Why do interest rates matter?

Interest rates are important, but not in the way that most people assume. To see why, it might be helpful to start with an analogy. Why does inflation matter?

If you ask the average shopper, they’ll tell you that inflation is obviously bad because the public has to pay more for the stuff they buy. But if you pick up any economics textbook, nowhere do they mention this factor as a “cost of inflation”. That’s because when people spend more on goods, other people earn more selling those goods. By themselves, higher prices are a zero sum game.

That doesn’t mean high inflation is not a problem—I believe it’s a very serious problem. But it’s not a problem for the reason that most people assume its a problem.

The same is true of interest rates. Most people (wrongly) think that low interest rates are good for the economy. Stock traders know better, and indeed stocks fell sharply on Friday, even as longer-term interest rates plunged.  As with inflation, interest rates can be important—but not for the reasons that you might assume.

I suspect that most people make the same mistake that shoppers make with inflation, reasoning from personal experience. Thus they might imagine how lower rates might make it easier to buy a house or car. But interest rates are a zero sum game. When one person pays more interest, another person receives more interest. 

I suppose you could argue that a change in interest rates might cause some sort of “redistribution”, although it’s hard to say exactly how.  When rates are low, the media complains that big banks benefit and retired folks with savings accounts are hurt.  When rates are high, the media complains that credit card holders suffer and big banks benefit.  But even if falling interest rates were to redistribute money to lower income people, that doesn’t mean it would help the economy.  Low rates also tend to reduce velocity, as people have more incentive to hoard cash.  As always, it depends on why interest rates change.  

Consider the following two claims:

1. It’s nonsense to speak of how interest rates affect the economy.  It would be like talking about how a change in oil prices affects the oil market.

2. It makes a lot of sense to speak about how monetary policy affects the economy.

If both are true, then obviously interest rates cannot be monetary policy.  So what is monetary policy?

3. Monetary policy is a set of actions taken by central banks that impact the supply and demand for base money, often with the goal of impacting macro aggregates such as prices, employment and/or NGDP.

So then what do interest rates have to do with monetary policy?

1. Prior to 2008, the Fed targeted the fed funds rate by instructing its open market desk to buy and sell Treasury securities.  This policy directly impacted the supply of base money, and indirectly affected interest rates.

2.  After 2008, the Fed continued to change the supply of base money (via QE), but also used the tool of interest on bank reserves to impact the demand for base money.

3. The Fed also impacts the demand for base money by affecting the expected growth rate of NGDP.  Faster NGDP growth trends to reduce the real demand for base money.

Note:  These three effects do not all work in the same direction!

This is the key point that so many people miss, even many economists overlook this problem.  (If you are familiar with the recent debate, the first two mechanisms are emphasized by Keynesians, and the third has implications related to NeoFisherianism.)

Because monetary policy affects interest rates in a complex and often contradictory fashion, it’s not possible to look at changing interest rates and make any inferences about the stance of monetary policy.  I suspect that the reason why America has never had a mini-recession (defined as unemployment rising 1% to 2%, and then falling) is because our central bank has often been confused about the relationship between interest rates and monetary policy.  When the economy tipped into a small recession, the Fed initially makes things worse by tightening monetary policy—creating a bigger recession.  They wrongly assume they are not tightening policy because they reduce their target interest rates.  But lower rates are not easier money—especially if the natural rate is falling even faster.

Suppose that central banks are gradually becoming aware of this mistake.  Then, if my hypothesis is correct, the US should begin experiencing minirecessions.  These will occur instead of normal recessions because the Fed will no longer be fooled by an obsession with interest rates.  The Fed will begin to focus more on a wide variety of financial market indicators, and also be more willing to adopt a “whatever it takes” approach to stabilizing market expectations of future aggregate demand (NGDP.)

It would take many decades of improved performance to be confident that it was not merely luck, so I won’t be around to see if my prediction comes true.  Indeed I don’t even know if the Fed has begun to see past the fallacy that interest rates are monetary policy, a necessary precondition for any improved performance.  It’s also possible that they could reduce the severity of the business cycle through a different policy reform, say the adoption of level targeting.

I’d guess that there is a greater than 50% chance that we are entering our first minirecession.  If so, I think it is likely that this period does not get labeled a “recession” by the NBER.  In that case, it would be our first ever soft landing.  And closely related to these points, it would be our first violation of “Sahm’s Rule”.

Alternatively, we might have a full-blown recession.  Either way, the next 12 months will likely be far more interesting than the past 12 months.  Here are my (highly unscientific) guesstimates:

1. Boom:  Unemployment peaks at 4.3%  — 5% chance

2. Minirecession: Unemployment peaks between 4.4% and 5.3% — 65% chance

3. Recession: Unemployment rises above 5.4% — 30% chance

I’d be interested in what readers expect—feel free to add to the comment section.  As an aside, these are my definitions; the NBER uses different criteria for defining recessions.  I define soft landing as at least three years of continued expansion without triggering high inflation, even after unemployment has fallen near cyclical lows.  We’ve never done that.  That would be a far more impressive national goal than repeating the 1969 moon landing in the 2030s.

PS.  This is kind of a cool graph: