Markus Brunnermeier

Edwards S. Sanford Professor of Economics, Princeton University

Paper: "The I Theory of Money" (with Yuliy Sannikov)

Andolfatto

Tell us a little bit about the question that you're pursuing in this paper.

The paper's main takeaways,
according to Brunnermeier:

  • The key takeaway of the paper is essentially that before the crisis we have three stability concepts: financial stability, bank regulators should take care of it; price stability, central banks should take care of it; and fiscal debt sustainability, that it government can pay back its debt, the government should take care of this. And we thought we could treat them independently and in silos.
  • This analysis shows that these three stability concepts are highly interconnected.
  • Monetary policy can avoid wealth redistribution that is driven by amplification effects and thereby reduce overall risk premia in times of crises.
  • Independence for central banks means to be protected from the other parts because there might be some financial dominance argument where the banking sector might make it hard for the Fed to control inflation.

Brunnermeier

Probably I should start by revealing what the "I" stands for. The "I" stands for inside money or intermediation. So it's all about reviving the "money and banking" field using modern tools. It became standard to treat financial stability and price stability separately: Bank supervisors take care of financial stability, and central banks take care of price stability. This paper shows how those two concepts are interlinked.

Andolfatto

So what are the main findings of your research?

Brunnermeier

This line of research is about the interaction between different stability concepts: financial stability, price stability and fiscal debt sustainability. The paper makes the case that credit and money created by the financial sector are key in understanding the amplification mechanism. One main finding is that monetary policy can avoid adverse wealth redistribution due to amplification effects and thereby affects risk premia. After a negative shock, banks try to shrink their balance sheets. This has two effects—one on the asset side of the balance sheet and one on the liability side of the balance sheet. On the asset side, banks cut back on extending new loans and try to fire-sell assets. With it the prices of assets go down, which hits banks' equity. Equally important, by shrinking the balance sheet, banks also reduce their liabilities, i.e., they create less inside money. So total money supply in the economy goes down. A decline in overall money supply causes deflationary pressure. Lower inflation hurts bankers' equity further because the real value of their liabilities rises.

chart

Andolfatto

This sort of story, it's a familiar theme I think in the history of economic thought. I think Irving Fisher's debt deflation theory rings a bit of that, and I'm aware of other people who have written here and there about the subject. Can you elaborate on what your paper does a little bit different? Are there any surprising results that these other authors have missed?

Brunnermeier

Indeed it's related to earlier research. In particular, to the Fisher deflation spiral. It's also related to Friedman and Schwartz who stress the decline of money supply during the Great Depression in the 1930s as banks went bankrupt. We combine the money view and the credit view within a single framework. More importantly, we go beyond a stock-flow analysis. Self-generated endogenous risk and risk premia play a major role in our approach.

So what do I mean by money view versus credit view? If you follow the money view, emphasized by Friedman and Schwartz, you try to offset the deflationary pressure that arises from financial instability. By doing so, you help the banks because they would lose out from this deflation. If you follow the credit view, which was, among others, pushed by Jim Tobin, you would like to restore the total credit flow to the real economy. So you try not only to offset the deflationary spiral, a la Fisher, but also the liquidity spiral on the asset side of banks' balance sheets. Banks' behavior is to a large extent governed by risk, and not only by the risk-free interest rate.

Andolfatto

Imagine a policymaker was armed with either the Friedman and Schwartz model or the competing view, the Tobin view. Do you think that the policymaker might misjudge how to optimally intervene in a time of crisis without your comprehensive view of the way these things interact?

Brunnermeier

To answer this question, I have to stress one distinguishing feature that I didn't highlight so far. Our analysis suggests a "bottleneck economic perspective." As a monetary policymaker you have to ask yourself: Where is the bottleneck? Which sector is financially impaired? Can monetary policy be used to address these financial impairments? If so, does it reduce overall risk?

Typically, the financial sector is always hit by a financial crisis, but other sectors are, too. For example, in Japan the corporate sector suffered from a debt-overhang problem in the 1990s when the two lost decades started. So you would like to conduct some ex-post monetary policy that helps the corporate sector. In contrast in the U.S. now, parts of the household sector are overly indebted. In that case, one would like to support the mortgage market—and with it the housing market—in order to repair the balance sheets of the certain households.

Andolfatto

You're not suggesting, for example, that the Fed should explicitly target certain hard-hit sectors of the economy. You're suggesting that the policies that they undertake naturally in a crisis—lowering the interest rate, quantitative easing—will naturally affect or help the most distressed sectors? What would be an example of a Fed policy in the present context that would, say, help the household sector?

Brunnermeier

Isn't this what the Fed is intuitively doing? The Fed's buying of mortgage-backed securities targets the housing market, which helps the household sector. In contrast, the corporate sector is not benefitting so much from that. Indeed, the Fed is concerned that the corporate bond market is trading at excessively high levels.

Andolfatto

So the present Fed policy of purchasing these mortgage-backed securities is, in fact, somewhat justified by your theoretical approach here?

