Richard Olsen: I think the whole issue of liquidity is something which people haven't yet sufficiently focused on, at least from my point of view.

Claudio Borio: No, I agree. Well, there is quite a lot of interest in market liquidity in the central banking community, even among market partitionists nowadays. Academics are a little bit behind, but that's ....

RO: (laughing) You're used to that!

CB: Yes, well, I mean perhaps because they are less exposed to some of the market developments, so after a lag when things become more important and there's increasing attention, then they start doing some serious work in the area. That's what tends to happen.

RO: But isn't there also a problem, just from an academic point of view, of really coming up with the appropriate definitions and then getting access to the right data?

CB: Well, clearly, there are questions with definition, but I think those are not really the key problems. Data, yes. Data is a problem, although I think it's a problem that, over time, will be reduced. There is quite a lot of data for the equity market. There is quite a bit of data on the bond market but much of it is in the hands of the regulators. For example, Italy, U.K., even the United States. Actually, I think there is a vendor in the U.S. but it's pretty expensive data. There is fair data for the foreign exchange market now. I think over time that is also going to be partly resolved because, of course, of the growth of electronic trading systems.

RO: What always concerns me, if you look at it from a data point of view, there are so many different types of liquidity because you have to distinguish between short-term, medium- and long-term liquidity. And it's very hard to just have one number that represents all that.

CB: I don't know whether I would make the distinction between short-, medium- and long-run. But I would clearly make a distinction between liquidity in normal times and liquidity in distress, in the sense that there may be situations in which liquidity in normal times is to be essentially plentiful. But, in fact, that's just a reflection of it being rather vulnerable, and perhaps there being too much liquidity. I think that's basically what happened in 1998.

RO: How did the issue of liquidity affect the Asian crisis, from your point of view?

CB: Again, it depends on what we mean by liquidity. I think that in many respects, the drying up of market liquidity in several market segments following the Asian crisis was more, it seems to me, a symptom rather than the underlying cause. In the sense that when it comes, for example, to the Asian crisis, I think there it was mainly the usual problems that you have in financial markets with people getting in over their heads, going in, and then all of a sudden realizing that there are problems, panicking, getting out and then, as a result, you had a quite severe drying up of market liquidity. Then you had, of course, the aftershocks. The interesting aspect of all this was that, in 1998 when we had post-Russia and at the time LTCM, there was a time when liquidity in the mature markets - not just, for example, in the FX markets of Asian countries - started coming under severe pressure. That, again, it seems to me was partly the result of the build-up of excessively risky positions in convergence plays, in the security value plays, the build-up of leverage in the system. An illusion, therefore, of liquidity with, for example, a number of hedge funds providing quasi-liquidity services, but without pricing the risk of those services correctly. At the end of the day, you then had the backlash - which meant that liquidity was dried up. The interesting thing about all this has been the persistence of this drying up of liquidity in a number of market segments. Now, if you look across markets, as I say, there are a number of factors; it's not just the fact of the turbulence of 1998. It was also a reassessment of the risk/return tradeoff in market making and the reduction of capital devoted to those activities. Changes in the market in hedging practices - partly as a result of those forces - and consolidation among financial institutions, which has also meant that capital devoted to these activities has been reduced, with credit lines capped, for example, in the FX markets and consolidated. As I said, the disappearance of some hedge funds that at least in normal times were providing liquidity or quasi-liquidity services.

And, then, of course, we have had the actual and prospective reductions in government debt, mainly but not only in the United States. This has meant that there has been a sort of drying-up of liquidity, or at least liquidity has been reduced in a number of government bond market segments of certain benchmarked securities. For example, in the U.S. there has been a move away from the 30-year bond towards a 10-year bond, and so on.

A lot of these factors are, of course, interrelated, and as a result we have also seen rather odd or unfamiliar behavior in short spreads which have now become a favorite instrument for hedging purposes as opposed to government bonds.

