Delegate biographies: Learn more about the panelists
| Executive summary
• At worst, the credit crisis will have a neutral effect on FX; at best, it will provide new opportunities in the sector
• The crisis has brought higher volatility, involving difficult adjustments to FX models
• High volatility is to be expected when it is not clear what central bankers are going to do when they are themselves uncertain
• Clients are moving away from traditional mandates towards greater allocations to absolute return
• There is room in different circumstances for both electronic and voice trading |
AP, Intelligence Capital Clearly the credit crunch is the big, generic story. How much of this spills over to foreign exchange in terms of banks’ risk appetite, their involvement in the industry, trading, your counterparty relationships, the services they offer you?
AB, GSAM I’m hoping that it could have positive implications: structured finance is shut and that was a big area of revenue for the banks. Now, the beauty of foreign exchange from a bank perspective is that it doesn’t use up balance sheet like securities businesses, so in that regard it should be seen as a potential growth area. So, far from banks retreating, perhaps they will commit more capital to this area.
DT, Insight Investment Well I go on history and going through a couple of these cycles, I haven’t really ever noticed much of a change. As Andrew says, foreign exchange is a transaction-based service and banks make their money through volume rather than having to warehouse risk, so I don’t see it necessarily impacting foreign exchange much. I think maybe there will be some knock-on effects with prime brokerage and we will see higher credit requirements for hedge funds. But if you’re purely involved in foreign exchange, unless you’re very highly leveraged, then the current crisis shouldn’t really affect you too much.
AP, Intelligence Capital About four or five years ago some people would have talked about the possibility of banks being partly replaced by the ECNs or the equivalent, with the buy side trading with the buy side. Will the credit crunch push more volumes out to the electronic platforms or will it have no effect?
GK, State Street Investment managers have embraced electronic trading for efficiency gains and as a means of slimming down the numbers of people that are required to trade and settle foreign exchange. I don’t think that they view electronic trading as a way of disintermediating the banks. One of the enormous benefits of electronic trading is that it has allowed firms to offer more tailored services, because they are not spending valuable time on quotidian back-office tasks such as booking thousands of trades. Machines do that better in any case. From a management point of view it allows you to focus on building better, higher-quality relationships with institutional investors. Electronic trading is not a threat to banks that have a deep understanding of the needs of their clients.
AB, GSAM But I don’t think we should kid ourselves, the liquidity on the electronic platform mainly comes from the banks and from the algorithms that the banks are running. So electronic trading has been a great help, but it’s dependent on the banks actually providing the liquidity for it.
The volatility environment
AP, Intelligence Capital The crisis has altered the environment from a low to a high volatility regime, with the added strong trend in the dollar and the commodity currencies. How have quantitative and qualitative strategies held up?
JG, JPMorgan AM I was surprised how long it took for quantitative processes to encounter a phase of more negative performance. Carry-oriented processes outperformed during the era of low and falling volatility and this continued into the initial phase of the sub-prime crisis. However, the scale of the spike in volatility, the associated unwind of short yen and Swiss franc positioning and the use of FX as a hedge against underperforming mortgage and corporate credit-related investments ultimately created the conditions for a more sustained phase of adjustment. From a qualitative perspective, I was brought up during a period where you were paid a premium to offset potential capital losses, not to augment potential capital gains, and clearly for the past few years the latter has been the focus. This has led to a bias to high-yielding assets over low-yielding assets, and valuations got to a point where they couldn’t be sustained. The unravelling of that was always going to be messy, and the more widespread use of models may have contributed to this. The flip side is that it makes qualitative processes much more interesting and the diversification that they bring to investment processes is going to be more telling. So, looking forward, I’m quite excited.
AB, GSAM It’s a tough environment for quantitative models, but the quantitative approach to currency management remains valid. Most quantitative models are based on sound economic rationale – things that do matter for currency markets, such as being long carry. The models also use other factors, such as valuation, momentum, and economic performance – all of which will be important in driving currency markets.
DT, Insight Investment Last summer there was a rush of money into high-yielding assets that pushed the premium and volatility on high-yielding assets down, which enticed more money in, because return over volatility was the measure. It’s a virtuous circle until there’s a reverse, and that has happened in credit markets in particular, which has ramifications in currency markets, because the carry trade is simply the currency reflection of the global credit cycle. If you look back over history, you do get periods where carry trades return negative amounts, or negatively, for sequential periods, months, even years. So what’s happening in the currency markets is the same as what is happening in other markets, but with different letters and expressions attached to it.
