The Dodd-Frank Act, market-based measures of systemic risk and stress tests

Viral Acharya interviewed by Viv Davies, 20 August 2010

Unfortunately the file could not be found.

Open in a pop-up window Open in a pop-up window


Download MP3 File (15.7MB)




View Transcript

<p><em>Viv Davies interviews Viral Acharya for Vox</em></p>
<p><em>August 2010</em></p>
<p><em>Transcription of an VoxEU audio interview []</em></p>
<p><strong>Viv Davies</strong>: Hello, and welcome to Vox Talks, a series of audio interviews with leading economists from around the world. I'm Viv Davies from CEPR. It's the 17th of August, 2010, and I'm talking to Viral Acharya, Professor of Finance at the New York University, Stern School of Business, about his current work on financial reform, and in particular, capital requirements and measuring systemic risk.</p>
<p>I began by asking Professor Acharya how important the recent Dodd Frank Act is, in the context of reforming the American financial system, and what he thinks are its strengths and weaknesses.</p>
<p><strong>Viral Acharya</strong>: I would say, by and large, this is clearly the biggest overhaul of financial reforms in the United States, at least since the Banking Act of 1934. There have been important changes along the way, but nothing where the policymakers, and to an extent, the industry, have made a concerted effort of this magnitude, to realize the law of the land, so to speak. On its strengths and weaknesses, it is perhaps not surprising that, given it was in the aftermath of the severe financial crisis we witnessed, it is focused on systemic risk. The risk that, in many institutions, faced together, it's hard to essentially let them fail, or the risk that, in a large institution, that is interconnected and is a critical player in the economic plumping, it's very difficult to have it fail.</p>
<p>So, on this front, I would say the Act is good in the sense that, perhaps for the first time, that it is explicitly recognized, and dealt with, in the financial reforms law, at least in the United States, that the purpose of the regulation is actually to deal with systemic risk.</p>
<p>So, in particular, it comes up with recommendations for how to designate an institution as a systemically important institution. It designs a systemic risk oversight council that's going to be in charge of its regulation of these institutions. And, by and large, proposes an array of requirements that these institutions have to meet, over and above what the provincial regulators might require of the small or medium size banks at large.</p>
<p>The other strength is that, in principle, the Act covers not just the traditional banks - I would say the commercial or the universal banks that are systemically important. But it also covers non-bank institutions, which are increasingly being called the &quot;shadow banks&quot;, who have not been regulated as much. This would include, essentially, insurance companies, hedge funds, investment banks, which are now all merged, and so on.</p>
<p>In principle, if the systemic risk oversight council finds any of these to be also systemically important, they could bring them under the purview of the legislation, and the requirements that they subject systemically important institutions to.</p>
<p>I would say, in terms of weaknesses, they are primarily three weaknesses, in my view. Which is that, while the Act takes systemic risk on board, it doesn't quite go all the way in actually dealing with it then.</p>
<p>The first and foremost is that, it does not require that the systemically risky institutions essentially pay upfront for the fact that they are systemically important, and ex post there is the likelihood that they might receive more forbearance than the others.<br />
Instead, what the Act says is that, in case they are to be failed, in case they get into trouble, they will have to be failed necessarily. Their creditors will be wiped out, shareholders will be wiped out, management will be replaced, etc. And, if at all there are any costs over and above this, they will actually be charged to other financial players in the system.</p>
<p>Now, an issue with systemic risk is that, one, there will be other costs, besides this player. And two, to charge this to other financial firms at that point, just seems a bad economic design.</p>
<p>It's basically, like, saying that when my neighbor's house goes on fire, I'm actually going to pay for the mess that it creates. And this is precisely the time that I might actually be more concerned about saving my own house, rather than actually paying the [inaudible 04:42] from cleaning up the mess in the neighboring house.</p>
<p>So, it just seems like postponing the problem to the next crisis, rather than actually requiring that systemically important institutions pay a penalty for actually being systemically important, they paid upfront so that the costs can actually reduce the burden later on. And also, they have lowered incentives to suffer these costs in the first place. I would say this, to us, has been at NYU Stern, in our book, &quot;Regulating Wall Street&quot;, I would say, has been one of the first and foremost weaknesses.</p>
<p>The second weakness, I would say, is that, even though the Act proposes a resolution authority, which is going to be more in the FDIC style, resolution going forward. It actually leaves out how to deal with markets that are systemic, and not just institutions.</p>
<p>To give you an example, each money market fund in the grand scheme of things, is rather small, but after the collapse of Lehman Brothers, a large number of these funds were experiencing runs at the same time.</p>
<p>In fact, the run on the repo markets, which caused the instruments to fail, was also, essentially, not run off any one individual, these are not just runs on individual firms, they potentially, you run the risk that the entire market may freeze all at once.</p>
<p>And, essentially, the systemically important markets, the money markets, the deposits from money market funds, sale and repurchase agreements, which are now a very significant part of the plumping of the financial institution, it's not just taking retailer deposits anymore. These markets have their infrastructure, how to deal with a run in these markets, has not been appropriately dealt with.</p>
<p>I would say the third thing that's &hellip; these two are the most important. One, because it's about controlling systemic risk ex ante by making these systemically important firms to pay up front for it. And the second is ex post, which is, how do you deal with this kind of systemic risk when it actually arises? I would say that these two are the most important weaknesses.</p>
<p>The third, sort of minor weakness that I think you might attribute to the Act, this essentially concerns the fact that it completely omitted certain important parts of the financial sector altogether. In the United States, that would primarily include the government-sponsored enterprises, so Fannie Mae and Freddie Mac, which contributed in pretty significant measure, to the housing credit boom. They are in conservatorship right now; they already bled to about $150 billion dollars of taxpayer's money. More might be required.</p>
<p>To be fair, there is a summit, in fact today, as we speak, in the United States that the Obama administration is going to vein on certain proposals in discussions with certain industry experts, academics, and so on. And they plan to propose some reform proposals by January of 2011.</p>
<p>But I think, too, we need to, while it's important to think of them reforming them separately, I think there is a number of pieces in the Dodd Frank Act that should apply, also, to the government sponsored enterprises. It's time to remove from them the special status, which is that the law, somehow, works separately for government-sponsored entities than it does for these private institutions.</p>
<p><strong>Viv</strong>: Viral, as you've suggested, it's widely accepted that systemic risk in the banking system needs to be contained. However, as you've noted in a recent paper, the systemic risk of an individual institution has not yet been measured or quantified by regulators in an organized way. Recent work that you've undertaken in this area has focused on criteria for determining systemic institutions and on measuring systemic risk. In particular, you've been looking at capital requirements and stress tests. Could you tell us some more about this?</p>
<p><strong>Viral</strong>: Essentially, there are two ways to go as far as measurement of systemic risk goes. One way to think about systemic risk is to entirely rely on regulatory assessment, as was done through the stress test in the United States in summer of 2009, or spring of 2009, and in Europe very recently. In this kind of an approach, regulators basically subject the balance sheets of financial institutions to some macroeconomic or market-based stress scenarios. For example; 10% unemployment rate, contraction in GDP of a certain level, housing price declines of 20% all across the county. And then get a sense of... because these shocks are correlated across balance sheets of institutions, see which institutions are likely to become undercapitalized in these times.</p>
<p>And then, either, have a plan to basically close down on those who are going to be troubled in such times, if they already look pretty much close to insolvency. Or, essentially get them to be recapitalized. Either privately or impose some discipline on them by saying that if they don't raise the private capital themselves, the government will do it for them at the cost of diluting the management and the shareholders.</p>
<p>The alternative approach... so this relies heavily on design of the supervisory stress test, valuation of losses by, or estimates of bank loses by bankers, and then their agreement with what the regulators think are reasonable estimates.</p>
<p>In contrast, you could also rely on some market-based data in order to extract measures of systemic risks. This is something that we have been doing at NYU Stern. In fact, we have a website on Volatility Lab run by Robert Engle, the Nobel Laureate in econometrics in 2003. Essentially on this website, we provide a daily ranking of the systemic risk of US financial institutions based on market data until that date, but in a projected or a forecasting sense.</p>
<p>So let me explain simply what we do. If you didn't want to rely on a regulatory definition of stress test, you could say a stress test is a scenario, say, in which the aggregate market, the stock market or the economy, collapses by 40 or 50%. This is what has happened in the Great Depression and is what we witnessed in many countries in the Great Recession.</p>
<p>Then you could ask the question: In this kind of a scenario, which firms&rsquo; equity capital, which financial firms&rsquo; equity capital, is basically going to get wiped out the most? Because, that would be one way of knowing whether they get undercapitalized.</p>
