Flintlock Capital

Caution informs ex-Tudor commodity trader’s approach

HAMLIN LOVELL, CFA, CAIA and FRM

The absence of animal metaphor leaves “Tudor Cub” lacking the resonance of “Tiger Cub”, but the alumni of Paul Tudor Jones’ Tudor BVI hedge fund are growing in numbers and stature – and might eventually eclipse Julian Robertson’s huge hedge fund family tree. One Tudor protégé is its former head of commodity research and portfolio manager Steve Mathews, who set up commodity trader Flintlock Capital Asset Management in February, 2010. Mathews was also selected as one of Tomorrow’s Titans: Blue Chip Managers for the Next Decade, in The Hedge Fund Journal survey sponsored by Ernst & Young last June. Paul Tudor Jones may have been the major formative mentor for Mathews’ trading style, but he also names three other key influences at Tudor: gas and energy specialist Jim Pulaski, Spencer Lampert and John Scranton.

Tudor was crucial for Mathews because, whilst his roles at Citicorp and Bear Stearns involved him introducing advanced new analytical techniques, they were focused on analysis and programming, not hands on trading. Tudor was his baptism into trading, although success did not come immediately. For his first four years trading for Tudor, Mathews preserved capital but didn’t produce profits. Paul Tudor Jones tolerated this learning process because money was not being lost (and because Mathews was also building a library of analytical and database resources for Tudor). It was only after technical, flow and sentiment inputs were added to fundamental analysis that performance began to accumulate – and with it, Mathews’ trading capital steadily rose to $85 million from $10 million initially.

Risk management
The philosophy of increasing risk in response to profits and reducing it following losses contrasts with the more aggressive ex-Tiger “doubling down” approach of averaging down on high conviction positions. The more cautious method has continued in the Flintlock fund, and is partly a consequence of the stop loss policy. Flintlock continues to use the hard, no argument, 1.5% position level stop loss agreed between Mathews and Tudor Jones, which was somewhat tighter than the maximum risk tolerances prevailing at Tudor. The Mathews BVI account (part of the giant Tudor BVI fund) never came close to its Value at Risk limit of 3% for a one day, 99% confidence level loss and currently the same VaR metric has been running at around 1% because it has been a challenging environment for trading commodities.

In fact VaR briefly touched zero at the end of October 2010 when they decided to tactically take some profits and go to cash. The measure of the challenge presented by 2010 was that the fund has been stopped out more often than Mathews would like. The high frequency of hitting stop losses will lead Flintlock to revise risk systems, and over time they envisage 99% one day VaR averaging near the 2% level just as it did in the Tudor account. The VaR is calculated from three and six month look-backs, using Investor Analytics alongside internal systems, with dedicated risk manager and MIT math/physics quant Justin M. Nelson doing stress and scenario testing to reach further back to gauge vulnerability to years such as 2008.

flintlock1

Best years
The best two years for the Tudor account were 2007 and 2008, both registering gross returns above 23% (17% net returns, given Flintlock’s current fee structure), which underlines the uncorrelated trading style. Investors will recall that 2008 was one of the worst bear markets ever for commodities with the Goldman Sachs Commodity Index ending down 70%. Mathews has not made a conscious effort to eliminate correlation to commodities: he takes both long and short views, but is not systematically biased in either direction over time. Yet what both years did have in common was high levels of volatility, which might lead investors to conjecture that Mathews’ style is long volatility. To some extent all commodity trading needs some volatility and can benefit from it, says Mathews, but he has never put on a trade based explicitly on a view of volatility. Plain vanilla options may be used to express directional views but emphatically not to take a view on volatility itself.

