Monday, May 25, 2009

Total Bond Market Index (BND) and IEF as Non-Correlated Assets to Stocks

Even if I am satisfied that an asset shows very low positive correlation with stocks, or even  negative correlation at times, I may not invest in it because I view my downside risk to be significant, whereas my potential upside is limited.  I have not lost anything other than a small dividend by selling BND last year. (BND is the Vanguard ETF for the Total Bond Market Index: Vanguard - Vanguard Total Bond Market ETF Overview) I had to ask myself.  Is the risk to the downside for BND significant and more than a possibility, even a probability if assessed in terms of a year or two rather than in a more limited time of months?  I am responsible to myself and have no choice but to act based on my personal assessment.  With the current expansionary monetary and fiscal stimulus, it is hard to envision that a security like BND has much left in upside stock appreciation, and a real possibility of losing value. And, I personally do not believe that the downside risk is being compensated with a 4% dividend. Others may disagree and advocate a total bond market index as part of an asset allocation now.  

A separate issue is whether the total bond market index  is a low positive correlation or sometimes negatively correlated, with stocks.    

The correlation between bonds and stocks depends to some degree on whether the stock market is in a bear or bull mode. Starting in October 2007, I would expect a strong negative correlation between BND and the S & P 500, and a small positive or negative correlation during many stock bull market phases, a conclusion recently confirmed in an academic study summarized in this blog. Stock and Bond Returns Correlation Variability

Easy monetary policy as now would most likely contribute to a negative correlation period. This would suggest by itself, without taking into account other relevant factors, a shift of some assets out of bonds and into another non-correlated asset class at or soon after the start of a bull market phase. The purpose would be to find another asset that would be a better hedge than bonds during such a period, even though bonds would still provide diversity to the allocation. 

The academic study referenced above used the Vanguard Total Bond Market Index, which has a lot of GSE debt in it. If you had the misfortune of being in one of a large number of bond funds which substantially underperformed that dumb index in 2008, then your returns between those bond funds and stocks would be a positive correlation, and a fairly strong one, at a time when you needed assets going in the opposite direction to your stock portfolio. I will never buy a bond fund again, and will simply stick with the one that I have now devoted to Tennessee Municipal bonds, more due to the fact that I do not want to fool with buying individual Tennessee municipal bonds, at least for now, maybe later.      

Similarly to BND, the Ishares ETF for the 7 to 10 year treasury bonds (IEF), iShares Barclays 7-10 Year Treasury Bond Fund (IEF): Overview, would have shown strong negative correlation statistics to the S & P 500 since the onset of this bear market. However, the current yield on this ETF is just 3.67%. MarketWatch.com Quote

While I made a small mistake in not including IEF with my other bond ETF buys made at the start of this bear market, I can not make a case for it now anywhere in my asset allocation. So, under static allocation, this kind of security would have a permanent home. I, however, can not ignore the risks associated with owning this security at the current time, including the risk of lost opportunity and the risk of a negative real rate of return after inflation and taxes. Since it is my personal view that I am not being compensated to assume those risks with a yield at 3.67%, and also based on my view that the downside risk in price is more prominent than further upside gains, I have no place for IEF in my portfolio at the current time, possibly that may change by 2011 or 2012.    

Then, there is a question at least in my mind of how good the negative or low correlation is between stocks and a 5 year treasury. This is a link to a chart showing the positive and negative correlation between stocks and the 5 year treasury over rolling ten year periods since 1938. Portfolio Solutions While the negative correlation has been strong during the recent bear market, as you would expect, most of time the correlation is positive, sometimes hitting the .25 to .50 range, particularly during a bull market in stocks. That actually makes intuitive sense if interest rates are falling causing investors to afford richer P/E ratios to stocks, so that a bull move would be occurring in both markets. This would suggest at a minimum a change in static allocation theory, requiring a move more into bonds soon after the start of a bear move (confirmed by the formation of an Unstable Vix Pattern), moving more when situational risk is more prominent, and then a move out of bonds, to some meaningful degree, upon the formation of a Stable VIX Pattern. Then, some of those proceeds from bonds sales could then find a home in an asset class with less positive correlation than bonds with stocks. In other words, a shift in asset allocation caused by market factors rather than any consideration unique to the individual.

There is also a product from Ishares for the 3 to 7 year treasury that could also be used in an asset allocation, but it has the same problem as IEF from my perspective. iShares Barclays 3-7 Year Treasury Bond Fund (IEI): Overview  The current yield on this ETF, IEI, is just 2%. MarketWatch.com Quote

ADDED 6:47 P.M. 5/25/2009:  One of the many failures of the static asset allocation approach, which puts an individual on some kind of mechanical "glide" path, is an unwillingness to adapt and respond to the market when some kind of response is the only prudent action.   Many of my posts mention one tool that I use in implementing a Dynamic Asset Allocation called the VIX Asset Allocation Model.   An understanding of that model is critical to understanding points being made even in a Post such as this one.  Some of the posts summarizing this model include the following:





This model would have required a shift out of stocks in the fall of 2007 caused by a Trigger Event in August 2007, with the money raised shifted into bonds, which would be viewed a likely candidate for negative correlation with stocks at that time. The model would also cause a shift back into stocks when a Stable Vix Pattern is formed lasting 3 months.  The major shifts into and out of stocks signaled by the Model have lasted for several years when the model was backtested to 1990 using the VIX data from the CBOE. So even when I am discussing the ETF BND as a non-correlated asset to stocks, I am integrating that discussion as it relates to my Dynamic Asset Allocation approach using the VIX model. It was in the fall of 2007- after the first Trigger Event- that I would expect to increase my holding in BND from the proceeds raised from stock sales. The same would have been true in the 1997 and 1998 Trigger Events. This would be done irrespective of my glide path or my age.  Although I face no situational risk personally, the interruption in the so called glide path, to make a course correction in the allocation to stocks particularly for those facing significant situational risks, is deemed critical by me when the VIX model starts to flash the red signals, as it did with three separate Trigger Events between August 2007 to February 2008.   

No comments:

Post a Comment