RO: One reason why, for me at least, liquidity is such a dangerous issue is the following: There are several layers on which markets operate, so if on any one of those layers there is a lack of liquidity, then it falls over to the next layer and has a negative effect there. So the interconnectivity of markets, combined with the fact that if in one market segment there's a lack of liquidity, it usually means it affects the next level.

CB: Yes. Well, I mean the reason policymakers are interested in these questions, besides the fact that what tends to tie these markets together are trading practices, is also of course related to their risk management practices. One of the key aspects over the last few years has been the extent to which institutions are relying more and more on markets to manage their risks. That, of course, is predicated on the assumption that the liquidity will be there. And to the extent that that assumption does not hold, then of course you can obviously have problems because the premise on which your risk management is based collapses.

RO: So you would say that in earlier days, how did people then act? Did they implicitly say "we don't hedge" and now basically people say "OK, if we deem our risk is too high, we use hedging tools." Is that it?

CB: That's a very stark characterization, but we have seen the development of derivative markets since the late `70s or so in particular. Most of the instruments initially were for the management of market risk, but as a result, of course, they were also generic in credit risk - aspects that, I think initially, were not addressed with the same focus and intensity. Then, in the `90s, mainly with the spreading of credit problems - some of which did occur in the derivative markets but many of which have very little to do with derivative markets per se - institutions have now become much more concerned about credit risk. But, at the same time, there has been - precisely because of that - the development of a credit derivative market which I think in the future is going to be one of the key forces changing the shape of the financial environment.

Therefore, in a sense, more and more of the risk management that would not only have been done simply by maintaining limits to on-balance sheet positions on non-tradable assets, now is being done by hedging through the derivatives markets as well. To the extent that this is happening and liquidity is not there, then you're going to be having problems. Effectively, that's what happened in 1998 to a considerable extent. And that's why people are also revising their risk management practices and trying to see how they can incorporate liquidity into the risk management system.

RO: From your point of view, how different in terms of transaction volumes are the different markets, including the credit derivative?

CB: I think that the volume in foreign exchange markets including derivatives is something about three trillion per day. Globally, fixed income markets markets are in the region of about 120 billion. Equity markets - I think it's 20 billion. But I would have to check these numbers.

RO: What very much interested me before is your statement that basically there's been a shift away from the bond market due to the fact of lower government debt. To what extent does this alone affect liquidity?

CB: It affects liquidity in the sense that as a result of all this, liquidity in the government bond market is declining, particularly in the United States. What's happening in other countries is still unclear. Unfortunately, in the case of Japan, they're having an over-supply of government debt. Government debt is acting as a kind of benchmark for the pricing and, indeed, also for the hedging. Market participants are looking for different markets, both for pricing and for hedging purposes. The expectation, of course, is that this crisis will continue, but that naturally depends also on political factors and what will happen to the deficit. And to the extent that they withdraw from the market, it is reducing liquidity. So this is the usual self-fulfilling aspect of liquidity.

RO: To what extent is another factor - that if there is less government debt around, your amount of collateral that is available shrinks - having an effect? Is that being studied?

CB: Yes, people are studying the impact of all this on collateral. Of course, central banks are very interested in this because they also use collateral in their operations, and with the shrinkage of government debt, the question is, well, what else can they take? There are historical differences across countries. In the United States, the central bank has tended to operate exclusively on government securities in its market operations. In Europe, a number of central banks historically have used more private sector collateral. So, these are factors that are worthy of examination and are indeed being examined.

RO: From your point of view, how do you best describe some of the market dynamics of market liquidity? What are the basic driving forces?

