FX Debate: How to choose the right style for alpha

Debate participants

JT, ClientKnowledge From the various sources of data available, we estimate that the overall market in foreign exchange is now about $2 trillion a day of activity, and of that approximately $500 billion a day is alpha-seeking. By alpha-seeking, we mean activity dedicated purely to seeking foreign exchange returns, whether it’s undertaken by hedge funds, by traditional money managers, as a pure asset class, or by sell-side prop desks unrelated to underlying client activity. Harriett, is there a clear distinction between currency overlay and alpha-seeking strategies at JP Morgan Fleming?

HB, JPMAM Our currency overlay business began in the 1980s and was linked to the assets that existed in the underlying portfolio. Clients wanted those specific exposures managed. Over the years, they asked us to start broadening exposure across a wider range of currencies, even though they didn’t have those currencies in their portfolios. That became what I would now call a long/short currency approach. By the end of the 1990s, more and more clients were thinking of separating currency beta – the strategic hedge – from currency alpha. This should be much more effectively delivered without the kind of constraints that exist when FX is linked directly to the assets. I would estimate that 75% of our clients have now changed to less constrained mandates, whereas 25% are probably still keen to have that link between their portfolio and the long and short positions that they can take. In five years’ time everyone will have de-linked.

AE, MGI I agree. The vast majority are alpha-seeking to some extent. It’s a gradual crossover, as clients are aware of the risks that pure currency trading entails and therefore they still impose a lot of constraints. I think it’s going to take five to 10 years to get there, but it’s certainly moving in the direction of pure alpha strategies.

PP, DynexCorp We have demonstrated to clients seeking portfolio overlay that with a beta strategy they limit themselves to only one side of the market. We convinced them that by pursuing alpha-only as a ‘pseudo-hedge’ – in using both sides of the market – their benefits from the currency market may be vastly improved.

TP, CSAM Over the last few years, currencies have developed as a separate asset class. They compete with other aggressive-style hedge fund alternative investments, looking to add between 10% and 20% a year as a separate allocation. So, that’s another development that may become more interesting in years to come.

CE, DrKW I think the picture is still mixed. Some clients will let you help them make more money by trying to use options, volatility, correlation or hybrid products. They’ll be open-minded because they understand that you have an advisory role and that you’re here to add value. Others will definitely say, ‘No, I’m just doing the bond or the equity and I really don’t want to look at the foreign exchange’. It is more difficult to convince those clients that they have a risk and that something can be done to mitigate it or benefit from it. SH, FFT&W Yes, despite the fact that people now talk as though it is proven that you can consistently make money in FX, there’s still a lot of scepticism about whether that is really true. Some equity managers fall into that category.

TP, CSAM It depends who you ask and when they’re asked. Between 2001 and 2004 markets trended and lots of investors made money. Then there are difficult years like last year and this year, and people make less money or lose money. But that differentiates those managers with solid performance in different market environments from those who are just trend followers.

SH, FFT&W In my view most foreign exchange investors don’t make money in non-trending environments. The people who do best in that environment are the counterparties. You do need some sort of persistence of trends to have a reasonable chance of generating good money.

TP, CSAM Sure, but you still need a number of different cycles to identify who is good and who is not good, rather than just one.

Choosing a style

JT, ClientKnowledge Which brings us on to strategy. Assuming that investors accept there are returns to be had in FX, how should they go about allocating risk to seek out alpha? PP, DynexCorp The basic choice is between systematic and discretionary. Systematic means mostly trend-following, which at times can be pretty simplistic. Traditionally, systematic traders outperform discretionary traders, but there’s been a recent shift. Over the last two years discretionary traders have been outperforming systematic traders, if measured over 12-month windows. It is uncertain if this tendency of better performances from discretionary trading will persist or if it will turn out to be cyclical.

HB, JPMAM I think you should allocate as much risk as you’re prepared to take to discretionary and as much risk as you’re prepared to take to systematic, and then leave the two things to get on.