<p>Now, intuitively, we know from the standard sort of finance paradigm, that firms that have high betas so whose assets or stock price fluctuations are very correlated with all of the market are going to be the ones who are going to tank when the market as a whole tanks as well.</p>
<p>Now, of course, the thing about systemic risk is that it's sort of a tail end. It's not just about smooth variations and so on. So you can refine this measure a little bit to calculate like a tail beta, or what the specific econometric measure we use is called as marginal expected shortfall, MES, which you can think out it as the mess produced by each financial firm.</p>
<p>And it basically tells you that when the market in it's worst outcomes, you econometrically estimate, using historical data and some projections based on it, which firms are going to suffer the most as far as equity losses are concerned.</p>
<p>Now equity losses are not enough because if the firm has very high leverage then a given equity loss is going to bring it down much closer to the brink of failure.</p>
<p>And so our measure essentially combines these pieces of information which is: How variable are you? How correlated are you with the market? How much therefore will you lose when the market as a whole suffers a dramatic downturn?. And is that capital loss that you suffer is going to be sufficiently large relative to your leverage let's say so that you don't have four percent tier capital requirement in that stress scenario.</p>
<p>Now, you can calculate this measure and then rank different firms based on this measure, you can in some sense see out of the total loss that is going to arise out of this scenario which firms are in fact contributing the most. The beauty of the measure in some sense lies in the fact that you can test it historically.</p>
<p>You can see, if you are sitting at the beginning of the crisis, which firm showed up as being systemically most important. If you were sitting as Lehman Brothers collapsed and wanted to do the US stress test going forward from that point, which are the firms that you should have looked at and expect it to fail the worst in the stress test?</p>
<p>What our research shows is that the measure does a remarkable job of actually picking who are the systemically most important firms. The largest commercial banks that got into trouble, especially Fannie and Freddie, all of these show up in the top ten to fifteen in the world ranking of US financial institutions. Because they have tail risk, they have high correlation with the market and these firms have extremely high leverages all factors that contribute to systemic risk.</p>
<p>I think going forward it's not that we need to make a choice between market based measures or supervisory tests. I think we need to go both. There is clearly a lot of information that supervisors can gather if they are systematic and thoughtful about it, relative to what the market already knows all has factored in.</p>
<p>But I think given that there is no way of knowing whether a stress test is being done rightly or wrongly or adequately. It will be useful for regulators to have some market signals that give them a sense of who should they be looking for. If they assess the firm not to be systemically risky, is that consistent with what the market data are telling them?</p>
<p>So I think it is very good to and fro discipline between market data and supervisory data which might make future assessments of systemic risk more stable.</p>
<p><strong>Viv</strong>: To what extent, Viral, do your ideas and recommendations on risk in capital requirement etc. differ from what's contained in the proposed Basel III proposals?</p>
<p><strong>Viral</strong>: I would say first and foremost, Basel III doesn't even have the strengths of the Dodd Frank Act in my view. The Basel capital requirements have not yet explicitly recognized that they should be about containing systemic risk of financial firms. The Basel requirements are still about increasing the capital buffers of each financial firm in isolation, ensuring that that firm doesn't fail on its own rather than trying to ensure that this firm has a sufficient buffer if it is more likely to fail then other firms get into trouble or the rest of the economy gets into trouble.</p>
<p>So I would say at the conceptual level, our measure of systemic risk and how it could be used to design capital requirements is very, very different. Of course, I think that not addressing systemic risks in anyway into the volume capital requirement makes them a very weak and potentially even a harmful tool in dealing with systemic risks. Because as we saw before the crisis, all institutions seemed extremely well capitalized from a regulatory standpoint. But clearly there were extremely correlated and very highly leverage.</p>
<p>So what happened? I think what happened is that the regulatory capital requirements were simply not capturing the fact that the system was extremely vulnerable at that point. The other thing I would caution about is that the Basel III approach seems to be of going for more and more and more capital.</p>
<p>Either now or five years down the road, more liquidity in one way or the other. I think as I said the key question is not whether you have more capital or more liquidity in an absolute sense. But, whether you have enough capital or enough liquidity to withstand stress scenarios of the economy when other firms are likely to get into trouble at the same time.</p>
