Podcast: The post-crisis world

financial-crisis

Central banks have taken on major new powers and reformed their thinking since 2008, but there is still more work needed to overhaul modelling approaches, says Yale University’s Andrew Metrick.

Metrick, the Janet Yellen professor of finance, says there is a new recognition of the importance of finance when modelling the economy, new sources of data on money markets, and a new appreciation of risks among many central banks.

“There’s a sense in that there are linkages across sectors and across countries, risks that need to be paid attention to by central banks, that they weren’t looking at before,” he says, in a the first episode of a new podcast series on post-crisis central banking.

However, Metrick says central banks have only partially gotten to grips with the new modelling approaches, warning the financial sector is still a “bit of an add-on” to many models.

On one end of the modelling spectrum he sees dynamic stochastic general equilibrium (DSGE) models, as important “workhorse” models for thinking through the real economy. On the other, he sees more complex agent-based models as useful for understanding dynamics and linkages.

But in between, Metrick suggests, there remains a gap. “I would say that the thing that is in between has not yet been created.”

Part of the “art of central banking” in the coming years will be to understand the need to balance different modelling approaches, he says, while central bankers and academics develop new modelling approaches to fill the gap.

To hear the full interview, listen in the player above, or download. CB On Air is also available via iTunes or podcast apps. All episodes in the series are available here.

Transcript

Hello, I’m Dan Hinge, news editor at Central Banking, and this is CB On Air.

We’re taking a slight deviation from our other podcast series to talk about post-crisis central banking. The nature of a central banker’s job has changed dramatically in the past 10 years, with new roles and responsibilities emerging, and new tools for tackling the myriad of issues that central banks are now expected to stay on top of.

To navigate this complex terrain, we fortunately have an expert guide, Yale University’s Andrew Metrick. Andrew is the first Janet Yellen professor of finance and management at Yale, and one of the minds behind a new course the university is offering on systemic risk.

This week we’re going to be taking a broad view of the post-crisis world, and asking what a modern central banker needs to know.

Andrew, welcome.

Andrew Metrick: Hello.

DH: To kick off, let’s keep things broad. What are some of the key responsibilities that central banks took on post-2008?

AM: Well, prior to 2008, central banks largely thought of themselves as being in the monetary policy business, and that’s not a big surprise. Especially in developed countries, there was a thought that financial crises of the type we used to see all the time in the past, were not something that happened to developed countries. Banking crises, that is. Of course we all got a rude awakening starting in 2009 and, subsequent to that, in many major economies, central banks were given new responsibilities for financial stability. Or some old responsibilities that they had for financial stability, they started to take a whole lot more seriously.

Where we saw that organisationally is now you’ll see a department or a division of financial stability in major central banks that you didn’t see before in many many places. Really only the Asian central banks, who got their wake-up call 20 years ago, had put such things into place with any regularity. But now around the world, and in the United States, for example, the Federal Reserve Board created a division of financial stability, which now has I think about 30 PhD economists in it, and other skilled staff members. And that division didn’t exist at all prior to the crisis. You see similar things throughout Europe and in many other places.

So the big picture is that financial stability, writ large, has now become a very central role, something you’ll hear discussed in many many meeting, and at the top of the mind of global central bankers.

DH: And how did central banks’ thinking change after 2008? I think it’s fair to say we’ve seen new models, new data, new perceptions about risk…

AM: Yes I would say all of those things. On the modelling side it is really quite remarkable how many of the macro models that had been used at the highest level did not even have the financial intermediation sector in there, with important real linkages. A lot of the world had gone over to more real models, and in these models the financial intermediation sector was often ignored.

What we’ve seen happen is both on the academic side and transferring over to central banks, a movement towards how we should actually put banks and other financial intermediaries directly into the models. Right now I’d say it is still a little bit of an add-on. We have a situation where we have taken a lot of the models we were using before, and just grafted on top of them, some sort of financial intermediation sector. I would say that is an improvement over where we were before, but we still really need a new paradigm. There’s a lot of smart people working on that, but it’s not yet done.

On the data side, I would say the biggest thing is the recognition that we knew very little about many, many important money markets that were driving a lot of the short-term wholesale finance in the world. They just weren’t on the radar screen of central banks. So central banks in the past were spending a lot of time, understandably, watching short-term interest rates that they targeted, and using the instruments they have had in their arsenal for decades, to be able to target them.

But there’s a whole other world out there of asset-backed commercial paper, of repo, of securitisation, that had been creating a lot of short-term money market instruments, which we weren’t paying as much attention to, and we just did not have great data on. Indeed, in the United States, some of the largest short-term money markets, while tracked in federal data, did not have the level of granularity that would enable us to see what was going on. We had statistical discrepancies in some cases between the demand and the supply side as great as one trillion dollars. You can’t really have that happen.

And so there have been efforts under way to get all of the data updated, particularly for looking at the types of markets that central banks need to be paying attention to for financial stability.

