Best Performing Hedge Fund Strategies This Past Decade

Based on the HFR indices, here are the top hedge fund strategies over the past decade. We can note that converts have had bad years but were top performers on exit of recessions (2002, and 2009). Distressed investing shows a similar pattern (#3 in 01 and 03 and #1 in 2004, then #4 in 2009). Emerging markets have been consistently good performers these past few years, except in 08. Systematic strategies have done very well in the second half of the decade, even in 2007 despite August of that year; of course, last year is an exception due mostly to trend followers.

(click on chart to enlarge. Source: dataforthoughts)

Palladium and Copper Retraced

Palladium and copper came back in line with their linear historical relationship. The palladium contract was at $462 on 1/20, as indicated in the white panel, and as of Friday had lost about $20, while the copper futures had barely budged.

Butterfly Retraced Friday

Here's the same butterfly as of early afternoon Friday. The reason the downside peak we had seen on a previous post does not show up is simply that this chart shows only market-close levels.

Butterfly on 1-Month Options on Mid-Term Swaps

The 50:50 fly on 1-month options on 1-, 2- and 5-year swaps is current 2.7 standard deviations from its 1-year average.

Palladium and Copper

March contracts on palladium and copper at 0.8 std devs from their regression line.

Follow-Up on RYL v. KBH

Since my post of the pair trade long KBH/short RYL, KBH jumped 14.4% from $13.80 to $15.80 while RYL increased 6.8% from $20.50 to $21.90. With a hedge ratio of 0.88, the gain is $2*0.88 - $1.4 = $0.36 per share shorted.

The Lost Decade of Equities - What About Hedge Funds?

The quick study that follows looks at the past decade (from 1/1/00 to 12/31/09), and compares the total returns of a balanced portfolio of classic asset classes with that of a basket of hedge funds. (I used the HFR Weighted Composite Index as a proxy for a well diversified basket of HFs.)

The "classic" portfolio I picked consists of:

• US Large Cap (Total return on the S&P500), with a weight of 30%
• Global Equities (S&P Broad Market Global TR Index), 20%
• U.S. Corporate bonds (JPM U.S. Aggregate Bond TR Index), 20%
• U.S. Government Bonds (Merrill Lynch 10-year U.S. Treasury Futures TR), 20%
• Commodities (Merrill Lynch Commodity Index TR), 10%

The table below shows the values, normalized to begin the decade at 100, of the different asset classes (indices), together with that of the HF index and that of the classic portfolio described above.



On a non-risk-adjusted basis, the basket of HFs does just a bit better than the classic portfolio. We also note that commodities had a phenomenal run, and that equities of all sorts indeed had a lost decade. Now, if we adjust for risk, we can see that the HF index does considerably better. In the graphs below, we add to the classic portfolio a varying dose of the HF index: 0% corresponds to the pure classic portfolio, and 100% is the HF index alone.

The first graph plots annualized returns against standard deviation of monthly returns:



Surprisingly, the HF index beats the portfolio by about 1 p.p., on average, each year, for a much lower volatility. If we look at Maximum Draw Down as another measure of risk, we get the same result:



In conclusion, despite the turmoil of 2008 and 2009, institutional investors will probably (and should) maintain a healthy allocation to hedge funds.

An Indirect Impact of the Housing Burst on the Economy: Through Workforce Immobility

I keep hearing that the low mobility of people between European countries is a challenge to the Euro zone. Now it looks like the U.S. may be suffering from the same issue, due to the extremely poor liquidity of residential real estate.

In a Bloomberg article today: "The ability to relocate for employment, which helped the U.S. recover quickly after previous deep recessions, is the latest victim of the housing bust. About 12.5 percent of Americans moved in the year ended March 2009, the second-lowest ever, estimates Brookings Institution demographer William Frey, after a 60-year record low of 11.9 percent the previous year."

The impact on the economy is not just theoretical. In the same article: "The stagnant workforce may raise the long-term trend rate for unemployment by 1 percentage point and lower economic growth 0.3 percent a year through 2012, said Michael Feroli, an economist in New York for JPMorgan Chase & Co. It has already contributed to keeping the jobless rate as much as 1.5 percentage points higher than would have been suggested by the depth of the recession, Peter Orszag, director of the U.S. Office of Management and Budget, estimated in July."

The Fed Wants To Absorb Excess Reserves -- Quick

The Fed proposed today to sell term deposits, a few weeks after setting up three-party reverse repos. Both are meant to absorb the banks' excess reserves after the largest monetary expansion in U.S. history. That expansion could cause high inflation, which the Fed typically fights by increasing interest rates. But would that work this time?...

Banks hold so much cash that they would still be able to lend to each other at very low rates even if the Fed started to increase its discount rate. To pump so much cash out of the banking system, the Fed seems to realize it has few means, so it began experimenting with the tri-partite reverse repos -- in addition to more traditional tools like the sale of Treasuries, MBS, and now, deposits.