<p>This latter question is never being asked. To just give a simple example the Basel capital rates give a one fifth capital charge to an AAA rated mortgage backed security compared to an AAA rated corporate loan.</p>
<p>Now historically up until this crisis, it seemed that mortgage default risk was smaller and therefore this might have been a reason to justify it. But the trouble is that once you do something like this, the entire financial sector starts digging at the AAA rated mortgage backed securities. Because now it has a one fifth capital charge, it offers a greater leverage.</p>
<p>Over time the share of the housing sector in the economy becomes larger than that of the corporate sector. Suddenly an asset that looked more stable in the past is actually now the most systemic asset in the economy. But the capital requirements remain somewhat naive or innocent about all of these things that are taking place in the background. And essentially year by year they start looking woefully inadequate as far as guarding against systemic risk is concerned.</p>
<p>Of course this is very familiar argument. I think Hyack was worried about this all the time. That regulation basically creates pockets of concentrated exposures of the financial sector. And so if you get something like an interest rate wrong altogether, it basically leads to a big mess. Similarly here if you essentially get the risk weight of an important asset like housing wrong - and in this case you're getting it wrong because you're not thinking about systemic risk or the collective risk of institution - then you basically get the entire financial sector digging at that particular aspect of regulation. And that is indeed where the crisis happens eventually.</p>
<p><strong>Viv</strong>: You referred earlier to the stress tests that were undertaken in the US in the spring of 2009. How do you think the recent stress tests undertaken by the European banks compare with that exercise? Were the European tests robust enough in your opinion?</p>
<p><strong>Viral</strong>: They seemed to have helped in the sense that... so I think they are probably better than the opaque scenario that we had before the stress tests were conducted. I know from the postings of Xavier Freixas on VoxEU that he thinks that this has actually been a good thing as far as understanding the stability of the Spanish cajas is concerned. But at least from someone who saw the stress test from outside, personally my assessment is that they could have been somewhat better. I liked the American model somewhat more for the following reason. One, the American stress test had a very clear recapitalization plan that was put in place along with the stress test. Which is the stress tests were a way of assessing the capital adequacy of an institution in a stress test. If they did not meet it, they were given a pre specified amount of time to raise the capital privately.</p>
<p>Failing which, the government would basically inject the capital at the cost of diluting the shareholders, firing the management et cetera, et cetera. Now both of these are very important. The reason why is the following. If you don't have a meaningful way of addressing the balance sheet of a weak institution, it almost becomes dangerous to go and announce in the market that, &quot;Oh, this institution is in trouble. But we don't have any tools actually to fix its problem.&quot;</p>
<p>In that case you can see now that regulators might therefore be reticent about the true extent of risks in the first place. Now the markets again won't know who is where, who has made what kind of losses, or is likely to make what kind of losses in a systemic stress scenario. You get the sort of laws of confidence and the downward spiral of holding, not getting enough investments, banks not trusting each other, etc. happening as we have seen repeatedly in this crisis.</p>
<p>In contrast, if you have a credible plan to address the balance sheets of weak institutions now the regulator is exonerated from actually withholding the information. The regulator will say, &quot;Yeah, I should go and give this information out. The firm is going to in fact give this information of any place to go and raise this capital. If they don't do it, I'm going to inject this capital into the firm anyway&rdquo;</p>
<p>So, I would say it was decisive to have undertaken the stress test even if one year later than the American stress test, perhaps because the sovereign risk, targets showing up severe stress signs in the fall off of '09.</p>
<p>I think as far as addressing the balance sheets of financial firms is concerned, there has been a little bit lack of decisiveness in the European regulators. Perhaps the situation is very complicated right now with the kinds of holdings of sovereign bonds that may be out there.</p>
<p>And to sort of make a point on that for example it's not at all clear why the holdings of sovereign bonds on the banking side of the balance sheet were not&hellip; Essentially, efforts were given to these sovereign bonds holding on the trading book side.</p>
<p>But clearly, if you are holding things on your banking book you might make losses there too. It seemed as though effectively the stress tests were assuming that, &quot;Oh, if you suffer losses from there they are only going to be mark to market losses. Because either we are not going to like this sovereign thing, or we are going to backstop your losses, or we'll allow you to readily swap these bonds into the central bank of the region for liquidity.&quot;</p>