Worst year
The worst year for Mathews at Tudor was 2003, when he ended down 6%. From a portfolio diversification perspective this was not a worry since almost all other hedge fund strategies had a strong 2003. Whilst Mathews has not shown substantial correlations to any conventional asset classes or hedge fund strategies, one of the highest correlations shown to any index has been dedicated short sellers, partly because they too suffered from the synchronised risk asset rally in 2003. A similar story applied in the first nine months of 2009, when Mathews flat-lined in the “rising tide lifts all boats” market backdrop. Mathews is open about his losing trades, and how they sometimes sowed the seeds for winning ones. A brief excursion into platinum in early 2008 on the long side, where he got stopped out quite quickly, helped to convince Mathews to reverse his book into a short bias targeting copper.

flintlock2Aluminium and copper
Aluminium was Mathews’ best ever trade. He started buying in 2006 and caught a wave of index flow. Mathews started 2008 bearishly positioned
based on deteriorating economic data, but tactically switched to a long stance before returning to his strategically short positioning later in the year. This nimbleness and flexibility is a hallmark of the Tudor trading style, where positions can be cut and reversed and flipped back again several times in the space of a week. Right now mean reversion players argue that aluminium is far too cheap versus copper given the substitutability of the two metals and their historical relative values.

Mathews concedes that there is merit in the view that copper looks technically “outlandishly overpriced”, but also cites his other inputs: the supply of copper is so much more erratic that a mine incident in Chile could easily cause a price spike, flows are dominated by someone who has cornered most of the market, and the first exchange traded fund backed by physical copper is about to begin trading. So Flintlock was reluctant to short copper in mid-December.

Liquidity risk is controlled in several ways. Although Tudor did some over the counter options trading, after talking to investors Flintlock decided against this and also avoids physical commodity trading. Any future return to OTC trading would be contingent upon investor approval. The manager only trades on exchanges, and only then up to 10% of open interest and no more than one third of average daily volumes. These limits currently translate into fund capacity around $1 billion. This does not restrict trading to only the largest exchanges: the fund trades on nine of them, including grains on three exchanges: Kansas City and Canada’s Winnipeg as well as Chicago. Malaysian palm oil, on the Kuala Lumpur exchange, is the most recent commodity to surpass Flintlock’s liquidity criteria. While Harvard mathematician and research analyst James Dollinger-McElligott covered ocean freight at Louis Dreyfuss, neither freight derivatives nor those in iron ore or coal are apparently anywhere near the liquidity criteria. European carbon credits are closest to matching the volume requirements, but Mathews has other concerns about this market. He believes that the cap and trade regime was modelled on the Great Lakes programme for sulphur dioxide emissions, which can be easily attributed to particular polluters because S02 is not naturally present. The problem Mathews sees with C02 is its presence in the ambient atmosphere, which means emissions from a specific company must be inferred from other data such as fuel use. This creates potential for an unpredictable regime and scope for political manipulation.

Proprietary analytics
One analytical tool recreated and modified from Tudor (Mathews could not take the software with him) is Flintlock Reconnaissance Elements or FREs. The label alludes to the military backgrounds of three staff members, two of whom were injured in service, including senior partner Peter Dawkins, erstwhile US chief of Bain’s management consulting business. These injuries entitleFlintlock to be classified as a Service-Disabled Veteran-Owned Small Business. The aim of FREs is to build a jigsaw picture of physical commodity markets based on Mathews’ global network of 2,500 contacts, and also overseas visits to places such as Chinese warehouses, which sometimes corroborates and sometimes contradicts analysis of data published from the Commodity Futures Trading Commission, the US Department of Agriculture, the Environmental Investigation Agency and the Minerals Management Service. Flintlock welcomes international regulator IOSCO’s plans to enforce better transparency in physical commodity markets, because this would make it easier for them to validate data from their own sources.

This physical data gathering can feed into calendar spread trades, such as a 2008 trade that bought longer dated copper at a big discount to short dated copper after a visit to China confirmed reports of mounting inventories. Copper was also the basis for another relative value trade, against gold in 2009. Mathews took the view that gold would benefit more than copper from the Troubled Assets Relief Programme contributing to dollar depreciation, so he bought gold and shorted copper. In early 2008 it seemed as if all the stars were aligned for a natural gas trade: supplies were tight, hurricanes forecast, technicals supportive and sentiment bearish (if most players are positioned on the short side, there is more chance of a sharp spike on short covering). For this trade Mathews took an outright directional view and bought call options, which were sold within a few weeks.

flintlock3

Directional views
Flintlock’s current directional forecasts for commodities are broadly bullish, partly as he expects China’s renminbi currency to continue appreciating – but this constructive stance could change quickly if the data change. Mathews believes that around 20% of the money created by quantitative easing is finding its way into commodity markets. While he is not certain of the transmission mechanism behind this, he does not expect the second round of QE to end before June 2011 at the earliest. The manager sees all currencies depreciating against gold, but sees more upside in silver partly due to an exchange traded fund. Whilst Mathews admits that negative real interest rates are positive for precious metals that have no yield, he also foresees scenarios where interest rates and precious metals rise in tandem. Yet this does not mean he is unambiguously bullish: the fact that consumer credit may be contracting faster than liquidity is being injected into the system leads him to see some risk of deflation. Again, this view could shift swiftly if the data do.

