First let me declare an interest. For ten (mostly) happy years convertible bonds were my professional life, in both research and trading for two large investment banks. I was fortunate to work exclusively in convertibles during the 'golden age' period which I date from the end of the bond market meltdown in 1994 to the 'mini collapse' in convertibles in 2005. I retain a deep affection for this 'asset class' and for the people who still work in it. But the lessons of this year's (2008) catastrophic collapse in convertible valuations and performance ought to be studied by all investors, because they offer fundamental insights into the nature of alternative investing today and have an application far beyond this narrow and esoteric corner of the financial world.
As at 12th November 2008, the index of global convertible arbitrage hedge fund performance had declined a staggering -49.8% YTD (HFRX Index). How could a supposedly 'market neutral' strategy have collapsed so spectacularly?
While there are many factors at play, I offer two key lessons which might be pondered by all investors in alternative assets:
Lesson 1: When Hedge Funds Become the Market, It's Time to Get Out
When I started in convertibles in London in 1995, hedge funds were peripheral players. The London convertibles market at that time was dominated by Swiss private banks and UK listed funds. Almost all funds operated a 'buy and hold', long-only, investment strategy. In Japan, most convertible paper was in the hands of quasi-public institutions, and across the pond in the US, asset managers such as Calamos, Froley Revy, Transamerica and Fidelity owned huge swathes of the market (again on a long only basis).
In this environment, the early pioneers of convertible arbitrage (as it came to be known) emerged and enjoyed significant success. European funds like BAII were early adopters, as was an unknown Harvard student called Kenneth Griffon, who claims to have founded Citadel for the purpose of trading Japanese convertibles from his dormitory room! These early arbitrageurs thrived because they operated on the fringes of a predominantly 'long only' market. They genuinely did hold an information advantage. Only they, together with some of the dealing desks who serviced them, had the quantitative models which offered some insight (I say 'some' for reasons that emerge in Lesson 2) into what the securities were really worth, split into their constituent parts of corporate bond and call equity option.
But of course success begets imitation. Fast forward to 2007 and the convertibles market is unrecognisable from ten years earlier. Many of the old players have disappeared, and those that remain have been completely eclipsed by the new huge convertible arbitrage hedge funds. Citadel, Highbridge, DE Shaw, GLG, CQS etc. are now utterly dominant on every continent; so much so that trading desks virtually give up covering the slow moving and infrequently trading long-only funds that used to support the market. And speaking of trading desks, no self-respecting investment bank is without a 'prop desk' group, comprising a couple of senior traders and (perhaps) analysts who trade many US$ billions of convertible paper for the bank's own account. These are indistinguishable from the bank's hedge fund clients.
Convertible traders of more than five years tenure can start a hedge fund and expect to raise $200-$300 million without too much difficulty. But, significantly, every fund that starts has virtually the same risk-management system, the same model suite and data feed package, supplied by the same tiny handful of software suppliers. Independent convertibles research hardly exists. What little there is comes from the very investment banks profiting from the hedge funds' frequent trading and is little more than marketing for the product itself.
Rather, as the genetic mutations of inbred European royal families did not manifest themselves immediately, this dangerously unstable situation prevailed for a number of years. Funds and prop desks shuffled convertible paper around amongst themselves and the collection illusion of value creation was maintained. A serious wobble occurred in 2005, when performance dived and many funds suffered redemptions, but this was passed off as an unusual and non-recurring event, rather than as the warning siren it should have been as to what could happen if the flow of funds into convertible hedge funds was ever seriously reversed.
The surprising thing is that hedge fund allocators never stopped to consider where the 'arbitrage' in 'convert arb' really was in a market where 85-90% of the paper was held by funds looking to exploit exactly the same 'mis-pricings', using exactly the same models and trading strategies. It was sometimes suggested that corporate issuers of convertibles supplied Alpha (i.e. outperformance on a risk-adjusted basis) to the market by issuing the securities below fair value. This may have been true in the mid 1990s, but like everything else in convertibles, competition for the lucrative fees associated with issuing these 'complex' products caused issuing banks to compete away whatever 'new issue discount' may once have been available – mostly snuffing out this Alpha source before their clients could realise it.
By 2005 it had become apparent (to those 'with eyes to see') that what was really driving the performance of global convertibles were not trends in volatility or credit spreads, but simply fund flows into and out of the convertible hedge funds. When fund allocators were allocating money, as tended to happen in January of each year, valuations richened and performance was flattered. The reverse tended to happen in the mid summer months. Investing in the 'convertible arbitrage strategy' became, in effect, just a bet that others would continue to invest in the same strategy.
Indeed there was evidence by the mid 2000s that convert arb was already 'arbitraged out'. Performance started flagging after the banner year of 2003, the global convertibles index returning a string of poor returns after that: 2004 (-0.14%), 2005 (-5.69%) and 2007 (-0.95%). In fact, according to HFRX, of the past five years convert arb has showed only one positive year, +9.57% in 2006.
Because of the structural instability of the market, when a sizeable negative event eventually did occur there were simply no marginal buyers of convertible paper. Just like a healthy gene pool, markets need participants with different orientations and perspectives in which to thrive. Hedge funds have been very active in oil and other commodity markets in the last couple of years, but here they have traded not solely against each other, but against OPEC producers selling forward, against airlines hedging their fuel costs, against large exploration syndicates, against a wide spectrum of people with totally different perspectives on the market.
This is also true in equities, where despite their higher trading velocity hedge funds are still marginal players against the giant long only savings institutions of the world. Hedge funds' total assets under management are just under US$2 trillion, versus the world's US$40 trillion market. In convertibles, by contrast, some estimates put their total ownership of the asset class at 90% by 2007.
