Your portfolio needs fossil fuels - Green Alpha is wrong.

OK - before I get in to this post - I should say that in full disclosure, I am a member of the Sierra Club - I think it is a great organization and I support a lot of their work, but this article that appeared on the "Green Alpha" blog - housed on the Sierra Club website is, quite frankly, a load of ...

Ok - let me back up a bit.   Green Alpha is a socially responsible fund management firm started by the Sierra Club to offer environmentally friendly investment options.   Depending on which product you invest in, you can pay 1-2% management fees (ouch!).

Back to the article - the author claims that investing in fossil fuel companies is a bad idea - from an investment standpoint.   Now I get that the Sierra Club thinks that reliance on fossil fuels is bad for the environment - but are they also bad for your wealth?

If Green Alpha claims that as a socially responsible fund, that they should not invest in fossil fuels - then that's fine, although I do note that they invest in Google, which, in 2011, used as much electricity as would be needed to power 200,000 homes.  But that's a debate for another day.

No, the reason green alpha thinks we shouldn't buy energy stocks is because they argue that Modern Portfolio Theory advocates a portfolio that is better suited to the 1950s than today.

Modern portfolio theory’s asset allocation models were made for and reflect a world where fossil fuels were the only imaginable primary power source.

This is utterly wrong.  MPT's asset allocation models were not made for any specific world.  Heck they'd work in a world where all power came from solar, wind or butterfly burps.   Modern Portfolio Theory states that the efficient portfolio is based on balancing the volatility and expected returns of all available investable assets.   The net result is that the weighting of each asset class in the "optimal portfolio" is that asset's market value.  MPT says nothing about green energy, or fracking, or hybrid cars.

To argue that MPT is somehow flawed is really a feeble attempt to justify an investment strategy that only invests in a subset of assets and then charges very high fees for doing so.

By investing in a subset of stocks, you are going to increase your risk without getting a corresponding increase in return.  In effect, you are giving up some diversification.   This might make you feel better about yourself, but it won't help the Polar bears, and it certainly won't help your retirement portfolio.  On top of that, you are likely to pay way too much in fees for this suboptimal portfolio.

My advice:  If you care about the environment - index (and yes, hold energy stocks), but bike to work from time to time, turn down your thermostat a bit and even donate to the Sierra Club.

How unbiased are financial advisors?

Not very -  according to a well crafted double blind study. (Full paper here).

The study found that too often advisors took clients portfolios and re-worked them in ways that would earn them the highest commissions.

My advice:  Index.  Fees and commissions are very detrimental to your wealth.

Dilbert on active management

via Newmark's door.

The winning portfolio?

Apparently a passive portfolio of 50% bonds and 50% stocks wins out in most markets.  

There are probably a few things going on here:  First bonds and stocks provide great diversification.  Second, because the weights are fixed, there is no market timing going on here.  Market timing, as we know is a fast way to loose wealth.  Third, the stock portfolio is indexed.

In otherwords: diversify, don't market time and index.  Simple.

SPY and GLD correlations and volatility.

Apparently the value of the GLD ETF (an Exchange Traded Fund that holds gold) has exceeded the value of the SPY ETF (which holds stocks from the S&P 500).  All this gold is held in bank vaults in London (which in light of recent riots may not be that sensible).  The total amount of gold held by the fund in these vaults is about 41. 5 million ounces, which is about 1300 tons.  That seems like a lot of gold.

In the article, a hedge fund manager was quoted as saying that
“Gold has become something of a near-perfect hedge for financial assets such as stocks,”
This got me thinking - I wonder what the correlation between GLD and SPY is?

Here are the daily correlations.  Note that GLD has been trading since 2004.
2004-2011: 0.0506
2008-2011: 0.0108
2010: 0.1961
2011: -0.2655

First of all we can see that over the long run, there has only been a weak correlation between GLD and SPY.  In 2010 however this correlation became positive and quite strongly so.   But for the first half of 2011 we can see a strong negative correlation.  While not a "near perfect" hedge, it certainly looks like GLD is moving against stocks.  The problem of course is predicting how long this movement will continue and at what point in time (if ever) the old pattern of a small positive correlation will reassert itself.

Looking at daily annualized standard deviations, (assuming 250 trading days) we see that GLD is a little less volatile that SPY.

2004-2011
SPY Std Dev:  22.6%
GLD Std Dev: 21.0%

2011
SPY Std Dev. 20.00%
GLD Std Dev. 15.24%

Surprisingly SPY is not as volatile as you might think - perhaps a classic case of investors anchoring their beliefs too much on recent market gyrations.

While I don't make prognostications, it does seem likely that a large amount of the demand for GLD may be coming from people moving money out of stocks.  If this is the case, then if and when the stock market starts to rebound, we should see a pretty rapid fall in the price of gold.