MP, Hermes Investment Management The high-volatility regime is going to be a difficult environment for models and it will persist for much longer than people think. The low-volatility environment between 2002 and 2006 was an easier environment, especially for carry models.
AP, Intelligence Capital But hasn’t high volatility traditionally been very good for systematic trading and low volatility been a problem for trend-following models? I guess it depends what we mean by volatility?
RC, Harmonic Capital Yes, because the issue isn’t volatility per se, it’s dealing costs. These have fallen as markets have become more efficient, but because of the fall in volatility, the ratio between dealing costs and volatility has meant that – whether it be a trend-following approach or a systematic approach that’s trying to say "this looks cheap and I’m shifting slightly" – you’re eating up your alpha when markets aren’t moving much. So I would say that volatility’s actually very good for trading. Dealing costs haven’t become higher because of volatility, spreads might have widened slightly, but volatility is a good thing. With regard to carry, there’s no doubt about it, for the last two years, up until June, the best-performing sort of blind systematic strategies that you can pick up in a research paper would be a carry-type approach. It’s consistent with a low-volatility environment. It’s also fair to say that carry tends to perform extremely well after you’ve had a setback because everyone panics out of it. However, the low-volatility environment was far harder, you end up going towards a boring strategy like carry.
Of course you can build something that will systematically deleverage you in terms of risk to carry in dangerous periods. Historically there are a number of indicators that we’ve used that are very good at that. However, in August those indicators were not relevant because there was wholesale risk aversion deleveraging. There was definitely a crowding of the trade, whether that was carry or whether too many people starting from a different perspective ended up with the same view.
GK, State Street That is exactly what we saw in the middle of last year, dangerous crowding into a single strategy, carry. In the 14 years that State Street Global Markets has been tracking investment flows into currencies, the biggest and most persistent short positions were in the yen in 1998 and 2007. To measure positioning we use a percentile ranking of six-month flows into forwards. Flow in the zero percentile would represent the biggest ever short position. Flow in the 100th percentile represents the biggest ever long. A flow below the 50th percentile would represent a building short position. The 2007 yen short persisted for 300 days and the 1998 short for 278 days. As we now know, August 1998 and August 2007 turned out to be momentous months in markets. What was interesting in both episodes is that institutional investors were unwinding their short positions ahead of the worst of the turbulence. As late as early July the yen short was in the 12th percentile [flows into yen higher on 88% of previous six-month periods]. But, by mid-August, the short position was closed at an average exit price we estimate at 119.
The yen short unwind had a corollary – a marked switch from a heavily pro-carry bias in FX portfolios. Another way we tracked this shift was through a simple rank correlation between FX flows and yield. At the end of May when there were entrenched short positions in yen and Swiss franc and longs in the Brazilian real and other high yielders, the correlation was close to 50%. A reading above 20% is statistically significant. In essence, yield was the only strategy in FX. But, as early as the start of July, that correlation had fallen to 9%. It wasn’t just the unwinding of the extreme yen positioning that was signalling a change in investor behaviour and style preference. There was a more general unwind of carry strategies that had started before the extreme volatility we saw in August. By then, unsurprisingly, the correlation between flows and yield moved sharply negative.
Structural trends
AP, Intelligence Capital Guy, more generally, what do you think about volatility? It’s clearly at a high level today, but are there structural trends in FX volatility?
GK, State Street Well, the notion that declining volatility year on year in currency markets is an irreversible secular trend has certainly taken a knock. Suddenly, the business cycle, credit cycle and cycles in volatility are fashionable again. There is some cyclicality and it is partly the flipside of crowding into consensus trades. As volatility falls, traders look at their VaR models and they tell them they can take on more risk. Lower volatility leads to greater risk-seeking and more crowded trades. That becomes a self-perpetuating cycle until something breaks it. This time it was caused by a dislocation in credit markets the like of which is unprecedented. Against that backdrop, volatility will remain high. However, one explanation for low volatility does have some credibility. The IMF’s Composition of official foreign exchange reserves (Cofer) database shows that central bank FX holdings rose from $1.8 trillion in 2000 to $6.3 trillion at the end of last year. That is a significant new force for stability.
JG, JPMorgan AM I just want to ask a question. If you’re constructing a model of medium-term FX volatility or a model of the structural level of volatility, what will you have in that equation? I mean, what do we think are the primary determinants? Obviously we know what the short-term causes of that were in terms of the trigger events last year but, rolling forward, what are the factors that keep it high, say, for the next two or three years and create a persistently different regime? Is it inflation volatility being higher, underlying asset market volatility being higher? I just wonder if there is actually a scenario where volatility in other markets remains high but the FX market settles down against expectation and we’re back into a sooner-than-expected carry trade.
MP, Hermes Investment Management Well, I think it’s difficult to imagine vols in other asset markets remaining high but FX vols declining. Another variable is monetary policy. If monetary policy is very transparent and it’s clear what it’s going to be, then you may well have a low FX volatility environment. When you don’t know what the central bankers are going to do, because maybe even they don’t know what they’re going to do, then that’s the environment where you could expect high volatility.
AP, Intelligence Capital One of the things I recall, looking at State Street’s custodial data, is how the flows were really very powerful when investors were long and wrong. When investors were long and content, the flows didn’t seem too predictive but when they were long and wrong, the direction of the exiting flows drove prices.
RC, Harmonic Capital I think that’s part of the way to look at risk aversion. It’s only possible to do if you’re a very, very large prime broker or custodian, but finding a concentration of views highlights the required condition for a burst of volatility.
AP, Intelligence Capital And it would imply, from what Guy’s saying, that because we don’t have very large positions at the moment, that in fact the source for sustained volatility in the foreign exchange market may not be there, but FX vol cannot decline ahead of a decline in bond and equity vols. You can imagine a trending market where vols are actually falling.
RC, Harmonic Capital Which is exactly where we’ve just come from: we’ve had two years of the dollar depreciating strongly and vol going down.
DT, Insight Investment But you need to have some sense that the worst is past before volatility can fall. So while there is continued asset price deflation in the US, I find it hard to believe that people’s confidence about the future is going to increase, and therefore people’s confidence in sticking to a view about the future will not be great, so you’ll still have these huge divergences.
AP, Intelligence Capital Also, currencies tend to continue trending until they provoke a change in policy: it’s the first rule of foreign exchange.
RC, Harmonic Capital So – for example - if dollar/yen goes to 80, it may well go back to 120 very quickly thereafter.
Current trades
AP, Intelligence Capital So, have the dynamics of the last six months cleaned out some of the more crowded trades like long Aussie so that you can safely go back long again?
GK, State Street There is a lack of consensus now. Some of the measures that we follow that track institutional investor behaviour suggest that the extreme bearishness we saw in the last quarter of 2007 and January is over. We look at the pattern of cross-border equity flows to anatomize five distinct investment regimes, each representing a different gradation of risk appetite. Back then we were stuck in the longest ever run of Riot Point regimes, the most risk averse. That is no longer the case. However, within most regimes we track, flows are consistent across asset classes. If equity flows are risk-seeking, we generally expect to see a similar pattern for FX flows. At the moment, equity flows are more risk-seeking than flows in foreign exchange. In FX, carry, a good proxy for risk-seeking, is still firmly out of fashion and investors seem more inclined to buy currencies with strong current account surpluses or competitive valuations, as measured by real effective exchange rates.
JG, JPMorgan AM The thing that concerns me about the quant environment is not carry underperforming, it’s how models cope with the vol of vol being high? Most models use vol as an input and other financial variables as an input, and it’s the volatility of those inputs that is going to create problems because of the number of signal changes you will get. It’s not just a case that vol is going to be high generically – the range will be higher – it’s that vol within that range will also be more volatile.
RC, Harmonic Capital You certainly can’t trade with slow models in this environment.
JG, JPMorgan AM Models with a buy-and-hold strategy may perform reasonably well, but they run the danger of trying to trade the mini cycles, and I think they will routinely get those wrong. That will create volatility for more fundamentals-based processes to take advantage of.