And so (at last!) to my first lesson from the debacle of 2008: if any market or product becomes substantially dominated by 'arbitrage' investors using the same analytical framework, no one can make any money (not genuine Alpha over a sustained period anyway). The strategy is 'arbitraged out' and it is time to move on.
Lesson 2: You Can't Model Everything – And What You Can't Model Can Kill You
Thomas Pynchon wrote of the V2 rockets that fell on southern England in the late stages of WWII that you never heard the incoming scream of the one that got you.
Prior to 2008, convertible traders and analysts did not sit around ignoring all risks; on the contrary, they worried incessantly. Would equity volatility increase or decrease (although this is a red herring – because the textbook is wrong and convertibles are not a 'long volatility' strategy, always performing atrociously in periods of severe equity volatility). Would credit spreads widen or tighten? Would dividends rise or fall, and which convertible issuers would be taken over for cash? Extremely sophisticated models were developed to model these parameters.
But the analysis was myopic. It completely ignored (or ascribed a probability of zero to) the two factors which destroyed convertible valuations in 2008: 1) the temporary bans on short selling and 2) a general collapse in the prime brokerage model which saw investment banks extend high levels of leverage to convertible hedge funds.
In my ten years in the product, I never once heard any investor or trader speculate as to the effect on the market of temporary ban on short selling. It just never occurred to anyone that it could happen. All convertible valuation models relied on the fundamental 'long bond / short stock' portfolio to work. Take away the latter and you have nothing. So on 19th September 2008, when the SEC and the UK's FSA regulators co-announced a temporary ban on short sales on a wide range of 'financial' stocks, the sky fell in on convertibles.
The FSA (sensibly) exempted those holding instruments convertible or exchangeable into the underlying shares but the SEC, crucially, did not. Either way, the damage was done: convertible valuations were decimated. Financial convertibles were rendered virtually untradeable in the US and everywhere convertible premiums were marked down sharply to ascribe a probability to further such bans occurring in the future.
Similarly, no one (to my knowledge) produced a systemic model which showed how much less convertibles were worth as the margin required to hold them at a prime broker increased. This was because in theory a convertible's (or any other security's) value should be unaffected by the cost of funding it. But as we saw in Lesson One, in a market where virtually all participants are relying on leverage to fund positions, the level and availability of funding is vital.
The relationship between prime brokers and hedge funds had been deteriorating all year through 2008, and not just in the convert arb space. As capital began to become constrained at the major prime brokerage houses (Goldman Sachs, Morgan Stanley, Bear Stearns, Lehman etc.) executives began to look again at their prime brokerage clients to assess what they were really getting in return for committing vast swathes of their balance sheets to fund their trades. The answer, in many cases, was: not enough.
The re-evaluation of convert arb accounts was well advanced before September, but it accelerated following the short-selling ban and the disastrous month of September, in which many funds were down in excess of 20%. New rules were hastily drafted, which substantially increased the amount of collateral required to hold convertible positions, and many smaller firms were told that their accounts no longer made economic sense for the banks and would be closed. The withdrawal and re-pricing of leverage to convertible arb funds has probably done as much to destroy the market as the short-selling ban.
Otto von Bismarck called politics the "art of the possible". In finance this might translate into modelling being the 'art of the quantifiable'. The problem with both the serious problems in convertibles this year is that neither really was readily reducible to 1s and 0s.
So Lesson Two is quite straightforward. Investors have got to understand the real risks involved in the securities they are assigning money to, and not just the ones they are told are the relevant risks by investment advisers. The problem that manifested itself in 2008 was really a failure of imagination. All convertible models model securities in a range of stock prices that stretch from zero right up to very high values. But analysts did not stop to think through exactly what the world might be like in the scenarios where underlying equities had fallen by 50 – 80% in a short time period. There is plenty of evidence that stock borrow rates tend to get prohibitively expensive as stock prices get close to zero, and this insight should have been applied more widely to take in the possibility of total short selling ban. But it wasn't.
Likewise, allocators should have been more imaginative in considering under which scenarios they might need to withdraw their capital (which virtually none of them can now do because almost all funds are gated against redemptions). It was as if the fund of funds blithely assumed that they and they alone would be the only ones needing to withdraw cash at any point in time.
Surely the only way to account properly for the risks in a complex product like convertibles where a lot of things have to 'not go wrong' over a very long period is to employ a realistic discount to what imperfect models might suggest is 'fair value'. Only by using such a methodology can the unquantifiable, but clearly very real, risks be discounted at all properly.
A Word on the Future for Convertibles
Nothing in this note should be read as a fundamental criticism of convertibles as securities. They are extremely useful to many companies as a capital raising vehicle and their pay-off profile between bonds and equities makes them a useful tool in the investor's armoury. But they are not magic. And they cannot simultaneously deliver Alpha to an infinite number of identical arbitrageurs.
The ultimate irony is that the convertibles markets of late 2008 once again closely resemble the market I stepped into all those years ago in 1995. Trading is very limited (agency only, for the most part), issuance is very subdued and once again the emphasis is on long only outright investment. Could it be that this re-orientation of the market back to its roots is sowing the seeds for another 'golden age' for convertibles? It is just possible, but I tend to doubt it. A long only fund cannot possibly command the 2% management / 20% performance fee structure to which the hedge fund industry is desperately clinging. My guess is that economics will win out, and at the slightest normalisation of the market the funds and their fund of funds sponsors will be back out there pounding the table for good old fashioned 'convert arb' with good old fashioned fees.
And who knows, it might even work ... for a while.
Stock position: None.