<p>All of these seem to be essentially saying that we don't want to address the root causes of the problem. We are just going to tide them over. But yet we want to get some information to the market to make them feel a little better.</p>
<p>So I'm left with a feeling that they didn't go all the way.</p>
<p><strong>Viv</strong>: How important is international coordination in banking regulation and financial reform? Are there fundamental differences, for example, between the US and Europe in terms of what is required going forward? And in this respect, do you think that European governments could be following the US lead in establishing laws for financial reform?</p>
<p><strong>Viral</strong>: I would say that by and large, clearly, some forms of coordination are very important, and even though I'm myself somewhat critical of the current state of Basel capital requirements, I think the process of going through the financial stability board to actually agree on a capital regulation, the G20 where some of these issues get discussed in some detail. I think these are important forums which, even if sometimes not as well functioning as one would like, I think that we set the stage for achieving a certain amount of common ground in the regulation. I think some areas where they become especially important are areas such as derivatives. This is because derivatives have the somewhat off balance sheet nature, which is that most derivatives when they are created have zero value at inception. But of course the leverage can potentially be very high because when risk materializes, one party may be losing a lot of value to the other.</p>
<p>These are done across country borders. They have relatively been opaque so far. So this is one area I would say, I didn't mention this but the Dodd Frank Act actually does a reasonable job in trying to improve the infrastructure of the OTC derivatives market the over the counter relatively opaque, unregulated aspect of the derivatives market.</p>
<p>By trying to bring at least some of these back on to banking balance sheets, trying to ensure that the more opaque, exotic varieties are capitalized better. Having both public transparency of price and volumes of trading but also almost close to full transparency for regulators as to what kind of counter party exposures are getting built up in these markets.</p>
<p>There are some things that could have been done better, but at least the spirit of the reform of derivatives in the United States, I would say, is probably one of the best done aspects of the Dodd Frank Act.</p>
<p>Now, a really worry, though, is that if one of the other financial centers, in it's bid to actually attract more of derivatives business over to itself, starts lowering the standards, either in terms of disclosures or collateral requirements, or requirements that derivatives that are actually clear on centralized platforms. If one financial center essentially sees an opportunity and decides it's willing to take that risk, it could essentially completely unwind the... it could blunt the edge of the US reforms. Because, financial business, even if not fully mobile on an overnight basis, has had shifts in terms of moving from these regions and where the laws are the weakest.</p>
<p>And there is no doubt that if, say, one financial center in Europe provides weaker regulation of credit derivatives and it is going to be cheaper for financial firms to produce credit derivatives there, eventually that is where the trading will take place.</p>
<p>Fortunately, the central bankers are talking a lot about coordinating on derivatives reforms. I know the New York Fed, the European Commission, the ECB, etc. are in talks about these things, the Bank of England. And the hope is that they will not actually go for a race to the bottom, but instead actually ensure that a minimum standard is maintained.</p>
<p><strong>Viv</strong>: Finally, Viral, before we wrap up, I'd like your take on a comment that was made recently in a speech at the Stern School of Business by the US Secretary of Treasury, Tim Geithner. He suggested that the recent reforms will fundamentally reshape the entire financial system. They'll require financial firms to change the way they do business, to change the way they treat customers, to change the way they manage risk and to change the way they reward their executives. Now that's quite a tall order. Do you think the banking system is really ready for such a significant cultural shift?</p>
<p><strong>Viral</strong>: That's a good question. My sense is they have been getting ready for it. I think it was to be expected that business could not just go on as normal after a crisis of this sort. Initially, it seemed that, in fact, the Dodd Frank Act might even get watered down quite a bit due to the lobbying attempts. I would say the end effort has not been as diluted as one would have thought perhaps a year back. Culturally, it does require a shift, and I think that shift has happened. I think industry knows that they face a fair bit of regulatory uncertainty. And I use the word &quot;uncertainty&quot; because that is what might be the critical thing to worry about right now which is that one aspect of the Dodd Frank Act, including of it's strengths, such as the derivatives reform, is that a great deal of discretion has been left to prudential regulators.</p>
