It took me a while to convince myself to write about “neo-fisherism” as a solution to low trend inflation.
The motivation behind neo-fisherism is relatively intuitive – we have observed nominal interest rates and inflation move together, and require an explanation that supports that stylised fact. Furthermore, the idea behind neo-fisherism is that there is the Fisher effect describes how inflation expectations must move when nominal interest rate goes changes – allowing us to keep our assumption of “neutrality” in the long run with a fixed real interest rate, rather than relying on changes in the neutral real rate of interest.
However, at face value this completely contradicts the conventional monetary policy set up, where we were taught that inflation rate and nominal interest rate follow the Taylor principle, so that when you cut the nominal interest rate, then inflation (and inflation expectations) goes up. Given this way of thinking, and the empirical regularity that inflation and nominal interest rates DO move together, does this overturn conventional monetary economics?
A well -respected economist , Professor Stephen Williamson, an advocate of neo-fisherian school, wrote on his blog post for St Louis Fed describing the idea of the one-to -one movement between interest rate and inflation , where the Fisher relationship is given as:
i = r + π
Here i is the nominal interest rate, r is the real interest rate, and π is the rate of inflation. If r could not change it is quite natural to look at this (approximate) identity and state that nominal interest rates and inflation must move together!
Although the argument behind the positive relationship between interest rate and inflation is not well-elaborated in the post, Prof. Williamson refers to the data as an existing evidence of this relationship.
He also explains why unconventional monetary policy tools are inefficient. Namely, he says that
- Negative nominal rates make inflation lower;
- Quantitative easing does not give a clear perspective whether it does anything significant referring this to Japan case;
- Forward guidance does not work so well, as promises not met.
However, the evidence of the relationship not backed up by theoretical argument may lead us to a trap.
Think of it from Phillips curve perspective, back in the day when the data were showing that low unemployment leads to higher inflation, economists started to target lower inflation viewing this as a trade-off. However, as soon as it happened, Phillips curve became useless, as the data started to move in opposite direction. This is an example of using a framework without a theoretical back up, solely relying on historical data as an evidence.
What is driving the positive relations between interest rate and inflation in the data?
One way of thinking about this relationship could be a third factor such as a demand shock which is missing from an identity like the fisher effect. It would be good to think if there is some primitive factor which can be affecting both inflation and interest rate that are not captured in that relationship.
Milton Friedman was a very smart economist proposing to think of economy from thermostat effect perspective. The idea behind his thermostat effect was that (I am paraphrasing the example from driving engine to the air-conditioner one) when the temperature outside changes, the air-conditoner automatically adjusts the temperature in the room to its pre-set level. At the end the temperature in the room keeps constant, even when outside temperature changes.
If we are measuring temperature using this thermostat, then it seems like the increases and decreases in the effort of the air-conditioner are doing nothing as the temperature inside stays constant – when actually there is a second (unobserved) factor that influences the temperature, the outside temperature, and the air-conditioner is functioning to cancel this out.
Now taking this example to the inflation – interest rate relationship. When there is a negative demand shock moving down a central bank responds to this by cutting interest rates. If they are able to do enough to prevent inflation changing, it will look like they are changing interest rates and it is doing nothing to inflation.
In reality central banks don’t have perfect foresight, so when there is a demand shock they respond in part – but generally inflation will also fall. As a result, their reaction function determines the negative relationship between inflation and interest rates that we observe! It isn’t that lower interest rates cause lower inflation, it is that a third factor causes both!
Neo-fisherism uses an accounting relationship and an empirical relationship to reach a conclusion – which are good things to use. But behaviour is missing – there is no attempt to truly answer why higher nominal interest rates increase inflation expectations (a point also made here). When we recognise that there is a series of “demand/expenditure” shocks in the background we are able to see how looking at the data and the accounting relationship alone may mislead us – and lead us to the dangerous conclusion that higher interest rates are the correct solution to a short-fall in demand.
It appears to be a result that is contingent on rational expectations – at a stretch such a model may work by driving an immediate decline in the price level following the increase in interest rates, with higher expected inflation in the future as “catch-up”. Adding these extra details starts to indicate that the model does not necessarily describing the experience of countries at the zero lower bound, and it is only in going through all these details that we can see where our model results may be fragile.
However, it does raise an important point – if there is a reason why inflation expectations decline then the nominal interest rate in the future will end up lower. As a result, trying to ensure that we credibly anchor inflation expectations is an important part of monetary policy, and potentially a short-coming with taking forward guidance to the extreme.