On oil Flintlock is moderately bullish but that does not mean he is a ‘peak oil’ believer – or that he sees a rapid return to price peaks briefly touched in 2008. Mathews argues that ‘peak oil’ is most popular outside the oil industry. As for 2008, he agrees with the view expressed by BP, Shell and OPEC at the time that $75 a barrel was, and still is the right price for oil. The rest of the run up to $147 was, in his view, caused by speculative financial flows, which even now could be adding $10-$15 onto the price of a barrel.

Forecasting, and frontrunning, flows into commodity indices is one area where Flintlock feel they have an edge. With between $180-$220 billion of the $350 billion face value of commodities traded on US exchanges accounted for by indices, these mammoth investment vehicles can move markets further and faster than more active day traders. Analyst Dollinger-McElligott was monitoring flows into commodity exchange traded funds whilst working for Paul Touradji. Mathews thinks that markets discount index buying and selling not just after changes are announced (before they are executed) but in fact many months before the official announcements are made. He claims that the 2011 rebalancing changes announced in November 2010 were alreadybeing anticipated by the markets as early as August 2010.

Analysis of legislative and political factors is another differentiator for Flintlock, and here the manager has actively researched the impact of proposed position limits on the big commodity indices. High oil prices should ensure continuing political pressure for limits on position sizes for commodity exchanges. A straightforward interpretation of CFTC proposals would make it impossible to operate the GSCI or the Dow Jones UBS Commodities Index at their current size. It is more likely that the position limits are redefined in terms of the ultimate end investor in the indices, which would itself be a complicated process. But even then some of the biggest pension fund allocators to commodity indices could easily find themselves breaching the proposed position caps. All of this means that position limits will not happen overnight without a transitional period and with multiple exemptions.

Index product launch
Flintlock will launch its own enhanced commodity index – Flintlock Advanced Commodity Index (FACI) – when seed assets of at least $95 million materialise for the new, daily liquidity, product. It needs this amount of assets to avoid either omitting or over-weighting the tightest contracts such as cocoa. This long only vehicle will differ from other jazzed up commodity indices in two respects: some of the enhancements will be discretionary rather than systematic, and the precise techniques used to enhance performance will remain proprietary to prevent front running by others in the market. But Mathews can disclose that he will be trying to optimise rolling of futures, and will also be seeking to exploit some seasonal patterns in commodity markets. Regarding the rolling, indices that roll futures sequentially sometimes find when forward curves are sloping upwards (in contango), the cost of rolling can outweigh positive performance from underlying commodities, as in 2006. Mathews will also mention that buying commodities in backwardation (downwards sloping calendar curves) and selling those in contango is a strategy supported by strong empirical evidence. In relation to seasonality, both harvest cycles and weather patterns are important for commodity markets. The FLACI(FACI) index will be broadly based on the same commodity universe as the DJUBSCI, which is more diversified than the oil-dominated GSCI. FACI aims to beat DJUBSCI by 3-5% per year.

Tudor has not yet invested because the firm is still in the process of rebalancing its own liquidity profile. However, Tudor does retain an option to take up to 25% of Flintlock’s capacity, although there is still $30 million of room remaining in the lower fees founder’s share class (charging 1.75% and 17.5% instead of the usual 2% and 20%) available on a first come, first served basis. The index product, of course, will have far greater capacity of up to $10 billion and will cost founder investors (the first $100 million) 0.5% in annual fees with subsequent investors paying 0.85%, with no performance fees. Additionally, the Index product will offer daily liquidity.

A priority for early 2011 is another fact finding
visit to China. Beyond that, Flintlock is placing increased emphasis on locking in additional keystone investors, with a particular focus on state pension funds.