<p>Now to some extent, you don't want the Congress in the United States to be writing the details of every single law out there. But what this also means is that, over the next three to five years, a large amount of rule making is going to take place. And currently, there isn't full clarity on what's going to happen.</p>
<p>To give you an example: It's not clear which segments of derivatives are going to cleared on centralized platforms and which segments are not going to be cleared. But if you were a derivatives boutique firm or, say, a large bank with a significant presence in derivatives, you would want to prepare for something like this, but you actually don't know yet how to prepare fully because there's uncertainty as to what the exact rules are going to look like. So I would say that this uncertainty is what, in some sense, the banking sector needs to be able to deal with. Now, unfortunately, this kind of uncertainty is not something they can hedge very easily, because it's an uncertainty that affects every single firm out there. So most likely it might just produce a reduction in the risk taking. They might hedge their bets a bit by just holding back on their capital and liquidity.</p>
<p>And this may be a somewhat tricky situation in a time such as this one when you're actually trying to kick start lending and kick start investments in the real economy. Now having said that, there are some aspects I think which might fundamentally alter the way banks are operated. One of the things Timothy Geithner mentioned in his speech was that even though&hellip;has not yet footed this idea openly, that at least the United States is considering the possibility of a two tiered capital requirement.</p>
<p>One will be like a minimum capital requirement, which is for normal times. And then an additional buffer that financial firms especially the systemically important financial firms will have to hold for withholding stress in a systemic scenario. This actually comes out of good economic principles for regulating systemic risk. In some of my recent work with Hamid Mehran at New York Fed and Anjan Thakur at University of Washington, sorry, Washington University of St. Louis.</p>
<p>We do find that indeed if you had to regulate systemic risk, this is what you want to do because you want to have an additional buffer that firms would have to in some sense post. If they want to take on systemic risk where they're going to suffer a lot in the stress scenarios. I think an analogy would be that this is like a deductible on insurance. And the more risky the party that wants to be insured is, the higher will be the deductible that they'll have to put up.</p>
<p>And they will lose this deductible in case things go bad. That's the sense in which it's an extra buffer that they have to put over and above the price that you have to pay for buying insurance in the first place. Now these kinds of changes, I think, will require a much more fundamental shift in the banking sector. So far the banking sector has been very focused on looking at returns on equity.</p>
<p>Returns on equity invariably go up with leverage when some parts of your leverage costs are essentially implicitly subsidized by the government, as they are in the case of large financial firms. Now as you get more and more constraints on your leverage, this means that becomes not an easy way for the banking sector to generate their equity profits. Which is what they get compensated on, which is what their shareholders care about, even if it is mor,e in the short term, an accounting measure of performance.</p>
<p>So, now for the banking sector to generate value therefore they can't just use leverage. They will have to increasingly use the real fundamentally value of the balance sheet of the bank which we would call as return on assets. They just can't increase their value by simply increasing leverage as easily and by correlating with others, for example, because with they're going to be bailed out in such times.</p>
<p>They will have to instead increase their profitability in a fundamental sense. Maybe they should be looking more at expansions to emerging markets that the growth hedging currently seems to be lying. Whereas from 2003 to 2007 they seemed excessively focused on funding housing in the US sector - an extremely advanced part of the financial sector. and one where it's not clear what the economic returns on this activity might look like over the long run. Even though it seemed very easy to leverage on the US housing exposure because of the variety of regulations that were in place.</p>
<p>So I think these kinds of fundamental shifts are likely to take place. I think that they are going to be for the good overall. Clearly capital should be chasing the highest economic return activities rather than the highest return on capital, taking account of weaknesses of regulation out there. So if we are able to address the weaknesses some of these reforms are proposing, we might just get a more sensible allocation of capital and overall higher economic growth globally.</p>
<p><strong>Viv</strong>: Viral, thanks very much for taking the time to talk to us today.</p>
<p><strong>Viral</strong>: Thank you Viv. It's been a pleasure.</p>

Topics:  Global governance International finance

Tags:  financial reform, stress tests, Dodd-Frank Act, BASEL III

C V Starr Professor of Economics, Department of Finance, Stern School of Business, New York University and Former Director of the CEPR Financial Economics Programme


CEPR Policy Research