GK, State Street A lot of money went into quant strategies during the course of 2007. That money is staying there, but people are working hard at reassessing the inputs for those models. Arguably, everybody was looking at the same things. If everyone has the same inputs, the same outputs are inevitable and that is what can lead to consensus trades and rapid unwinds. It is a good time to start assessing new and differentiated inputs, and I would certainly include fact-based behavioural measures in that mix.
JG, JPMorgan AM This is not a new game. The quantitative process has been around for a long time and people are always looking for different inputs. One iteration was to identify early warning signals, for example.
AB, GSAM Ultimately, though, the same things drive currencies, regardless of the environment: interest rates, current account differentials, monetary policy, growth, the risk environment. What does seem different right now is that the environment switches frequently, and the sell-offs in risk are abrupt, and they reverse equally as abruptly, and then we get big rallies in risk associated with the squeezes in equity markets.
RC, Harmonic Capital We’ve all been through several of these environments before. You don’t know what’s coming next – there could be a bank crisis this afternoon – but it’s not very different to other crises, that’s why volatility’s gone up and it’s not going to go down in the near future. I didn’t enjoy August certainly, but the previous two years were a much harder environment for currencies because there wasn’t enough going on. In terms of volatility, we haven’t entered a dangerous environment, we’ve returned to a normal environment!
Manager mandates
AP, Intelligence Capital What do your typical clients look for when they’re looking at a currency manager? Has that changed?
RC, Harmonic Capital In hedge funds base, absolute return. You can look at correlations and benchmarks, but at the end of the day have you made Libor plus 4%, for example?
JG, JPMorgan AM Yes, I think we’ve definitely seen more interest in higher tracking error products, actually greater risk-taking in the UK market and a little bit in the US as well. At the margin there is clearly interest in the potential for alpha and a higher risk utilization within the asset class as well. So I think that’s quite positive, but obviously we’d like to see more!
DT, Insight Investment Clearly one development is the move away from traditional mandates and having greater allocation to absolute return. And currency hasn’t yet taken as much of that absolute return space as other asset classes.
AP, Intelligence Capital Do you struggle with what is the benchmark for a currency fund?
DT, Insight Investment I think what you struggle with is that there is no beta, there’s no coupon, no long bias. So you have to use other arguments about structure, risk premium and diversification. There’s not even the type of structural story that you can find in commodities, which also lack beta, because you can’t say, for example, that the Chinese use of foreign currencies is going to go up by 80%.
AB, GSAM There is a story in that currency is an inefficient asset class, in the sense that if you look at equities or fixed income, then most of the participants in those markets tend to very much be profit-maximizers. Whereas if you look at the currency market then most of the participants are actually taking part in the market just to facilitate some underlying business; as a result of that, us active currency managers can hopefully extract some of those inefficiencies.
AP, Intelligence Capital Yes, that’s quite unique to currencies. We have a large number of players who are involved in the market, irrespective of any notion of valuations. Of course that should create more opportunities for foreign exchange than other markets.
RC, Harmonic Capital Yes, most people in currencies have got no alpha expectation. It does create opportunities and volatility. Currencies are the ultimate trading market. Maybe you can have a carry bias in currencies but there’s no intrinsic beta that you can lock in.
Discretionary strategies
AP, Intelligence Capital So how do the discretionary investors here deal with this environment?
AB, GSAM The key is in the risk management: trying to make more on the trades that you are right on than on the ones that you lose on. That’s got to be the key for all us discretionary managers, because you’ve got to be very lucky to get more than 50% of your trades right.
DT, Insight Investment We are talking about portfolio construction, and I agree with Richard: in a sense we are all using models in order to make sense of all the data. The difference with the qualitative approach is that the models are not fully formed and documented. Everyone uses models; it’s then a question of how you manage risk. That starts with a view of how much volatility you are prepared to accept in your portfolio. And my experience has been that we have
ex ante expectations that we can have a lot less volatility than is realistic for the returns required. So one has to accept a little more volatility and manage that using derivatives, otherwise you’re stopping yourself out – stop loss is typically a good way of underperforming.
RC, Harmonic Capital To me that is one of the benefits of being quantitative. I can follow 22 markets in real time and try to focus on picking up lots of pennies, if you like, rather than relying on getting one big bet right.