In terms of the risks side, our thinking is in many ways ahead of our data, or rather the data is being gathered, but the thinking requires historical data that we will probably never have. In many many places we see models that are heatmap kind of models of risk. Those heatmap models of risk will have data from a variety of sectors, and an attempt to see how those things interact with each other is mostly a modelling exercise, since we don’t have very long time series to do it all empirically. But there’s a sense in that there are linkages across sectors and across countries, risks that need to be paid attention to by central banks, that they weren’t looking at before.

DH: You touched tantalisingly there on the models, which is something I’ve been digging into in the other podcast series. Can I push you to expand a tiny bit? There has been this debate over whether we try and fix DSGE models, or do we move to something completely different, maybe agent-based models. How do you see that panning out? Do we just need a greater plurality of models?

AM: I would say that, look, we’re never going to get rid of the [DSGE] models, the DSGE models are going to be good workhorse ways to think through comparative statics on the real side, and they probably will always be part of the toolkit.

I’m not sure, however, whether or not those types of models are ever going to be able to properly incorporate what’s going on on the financial side, in part because it’s not what they’re optimised for. It will always be a bit of a – adding epicycles, adding things on to those models that they’re not really built for.

At the other extreme, perhaps, I think agent-based models are very useful for giving people ideas for what might happen. They’re very good ‘what might happen’ types of models, in the sense that the model sometimes teaches you about linkages or dynamics that might not have occurred to you outside the model. They’re not, however, I think, ‘what will happen’ kinds of models. They’re good for expanding the way we look at possibilities, but I don’t think they are designed to tell us, or predict for us, what exactly is going to happen. That’s just the nature of those beasts – they are on purpose, very, very complex, with complex dynamics, and you have to make a whole lot of assumptions about exactly what agents you’re going to model, what their linkages will be.

So, we have assumption in all models. In some cases the assumptions are simple so the models can’t get us very much, don’t necessarily tell us the full range of linkages that are out there. Then when you have very complex models on the other side, you can end up with the opposite problem. So I think what we’re going to end up with, as you suggest, is a plurality of models. The agent-based models will have a place, since they help us to see things we might not have seen otherwise. The models in fact can teach us things. DSGE models will be there because the DSGE models really help us on the real side, and we have a lot of experience with interpreting them.

But we’re going to need something that’s in between, and I would say that the thing that is in between has not yet been created. We see a lot of efforts that are going on, and here the incentives to get these things right is very high, on the academic side and on the central banking side – people trying to put together models that have financial intermediation more embedded in them, but that are more traditional representative-agent, or perhaps specifically heterogeneous agent models. 

So I think where we will go is some combination of these things, but perhaps less confidence, even less confidence than we had going into the crisis, that one more or two models is going to give us all our answers. The art of central banking, I think, will stay very alive going forward.

DH: Interesting, thanks. So, in this brave new world, what kind of skills would a central banker need that they didn’t need – or at least they didn’t know they needed – before the 2008 crisis?

AM: Well, here’s where things get interesting, I would say, because prior to the crisis, there was a tremendous need for traditional macroeconomists, and monetary economists, and that’s what central banks needed to gear up on, needed to tool up on.

Now there is much more of a demand for financial economists to interact with the macro and monetary economists, and to have more of a sense, as we talked about on the modelling side, of the linkages between the real side and the financial side. But secondarily, have an idea for how the financial system, independent of just what’s going on in short-term interest rates, how the financial system, through the creation and the management of a wide variety of money-like assets, can lead to macroeconomic effects and overall destabilisation of the entire economic system.

So specifically for skills, there is more of a need to connect the understanding of monetary theory and policy with the payment system, and central bank operations. So those two things, they would often talk to each other not all that much, they were already in central banks, they needed to talk to each other more. Similarly, on the supervision and regulation side, many central banks have those functions, but they were oft-siloed, and now we’re going to need much more communications between what the examiners on the ground are seeing, and what the monetary policy-makers are doing.

Then, finally, there’s a whole brave new world of thinking about all these financial stability measures and macro-prudential tools. A lot of things that just didn’t exist before, things like countercyclical buffers, which are going to really require co-ordination and communication between the supervisory, regulatory sides, and the monetary policy sides. 

So, I think it’s gotten more than twice as interesting as it was before. If you had to know x to be really effective before, I think you have got to know more than 2x now, and that we’re seeing a lot of efforts by the central banks internally and through their hiring, to build up that capability. But given how much they know about the first x, we’ve still got a long way to go before we learn everything we need to know about the newer parts.

DH: Exciting time to be a central banker.

AM: Yes it is.

DH: Andrew, thanks.

AM: You’re welcome.

  • LinkedIn  
  • Save this article
  • Print this page  

You need to sign in to use this feature. If you don’t have a Central Banking account, please register for a trial.

Sign in
You are currently on corporate access.

To use this feature you will need an individual account. If you have one already please sign in.

Sign in.

Alternatively you can request an indvidual account here: