| |
Welcome to Variance Capital's monthly newsletter covering key issues of the global hedge fund industry.
If you wish to contribute or comment on a subject in the coming weeks, feel free to
drop me a line.
Virtually yours,
Martin
 Disclaimer
|
|
 |
| |
View from the battlefield
- Assessing the Turmoil in Quantitative Trading in August 2007
|
|
 |
Rishi K Narang is the founder and portfolio manager of Telesis Capital, a quant-focused fund of hedge funds based in California. He has spent over 11 years in the hedge fund industry. Telesis is a boutique firm, launched in January 2005, with five investment professionals and approximately $300mm under management.
The performance of quantitative strategies has been in the press of late, as the outsized losses that have accrued to investors in “black box” strategies have led to the usual spate of Monday-morning quarterbacking from various pundits and soothsayers. This analysis is intended to shed more accurate and comprehensive light on the disappointing, and virtually unprecedented, losses of quant traders from late July through early August 2007. We also discuss our thoughts looking ahead for these strategies.
Misconceptions
The press and many investors have badly misunderstood the events that transpired during a roughly two week period from late July to early August 2007. The three main misconceptions we have seen characterize the losses of quant strategies during this period as being attributable to three causes: an increase in market volatility, a decline in stock markets, and the sameness of quant managers. Markets did indeed experience a substantial increase in volatility over this period, as shown in the chart on the right. |
 |
However, the level of volatility reached was still significantly lower than it was during the best years of quantitative trading (1999-2002, for example). During those golden years, the average rolling 22-trading-day volatility of the S&P 500 was almost always over 1%, whereas this was roughly the peak daily volatility of the S&P during the most recent stress period. The increase in market volatility was associated with larger moves in the stock market, including some large negative days. However, for the period from August 1-9, the stock market was almost exactly flat, while unleveraged market neutral strategies experience an unprecedented 8% decline, as shown below. Finally, the notion that all quant managers are the same, or hold roughly the same portfolios is as ludicrous as the notion that all discretionary managers are the same, or hold roughly the same portfolios. Empirical data relating both to underlying holdings, and strategy returns, even during this period, do not support this contention at all. Bernstein Quantitative Research published a piece in January 2007 demonstrating that the rankings of stocks by almost two dozen of the largest quant managers were actually very differentiated.
A similar piece is slated for release by Lehman Brothers’ research in the coming weeks. Our experience in owning separate accounts with managers is that an amazingly high degree of “position netting,” which is when one manager holds a name long while another holds the same name short, takes place among quants. Furthermore, returns correlate at an average of 0.1 among our quantitative hedge fund strategies, even among those using ostensibly very similar sets of data and analysis techniques. During the stress period also, some quantitative trading styles experienced significantly greater-than-usual upside, and in fact, the spread from the best performing quant fund in our portfolio to the worst performing during August was approximately 50% (i.e., we had one manager up 25% and another down 25%), nearly five times the average spread previously. So it seems that, despite the convenience gained by simplifying the reasons for what happened, we must look to more nuanced causes. There was no monolithic “quant debacle,” but rather a concentrated loss that impacted a specific type of trading, and reached beyond quant trading, also hurting discretionary stockpickers with a value orientation.
“Complex Black Boxes”
Quantitative trading strategies are widely referred to as complicated “black boxes,” conjuring images of a Rube-Goldberg device hidden behind a thick wall of secrecy. However, the most severe losses in quant trading in August are attributable to the very simplest aspects of these strategies. Indeed, it is highly likely, as we will see later, that the troubles of these quant traders came about precisely because their strategies are easily understood by anyone who’s ever paid any attention to Graham, Dodd or their finance professor. Like most decision-making processes, quant trading strategies are input-output models. They take inputs and, using a consistent process, they give outputs. Many quant strategies, and more to the point those that were hurt badly in July and August, attempt to forecast the relative attractiveness of some stocks versus others. In so doing, these “relative-value” strategies must answer many questions, two of which are: what makes a stock more attractive, and relative to what?
In answering the first question, quants turn to the same kinds of principles that any solid fundamental analyst consider. For example, some strategies (more to the point, the ones that were hurt most badly in August) tend to buy stocks that are cheap, given some set of ratios related to information on publicly filed financial statements. Determining attractiveness frequently relates to some pretty well-known factors, such as Price-to-Earnings, Price-to-Sales, and Debt-to-Equity ratios. Some years ago, quants learned to add to the mix some notion of growth or price momentum, to avoid potential “value traps.” These are hardly complex ideas, but simplicity does not imply sameness. Quant strategies in aggregate employ an array of factors every bit as broad as those considered by discretionary traders.
More differentiation and complexity tends to come from answering the second question: to what should we compare a given stock? Should we compare a high-flying internet company to a staid retailer, or to another high-flying internet stock? Many quant strategies take the latter option, forecasting the relative attractiveness of stocks relative to their peers. A great deal more can be said on this subject, but it is not the focus of this analysis. Nonetheless, in more normal market environments, subtle differences between the way managers specifically design and implement their strategies can lead to large differences in long-term performance. Better managers, ostensibly, make better choices in these areas of subtlety, leading to their “edge” over lesser competitors. But in the exercise of forecasting stocks, for the most part, things are refreshingly simple. Certainly, complexity was not the culprit in the recent losses. Rather, the blame here can be laid definitively at the feet of the humble P/E ratio (and its cousins).
Probable Causes of the “Quant Debacle”
In our view, several drivers coincided, leading up to the disastrous recent performance of value-oriented quant strategies. The first is size. The kind of multi-factor stockpicking approach described earlier appeals to blue-chip asset gatherers because its low turnover and longer term investment horizons seem ideal for the placement of large sums of capital. In aggregate, it is likely that hundreds of billions of cash has been invested in quantitative long/short funds and bank proprietary trading desks, much of this over the past two years or so. These kinds of strategies probably control over $1 trillion in gross positions (the value of longs and absolute value of shorts added together). The significant majority of these positions are held in US securities, because the deep liquidity and large number of large cap stocks to choose from allow for sufficient diversification to support the great deal of capital at work.
The second driver seems to be that many of these operators had already been suffering sub-par returns for some time. Many big-name funds with a US focus have been flat or negative year-to-date before the troubles in August. It’s clear enough that, if one is betting on value in the US, it hadn’t been working for awhile. Much interesting analysis can be done on this fact, but for the sake of brevity, we leave that for another day.
A third cause, critically important, is the widespread practice, among banks’ proprietary trading desks and multi-strategy hedge funds, of cross-collateralizing many strategies against each other, including credit trading and longer-term quantitative equity strategies. Credit-based strategies have proven several times in the past 10 years to be far less liquid than they appear during “normal market conditions,” and in July we saw some of them experience spectacular and sudden losses. We know that these strategies were cross-linked with the (seemingly) far more liquid long-term quant strategies already described. A related, but technically distinct, driver of the losses in quant trading risk-targeting, whereby risk managers target a specific level of volatility for their funds or strategies. They hope achieve this “constant risk” by adjusting leverage inversely with the amount of risk their portfolios are taking. The most common tool for measuring the amount of risk in a portfolio is VaR, which incorporates measurements of the historical volatilities of individual instruments, and the historical correlation of these instruments. With models such as these, risk is computed to be higher when market volatility is higher, and/or when correlations among individual instruments are higher. Note that these two phenomena can be causally linked, in that markets tend to get more volatile precisely because they are being driven by some risk factor that also leads to higher-than-normal correlation among individual instruments. As a consequence of using models such as these, the amount of leverage being used in a wide variety of strategies has gone up dramatically in the past few years, since volatilities had been near historic lows until the summer.
What Happened?
It appears that some large multi-strategy hedge fund(s) and/or proprietary trading desks began to de-leverage their portfolios in late July, in response to poor performance in credit strategies and/or higher market volatility (implying lower leverage targets through VaR readings). The de-leveraging began with quant/long-short trading in the US, the most liquid strategy at hand, which also happened to have been underperforming (another common practice among multi-strats and props desks is to reduce exposure to underperforming strategies). As they sold off their longs and covered their shorts, these stocks began to move in adverse ways both to their remaining holdings and to those that overlapped among other managers’ portfolios. De-leveraging led to losses, which led to more de-leveraging and more losses. In short, we witnessed an all-out liquidity crisis in the most liquid markets on earth.
The first wave of liquidation in the US was largely confined to portfolios of cheap stocks on the long side and expensive stocks on the short side. This selling continued and accelerated through the first few days of August, but the impact was confined to a relatively narrow (albeit very crowded) subset of the quantitative trading universe. But the wave eventually became a tsunami, and by August 8-9, virtually every style of equity trading (including many that actually tend to take opposite positions) in most developed markets suffered incredible losses. Still, there was a distinct association between crowding and the magnitude of losses. Stockpicking strategies in the US were the most impacted, followed by those in Japan. European portfolios managed with the same kinds of strategies actually fared less poorly, simply because less liquidation happened there. Still less common geographical regions, and not surprisingly those where the least capital can be deployed, were virtually untouched.
So it was that we saw strategies that are normally completely uncorrelated with each other all lose money together. Note that there is little overlap between the strategies that lost money later and those that set these events in motion. When quant traders see their models behaving in an inexplicable and perverse manner, they tend to liquidate positions first and ask questions later. Perhaps the greatest irony in all of this was that smaller, more boutique quant traders, engaged in less commonplace strategies that had minimal overlap with the more conventional and larger-scale institutional quants, ended up experiencing losses and liquidating their portfolios only very late in the game. As a result, they ended up needing liquidity at the tail end of an already-massive de-leveraging.
A bit more should be mentioned about why so many managers reacted in the same way, namely by de-leveraging and liquidating positions. Early August was a period of exceedingly perverse behavior. Not only were tried-and-true factors not working, but they were actually working negatively. And because the primary discretion employed by a great many quant managers is the decision of when to unwind positions in the event that the models are behaving badly, many quant managers did exactly that, leading other managers with any overlap to experience losses and to de-lever themselves in response.
Looking Ahead
I have been investing in hedge funds for almost 12 years now, and have been involved in quantitative hedge funds for most of that time. No one I know has ever witnessed anything in quant equity trading like what we experienced in August. Obviously, this event raises valid questions for investors and practitioners. First, was there an underlying cause of the poor performance in the US prior to August? If so, has that cause been addressed? If this cause has not been addressed, it is not clear that the poor performance of these strategies will abate anytime soon. We have seen what appear to be smaller liquidations and continued underperformance of longer term quantitative strategies in early September, but the root causes of these observations remain to be properly explored. Second, given that the liquidation was set off by a crisis in the credit markets, which are scarcely rare phenomena, it is entirely plausible that this kind of aberrant behavior may go from being a once-in-two-lifetimes event to being a once-in-five-years event. If so, what is the optimal strategic allocation to quantitative long/short strategies? Perhaps it can be considered like merger arbitrage, which invests in liquid securities, but cyclically is crowded by many practitioners and relatively low barriers to entry. This, too, is an open question, the answer to which remains to be seen.
With that being said, we believe that the issues going forward are largely concentrated to a fairly narrow, albeit still very crowded, range of strategies. We are focused almost exclusively on quantitative trading strategies, and we have never been more bullish on the environment for the types of traders in whom we invest. With some care in the selection of strategies and managers, we believe the quant space remains among the very best options for an investor in a volatile market environment. We further believe that this is a niche in which specific domain expertise and the broad diversification offered by a fund of hedge funds proves well worth the much-maligned “extra layer of fees.” |
|
| |
| |
| |
| |
|
 |
| |
Have GTAAs fulfilled their promise?
|
|
|
Whether fund managers can consistently add value to a portfolio through Global Tactical Asset Allocators (GTAAs) is still a controversial topic with the investment community. Clearly, markets like the ones we’ve recently encountered (which are liquidity as opposed to fundamentally and rationally driven) tend to be difficult for GTAAs, bringing the debate to the forefront once again.
Originally, GTAAs were designed to capture relative value opportunities across countries and asset classes but they have evolved from their early days and embody many different types of strategies. August performance data is a good case in point. Although volatility spiked up during the month, some GTAAs fared quite well (or at least didn’t lose any money). Others on the other hand, weren’t quite as lucky and had drawdowns that went above and beyond the acceptable.
By way of quick background, GTAA models are quantitative in nature and based on sound economic principles. Their sources of return depend on identifying price/value discrepancies using instruments and markets that are global in scope. The strategy typically invests in many different small size bets. The main advantage is the ability of the managers to invest with autonomy and flexibility across multiple sectors and markets. It is very liquid with a disciplined win/loss ratio giving it increased odds of driving performance through time.
There are 2 main identifiable classes: GTAAs which trade intra-asset class and those who trade inter-asset class. Today, for some funds with 8 different modules, only 1 will be dedicated to cross asset class trades. In each class, there are momentum, value and reversion to the mean players to name but a few. Value players can simply look at a couple of metrics (i.e. low P/Book) to determine which markets to go long and which ones to short whilst others will analyze the equity risk premium (i.e. equity yield vs. bond yields). Value investors may also use momentum indicators to avoid value traps.
Unfortunately, quant strategies have quite a bit to answer for. The funds that lost very little (or are flat) in August have barely returned the risk free rate this year whilst some funds shooting for high (or higher) returns suffered greatly. In the end, the holy grail of high Sharpe and information ratios that could be customized to an investor’s risk/reward profile, has been somewhat tarnished. Gone are the days where investors can customize a GTAA overlay onto their beta exposure on a pure risk/reward basis.
This may have to do with the difficulty of finding independent bets. A bit like a casino, a good quant manager works with a modest edge over the other players (if his edge is too strong and obvious, the market will very rapidly price it away). The “house” wins most when the amount of players at the table is high, the number of tables is also high and each player is betting a lot. It’s a bit the same for quant managers where the odds of winning increases with the amount of differentiated bets (i.e. large amount of players at the table) and the frequency of those trades (i.e. high amount of bets). Essentially, the higher the frequency and the larger the amount of differentiated bets, the higher the Sharpe.
At this juncture in the liquidity cycle though, there aren’t that many independent bets (equities, bonds and commodities) in that many regions (Asia, Europe and the US) a manager can play. Not only has globalization created unforeseen relationships between asset classes but when markets correct, correlations all go to 1. Essentially, quant managers are working with a 3 by 3 grid without much independence between the parts. Bonds are globally correlated and equities haven’t shown that much decoupling yet which makes it difficult. Again, when the amount of tables (and players) is low and each player limits the amount he or she plays, the odds of success are quite low.
Returns are therefore quite cyclical and the Sharpe may oscillate between periods of low and high market correlations. Dimensions are limited at present but the degrees of freedom are increasing. As per my previous comments, sources of profit for GTAAs can vary greatly (momentum, mean reversion, value) and over time, as market correlation abates, the value proposition can only be enhanced.
Another recent problem relates to the fact that GTAAs will often be confronted with gap risk. This is always hard to assess, but playing the carry trade (between high yielding and low yielding currencies for instance) is a bit like shorting volatility. It’s fine for a while until there’s a spike and you can lose your shirt. Going long the higher risk asset classes to short the lower risk assets not only flies in the face of a well thought through risk based approach but is compounded by the crowded nature of the trade. Unlike factor based models, quant models using macro inputs don’t offer too many barriers to entry.
It’s a new type of risk management companies have to consider. Although managers may feel comfortable with the liquidity of their markets (and instruments they employ), when everyone is moving in the same direction, liquidity seems to be an elusive concept. Unfortunately this is still very qualitative in nature and hard to assess in a model (i.e. hard to quantify haw many market participants are in the trade and what leverage they employ etc…). August has shown how important it is for managers to try and quantify this new type of risk.
« Either the fund isn’t very transparent and we’re led to believe the mangers have found some type of magic formula to generate 20% returns per annum, or they’re open with their process and it becomes quite apparent the models are quite replicable” confided an allocator in Geneva. “ Inputs such as GDP growth, PPP, Yiled curve slope, historical vs implied volatility, momentum, inflation (i.e. sound principles combined together in a mean reversion strategy) seem to be the norm.” he goes on to add and concludes” Because of this, we don’t know at this stage if we want to be exposed to quants”.
At this stage in the cycle, it may be hard for macro funds (quantitative or discretionary) to fulfill their promise of achieving high Sharpe ratios. Allocators should clearly keep a close eye on this space but tread carefully if the allocation to this space is to be increased.
|
|
| |
| |
| |
|
| |
Latest on the European credit markets |
|
| |
European leveraged credit markets have traditionally been driven by a multitude of events: the macroeconomic environment, balance sheet and income statement fundamentals, supply and demand technicals, and somewhat by sentiment. Defaults were idiosyncratic and in some case industry wide (i.e. the telecoms sell-off and default of 2002-3).
Now it is “all-sentiment, all-the time”: illiquid markets driven by fear of contagion from the US are the main drivers of price.
“One of the arguments I’ve been making for years, without evidence to back me up, is a decoupling of the credit markets between the USA and Europe. Sadly for me, the beta is still one (1)” claims James Ward who runs AXA’s European high yield funds in Paris and leads their global high yield investment strategy. “The reasons for decoupling for me are simple: economic fundamentals between the USA and Europe are different, with the USA containing a higher degree of risk in credit cyclicality. Europe’s value-added export economy of both goods and services has a steadier stream of cash flows because it is more robust and not driven as much by domestic consumer sentiment.”
The current negative credit sentiment in Europe was also exacerbated by seasonal illiquidity: it is no secret that traders have their books flatter in the summer and left for the beach. It’s quite normal that no real money or hedge funds wish to add risk in the current environment of increased volatility unless they have really done their home work (and those that do will easily outperform in 12 to 18 months).
In any event, when traders come back from the beach it was still the same story: no liquidity. Europe did what Americans did once again: we all waited for the FED.
The current dichotomy in the markets is unprecedented, and some would argue that liquidity is being arrested on both the buy and sell sides by the presence on the dealing and portfolio management desks of risk managers. Risk managers called this crisis incorrectly and are now solving the problem by sticking their noses further in the business side of things. The standard joke within market participants is how many feet away from you is a risk manager now?
In addition, liquidity for investment grade financial paper has fallen off the cliff. Dealing desks, operating in this much uncertainty, have driven short-rates very high which only acts to exacerbate liquidity: when an investment grade desk trader is getting his financing at LIBOR+50 there isn’t a trade he can take without negative basis, so he doesn’t trade.
Capital Markets bankers now are in an unusual position. Instead of hitting the road to pitch new deals, they are hitting the road to see end customers (typically a job for brokers) and gauge sentiment for purchases. In the USA, non-deal road shows like this are trial balloons, where the bankers try to feel out the market for what is possible to be placed. “Would you consider a double-B credit, industrial, cyclical to be sure, but non-acquisitive, in a senior secured floating rate note, and would any of your investment grade friends be interested as a yield enhancement?” is the standard line I’ve heard from eight bankers in fourteen days” states Ward.
“I don’t blame them” Ward goes on to add. “The size of the hung bridges is huge, approximately US$500Bln globally, of which easily US$80Bln was destined for Europe, most of it on the leveraged loan side. But with the CLO market imploding on the back of the toxic environment for *any* structured product there is no bid. Of that US$80Bln, 80% was destined for the loan market, and 65% of that was destined for CLOs, and the rest for leveraged credit funds (read hedge funds). Sadly for the banks, the collapse of the European CLO market is operative: these hung bridges cannot clear until that market restarts or the bid returns in some other form.”
One form the banks don’t like is the stressed/distressed hedge fund bid. “We were invited to participate with 20 other platinum accounts to offer a bid for large blocks of loans from a large, Multi-National I-bank that is making headlines daily. We along with (we have heard) ten other accounts offered bids of 90 cents on the dollar, and we were refused.” claimed another manager who preferred to stay anonymous.
“95” is an operative number for the banks. If they can reach that threshold for an auction on the hung bridges, the remaining inventory does not have to be marked-to-market. This is important, because it allows the banks to remain with the capital reserve levels they have now. But if the loans are sold off below 95, the wheels come off the wagon. The bank has to mark their loan portfolio accordingly and the capital reserves are increased. The higher the reserves, the less flexibility the bank has to work out the loans. The bank basically becomes its own distressed portfolio, and an extremely under-levered one.
So we have an impasse: there is a bid for performing, senior, secured loans at 90, but the offer is at 95. The market is not clearing.
Where did this situation come from? Oddly, the seeds of this impasse were written three years ago, as the public/private leveraged loan market in Europe began to experience exponential growth. Leveraged deals that typically had parts that were based on loans, and larger parts financed with high yield bonds began to have the opportunity to only be financed in the loan market at very favourable rates. Standard deals were +150, +175, or +200, so firms like Scwheppes-Orangina by-passed the high yield bond market completely and simply financed their structures with loans.
But these deals had a strange undercurrent; as the loan market proved to be robust, all-loan structures became more and more leveraged. Deals that were once three and four-times levered through bonds became deals that were eight and nine-times levered through loans. Leverage levels through senior secured loans at this multiple are unprecedented.
“It puts the world in a curious inversion” concludes Ward. “In the current environment many real money leverage loan desks prefer subordinated unsecured high yield bonds levered at four-times to the senior secured loans they can get levered at eight times. The spread differentiation between the leveraged loan and bond market reflects this, actually being historically relatively tight in relation to the security levels underlying. The simple fact is the average leveraged loan in Europe is levered eight times at the corporate level (before structure or portfolio leverage) while the average high yield bond in Europe is levered only 4.5 times at the corporate level.”
|
|
| |
| |
| |
| |
 |
| |
Shorting Japan between now and March ’08 is a dangerous proposition by Gavekal research
|
|
|
In recent meetings with clients, we have fielded a surprising number of questions on Japan. And each time, we could not shake the feeling that investors on the other side of the table were ready to throw in the towel on the perennially underperforming market: Japan bulls who bought the Topix Index two years ago have made a measly +12.4% in total returns. Over the same period, Hong Kong bulls registered gains of +81%, Korean bulls +79%, and Singapore bulls +86%. No wonder investors are getting sick of looking at Japan. And of course, the following spate of news has also not helped:
* An uninspiring political situation: PM Abe failed to carry through on the positive élan of PM Koizumi. And frankly, it seems likely that the 71-year-old PM Fukuda, who has just reappointed 13 of the last 19 cabinet members, will prove to be little else than your average LDP operator. It seems highly unlikely that the political spark to set the Japanese market on fire will come from the political side (see our August 28th & Sept 12th Dailies).
* Japanese equities tend to be rather cyclical: And at a time when global growth is weakening, it may not be an opportune time to be tied to market heavy on autos, machine tools, shipping,...?
* Companies have by and large failed to monetize a weak Yen: In the last two years, the Yen has dropped –19% against the Euro and –14.3% against the Won, and yet Japanese margins have not risen nearly as much as they have in Korea or Germany. How can this be? Do Japanese producers not care about expanding their margins?
* Japanese Banks are still underperforming: This does not imply a decisive turnaround is coming (see p. 2).
Despite these longer-term headwinds, however, Japan feels as if it is due a rebound. And this rebound could come from:
* As we approach year-end, managers will be tempted to mark up holdings. With that in mind, we should note that the Mothers Index was bid up +8.1% yesterday (it has been the dog to beat all dogs for the past two years). At the same time, investors who missed out on the Asia boom will be tempted to come in and buy assets that have not yet reached stratospheric levels.