Brunnermeier

It is to some extent justified, even though there is a lot of redistribution going on toward households who don't really need it. So it's probably to some extent not targeted enough. The problem is, the more you target certain sectors, the bigger the moral hazard problem becomes.

Andolfatto

Moral hazard, can you explain that?

Brunnermeier

Moral hazard means that people distort their behavior in an undesirable manner. If you know whenever something goes wrong the Fed will jump in and help you out, then you will not behave prudently anymore. You just buy, buy more expensive houses and drive up house prices. So, as a government or as the Fed, you can't help out too much because otherwise you depress risk premia in a boom phase too much and sow the seeds for the next crisis.

Andolfatto

So there's a little bit of a tradeoff here, to trade off these bad incentive effects owing to this moral hazard effect versus the insurance effect of recapitalizing the household sector, in this case that was really quite shocked—the drop in the real estate prices, these distressed homeowners with the mortgages that are constraining their spending habits.

Brunnermeier

Yes.

Andolfatto

You speak as if this is a policy that the Federal Reserve of the United States or indeed in general central banks around the world might wish to undertake. But the idea of redistributing wealth in the economy is a politically sensitive issue when it comes to a central bank, especially the U.S. Federal Reserve. This sounds like it's a job that at least in principle is ideally suited for the fiscal authority. There's a body of elected representatives and then the fiscal authority redistributes wealth all the time. This is partly what their job is to do. The Fed, to the extent that it likes to maintain independence and be politically detached, refrains from overt acts of redistributing wealth. Is there some reason for why you believe realistically the Fed should be doing this job as opposed to, say, elected members of Congress?

Brunnermeier

I would distinguish between actively redistributing wealth and stopping redistribution that arises from inaction. I don't think the Fed should get into the business of actively redistributing wealth. However, the Fed should not repeat its mistakes from the Great Depression: By not doing enough, large redistribution occurred, which ultimately led to an overall wealth destruction. The aim is to switch off this redistribution rather than actively doing some redistribution. This reduces risk premia and endogenous risk and thereby stimulates the economy.

Andolfatto

So the Fed is there to prevent an unintended redistribution of wealth that if it was left untended would actually have adverse macroeconomic consequences, potentially adversely affect the economy as a whole?

Brunnermeier

Exactly. It's not a zero sum game. It could be that everybody's worse off if the Fed is not intervening, like in the Great Depression.

Andolfatto

You alluded to a notion of an economy where, if in a time of relative tranquility like during the Great Moderation, the overall risk exposure to the economy may not decline because private actors in the economy—banks, financial managers, whatever—might expose themselves naturally to more risk. They become more tolerant of risk so that the overall risk in the economy might not diminish. Elaborate on that a bit, if I've characterized it correctly or not.

Brunnermeier

In this analysis, we can distinguish between two forms of risk. One is exogenous risk, risk which comes from outside because fundamentally certain projects are risky, and endogenous risk, risk which is created within the system, so it's self-generated risk. It is this self-generated risk that leads to redistributions. The Fed's aim should be to reduce the self-generated risk in the system due to amplification.

Andolfatto

Implicit in the statement you just made was the idea that, left to its own devices, the free market economy would generate excessive amounts of risk.

Brunnermeier

That's correct. This is one surprising result we found in our formal analysis. We call it the "volatility paradox." As one lowers exogenous risk, the endogenous risk doesn't decline. Even as exogenous fundamental risk vanishes, endogenous self-generated risk remains high.

Andolfatto

But why can we not rely on the private sector to generate just the right amount of risk? Surely zero risk is also not desirable because there are tradeoffs.

Brunnermeier

Markets are not perfect. The problem is excess liquidity mismatch since long-term risky projects with low market liquidity are financed with short-term debt. In an incomplete markets setting, each individual market participant does not internalize the pecuniary externalities he imposes on others when he levers up more.

Andolfatto

What is causing the banks to overexpose themselves to this type of risk?

Brunnermeier

It is crucial for a functioning economy that the financial sector take on some risk, some fundamental exogenous risk, but one should limit the endogenous risk which is self-generated risk in the system. The latter results from the fact that banks don't fully internalize that their levered risk taking has negative repercussions on others as well. So for example, when the banks lever up too much, they will be forced to de-lever after a large negative shock. They will try to shrink their balance sheet. This depresses the asset prices, which has some negative repercussions on other banks, say you. Importantly, as each individual bank builds up leverage in good times, it doesn't take fully into account the negative externalities in times of crisis on others. These pecuniary externalities lead to inefficiencies in an incomplete markets setting.

Andolfatto

Would this be an argument perhaps for these banks to coalesce, to conglomerate and form one big bank?

Brunnermeier

To some extent you see it in Canada. In Canada you have very few banks, and they didn't shrink their balance sheet so dramatically. In the U.S., you have a competing banking landscape that shrinks the balance sheets much more dramatically. But I would not argue that there should be a collusion among the banks. This comes with severe additional side effects.

Please note that the content in this section is based on the interview but should not be considered a transcript.