GDP or Cap weights?

Fama and French discuss whether you should weight your international equity portfolio based on country GDP or on market cap.   Clearly market cap makes the only theoretical sense.   Country GDP is just arbitrary.

Beta chasing

Via my colleague, Craig Newmark.  Apparently 80% of quants use beta as a key factor in stock screens.

The idea behind "chasing beta" is that high beta stocks have higher expected returns than low beta stocks.  Therefore, if you are chasing absolute returns, then high beta is the way to go.  

The idea does have a few flaws.  First it assumes that there is a direct relation between historic beta and future returns.  While academic studies have shown this to be true, the magnitude of the relation is below what theory predicts.  Second, assuming that you can juice up your portfolio returns with higher beta, also assumes that your benchmark is not risk adjusted.   A risk adjusted performance measure (such as alpha etc) would strip away this higher beta effect.  Finally, why not just buy the low beta stocks (which by implication are being ignored) and then lever up?

Diversification by Omission?

Ken French talks about the effect of omission on diversification.
Key point - its not just all about correlations.  Raw variances matter too.

The idea of bets by omission is a really important one and something I recently talked about with Ron Elmer who is the author of "The 401K Cook Book".  Ron's point is that many portfolio managers who are benchmarked against the S&P 500 hold only a fraction of the 500 stocks in the index.  Perhaps the stocks that they hold are the ones that they feel most strongly about.  Or perhaps they just don't have the man power to examine all 500 stocks.  Either way, the stocks that they don't hold actually represent a bet against those stocks.  In effect they are underweight those stocks.

We should be very careful to think about what we are not including in our portfolios because these "omissions" actually represent unintended bets against those securities.

What is the optimal global weighting system?

Portfolio theory states that if we believe market prices are correct, then the optimal weighting system for a portfolio should be based on market values.   Fama and French tackle the question of what happens when you factor in local market distortions that may provide a home market bias.

The answer is messy.  But at the end of the day, diversification is what counts.

A couple of reposts from the Finance Professor Blog...

Just working through some of the starred postings in my RSS reader...

The Finance Professor  blog has a nice video of Harry Markowitz talking about portfolio theory and another one featuring Gene Fama Jr talking about asset pricing.

Well worth a view.

Pension fund return assumptions.

Pension funds have to make assumptions about the returns that they expect over long horizons. These assumptions then form the basis for how much money the fund should invest. If the fund is run by a state, these assumptions then directly impact the state budget.

Pension funds are pretty much free to come up with whatever expected return they choose. Of course this is a decision of monumental importance. Set it too low and you have to raise taxes or cut current expenditures in other areas of state government. Set it too high and you end up with an unfunded liability.

An MBA 523 student of mine forwarded me this article from the WSJ.

The article reports that most states are using an 8% expected return and then goes on to suggest that this might be too high.



I think it is too high. Lets look at some basic numbers..

Assume that a pension fund is 50% in T Bonds and 50% in Equities.

The yield on a 30 year T Bond is about 3.9%.

We can solve for the expected return on the equity portfolio using a little portfolio math...

8% = R(TBond)*0.5 + R(Equity)*0.5
8% = 3.9*0.5 + R(Equity)*0.5
Solve for the R(Equity) = (8 - 3.9*0.5)/0.5 = 12.1

Based on my assumptions these plans are working on an expected return on equities of about 12.1%. Given a risk free rate of 3.9%, this implies an equity premium of:

(1+0.121)/(1.039) - 1 = 7.9%

The historic equity premium has been 6%. To assume near 8% is purely delusional.

However, it could be that I have my portfolio weights wrong. So lets assume a more modest equity premium of 5%. This implies an equity return of 1.05*1.039 - 1 = 9.1%

We can now solve for the portfolio weights.

8 = wRF*3.9 + (1-wRF)*9.1

Doing a little algebra, we can solve for wRF which is the weight in the risk free bond (in this case the T Bond).

We find that wRF = 21%. This implies an equity holding of 79%.

Again, for a public pension fund this seems far too high.

What can we conclude...
1. State governments are just kicking this problem down the road. Eventually taxes will have to go up to fund these plans, and/or benefits will be cut to retirees.

2. If you are in one these plans, you should be funding a 401-K or other plan as a supplement.

Finally: What return assumption are you using for your plan? Over optimistic investors end up living with their kids in their retirement.

Want to manage a portfolio? Go and work for Google.

Google has built a trading floor to manage its $26 billion of cash.

Fama and French on portfolio optimization and the equity premium.

Fama and French talk about how to figure out the equity premium to use in a portfolio optimization problem. Excellent stuff and required reading for any finance MBA students!