* The economic data is looking better: Japan’s money supply growth is back in positive territory (see p. 2). And in August, export growth to Europe (at +15.6% YoY ) and China (+23.8%) made new highs. Of course, demand from the West is likely to slowdown; but we believe that continued growth in Asia could pick up a lot of the slack (see Can Asia Decouple?).
* Japanese equities now offer attractive valuations. As a result, we are starting to see a surge in share buybacks. In fact, in August, Japanese companies bought back US$6.5bn-worth of shares, a 2-year high.
All in all, we remain convinced that there are more exciting places to deploy capital than Japan. But nevertheless, Japan could see a nice end-of-the-fiscal-year rally; so shorting Japan between now and March seems dangerous to us.
|
|
| |
|
| |
Oil, OPEC and the US dollar by Gavekal research
|
| |
It is a widely accepted truth that the moves of the US dollar will affect the price direction of commodities. In fact, in recent months, the OPEC president has come out publicly with comments on the US dollar’s weakness, highlighting that this was a cause of concern to the Cartel because: “this is having a significant effect on the purchasing power of oil-producing countries.” So how does the weakness of the US dollar affect OPEC purchasing power?
The chart below shows two price curves for OPEC. The blue line represents the OPEC basket price as quoted. The mauve line shows the basket price adjusted by the US dollar index (traded on NYBOT). In other words, the blue line represents the basket in nominal terms, while the mauve line is in real terms.
To think in real terms is certainly not new to investors, and given that OPEC’s revenues can form up to 90% of GDP in some of the member countries, the strength or weakness of the US dollar will have tremendous impact on their respective purchasing powers. Simply put, OPEC countries receive dollars for their oil exports, but are increasingly paying for imports in Euro, Yen, Won etc. The US$ index is thus an approximation of the purchasing power of OPEC.
The fact that OPEC has come out openly to state its mode of thinking is, we believe, quite significant. In 2004, the idea of pricing oil against a basket of currencies was briefly floated by some of OPEC’s economists in Madrid. The international community quickly rebuked the idea, and nothing came of it. Yet the rebuttal did not mean that OPEC had necessarily abandoned the idea. One simply had to think in real terms and apply the strategy anyhow. We believed that this is precisely what OPEC has been doing and we now have confirmation. Indeed, OPEC’s new calculations take into account the US dollar’s strength, as well as the members’ inflation, to derive a real price. We shall leave the inflation bit out of the equation (although it is relatively high in the region, about 14% for the UAE alone). Why? Because in a booming region a lot of input costs are linked to the US dollar (e.g. steel, cement, raw commodities, etc), and we do not want to risk double-counting. So even if OPEC doesn’t price oil in any other currency, the methodology is now officially weighted against such a basket of currencies, including members’ inflation.
The chart below shows the difference between the nominal basket price and the real one, plotted against the US dollar index:
Unsurprisingly, as the US dollar weakened, OPEC lost purchasing power rapidly (almost $15 against the nominal price, in this month alone). The chart below plots the US dollar index, in descending order, against OPEC’s FX gain or loss:
As can be clearly seen in the chart above, there is a remarkable consistency between the OPEC basket price and the US dollar’s movement (in fact, the overall correlation is around 0.85). And given this high correlation, we can state that the oil price itself is partly a function of the US dollar’s direction. We can then conclude that the direction of the US dollar index has a tremendous impact on the real OPEC basket price and, from there, on OPEC’s supply strategy.
In other words, when the dollar weakens, OPEC will seek to keep the market as tightly supplied as possible, so that THE resulting oil price increase will offset the FX loss. Conversely, when the US dollar strengthens, OPEC will tend to relax its grip, because the FX gain will easily make up for the outright oil price decline.
