Did you know that Australia once had a double dissolution election where the trigger was the conduct of monetary policy? It was our second double dissolution election, in 1951 (we are currently looking at our seventh), and the question at hand was the management of the Commonwealth Bank, which later in the decade had the Reserve Bank cleaved from it in an unrelated reshuffle.
Monetary policy matters. It has mattered for a long time. Tight monetary policy played a central role in the depths of the Great Depression. Even as far back as the early 1700s, monetary decisions caused the industrial sector of France to contract by 30%.
It would be quite a unique historical episode if monetary policy were ineffective. Some claim we are in that world now.
Let me explain why I think they are wrong.
First, let’s distinguish between two different types of economies: those at the zero lower bound for nominal interest rates (or near the lower bound, whatever it is, given some economies have negative rates), and those that are not.
Those not at the zero lower bound can still cut interest rates. I’m not sure why people think that the normal channels will not work properly. You often hear the claim that 'an extra 25 basis points (the amount of a usual interest rate cut) won’t have any effect'. Why would it have any less effect than usual? When I’ve looked at market and interest rate reactions to policy rate surprises lately, I haven’t seen any change in the effects of policy.* I don’t understand what evidence people are looking at when they say interest rate cuts don’t work anymore.
Another refrain I often hear is: 'if companies won’t invest at these low rates, they won’t invest at rates that are even lower'. But it is unlikely that the direct effect of interest rates on business investment has ever been strong. So, again, it does not look like the effects of monetary policy have changed.
Economies stuck at the zero lower bound can implement unconventional monetary policy, such as quantitative easing. Unconventional policy has been effective, as outlined by the Peterson Institute’s Joe Gagnon in a wonderful seven page brief on the topic. Gagnon surveyed studies that looked at the effects of quantitative easing, and had this to say:
By and large, this research has attracted little attention from the public or even the financial press. Studies overwhelmingly agree that QE does ease financial conditions and there is no reason to doubt that it supports economic growth. QE can be especially powerful during times of financial stress, but it has a significant effect in normal times with no observed diminishing returns. Rarely, if ever, have economists studying a specific question reached such a widely held consensus so quickly. But this consensus has yet to spread more broadly within the economics profession or the wider world.
A stark example of the success of unconventional policy has been Japan…so far. Below is a graph of Japanese inflation and an indication of when quantitative easing began in Japan. We have gone from outright deflation of around 1% a year, to inflation of 1% a year (the spike up and down in 2014-15 is the effect of a tax increase, ignore it). Some will counter that this is just the effect of a depreciating exchange rate, but wage inflation is up too, and so is the GDP deflator, so it’s not just the result of more expensive imported goods.
So the Japanese experience is encouraging. One problem, though, is that they want to get to 2%, and that line has stopped moving up. That should call for more stimulus, but the Bank of Japan shied away from that during its last meeting. That left some observers, like Joe Gagnon, perplexed and worried that the BOJ does not have the will to get that line to where it needs to go.
Alright, so if monetary policy still has an effect, and if monetary policy settings are stimulative, why don’t we see more growth? One issue, I think, is that policy is not as stimulatory as it might otherwise be because the 'natural' interest rate has fallen. The 'natural' interest rate is, basically, the interest rate the economy would face in normal times. The IMF wrote about a falling natural rate in its April 2014 October World Economic Outlook. If the interest rates in normal times are lower than in the past then expansionary monetary policy would also have to be associated with lower interest rates (and more unconventional policy) than in the past.
You could think about it in the following way. Economists often talk about something called the IS curve which, in a very crude way, shows the level of output you get for a given interest rate. What the IMF is saying is that the IS curve has shifted down, shown in the diagram below as a movement from r to r*. So the same level of interest rates do not give you as much output as it did before. That is not the same thing as saying monetary policy is ineffective. The effectiveness of monetary policy is given by the response of output to the interest rate, or the slope of the IS curve. The IMF, and I, think the IS curve has moved down, and there is evidence for that, but there is little evidence that the slope has changed.
In the past, people have thought monetary policy was ineffective. I wrote a piece last year discussing how some thought the high inflation of the 1970s could not be solved with monetary policy . During the Great Depression, Federal Reserve monetary policy was influenced by the flawed 'real bills doctrine', that led policymakers to conclude that monetary policy was easy, and not much could be gained by a loosening of it. Those who doubt the power of monetary policy have typically been wrong. I’m yet to be convinced why this time is different.
*For the wonks, there’s a subtle reason why cuts when interest rates are lower should give you MORE bang for your buck. One of the channels of monetary policy is the wealth channel. When asset prices go up, people feel more wealthy, and they spend more. An interest rate cut of a given size, say 25 basis points, should deliver a larger asset price response when interest rates are lower. Take, for example, the Gordon growth model for pricing a company’s stock. The divisor in that equation is the interest rate (well, cost of capital, but for argument’s sake, say it is the risk free interest rate) minus growth. The lower the interest rate, the larger proportional change you will get in the divisor for a 25 basis point cut, and the larger the stock price response. There’s a number of issues I’ve abstracted from in this example, for example, the interest rate in the model is fixed into the infinite future, but in general they shouldn’t change the conclusion.