AE, MGI The big difference is that the discretionary manager cuts risk more quickly in markets where they’re not making money, and the trend-follower takes longer, depending on what model they’re watching. So, sometimes you see bigger drawdowns. Equally, when they get the markets right you see greater returns. A mixture of both is a very powerful combination.

TP, CSAM Diversification of styles is important and it should be implemented at a portfolio manager level, where a firm has different styles of applying currency analysis and currency management. Clients can diversify by selecting different styles of managers or a mix of managers who have a diversified approach.

HB, JPMAM As a provider of FX return products we have got to be in that blended space to have a long-term, sustainable business model. The problem with having a very specialized style is you can do exactly what you say on the tin and still have a very sharp period of drawdown. When you get a drawdown in an alpha strategy, that’s when the client shuts you down, even though it just happens that your style’s going through a tough period. It puts your whole business on a very risky footing.

CB, SG CIB There are many different ways to seek alpha in a currency market. We’re seeing more and more hybrid traders and hybrid portfolio managers. That is to say, using a model but with a certain degree of discretion,and then adding options on top of that. Their goal is to combine all three to generate alpha.

PP, DynexCorp What is the optimum number of managers? Six? 10? 25?

AE, MGI Most clients are looking for two or three managers for every mandate.

SH, FFT&W Particularly if the managers in question are operating diversified styles within what they’re doing, which many managers are now.

CQ, UBS When we talk about big-plan sponsor allocation you tend to see three or four overlay managers. When it comes to funds of funds, you see between 10 and 15 managers. It’s interesting, because if you market a foreign exchange fund of funds to an equity investor they will say, ‘Only 15? That’s not diversified’. In the equity world, they need 25 or 30 managers.

HB, JPMAM I believe that is because the correlations between currency overlay managers are quite low, even in common styles. That leads to a proportionality in terms of the marginal benefits of increased diversification that means the curve is lower.

PP, DynexCorp Out of these three to six managers, you would probably retain those with the longest history, wouldn’t you? Let’s say the top six change. Do you have to pick somebody new? Maybe this should be a two-step or two-level procedure. Surely the managers’ length of experience would influence whether to retain him or not.

AE, MGI If you’re someone new to the overlay space and you are quite a conservative investor looking at currency management, you’ll probably lean to the firms that have the longest proven track record.

HB, JPMAM There are studies of hedge funds that have discovered that you must invest in the first three years with a hedge fund to get the best overall performance. I don’t think that applies in currency management, because you want to see how people did in October 1998, for example.

JT, ClientKnowledge And within these styles what are the other issues investors should look out for?

SH, FFT&W Leveraging is interesting, because people provide information ratios on currency and then adjust for changes in leverage into their portfolio. So, often the information ratio doesn’t actually bear a lot of relation to basis points or percentage of returns.

HB, JPMAM That’s true. We find a lot of managers just optimize the models from the recent time period. So, if they’d optimized a model for 2002 to 2004, the model would have a natural bias towards long euros. How they formulate the model is an issue if they’re restricted to information from the last two or three years.

BH, Deutsche Bank Lots of models are based on data mining. But there’s a fad for quantitative approaches.

TP, CSAM A quant model may only be looking at the moving averages and trends over the last few years, but there are also quant models that look at a number of microeconomic variables in order to forecast currency moves. That would not necessarily be affected by what a spot has been doing, but by a number of other factors, such as the risk appetite and yield spreads.

From 1999, the market started positioning away from individual discretionary views towards more functional specialists: quant specialists, flow specialists, technicians, micro economists. That helped the buy-side provide the research for the different functions to come up with an overall view. Our model is a blend of quantitative factors, looking at a number of macroeconomic variables rather than just at the past. For example, our quant model had a difficult time in the first half of last year and what saved our performance was the fact that all the other variables were moving differently from the quant model, so the overall performance was diversified.