From the chart below, we can see that this relationship also holds true for the international price benchmark for crude:
1- OPEC’s Real Price Target
So we now know that OPEC targets the real, not nominal, basket price (even if its stated aim is only to ensure supply/demand balances). In January 2000, OPEC announced a price target of $25 per barrel with a $6 variance, just as world oil benchmarks were pricing in the vicinity, and the Euro was trading at 1.000 to the dollar. The basket, whether nominal or real, remained broadly within that range until Q1 2004:
For the sake of consistency, we will keep working with the real price. Starting in Q1 2004, the oil markets entered a powerful rally. This rally kept going long enough to make OPEC’s range de facto redundant. To date, OPEC has yet to declare the price in a new range. However, given our analysis and understanding, we should expect OPEC to have a price in mind, and that this price is in real terms. So how can we find out this new price range?
To discover what price OPEC is targeting, we have to look at the price distribution of the real basket price from the time the rally started in Q1 2004. After all, given that the world has come to accept that oil prices have entered a new and higher range, it makes perfect sense to apply the same thinking to OPEC’s new
price target. This exercise yields the following chart:
As shown above, we find a mean real price of US$44 (median at US$47), with a US$10 standard deviation. For the same period, the price distribution of Brent had a mean of US$57, with a US$12 standard deviation, while the US dollar index had a mean of 86 with a 3-point standard deviation. As we see it, this implies that
every full point change of the dollar index brings a US$1.66 change in the price of the real OPEC basket, or US$2 for Brent. Having said that, changes in the oil price are obviously more than just changes in the US dollar, and in any case the relationship is bound to be exponential rather than linear (i.e. fear that the loss of
purchasing power will increase as the US dollar goes down, speculative activity will increase, etc).
This is precisely what the last chart shows. Over the past three years, as the US dollar declined, oil prices entered into a powerful rally, as the impact on OPEC’s real basket price became much greater:
During the 1990s, with oil prices around US$20, any full point change in the US dollar index would have had a 30 cents impact on OPEC’s real basket (in fact, the relationship was surprisingly symmetric). However, since Q1 2004, the US dollar’s impact has changed to US$1.66 per barrel; an added explanation would thus be to say that, as prices increase, so does the need for compensation.
Maybe we should listen carefully to OPEC’s own words: “A price of US$60 to US$65 is appropriate for consumers and producers…”, Hassan Qabazard, OPEC’s head of research was recently quoted as saying. According to our previously stated rule of thumb, this would add up to US$47 for the real basket price.
In the end, this means that any forecast on oil now seriously needs to incorporate how the US dollar will do. This consideration is particularly important today given that the US dollar is at a very important technical crossroads.
But of course, it will not be the only determinant. We have recently argued that, short of a disastrous hurricane season, the oil market’s upside is very likely over. One of the reasons is that the market is now once again fully backwardated (and this should remove all incentives to pay for storage). Inventories are ample in the US and OECD, while gasoline, which had been major factor behind this summer’s rise in crude, is now in full downtrend. Finally, speculators, despite record length, did not manage to break a new high in crude, and liquidity conditions are now undeniably tighter than they were just a few weeks ago.
Overall, we like to think that an upcoming correction in energy prices will further help the US current account deficit to improve, which could ease some investors’ concerns on the US$. And as the US$ rallies, OPEC’s concerns price concerns will ease and oil would then be able to ease further.
|
|
| |
Recommend this newsletter!
DISCLAIMER
Unless you have a personal relationship with us, you are being sent this e-mail as a representative or employee of an institutional investor. This email and its contents are confidential and may only be used by the intended recipient. Under no circumstances may a copy of this email be shown, distributed, copied, forwarded, transmitted, or otherwise reproduced or given to any other person. If you received this message in error or are not the intended recipient, you should destroy the e-mail message and any attachments or copies. Please inform us of the erroneous delivery by return e-mail. Thank you for your co-operation.
Variance Capital Management is a consulting company and this e-mail and any attachment cannot be perceived as an act of solicitation. No information shall constitute, or be construed as, an offer to sell or a solicitation of an offer to acquire any security, investment product or service referred in any jurisdiction where it is unlawful or where the person making the offer or solicitation is not qualified to do so or the recipient may not lawfully receive any such offer or solicitation. It is the responsibility of any prospective investor to inform him or herself of, and to observe, all applicable laws and regulations of relevant jurisdictions.
|
|
 Delivering digital for the financial industry |
|