GK, State Street One of the keys to currency performance is identifying the style that works at a particular moment in time. We have identified five styles that we think describe what drives currency returns using the correlation between FX flows and measures of competitiveness, current account, yield (carry), momentum and equity flows. The core of this process is switching between those styles. But, those decisions are also informed by what the investment regime is and whether investors are risk-seeking, risk-averse or somewhere in between. The combination of the Regime map and the style indicators are a very good guide to the investment environment.
RC, Harmonic Capital In that sense, the quantitative approach is very similar to a discretionary approach. In both you build models and in both you find that certain models work in certain environments. The difficult thing to do is to decide what the environment is and therefore which models you should be trading. And I came to the conclusion after many years of making money with the models and making a little bit more money by trying to turn them on and off, if you like, as a discretionary manager backing them, that I was actually reducing the Sharpe ratio, because I was adding more volatility than alpha. So I came to the conclusion that you should spend less time trying to pick the environment and more time trying to find independent models that are alpha producers over the long term.
DT, Insight Investment Identifying the environment, though, is the holy grail isn’t it?
JG, JPMorgan AM That’s why you have to have diversification, you have to have models that will succeed in different regimes.
MP, Hermes Investment Management Well, active management, in my view, is trying to find which model is the right one at a given time. If you get that right more than 50% of the time, that’s the key to alpha generation.
RC, Harmonic Capital As long as you risk manage the times when you are wrong.
Risk management
AP, Intelligence Capital You’ve all identified risk management as a key differentiator of performance. Can you elaborate?
RC, Harmonic Capital Risk is not delivering what you promised. It’s not volatility: nobody cares about volatility if you’re making money; and they don’t care about what your average volatility was last year, they care about how much you lost in August or how much you made in April.
AP, Intelligence Capital The key surely is time horizon? Take mark-to-market. What happens if you are managing money for someone with a five-year time horizon, but the risk management model you’re using produces a likelihood of them losing money tomorrow?
DT, Insight Investment I agree, and that would be a typical mismatch between the real needs of a lot of clients, which is for long-term alpha generation and the tendency to choose managers on the basis of month-to-month performance, which is actually a random variable.
Liquidity issues
AP, Intelligence Capital What are your thoughts on liquidity, especially with regards to emerging market currencies?
GK, State Street One of the key trends in fund management over the last five years has been a move to less constrained investing, both in terms of the ability to go short as well as long and the latitude to add markets to an investable universe. Fund managers are demanding liquidity in markets where previously it was limited, and it is our job as a provider to meet that requirement. Most investors are searching for new currencies to invest in, and the explosion in the liquidity in FX has been led by the periphery as much as the core. I don’t believe that the volumes that are going through the majors are significantly higher than they were in the last couple of years, but the volumes going through the emerging markets are significantly higher.
AP, Intelligence Capital Dale, do you think about liquidity when you’re thinking about foreign exchange positioning?
DT, Insight Investment I think perhaps I should think about it more. For example, last summer we had trouble with the Kiwi and it makes you think. Liquidity is better than it has been historically, and it’s not something that is a big factor in our decisions because the slippage of a point or two is not going to make or break what we do. I think where some of the high frequency intra-day guys have come unstuck perhaps is that widening of spread, which in a micro timeframe has hit them quite hard. But for us it’s not so much of an issue.
JG, JPMorgan AM I wonder if we’re still in a transitional phase. It’s similar to the story in emerging markets: it’s very surprising how easy it is to get stuff done, and I wonder whether in both emerging markets and in terms of wider liquidity we simply haven’t seen the shoe drop yet.
Electronic platforms
AB, GSAM One thing that was noticeable about last summer was that if you went directly to the dealers, there was liquidity. But the liquidity in the electronic trading platforms really switched off. So if you were used to doing large portions of your transactions online, then that would’ve been a problem.
AP, Intelligence Capital Did that make you think twice about your use of those platforms? You want a platform you can rely on.
AB, GSAM Yes, but it reinforced the importance of maintaining existing bank relationships as well.
RC, Harmonic Capital Yes, that’s definitely the case and we knew that before it happened. If you’re doing small bits and pieces, then electronic’s an efficient way to do it, and if you’re trying to perhaps do lots of trades, trying to build up a big position very gradually, maybe you can do it under the radar electronically. But if you need to do a large trade in normal conditions or even a small trade in difficult conditions, you have to do it through voice.