In my view, the Fed acts like admitting their main inflation-fighting weapon (rates increase) would be ineffective if it had to be used today, and that they're running out of time and out of tools to fix that weapon (i.e., to absorb excess reserves).

KB Home and Ryland

The equities of KB Home and Ryland are strongly correlated. Still, market close on Dec 22 was an outlier.



When overlapping the two plots, we see that indeed Ryland suddenly became rich relative to its comp KBH. It was an opportunity to short RYL and go long KBH, with the hedge ratio given by the slope of the regression line above.

Pershing Square and Hovde Fight Over GGP

Two hedge funds, Pershing Square and Hovde, have been arguing over the value of GGP. Ackman claimed today that the GGP (new ticker GGWPQ) are "undervalued against its peers." Comparing the share price with that of Simon Prosperties and Boston Properties do not support his claim, however.

The Fly Already Retraced

Butterfly on 2-, 5- and 10-Year Swaptions

The 1:2:1 fly on 3-month options on the 2-, 5- and 10-year swaps reached levels not seen in 10 years. Definitely a trade to put on.

Spreads on IG Bonds Remain Elevated Relative to Leverage

Morgan Stanley's "2010 Outlook" report shows that, even if spreads on investment-grade bonds have significantly tightened, they remain elevated given the leverage corporations have taken.

USD-Funded Carry Trade: Return of 27%, YTD

Interesting stats that the FXFB page on Bloomberg offers. A simple buy-and-hold strategy of being long the 3 most-yielding currencies of the developed world, in equal weighting, funded by borrowing U.S. dollars, would have returned more than 27% year-to-date this year, with a Sharpe ratio of 1.37. Of course, this is only for the last 12 months, and we know how the Yen carry trade ended... (The actual return is a tad higher since the tool computes the excess return over the risk-free rate.) The screen shot below shows monthly returns and is thus as of Nov 30.

CRE CDOs: Events of Default Triggered on Senior Notes for Missed Interest Payments

On Dec 8 (sorry for the delay!) Standard & Poor's lowered its ratings on 12 classes from LNR CDO V's series 2007-1 (LNR CDO V), which is a commercial real estate collateralized debt obligation. The downgrades "reflect liquidity interruptions to the transaction" after the outstanding classes did not "receive interest according to the trustee
remittance report dated Nov. 23, 2009." The most senior tranches, Classes A and B, are non-deferrable interest classes and they have each missed interest payments, which "resulted in an event of default (EOD) under the transaction's indenture," so S&P lowered the ratings on these classes to "D."

Ironically, the ratings on the 10 subordinate classes were lowered to a relatively higher grade of CC, because the interest due to the classes may be deferred for many years following the EOD. S&P indicated though that they "believe these classes will likely experience principal losses [..] due to principal losses on the underlying commercial mortgage-backed securities collateral."

Smurfit Stone Bond: The Impact of Mutual Funds

We know HY bonds have had a nice run since March this year, mostly due to retail mutual funds. The Smurfit Stone bond pulled up in Bloomberg below is no exception. It is a favorite of credit hedge funds, but its major holders have been U.S. mutual funds ("MF-USA") and insurance companies ("Sch-D")



Looking at historical holdings confirms MFs have been adding to their position on this bond. One fund increased its holding 10x in Q2!

Exxon's Acquisition of XTO...

... simply brought XTO's price in line with its peers. For example, we had noticed on Oct 20 the dislocation w.r.t. Anadarko and suggested a long/short stat arb trade. The long position on XTO pays off about $3/share, and the short position on Anadarko pays about $6/share.

Safeway and Kroger

Safeway and Kroger dropped like a rock after their downgrade by UBS. But Kroger fell more, relatively, than Safeway, resulting in a data point really far from their linear historical relationship (R2 = 72% over past 12 months).

Hedging Corporate Bonds With Other Issuers

Today's post is about hedging corporate bonds with the bonds or CDS of issuers in the same industry, taking paper companies as an example. First up: Smurfit-Stone and International Paper. The two bonds below have slightly different maturities and slightly different coupons, but their prices follow a nicely linear relationship:



The actual relationship may in fact be slightly convex, or simply the bottom left data points, which date back to late 08/early 09, are anomalies due to panicking post-Lehman markets. So the IP bond may be a good hedge, but of course shorting a cash bond is not always easy, so buying CDS protection on IP would be easier if the relationship with the Smurfit bond is robust -- and it is:



The relationship is similar with a comparable bond of another paper company, Temple-Inland:



This makes Temple-Inland another good candidate for hedging the Smurfit bond.

