Andrew Haldane is the Bank of England’s Executive Director for Financial Stability. I recently talked to him about the crisis of contemporary economics and the way forward.
You are vocal critic of mainstream macro economics. How does this square with the tradition of central banks who historically have been very conservative institutions.
I don’t know if what you say about conservatism in central banks is actually true. I can see why people say that central bankers want to have ideas tested in a laboratory before they bring them to the world. In recent history, many ideas first bubbled up in academia and then eventually migrated into the policy sphere. Monetarism is one example for this, and the whole notion of adhering to policy rules grew out of academic literature. But during my time here there have been a number of examples where the causality has been reversed. In the current crisis, we looked for guideposts from academia but they were relatively thin on the ground.
Does this mean that academic economists are behind the curve?
All of us have to learn from experience. If you work in a policy institution, you have got less time to learn because you have to handle the experience. You can’t afford to come back to the issue in five years time.
You are closely working with scientists from a number of other disciplines like biology, physics and epidemiology. Why’s that?
I am just trying to make sense of the real world. The workhorse model we used in economics is no longer working properly. Hence we are compelled to think innovatively, to seek insights from disciplines beyond our own. We are looking for systems behaviour that matches what we can see happening in the real world. In some ways we are just borrowing ideas from other disciplines.
What are the issues of contemporary macro economics?
With the benefit of hindsight, we built an edifice, a set of models, that were quite peculiar in the assumptions they made . For example, we drifted away from the notion of having a multiplicity of equilibria. We also forgot that we could even get stuck in the wrong equilibrium, which is one way of making sense of where we are today. We drifted away from the notion that systems could become dynamically unstable when stretched or stressed too much. It is hard to think of a system outside of economics, whether from the natural world or from social sciences, that doesn’t have a multiplicity of equilibria and doesn’t behave peculiarly when stretched or stressed and can which be destabilising. However, in modern macro, the notion of equilibrium was a singular stationary, stabilising concept.
Can you give a concrete example, please?
Back in the 20s, when macro was beginning to find its feet, Knut Wicksell used the metaphor, when explaining economic systems, of the rocking horse. Imagine you are hitting a rocking horse with a stick: The horse moves according to a pretty regular pattern. That metaphor is a fairly accurate summary of how we approached economic modelling theoretically and empirically for a century.
If the economy doesn’t resemble a rocking horse, what’s a more appropriate description?
Instead of beating a rocking horse, imagine hitting a pack of wild horses with a stick. What happens next? Well, the truth is: you don’t know. One possibility is nothing happens at all. Maybe the horse you beat runs off in one direction. Most likely, however, the horse runs off scared and panics all the other horses who then also charge off in a direction that is impossible to predict ex ante.
This sounds pretty chaotic.
Exactly. We effectively have a quasi chaotic system driven by interactions between living , breathing, interacting agents where a multiple of equilibria are possible. Some of which are good, some of which are bad.
Is it possible to make sense of such a system at all?
Yes, to a certain degree. But economists have to recognise two things which are not currently embedded within the workhorse framework. The first is heterogenitey of the people in the system. The second important insight is that what people do depends a lot on what they think.
Expectations are already built into macro models, aren’t they?
That’s right, but not in a satisfying way. Rational expectations theory says that what people think depends a lot on what’s happening. Basically, expectations are seen to mirror fundamentals. For me, it’s sometimes the other way round. Fundamentals are driven by expectations. Happiness, sadness or panic are real emotions. Even if they are based on nothing in particular, they will drive behaviours and outcomes.
… like the Keynesian Beauty Contest, where you are trying to guess the taste of the majority.
Exactly. That’s a very good example. We need to recognise that uncertainty and panic and all the rest of it can become self-reinforcing and self-fulfilling. The same is true for complacency of course, such as during the Great Moderation.
How did your own perception changed during the crisis? Have you always been sceptical?