Volatility strategies

JT, ClientKnowledge What are the volatility trends and how can alpha-seekers benefit from them? CE, DrKW Structurally, the currency client base is selling volatility. Corporate hedgers buy structures in which they sell volatility. Model-driven hedge funds in options mainly sell off volatility: they pick eight to 10 currency pairs, sell eight of them and buy two of them, so they are net sellers. The overlay model guys that do options are also sellers of options. The exotic market is becoming more and more liquid, and I see an enormous number of binary/digital trades going into the market with a binary risk at a very aggressive price. And, when those exotic structures enter the market, they sell off volatility. As structures get closer to maturity, the volatility that the market-maker needs to sell to have a flat position becomes more important. The closer they get, the more vega, the more volatility goes down.

HB, JPMAM The lovely thing about volatility selling strategies is the negative correlations with technical system strategies. Obviously, if you’re in a trend, that tends to be a period when technical strategies are doing well, and it’s probably a time when writing volatility is not paying off. Similarly, if you’re in a very range-bound period, lo and behold, technicals do badly, as we’ve noted for the last couple of years. But writing volatility strategies offsets that. So, there’s a strong business need for technical managers to diversify their drivers. And one of the nice things they can do is sell volatility; and so the correlations play into that equation.

AE, MGI Listening to this, volatility seems an easy one-way bet. In that case, why have returns in foreign exchange not been particularly good in the last two years?

CB, SG CIB In general I would say that options users, especially sellers, have made money over a sustained period of time until a market crisis has occurred. But alpha generation in a systematic volatility strategy can contain a discretionary element. Let’s say you have a model that tells you to sell 70% of certain currency pairs and buy 30% of other currency pairs.You will need to manage the gamma of each currency pair differently by using a discretionary or a model approach.

CE, DrKW I think it is important not to overstate the idea that structurally vols have come down. Volatility means you have movements around a mean. It’s just a function of how spot behaves during that trend. And historical volatility is very much in line with implied volatility.

CB, SG CIB Yes, though sometimes historical volatility can override implied, and when this happens, you often see alpha-seeking players entering into volatility swaps or correlation swaps.

AE, MGI Even investors who are not trading vols need to remember that volatility traders can distort the spot market and affect other strategies as volatility-related gamma traders; and short-term model activity can create a lot of false breaks of technical levels. Often prices then revert back into ranges. Trend-followers or range or momentum players can get sucked into moves that aren’t actually there.

Innovation and growth

JT, ClientKnowledge A lot of the growth in alpha-seeking behaviour could not have occurred without changes on the sell-side. How far has the sell-side facilitated what we’ve seen in the last five years? BH, Deutsche Bank A lot. Banks have provided the infrastructure to allow many more investors, including hedge funds, to be able to trade foreign exchange as an asset class. They have also been heavily involved with the e-trading platforms that have made it a lot easier to transact in currencies and that have made the market more transparent.

CB, SG CIB This is particularly true in derivatives, where clients have been extremely demanding in terms of research, historical databases of prices and volatilities and looking for new products to enhance their portfolio.

AE, MGI I agree. Look at emerging market options where the market – especially the exotic option market – has really developed. Often a larger hedge fund can’t do large volumes in these spot markets because the slippage, especially on a stop-loss exit, can be huge. By positioning through an option structure, the investor can effectively outsource the delta management to the bank, and hence get access to that market in a substantial and significant way. From the hedge fund’s point of view, the bank therefore enables the investor to manage risk and calibrate the downside on positions in what can be more illiquid markets.

CE, DrKW Banks have enhanced their role as risk advisors. They have helped their clients by pointing out exposures and opportunities that they might have missed. For instance, in the late 1990s and early 2000, there were large foreign direct investment flows into the US through mergers and acquisitions activity. Often the acquirers would overlook their FX risk arising from the long dollar position. Banks have also pointed out to private equity firms their currency exposures and ways to manage them effectively.

AE, MGI The banks have had a huge effect; whether it be the website enhancements, volatility analysis, option product analysis, or exotic option pricing and education. In addition, the whole research effort attributed to FX has increased considerably. Banks are increasingly providing analysis toolkits to assist investors in generating alpha from currencies.

CQ, UBS The banks’ investment in technology, prime brokerage and electronic trading has made price dissemination much better. Spreads are getting tiny, so banks are also trying to find out how they can help add value to the managers. With the hedge funds for instance, it’s clear that systems and data provision for building their own models are very important. We also see that people are now requesting things such as help in raising assets, introductions to new types of investments and access to retail, rather than institutional, investors. So, it’s a partnership that is evolving over time.

BH, Deutsche Bank One of our most popular products this year has been what we call FX Select. It allows alpha providers to go onto a platform from where the alpha consumers can pick the alpha providers they want. So, we perform an intermediary role that allows all the buy-side to come onto the platform and then consumers to access that alpha.

CE, DrKW And let’s not forget one of the most important ways in which banks have helped clients seek alpha – providing liquidity. A recent innovation has been banks seeking liquidity directly from alternative investors. If you need to buy X amount of a financial product, instead of going through the interbank market, banks can source their liquidity from these alternative investors who have the opposite interest. The advantages are confidentiality, because it doesn’t go into the market, and cost, because you don’t pay a brokerage. The slippage is minimized too because you do it at a very efficient price.

Choosing a counterparty

JT, ClientKnowledge From our research, we see a growth in number of asset managers using transaction cost analysis in foreign exchange, but overwhelmingly the term is simply used to refer to bid offer spread, not anything else. Do the buy-side use some form of transaction cost analysis?

CQ, UBS Definitely. Cost analysis will become more and more important, not only for currency managers, but for anyone involved in foreign exchange. Transaction cost analysis won’t become something provided at a price, but I think it will be used as an instrument to highlight low cost execution and to highlight other services related to it.

TP, CSAM We have a counterparty scorecard that looks at five different factors. We then allocate the volume of business proportionately across the liquidity providers, depending on the scores. But no one factor alone channels more business into one area.

AE, MGI If you want to trade, would you go to three providers and ask them for competitive quotes?

TP, CSAM No. We would probably go to one provider and get a better fill in the market.

SH, FFT&W So, if a mega bank came in and always gave you a tighter spread, what influence would that be in driving you towards transacting with that bank?

TP, CSAM It would depend on what other value we get from that provider. Trade by trade, of course we go to one provider, but overall, over the years, we tend to share the volume of business across our liquidity providers, depending on our five factors.

HB, JPMAM We do things slightly differently in that we have four traders and it’s their job to get best execution for our clients. We monitor this by looking at what mid-market was at the point that we went in. We then have a table of different expected transaction costs for different sized trades and we measure how well they did against that particular trade. We’re fiduciaries and that’s what we’re supposed to do. But we also do a more qualitative ranking that helps us differentiate between banks. We give them grades A to D in terms of how well they’re doing on service and research, then we give them the feedback every six months.

PP, DynexCorp Do you think that some liquidity providers are consistently better than others? We tend to think so.

SH, FFT&W Haven’t you noticed that it varies? It goes in cycles of 18 months, so it’s not static over five years.

CQ, UBS There are certain banks that have certain strengths in certain markets, but frequently now, banks go into client meetings and say, ‘If you want to trade this currency pair, you would do well with us, but another bank would do better with that particular pair’.

CE, DrKW Absolutely. We only offer books where we have strong research, and if a client is looking for everything we tell them it won’t work. It’s tailor-made, it’s got a cost and we expect a return on it. It doesn’t mean that you don’t have to be transparent, it just means that you’ve got to be very focused.

AE, MGI When you say return, do you mean that when they do an option trade in your particularly strong currency, you would expect to be shown the opportunity to quote on that deal?

CE, DrKW Yes, exactly.

CQ, UBS It’s also important to know why you’ve missed a deal. Banks need to know when they’re being weak in a certain currency. Feedback is important.

AE, MGI If you consistently have fantastic research but miss every single time you’re asked to quote, then there’s an issue there.

CE, DrKW When you provide a service that stems from pricing, liquidity, research and back office, as a manager, I find it very difficult when asked to quote competitively and in a large size, that your client would pass on 0.05 volatility price in the three months.

AE, MGI It’s unfortunate if another bank comes in and prices at 0.04.

CE, DrKW No, because in 0.05 volatility you’re talking about something that is a pip wide.

AE, MGI We cannot go for 0.05. We have to go for 0.04. We will always seek the best price. If, however, we are trading on the idea of a counterparty and they are out by a small margin, we will give them the opportunity to match the best price.

JT, ClientKnowledge Thanos, would you give an extra pip for a bank that was providing research?

TP, CSAM Our guidelines say best execution, and that combines everything. When you go to do the trade and you allocate the share of the business to a provider, if that provider is not giving the best price, then of course you can change, as long as over the year we have the share of business allocated accordingly. If, for example, we decide to do a trade in the spot market, we’d execute that trade with the given provider that seems to have the best price at that point in time.

JT, ClientKnowledge So, we still examine the price in and of itself rather than looking at the ancillary costs related to the execution. Your custodian will typically charge you a booking fee for booking a trade. There is a size of trade below which it is cheaper to trade directly with the custodian, but at a wider spread than it would be to trade away.

SH, FFT&W I’m not sure about that. In that situation it’s not in your interests, because there’s some differential pricing with the custodian, which you would point out to the client. The cost comes out of performance, and the fund manager is very sensitive to the performance.

Get the best from banks

AE, MGI We rate you on sales, trading, research and pricing. How do you rate us and how does that affect the pricing we get? CE, DrKW That depends. Some advisory work can be very intensive and we need to be rewarded. We tend to look at what research we provide investors with – how many investor trips we’ve organized for them, how many conference calls with various government officials and so on. We may come to the conclusion that it doesn’t fit both parties, that the relationship is not a symbiotic one.

CQ, UBS There’s also an issue of differentiating what you offer. Different clients want different things, so you need to understand what they want and how they will reward you. Some will want only liquidity and pricing, so there is no point in trying to push research. If a client requests a certain number of services you provide, you expect the client to be happy and reward you. If the client doesn’t, you first need to ask yourself if you are providing what you were asked to, and then examine the client relationship.

AE, MGI That is not my experience. Harriett says that every counterparty visits her every six months, I can only think of two heads of sales that come and ask me whether we are happy with the sales trading relationship. It would be extremely positive, especially if you got feedback. I’m happy to talk to any of our banks to identify how to optimize our working relationship.

CB, SG CIB Our business model is a sales trading model, so we value the information just as much as the deal itself. Our approach is very much a partnership with our clients. In exchange for a very tight execution, we expect information. We leverage on this information by taking positions on the market.

AE, MGI That’s the model I was brought up on.

CB, SG CIB We think it’s the best way to provide quality execution and a good level of advisory to our clients. We try to gather information and that’s where we get the hedge and make the money. It’s taking positions on the flows of all the information we gather. It’s therefore very important to talk closely with our clients and monitor their performance.

You know when a signal comes in from a system guy that they all pretty much have the same model with the same origin. If you combine that with the flows from the other portfolio managers (global macro, real money, etc.), basically you have a good overview for the next 20 or 30 pips.

AE, MGI So, to make money in the foreign exchange market you must receive as much information as possible from professional clients.

CB, SG CIB Exactly, and we’re just trying to capture the next 10 or 25 pips. I’m not saying it always works, but so far, so good.

The future

JT, ClientKnowledge Given the changes we’ve seen over the past five years, what can we expect as the market continues to develop?

HB, JPMAM I believe alpha is not an infinite commodity, but that capacity is much larger than what we are currently managing. I’m optimistic that we will see continued growth in the developed markets and, further out, we’ll see substantial growth in the liquidity of emerging markets as well.

CE, DrKW I think that the market will become increasingly automated. We already have risk positions that traders do not risk manage themselves. These are traded automatically by electronic algorithms. I also do not believe FX is going to be concentrated in a handful of banks, for credit reasons. Additionally, to diversify credit, banks will also have to take into account transaction costs.