CB: If we take it from a metric level, we define liquidity roughly as ability to pay at short notice with little impact on prime, i.e., lots of value. I would say there are basically three factors that would affect liquidity. First, there is the assumption you must have that there are differing views of the value of what is being transacted at that point in time. We tend to think of it as the present value or whatever of a piece of equity and so on. But, of course, there is a user-specific value which is the opportunity cost of holding that asset. For example, if you have liquidity needs, for you that asset would have a different value than someone who doesn't have liquidity needs. So, that's the first aspect. The second is that, of course, you need to be willing to transact at that price. For example, you would not add much liquidity, if they're very uncertain about the value of what is being traded, so that you will tend not to be very active. And, of course, you have to have the opportunity to transact at low cost, and all this has to do with access to information, access to the trading infrastructure, particular competition in the industry, and so on, for every position, etc. I tend to think these are ultimately the factors that drive liquidity. But what tends to be true in normal times can be quite different in stress times.

For example, what I have in mind is that in normal times you will basically think, okay, you've got a good infrastructure, you've got lots of market participants with heterogeneous beliefs and so on. You also have good access to information and that also tends to generate liquidity, not just the degree of transparency but what kind of transparency one is talking about, but I'm leaving those aside.

At some point, the situation changes and factors like your ability to transact certain volumes requires you to have access to liquidity. For every one who sells an asset, there's someone who buys it. So where does the something to buy come from? The derivative instruments that are being used these days imply that in normal times we'd have low settlement volumes, but under stress you tend to have very high settlement volumes but actually in a non-linear fashion. In abnormal market conditions, factors like concern about the standing of your counter parties have become important because we know that not all assets are traded on a DVP basis, delivery versus payment. Indeed, this is not the case for foreign exchange transactions yet. And, of course, even if they are transacted on a DVP basis, you may need to get the funds in order to be able to buy the assets. So, very much like the case in institutions, in the case of markets, you could have a withdrawal from transactions that is related not so much to the fact that you may think the asset is worth "x" or "y" but simply because you're concerned that your counterparties will not be solvent. And this kind of withdrawal, which tends to happen in acute episodes, is perhaps the most damaging because here it's difficult to see how you can get equilibrium because, regardless of whether you think that the asset is undervalued, you will be withdrawing from transactions.

RO: And then there is a self-feeding process.

CB: That's right. Let's look at what happens when we talk about systemic risk in banks. You judge what your counterparty credit quality is because you don't know the exposure of your counterparty to his own counterparty, and so on. The same can easily happen in the context of markets. And we have seen this in episodes of distress, like for example at the time of LTCM. So I think that sometimes the stark distinction that is made between markets and institutions is wrong. I think that may be true in normal conditions but clearly it is not always true in abnormal conditions, particularly when markets come under severe stress.

RO: But isn't there the added factor that the market maker basically takes on positions and incurs losses for a period of time just to maintain market liquidity?

CB: Well, if you can! I'm talking about situations in which market makers would find it difficult to sustain that. And then the question, of course, depends very much on what kind of risk management systems the market makers themselves have.

RO: From my point of view, the question closely related to the point that you've raised is how does technological change affect all this?

A: That's a very good question. If you go back to the kinds of fundamental factors that might drive liquidity ...if you believe there is reason to trade, willingness to trade, and ability to trade at low cost, then clearly technology has an impact on all of these factors. For example, I might argue that technology reduces the dispersion of belief and in that sense, it may actually reduce the natural liquidity that exists in the sense that people have access to the same type of information and they may elaborate it in similar ways. On the other hand, you can argue that - precisely because people can access a much wider set of information - it would tend to increase liquidity in normal times. The fact, of course, that they get access to information much more speedily means that, other things being equal, they would be more willing to trade, basically because they will be changing views more often and so would be willing to trade more often.

