a car. He compared money creation by the Fed to pushing on the gas pedal. On flat ground giving the engine more gas makes the car speed up, so giving the economy more gas should cause it to accelerate as well. But giving the engine more gas by pushing on the pedal may allow the car to slow down when climbing a hill if the engine doesn’t get enough gas to overcome the gravity involved in climbing the hill.

Friedman’s point: interest rates tell us nothing about whether Fed policy is too tight or too loose. Only the speed of economic growth can tell us that. Low interest rates may be too high if the economy is climbing a steep hill, like a recession. On the other hand, high rates may be too low if the economy is speeding up, as it does near the end of an expansion. The grandchildren of Friedman use that analogy to argue for nominal GDP targeting by the Fed instead of price targets. The problem with the Fed’s driving strategy is that it never knows if it is climbing or descending a hill or how fast the economy is growing at the time it makes its policy decisions because of long lag times from policy decision to its impact.
Remember that mainstream economists gave up on fiscal policy as the gas pedal for the economy decades ago because of long lag times. There was the lag in recognition of the problem because the current data is always from the last quarter and most economists want to see two quarters of data before deciding there is a problem. Then Congress created another lag while it debated. Finally, the federal government took a long time to implement new legislation, thereby creating even greater lag.

Market monetarists, such as Scott Sumner, claim there are no lags to monetary policy. Prices respond instantly to changes in Fed policy. That may be true with overnight rates and bond or stock market prices, but it’s not true of the rest of the world of finance according to a recent paper from the Bank for International Settlements, “Market volatility, monetary policy and the term premium.”

The goal of the authors was to investigate “the role of (option-implied) stock and bond market volatilities and monetary policy in the determination of the US 10-year term premium.”

They discovered that Fed policy effectiveness before and after the Great Recession changed: 

... that an unexpected loosening of monetary policy – through a cut in the federal funds rate in the pre-crisis sample or an increase in bond purchases post-Lehman – typically leads to a decline in both expected stock and bond market volatilities and the term premium. However, while conventional monetary policy boosts economic activity in the precrisis period, bond purchases are found to have no statistically significant real effects postcrisis.

Those are important findings, but for me the best part was their analysis of “impulse response functions” beginning on page seven. These functions tell us how a one time “shock,” or a change in monetary policy, will affect different variables over time. Time is missing from mainstream economic analyses. The graphs in the paper show the level of change on the left, vertical axis with the time until that change happens on the horizontal axis. Regardless of the variable, the graphs show considerable lag between the “shock” and the maximum impact on most variables. In many cases the lag is 12 quarters, or three years.

Take the case of the impact of an interest rate reduction on industrial production (IPUSYY) shown in the first graph on page eight. Production worsens in the first year after the shock and doesn’t turn positive until 14 quarters, three and a half years, later. It reaches peak effectiveness at five years.

Also note the two blue lines, one above and one below the red line that represents the response variable. The blue lines are the “confidence intervals.” They show the range of results that might happen instead of the result indicated by the red line. There is a lot of wiggle room there.

Mainstream economists tell us that the Fed can control the economy through its monetary policy, but the reality is that the Fed doesn’t know the full impact of those policies for up to three years after the policy decisions are made. That would be comparable to driving a car at 80 mile per hour and having the lag between pushing on the gas pedal and the engine responding of say an hour. Or maybe the driver stomped on the brakes but the car didn’t respond for two hours.

Friedman tried to warn economists about using monetary policy to fine tune the engine of the economy because of the long and variable lags involved. Try giving the keys to a Corvette to a sixteen year old boy and telling him not to drive it. Neither can you give power over the money supply to any human and expect any kind of restraint at all. Hubris will always win.

It could be said that the Fed proposes, but the real economy disposes. Austrian economists have often said that the Fed may ignite an expansion but the real economy ends it. Fed policy always has repercussions. The forces set in motion by monetary expansion appear to work for the growth of the economy in the short run, but in the medium term, they morph into monsters that destroy. For investors, that means we need to pay attention to the real economy and not just what the Fed is doing.

Now that I think about it, the car analogy with such long and variable lags does a good job of explaining the economy for the past century since the creation of the Fed.