JT, ClientKnowledge And will larger investors become genuine market-makers? Will a limited number of very large alpha-seeking buy-sides lead the corporate and general business to the traditional sell-sides, but make prices to other currency managers and to other banks?

SH, FFT&W I don’t think you’ll get the big currency managers making prices, it’s not their job to be liquidity providers in that way.

HB, JPMAM Well, we don’t have any capital of our own. That would be the major drawback. It would mean I’d have to go to all of my clients and say, ‘I want to start putting some of your capital into this’.

TP, CSAM And then of course you have the credit risk of all the other currencies.

JT, ClientKnowledge Yes, a prime-broker intermediated environment – that is, a model where participants can see a wide range of counterparty prices (generally anonymously,) based on their credit relationships with their respective prime broker, like EBS Prime or like Hotspot – takes away the credit risk. I’m not sure what the proportion is on EBS Prime. Hotspot is moving towards a majority of the matches being between two buy-sides. That’s another area of debate.

SH, FFT&W If buy-side fund managers are more educated, is there still good money to be made out of the customer business from wide spreads? There was an impression that an equity fund manager doing a trade in equities paid a 1% spread.

CQ, UBS Yes, we’ve done some analysis actually and the figures we could get on foreign exchange execution for custodians are quite interesting. They go from 20 to 70 basis points. I think 70 basis points is what an equity manager gets paid and we thought that as we didn’t have access to a lot of accounts maybe there was a bias. But there are four or five companies doing these analyses and they came to exactly the same conclusion.

CE, DrKW And for largest custodians it’s a very, very profitable business.

CQ, UBS When it comes to the asset owners, the plan sponsors, I agree with you. But where the asset owner is the asset management company, I disagree with you, because the foreign exchange tends to be on the back of an equity order. So, you buy Vodafone, you buy IBM, or whatever. It’s a small trade, you have loads of tickets, and, if that particular asset management company doesn’t have a centralized foreign exchange desk, there’s a lag in the execution. So, you’re not only waiting one or two days for that foreign exchange to be executed, there’s also impact on the price.

HB, JPMAM We do everything third party, just for the record. But I can categorically state that us doing everything on a competitive basis is not a service that anyone values. No one will pay us for doing them. We’ve got to do it, obviously; it’s incumbent from a regulatory point of view, but it’s not anything that anyone seems to value.

TP, CSAM That’s right, and in CSAM you tend to find that some clients prefer to deal with their custodians.

HB, JPMAM Yes. ‘Oh, the custodian will do it for free,’ they say.

CQ, UBS In the equity markets everyone is talking about best execution, trading commissions and so on. This has to reach the foreign exchange market. Everyone says: ‘Oh, the custodians don’t charge anything’, but actually they do and it’s very easy to prove it. If you tell your custodian you no longer want to execute foreign exchange with them, they will charge. They will change the package. The price you pay for your fund accounting or whatever will change because they’re giving up that revenue and it has to come from somewhere else.

AE, MGI Foreign exchange prime brokerage is going to get bigger and bigger. Everyone over the next five to 10 years will probably move towards a prime broker’s side, which is why there’s so much money. The banks that started early have probably benefited from the efficiencies of foreign exchange prime brokerage.

CQ, UBS We found that. We did come later than other providers, but it’s really one of our biggest growing areas.

HB, JPFAM We do everything on a competitive basis with third parties. And we do third party and give up to the custodian, without having the prime brokerage. Obviously, you have to be a large manager to cope with that system. But as a customer, it’s best to have a custodian and a fund manager who does good third party execution because then you’re not paying the additional costs of prime brokerage. If you haven’t got a custodian, prime brokerage is your best route, and similarly within any kind of leveraged fund.

CE, DrKW Surely, with electronic foreign exchange you could negotiate with custodians, tell them, ‘Okay, we understand you’ve got to make a living, however this is an electronic platform’. You can agree not only on the spreads, but to spreading according to a volume, and it would avoid those 20 and 70 points Cláudia mentioned.

JT, ClientKnowledge You’ve all been very generous with your time and your thoughts and I thank you very much.