Another aspect about the impact of technology is the fact that competition is being channeled through these technologies, which means liquidity is being split between different electronic markets. We don't know whether this is a permanent condition or a temporary condition. And then, again, there are all the questions about the possibility of fragmentation of liquidity, in the sense that you would trade the same or very similar assets through different platforms. The question is, how would that operate, particularly under stress conditions? In a way, these issues are not conceptually different from those that were raised at the time of the stock market crash of 1987. So, I don't think that we can make general statements. The only possible general statement is the extent to which, if technology tends to shift the balance away from dealer-driven systems towards a market maker system - quote-driven as opposed to order-driven, we're back to the old story as to which of the two markets provides better liquidity. And, again, making a distinction between what might be true in normal conditions and what might be true under stress. I remember at the time of the 1987 crash there was an acute debate as to which of the two types of systems had performed better under stress, but I would say that perhaps the prevailing view is that liquidity in a quote-driven system might be more resilient to stress because of basically what you had said before: the fact that market makers have an incentive to provide the market, as long as they have sufficient capital, a cushion to do so. Their incentive is that they have better information so they have less concern about getting stuck with over-valued assets. For these reasons, people would say those markets are better and, to the extent that we're moving towards order-driven markets, then we're going to have less liquidity, both in normal times and in stress conditions.

Another alternative view is what I said before, that this is true as long as market makers themselves do not come under stressful conditions. Secondly, I would say that as long as they use risk management tools that are in line with their market-making function which, perhaps, we can take for granted.

RO: I have a general question. With the globalization of investment strategies and strong trade flow growth, and at the same time technology with its impact, shouldn't liquidity of markets be far greater relatively speaking than it is today? I mean, my claim is, relatively speaking, we did not have a growth of liquidity and it actually contracted.

CB: Where it might have contracted, I think it is very much for the reasons that I mentioned before: the LTCM crisis in 1998, the recent tradeoff in market making, consolidation among financial institutions, disappearance in some places of those who were providing liquidity services, the excellent prospective reductions in government debt. I mean, all of these factors are indeed playing a role. Still the question is very much in our minds how temporary or permanent all of this is going to be, and whether it's an issue that one should be just interested in, follow closely, or be more concerned about.

RO: From your point of view, what are the measures which your institution is taking to address the issue of liquidity?

CB: My institution is taking a close interest, and so there are a number of investigations into various aspects of market liquidity. I mean, you can simply turn to their web site and you will see that over the last few years there have been a number of studies prepared, in particular by the committee on the global financial system, that have looked into aspects of market liquidity. They include the impact of market dynamics and so on of availability and use of collateral.

RO: Well, some institutions are already trying to compute liquidity adjusted VAR, taking into account the size and perhaps even the potential changes in the size of the spread, the bid-ask spread, etc. There are a number of attempts, again, on the part of institutions to try and integrate liquidity risk in overall risk management, and also to integrate not just market liquidity risk with market risk but to also take into account the relationship between market liquidity risk, market risk and what I would call cash liquidity risk. They are all part of risk management. And there are questions as to how you can do that in a reliable way.

CB: These are indeed questions. What can the community do with their policies? Well, first of all, as I said before, technical systems are very important. They have to function. And of course there have been innumerable steps taken, particularly by the committee in terms of trying to produce systems that are more robust under stress; for example, the move to transparency about risk profiles of individual institutions. One could have argued, for example, that had there been more - well it has been argued, correctly argued - that had banks been more aware of the kind of counterparty risks that they were incurring vis a vis LTCM and other hedge funds, perhaps they wouldn't have lent so much and they would not have observed this enormous backlash that took place in 1998. This is described in the BIS report that looks at developments during the period. And a number of steps have been taken to improve transparency or are in the course of being developed.

One of the questions that I think we haven't quite grappled with, again, in this area of transparency is how we move from transparency about individual institutions to transparency in the vulnerabilities of financial markets. And, therefore, for example, how can we spot potential vulnerabilities in market liquidity? We're still a long way from having answers to that. There has been a beginning of an investigation. Again, if you look at the web site and the BIS report and issues about, for example, how to aggregate the risk profiles of individual institutions. And then how to map those onto trend strategies and risk management systems where we can get a feel for the kind of feedback that certain actions would have on market liquidity and credit exposure once the financial system is hit by some event. I think it's very speculative, but unless we do that I think we will hardly be able to assess market risk on a global level.

RO: I fully agree. So, you've given us a lot of homework. I really want to thank you for this very interesting interview.