AP, Intelligence Capital Guy, is it viable for the market that in quiet times people use the electronic platforms and shun the dealers, but in rough times when they need liquidity they call up the dealers again. Can the dealers survive that way?
GK, State Street Well I don’t agree that people only want to use the voice when it’s difficult times. Our experience has been that clients like the dealer service in a wide variety of different situations, and that cuts across all the investment management communities: they value the relationship that they have with their key counterpart banks. And the rise of electronic trading hasn’t changed that. There are segments of the market that want price and price alone, but that isn’t true of the majority of investment managers who want more than that. They want a wider definition of best execution, which includes pre- and post-trade services.
DT, Insight Investment What these electronic systems are great for is the large number of small trades that otherwise clog up people’s time.
GK, State Street They provide efficiency, but they’re not the end solution.
AP, Intelligence Capital Momtchil, are you an active user of electronic platforms?
MP, Hermes Investment Management No, because I don’t have a lot of small trades. I will do probably one trade a day or one trade every two days, and for me it’s enough to use voice. And I agree with what Guy said: investment managers don’t just want price. We want additional services like information about what’s going on and ideas generation.
DT, Insight Investment Exactly. You need a good relationship with the salesperson who can give you ideas, help you if there’s a problem, facilitate what you want. And I think to some extent the clients that have gone down the route of computerizing everything and diminishing the role of salesforces have suffered.
RC, Harmonic Capital Well we’re a systematic hedge fund and we use human beings to do the execution, because you’re in control.
Leverage
AP, Intelligence Capital What about leverage? Leverage has clearly been a big thing in the credit markets, and people are always talking about FX carry trades. Has leverage been an issue in the way the foreign exchange markets behaved this year? Have you been worried about deleveraging in foreign exchange positions in the foreign exchange market?
AB, Goldman Sachs AM Not so much this year. I think there was a huge amount of deleveraging last year, first of all in August and then anyone that didn’t delever in August probably did in November. There hasn’t really been a great deal of leverage built up since that period. So investor positions in currencies are relatively small at the moment, especially if you think that the big quant funds were really the dominant players in FX over the past few years and I can’t imagine that they’re back to being fully invested at the moment.
DT, Insight Investment No, you can look at the performance, the price series for some of the big quant funds, it was very striking: the volatility just collapsed, which suggests that leverage disappeared.
RC, Harmonic Capital There are two types of leverage. There’s size of positions and there’s conviction on trades. And vol’s gone up so leverage should go down if you’re going to run the same risk as before, so leverage will naturally shrink. But I do agree with Dale that P&L volatility’s gone down, which suggests that investors have pulled back. But I think it’s also to do with the fact that there is more going on in the majors now. A lot of the size of moves that happened in August made them extreme in a P&L sense. Over, say, 10 years, they’re not unique outliers, but certainly they were extreme in a P&L sense with the other two or three events because of the concentration of moves. So the euro hardly moved in August, but, as we were saying earlier, the Kiwi was a massive mover, and of course a lot of people deleveraged in that environment. Whereas now the moves are right across the G10 currencies. So if you’re diversified in your trading, you’re actually going to see less portfolio volatility. I think it’s more to do with diversification that P&L has been less jumpy.
DT, Insight Investment It’s interesting that P&L vol has collapsed way beyond what you’d expect from the reduction in leverage consistent with rising currency vol.
Volumes down
AP, Intelligence Capital How have your volumes of trading changed given the current environment?
JG, JPMorgan AM I think you have seen a collapse of active positioning in our markets and that has taken volumes down. I speak to people like Guy and in other institutions, and one thing that really surprised me is how not only did you have people collapsing their risk levels of active positions, but they’ve maintained them at low levels of risk as well, despite the fact that we’re all quite excited about the opportunities out there. So risk appetite is low and the majority of managers are less willing to put positions on now than they were.
GK, State Street I would reiterate that we are no longer in a deeply risk averse Riot Point regime. But it’s also true that FX market participants have not rebuilt confidence to the same degree as is becoming evident among cross-border equity investors. In FX markets investors are picking individual stories and sizing their positions carefully.
Given both the fragile state of confidence and the increased cost of leverage it will be a little while yet before we see a return to outright risk-seeking. Falling interbank rates would be a more reliable indicator of a turning point than policy rates.
