Daniel Murray (DM): From the global financial crisis to Europe’s debt crisis, the Covid pandemic, the invasion of Ukraine, the spike in oil prices and soaring inflation, the past few years have put central banks to the test. During this time, they had to exhibit agility while showing that they are stable, reliable and trustworthy. How difficult is it for central banks to achieve this balancing act when setting monetary policy?
Don Rissmiller (DR): Let’s start with first principles. Monetary policy should be a symmetric policy that moves growth around in time. It should not seek to change the trend growth rate in an economy, which is usually influenced by fundamental factors such as population growth, productivity, education or infrastructure. Instead, monetary policy should aim to “push and pull” the economy in order to stay as close to the trend growth rate as possible – avoiding depressions and hyperinflation.
DM: So that requires some agility?
DR: Yes, but I believe that is also important to frame the focus of central banks. I worry that monetary authorities have been asked to do too much. Take the US Fed, for example: It has a dual mandate of price stability and maximum employment. It is difficult enough for the Fed to tackle inflation and unemployment – but it has also been tasked with managing financial stability and international conditions, among other factors. I think that concentrating on its core mandate would be the first step towards becoming agile enough to move growth back and forth in time around a trend.
DM: And Stefan, as a former “insider” with real-time experience of this topic at a very senior level, how do you think central banks can achieve this balancing act going forward?
Stefan Gerlach (SG): What really matters for the economy is how long bond yields move. If a central bank is very agile – moving short-term interest rates up and down as data comes in – then financial markets are going to say: "It has now raised rates but it is probably going to cut them again in six months or in a year's time because it tends to move rates up and down all the time." That means that long-term interest rates and long bond yields tend to become insulated from the monetary policy decisions of central banks. Viewed from that perspective, central banks don't want to be too agile.
DM: So it is preferable for them to move quite slowly?
SG: Yes, I believe that one reason why central banks tend to make gradual changes to monetary policy is that it gives them more “bite” with regard to long bond yields. They might typically raise interest rates several times in a row in the same direction and then cut them several times in a row – they don’t hike rates by 200 basis points in a single move. In short: Being too agile can create problems when you set monetary policy.
DM: Let’s now turn to the risks that central banks face at present, including the risk to their independence. What do you see as the main pressures they face in a world marked by enormous uncertainty and change?
SG: It is very hard to identify a “pecking order” of risks – there are simply too many of them. Of course, there are the obvious risks that currently need to be addressed – such as cyber risks, environmental risks and fiscal risk, as well as the risk to central bank independence – but what we have noticed over the last few years is that certain risks that were no longer on our radar have re-emerged.
DM: Can you give us some examples?
SG: If we think about Covid, it was the first global pandemic since the Spanish flu outbreak of 1918. Then in 2008, we saw the first major global economic downturn since the Great Depression of the early 1930s. And with the invasion of Ukraine, we witnessed the first major armed conflict in Europe since 1945. No one thought that these things could happen. This shows why central banks need to have systems in place that increase their resilience while also giving them flexibility – they simply don't know where or when the next crisis will come.
DM: Don, what can you tell us about the risk picture in the US?
DR: I would say that a concern in the US right now is the risk of an AI bubble or tech bubble – and there may be limits to what central banks can do about it. The San Francisco Fed did some research on the housing bubble ex post. Looking back, it said: "What would the Fed have needed to do to stop the housing bubble in 2005, 2006 and 2007?" They found that it would have been possible to prevent it from forming – but the Fed would have had to start raising rates in around 2002 and they would have needed to go up by around 800 basis points, which was never going to happen.
DM: So what does that mean for a potential AI bubble?
DR: I would say that if the bubble risk exists, it is going to be very hard to manage it through monetary policy. It is possible that a macroprudential policy, maybe some sort of regulatory policy, could help – but even then, I am not sure you want to take that approach. We need to be very careful because almost any innovation of substance in the past probably had a bubble. Let’s not forget that at one time, there were too many railroads, too many car companies, too many airlines and too many internet firms. And I am not sure what the central bank can do but this matter is certainly a concern that will get a lot of attention going forward.
DM: And continuing on the topic of AI, what are your thoughts about its potential long-term impact both on markets and on the economy more generally?