Luckily, I had done some thinking about some of these issues before the crisis. I started to work on network models with interactions in 2003/2004. That is not to suggest that I got everything right. That’s certainly not the case. But those models showed that financial integration was a double edged sword. In the good times, when shocks are small, financial integration serves as a shock absorber, because risk was spread around the system. But if the shocks are sufficiently large, it flips around, and the system becomes a shock transmitter. You end up with a world that was very robust and yet fragile at the same time.
How were these results received by the discipline?
Prior to the crisis, this work did not arise particular interest – including by me, to be honest. I thought these results were interesting, but I was not like crying that the sky was going to fall in on the scale that we saw only a few years later.
What kind of policy implications do these network models have?
There are three different implications. First, we need better surveillance and data to understand interconnections in the financial system. Stress testing is also necessary for figuring out where the cliff edges are and how big they might be. The second insight is that you need to know where the key nodes of the system are. They require greater protection rather than less protection. This is a epidemiological lesson about vaccinating the super-spreaders. Hence we are working on the “too big to fail” phenomenon and are seeking to put in place extra insurance for systemically important players. The third thing is that you might want to reconfigure the system to make it less fragile in the first place. There need to be redundancies in the system – circuit breakers, for example, that prevent the spread of contagion. All three have now found their way into the responses of policymakers during the crisis.
Hence contemporary macro has been rendered useless?
No, I would not want to dump all that has been done. There is still a usefulness in the existing workhorse models, suitably augmented with some of the things I’ve discussed. Those models need a more imaginative introduction of expectations, a more imaginative introduction of heterogenous agents, and a more imaginative introduction of financial frictions. That said, it would be a real mistake to confine ourselves to that one class of model. We need a more plural approach to the way we think things through. So I would like to see much greater intellectual investment in different frameworks.
There are lots of non-conventional avenues which have historically been niche and need to become more central. Agent based modelling is just one example. This approach obliges you to take very seriously heterogeneity of agents, interactions between agents. Another important area is to increase the importance placed on studying the history of economic thought.
Contemporary macro economists seem to be rather reluctant to join this journey, aren’t they?
Yes, that’s right. There are huge amounts of inbuilt inertia in academia. I don’t see that as a criticism, more of an observation. It’s human nature: People don’t like re-writing their lecture notes, especially when they haven’t got any new notes to write. That’s not a surprise. The incumbent academics and policy makers have too much human capital tied up in the project and are too old – like me – to drive it to the next place. I’ve spent 20 years accumulating that human capital, too. Most of my career has been spent writing papers using this stuff. I’m culpable too. There is a sunk cost for me in all this as well.
Do you think that the change will happen despite the reluctance of the establishment?
Yes, I do. I think the profession is in the very early stages of a real paradigm shift. In line with what Thomas Kuhn told us, it takes a while, it’s almost a generational thing. It will start with your conventional framework not really doing the business, than a fledgling movement from within to begin changing things. I think that’s happening.
Who is driving the change?
Above all, it’s the new generation of students. People who might even have been drawn to the discipline because of the crisis. They want to make sense of the world. They are the change agents in this paradigm shift. For example, Greg Mankiw’s students wrote him a letter saying “Sorry, but what you’re teaching us does not help us to make sense of the world.” That’s an example of the quiet revolution taking place.
What role do policy makers like the Bank of England play in this situation?
I hope we can do a bit to nudge that along by saying: Look, these are the pressing problems at the moment, these are the phenomenon we struggle to make sense of and the current tools are not doing the job anymore. I’ve certainly tried to do this by putting in play some ideas that might be seen as radical. People are drawn to radicalism because what you have doesn’t work. It’s no more than doing our jobs, which is trying to make sense of where we are.
A lot of critics, especially those outside of the discipline, think that economics has become too mathematical.
I can see why people say that and I think our quest for analytical beauty did get in the way of reality for quite a long time. However, rigorous model based approaches definitely have their place. But they must not blind us to reality.