Debt Distribution and the Term Structure of CDS Spreads

Realogy has a massive wall of debt to refinance by 2013:



Not surprisingly, the cost of CDS protection on the firm's debt peaks at that horizon. If the company can pass that milestone (i.e., refinance this debt before it matures), the markets consider the firm will then have an increasing chance to survive:



Same thing with First Data: The firm has a huge revolver maturing in 2014. The bonds maturing in 2015 are not small either ($7bn), but half of it is PIK-able, i.e., its interests can be paid with more debt.

That explains the usual term structure of CDS protection on First Data: it’s most expensive over the next 5 years, when the 2014 revolver matures. Markets seem to think that, if the company can survive that, they should survive longer, so the cost of protection decreases.

CDS Protection on Campbell

In a previous post, we noticed that CDS protection on Campbell Soup is abnormally expensive relative to that of HNZ. It is not consistent with the yield on its debt either: the graph below shows that the yield on similar-maturity debt (in white) has tightened to 2.38% even though the CDS (in amber) kept widening out:



I did some casual research and found out that fundamentals on the company seem pretty good. Below, and except from recent Barclays research:

Campbell Soup Beats Consensus on F1Q and Raises FY Guidance. Despite relatively weak top-line performance partly due to difficult y/y comparisons, Campbell Soup beat consensus expectations for F1Q earnings and raised its full-year earnings and revenue guidance. Campbell 's debt balance stood at $2.905bn, up from $2.624bn last quarter and $2.756bn a year ago. However, the company contributed $260mn to its US pension plan during the quarter. LTM leverage now stands at 1.8x, up from 1.7x last quarter.

Follow Up on the Swaption Fly

The 1:2:1 butterfly on 3-month, 6-m and 12-m options on 1-year swaps did retrace from its low of -27 bpVol on Nov 20 to about -11 bpVol yesterday.

Follow-Up on the VIX

The VIX did continue to drop -- until Dubai scared the markets on Thanksgiving.

Follow-Up on Masco vs Mohawk

The P&L Wednesday would have been 0.263 * (42.40-41.36) + (14.39-14.09) = $0.573 per short share, or equivalently per 0.264 long share. The P&L today is 0.263 * (41.80-41.36) + (14.39-13.70) = $0.80 for the same pair trade.

Consummer (Food) Staples

In this post, we focus on consumer food staple companies: Campbell Soup (CPB, in pink below), HJ Heinz (HNZ, in orange), ConAgra (CAG, green), Sara Lee (SLE, blue) and General Mills (GIS, yellow). We can see that their CDS's tend to be very close, except for GIS in the first half of the year, and for HNZ, which tightened significantly lately. (The graph is normalized in that all time series start at 100.)



Their equities show similar relationships: they're highly correlated, and the stocks of HNZ, GIS and CPG have had the same return since the beginning of the period under consideration. (The graph is normalized again.)



We can notice, among other things, that the divergence of the CDS on Campbell Soup and HJ Heinz is surprising given that their equities have moved in lockstep. If we "zoom" on the companies, we see that indeed the costs of their CDS's are tightly correlated, but that today's market is an outlier:



This would suggest the following trade: go long (sell protection on) Campbell Soup while shorting (buying protection on) Heinz. Of note: CPB has been a favorite short of hedge funds lately, which may have pushed the cost of its protection higher than justified by fundamentals.

Gold Holding by the Gold iShares ETF

The amount of gold held by the gold iShares (shown in amber) has recently increased, but still has caught up with the increase in the value of the ETF shares (in white). The graph is normalized so that both time series start at 100. The bottom graph shows the ratio of the two series; its peak illustrates the fact that the gold reserves of the ETF are low by historical standards -- but of course, the value of that reserve has increased, which is not reflected here.



Another way to look at the same data, since the beginning of the index on Bloomberg. Gold holdings, in weight, increased by almost 20% while shares jumped more than 34%. Gold holdings are in Troy ounces.

Pair Trade: Masco vs Mohawk Industries

Masco and Mohawk Industries have the same credit worthiness, as illustrated by their CDS graphs:



Their stocks are highly correlated, too: the R-squared of the regression is 92%. However, Masco's stock is far (several standard deviations) from the value implied by linear regression, which would be around $11.



In this pair trade, since the slope of the regression line is 0.263, we would be long 0.263 shares of Mohawk for each share of Masco shorted.

The Same Swaption Butterfly

This is the same swaption butterfly than the one posted Nov 10, which retraced nicely toward its historical average and gave a nice profit. We have a new entry point today.

Cano Petroleum

Resaca Exploitation announced their acquisition of Cano Petroleum on Sept 30. Interestingly, Cano's stock skyrocketed 3 weeks before the deal was announced publicly, on unusually large volumes... Perhaps the SEC will want to have a look?