DR: I think there are a lot of concerns about how AI will affect employment – but that is the case with any productivity improvement. Back in the 1800s, people worried about how industrial developments would affect workers, and the Luddite movement at the time opposed technology. But if the technology is real and substantial, it should create new jobs. The big productivity booms that we have seen over the last few generations, such as the boom in the 1950s to 1960s, were not bad for employment. And think about the internet revolution of the 1990s – it didn’t harm overall employment. Of course, some jobs might be lost. For example, it was tough being a whale oil salesman when the shift to electrical lighting happened. And it was tough being a repairer of carriage wheels when the automobile first came into use – but new jobs tend to crop up anyway.
DM: And what about the immediate effects of AI on employment?
DR: Right now, the impact of AI is being felt most clearly in a few industries – such as the technology sector, the financial sector and business services. Interestingly, the effects seem to be greater for younger workers rather than their older colleagues and this very odd dispersion is creating a lot of angst, which needs to be addressed.
DM: Stefan, do you think that AI will have a lasting impact on productivity or do you think it's more of a zero sum game?
SG: If we look at productivity growth from the 19th century onwards, we see that while there were variations from decade to decade, average productivity growth remained at around the 2% or 2.5% mark. That was the case despite all the major innovations of the last 150 years – such as the arrival of electricity, radio technology, television, aeroplanes and IT. From that perspective, it is hard to imagine that AI will have a large and permanent impact on productivity growth. It seems more likely that we might see a decade of relatively high growth, which will then slow down somewhat.
DM: Why is that?
SG: There are some structural drivers that suggest growth might be lower in the future. For example, the birth of new ideas is correlated to the population growth rate – so if the population growth declines, you have fewer people and you are also likely to have fewer good ideas. Another key driver is education: If you go back 100 years or so, few people were very well educated, but that has changed over the last century – with a big impact of productivity. However, the scope for even further education-driven productivity gains is now limited. Perhaps AI will offset that to some extent, but overall, I am sceptical about the idea that AI will lead to a sustained increase in productivity growth over time. History suggests otherwise.
DM: To what extent do you think the work of central banks is properly understood by the public in general?
SG: I see parallels between the work of central bankers and the work of meteorologists who give the weather forecast at the end of the news report on television or radio. Meteorologists talk about the current situation – for example, they might say: “It has been cold over the last weeks but we hope it will brighten up in the next few days.” Central banks essentially do the same thing. They might say: “Inflation has been a little bit low, but we hope that with interest rate cuts, it will pick up next quarter or next year” for example. So there are some similarities in their approach and how it is viewed by the public.
DM: Can you expand on that?
SG: Metrologists develop large mathematical models of the atmosphere to understand how it changes over time and they try to figure out wind speeds, temperatures and other key criteria, while central bank economists spend a lot of time and resources developing models of the economy and how it functions. In both cases, a lot of work goes on behind the scenes before they present their respective forecasts. I don’t think these underlying processes are always properly understood by the public.
DM: Let’s move on the question of how central banks communicate. In the years since the global financial crisis, this topic has attracted a lot of attention – also in the context of discussions around central bank independence. In your view, how important is their approach to communications in today’s world?
SG: It is important from the perspective of accountability and independence. I believe that if you are independent, you need to be accountable. And to be accountable, you need to tell people what you are doing. In recent years, much of the debate about central bank communications has focused on forward guidance, which was introduced because policy rates hit zero and it looked as if central banks have become impotent. They couldn't lower long-term bond yields any further. And then they took the view by making promises or assertions about future policy moves, they might still be able to push down low bond yields. That probably worked to some extent but the issue today is central banks no longer have zero interest rates. So I think we will see much less forward guidance in the years to come.
DM: Don, what is your view?
DR: If we look at the Fed, I would say that in the past, observers spent a lot of time trying to figure out what it was doing after the fact. The Fed didn’t always issue FOMC statements in the past – and when they were published, there was usually little explanation of its monetary policy decisions. So the way the Fed communicates has certainly evolved over the years. Today, I would be more concerned about the interaction between fiscal policy and monetary policy and problems relating to fiscal dominance than the central bank’s approach to communication. I think as long as the Fed maintains its independence, the flow of information will continue.
DM: Thank you both for sharing your views.