Market Neutral ETF plays..

One can argue whether Fundamental Indexing really just exploit other types of alternative betas, or is delivering Alpha. One could also decide not to get to academic and just be happy we now have the oppertunity to construct Market Neutral trading strategies. These can be found accross many ETFs, and may be due to difference in index weighting ie. market cap vs. equal-weight or fundamental weighting the same basket of stocks.

RevenueShares fx. weight the S&P 500 shares based on revenue, where as S&P 500 is market-cap weighted.

Below some evidence of RevenueShares Large Cap ETF (RWL) out-performing S&P 500 (SPY). Same works in small-cap land.. RevenueShares Small Cap ETF (RWJ) against iShares S&P Small Cap 600 ETF (IJR).

OIL ETF&ETNs

Building commodity tracking ETFs can be difficult. People argue that such ETFs build on stocks like the CRBQ from Alps/Jefferies are not a commodity pure play. On the other hand ETF build upon futures has leakage at each roll-over during contango.

Below is a graph of different ETF and ETNs build to track Crude Oil. What I did was to index the evolution of the individual ETFs and WTI crude since 02.01.2009. You will see that the DBO seem to do a better job at tracking WTI than the more well-known USO or USL. (The ETN OLO is on par with DBO but has a higher expense ratio and ETNs has counterparty risk.) I will be interesting to see if this will reverse during backwardation.

Expense Ratio:

DBO 0.5%

OLO 0.75%

USO 0.45%

USL 0.60%

Disclosure: Long DBO

Emerging market Correlation – UPDATE!

Please find below an update of the correlation matrix on emerging markets ETFs vs. Nasdaq and SP500 proxied by the ETF QQQQ and SPY.

Sirens

Much like Circe warned Odysseus of the dangers of the singing Sirens, we are once again warned against quants and complex mathematical formulas like gaussian copulas, or sophisticated structures build upon them. In hinsight it seem remerkable that quants could get so powerfull in the short span between LTCM folded and Lehmann. The profilation of securization and the hedge fund sector surely has meant an increasing demand for these highly interlectual academic minds, some of the best going to Rentec running black-box schemes, we will never fully comprehend.

However the common denominator in financial crises through time and going forward, is the human nature of greed. We may have specific aspects like the role of incentives and how they shape behavior, complex models taken to far as in the case of CDO-square and cubes, dodgy rating agencies, or accounting mark-to-model guidelines working as catalysts. But it will always be greed, working as the fundamental factor.

So because we are now capable of creating complex constructs beyond our controls, it seams sensible to address the fundamental issue of systematic risks.

Just as a diet is a form of regulation to keep you from eating too much, financial regulations must prevent individuals and institusions from doing what they would otherwise like to do. Like when Odysseus passed the Sirens, we devise regulations to temper natural human tendencies.

But i doubt it will work..greed is the most agile, forcefull energy in the financial construct, and it will eventually flow to and destroy the weakest elements of our systems..by design.

Illiquidity framework

Im struggling with developing a framework on how to access the optimal allocation to illiquid stale priced asset in a multi asset/factor portfolio.

Fundamentally I have found evidence that the illiquidity premium for holding a given factor exposure in a illiquid form, like private equity, real estate, forestry does not compensate for the extended lock-up. Or in real option terminology the value of flexibility is higher, than the return spread illiquidity provide. Now assuming im correct, the only two argument for illiquid investments, is if they:

A.            Provide access to exotic beta exposure, that is not available in liquid form and that may benefit the portfolio as a whole from a diversifying standpoint.

B.             The stale pricing of such illiquid assets, may help to limit drawdowns on the portfolio.

I have earlier consulted with Glyn Holton from Riskchat.com on how we set the risk assessment on illiquid assets equal to liquid assets, so we can compare risk Apples-to-Apples ie. try to back out the autocorrelation, correct for lock-ups via a factor equal to SQRT (t), but Glyns perception is that we should not, as illiquid vs. liquid investments are fundamentally different.

I have thought of using the VaR concept in a ALM framework, using the first or fifth percentile of the funding surplus for illiquid investments, but how do you guys see it?

Europerformance rating on Emergin Markets funds

Weighted symmetry

Weighted symmetry

p – Let P define the probability of obtaining a return above the threshold θ (here defined as zero).

ω – Let ω define the ratio of the expected value above the threshold, vis-à-vis the expected value under the threshold θ. Meaning if we set the threshold to zero return, then a value of ω above 1, implies that the average value of positive returns is higher than the average value of negative returns.

From Mar 10 1999 until May 24 2009

Take a look at the Argentinian Merval

p= 0.5163934

ω = 1.3250000

Rolling weighted symmetry

Below i have plottet the evolution of the parameters p and ω.

It is seen that the probability of a positive return is higher than the probability of a negative return. (p>0.5) and that this probability is increasing over time. Going from 0.51 to 0.55.

Further the expected value of these positive returns are higher than the expected negative return, and actualy decreacing over time. From 1.575to 1.255. Now if only i had access to some managed account data…:-)

 

Then take a look at QQQQ and DIA.

QQQQ

p = 0.5081967

ω = 0.9250000

 

DIA

p = 0.5327869        

ω = 0.8450000

 

Emerging markets Vs. QQQQ

In relation to my responce post on Seeking Alpha

Keep in mind that many emerging markets are somewhat more volatile than developed markets. This will bias any mean-variance portfolio optimization ex.ante. I suggest to treat it as a seperate asset class, and apply covariance shrinkage a la stein estimator or PCA on this sub portfolio. 

correlation of QQQQ in the past 3 month in relation to

EWY – iShares MSCI South Korea Index

FXI – iShares FTSE/Xinhua China 25 Index

EWZ – iShares MSCI Brazil Index

EWT – iShares MSCI Taiwan Index

EWM – iShares MSCI Malaysia Index

ECH – iShares MSCI Chile Investable Mkt Idx

AFK – Market Vectors Africa ETF (AFK)

As can be seen from below summery EWY seem the most Volatile (marked with Red arrow)

Anatomy of a Bailout…

ETF Portfolio

 

I organize my portfolio into a Core and ring of satellite portfolios. I let the allocation to the individual sub-satellite portfolio be a function of a VPPI rebalancing algorithm. Here is a look at a Core as of 09.03.2009..

Because the Core is a stable longterm Core portfolio, I can issue Covered Calls on the individual ETFs. The volatility is still relative high, with VIX at 50ish as supposed to all-time high 80, so the premium is relative high.

The inner ring is a (not so naive) 1/N allocation, while the outer ring is an optimization based on taking the lowest denominator date, (28.02.2008), bootstrap and shrink the covariance matrix. I have done a optimization in the outher ring based on Conditional Value at Risk (Cornish-Fisher) due to the current turbulent markets. 

I realize that optimizing according to CVaR is suboptimal in the long run, but once we have a serie of regime shift signals, I will shift to a different algorithm. Below a display of current the VaR (Cornish-Fisher)  surface.