Bid-Ask on the CDX IG Index

The bid-ask spread on the CDX IG index (shown in amber below) dropped to its lowest level since 07. Interestingly, the VIX (shown in white) seems to be lagging, as usual, so we can expect VIX to catch up and drop. The bottom graph shows the VIX/bid-ask ratio, which just jumped but presumably will revert to its historical mean.

Follow-Up on the Swaption Butterfly

I hope you took your profit on Friday: the 1:2:1 fly on 3mx1y /
6mx1y / 1yx1y swaptions retraced, as expected.

Worst Performers among CDX Sectors

What a House Bill can change... Among massive tightening (even some homebuilders tightened, e.g. Toll Brothers), pharma and healthcare were the two sectors that widened most.

Follow-Up on the Lockheed/General Dynamic Trade

As of 12:20pm today, Lockheed was at $74.95 and GD at $67.05. The P&L on the pair trade since we initiated it on October 28 is:

• Gain on LMT = 74.95 - 68.32 = $6.63/share

• Loss on the short GD position = 67.05 - 64.30 = $2.75/share

The slope of th regression was 0.869, implying being long 0.869 LMT share for each GD share being shorted. Profit is thus 0.869*6.63 - 2.75 = $3.01 per GD share shorted.

1:2:1 Butterly on 1-Year Options with 3m/6m/1y Expiries

The 1:2:1 butterly on swaptions on the 1-year rate at 3m, 6m and 1y expiries is at its second most extreme level in 10 years -- the first one occured in September this year, and was follow by significant retracement. Expect the retracement from the current outlier to be similar, thus offering a significant potential for profit.

Is the U.S. Fed Fund Rate Too High or Too Low?

Is the U.S. Fed fund rate too high or too low? The Taylor Rule is the standard rule of thumb. With the original values, given the economy, the Fed Fund should be negative 2.25%! The screen shot below reflects the original values given by Taylor to the different coefficients. We can still his estimate (shown in blue) is pretty close to the historical values (in white).

According to Goldman Sachs (cited here), this implies that “Since the Fed can’t lower rates to less than zero, the Taylor rule means the central bank has to pump money into the economy through other methods, such as purchases of Treasuries, mortgage securities and agency bonds.”

Using better coefficient values (i.e., so that back-testing shows lower error; alpha and beta = 0.30), the Fed Fun rate would not increase before September 2010. Interestingly, that’s about what Eurodollar futures also imply -- see this older dataforthoughts post.

The Yield Curve as a Leading Indicator of Recessions

The Fed of NY just published its latest stats on the probability of a recession in the next 12 months according to the slope of the yield curve. According to that model, this probability is now almost 0:

Peabody and Commercial Metals Co.

The equity of Peabody has jumped relative to that of Commercial Metals. The graph below shows both plotted in the same graph (top pane), and their ratio (bottom pane):



This phenomenon est recent: the graph below shows that the pair has been diverging from its historical relationship. The blue points correspond to that last 30 days or so, and the yellow points to the previous 5 months.



The cost of protection on Peabody, however, is higher than that on CMC:



This may indicate that Peabody's equity is rich relative to that of CMC. (We can't however exclude the possibility of a change of correlation regime; we would need to look deeper at fundamentals and at recent news to exclude the option.)

Alternatively, these observations may tell us that market players expect a takeover of BTU, which would explain the sudden premium on its equity and its relatively higher CDS spread.

What Banks Charge Consummers -- Continued

From a Bloomberg news release today:
"The rate on a five-year auto loan is 4.49 percentage points higher than on a similar-maturity certificate of deposit that a bank can sell to raise cash, according to Bankrate.com data. That’s up from an average of 2.05 percentage points in the five years through 2007. Spreads between so-called jumbo 30-year mortgages and 10-year Treasuries average 2.71 percentage points, up from 1.58 percentage points in the same period."

What's the Future of Rating Agencies?

It feels like the proverbial rats are leaving the rating agencies’ ship:
Oct 31 -- Warren Buffett's Berkshire Hathaway has reportedly lowered its stake in debt ratings agency Moody's by 2.9% this week.
Nov. 2 -- Fimalac completed the sale of a 20 percent stake in Fitch Ratings for 300 million euros.

Pair Trade of the Day: Lockheed vs General Dynamic

The costs of CDS protection on Lockheed and General Dynamic are strongly correlated and currently are equal, indicating that credit markets do not consider one firm as more as risk of a major event than the other.



Their stocks are strongly correlated too (R-squared of 0.848, i.e. a correlation of 92%):



However, today's market (shown as a red star on the plot) is an outlier: GD appears way too rich relative to LMT. I would short GD and get long LMT in proportions guided by the slope of the regression line. The upside is about $15 per LMT share (if GD stays around $65, LMT should revert from ~$70 to a regression line level in the mid $80's), or equivalently $15 to $20 per GD share on the short side (if LMT stays around $70, GD should, according to this linear regression, revert from $65 to its trend line in the high $40's).