A Quant Geek’s Style of Value Investing

What is the best thing about saving / investing? It is compounding. Capital invested in anything with a positive IRR (internal rate of return) grows exponential rate as long as the IRR remains positive). That doesn’t necessarily mean a hockey stick growth. That doesn’t even mean you can get rich quick just by compounding. But the fruits of compounding, like any dessert, taste best when they are refrigerated long enough.

How is compounding relevant to an investor? Can it be a strategy on its own? Can you put money in fixed income and hope to get rich based on the power of compounding alone?  Does it always make sense to redeploy profits hoping that it will get compounded?   

Let’s start taking one some of these basic questions. Let’s start with fixed income. It probably has the most predictable IRR, at least more predictable than investing / trading in stocks / real estate / precious metals  etc. The problem is the IRR of fixed income investments is seldom positive when you take inflation into account. With fixed income, what you have on hand is an investment that has a highly predictable, but negative IRR in real (inflation adjusted) terms. That’s recipe for exponentially decaying capital! But as long as the nominal (not inflation adjusted) IRR is positive, it is better than keeping cash below your pillow, unless you happen to keep your money in a region of negative nominal interest rates (like Japan of late).

What about equities? Does the idea of keeping profits on the table forever and letting them compound make sense there? There is no word called forever in investopedia! There is nothing that makes sense all the time. Market indices have been cyclical in the last 4 decades and there is no reason to believe that it’s going to give up this behavior and will go up one way. Central bankers keep blowing bubbles and inflating them. The bubbles eventually burst and it would be time to blow new ones again. Even specific sectors will have their own bull phases but these bull phases don’t last forever.

So, under what circumstances does it make sense for investors to let compounding do its job? Qualitatively, the following conditions need to be met.

• When the expected real return from the asset class is positive.
• When the expected real return from the particular asset is positive in the next 5/10 years, given the current valuation.
• The probability of #2 happening is pretty high.

When it comes to diversified equity indices, the one key metric that can guide investment decisions is the valuation of the index. You can use P/E or forward P/E or CAPE ratio as a metric of valuation. I personally prefer CAPE. Here’s how I would use it.

• Test the CAPE time series (of the index in question) for mean reversion property. Something like the Augmented Dickey Fuller Test can be used.
• If it is not mean reverting, it’s useless for our purpose. If it is, regress this series against the time series of forward returns of the index (let’s say using 5/10 year horizons). It’s a regression of CAPE ratio of today vs the return the index has generated 5/10  years down the road (with the benefit of hindsight). See if there is a negative correlation and the quality of the regression (R squared) is good.
• If it is good (I would take 70% as good), I would use the regression to determine what I call the point of no return, for the time horizon of my interest. It is the CAPE level at which, we can expect the nominal returns from the index to be zero (based on the regression). If CAPE goes any higher, the nominal return going forward would be negative. I would also calculate another number, which I call, the point of no fun. This is the level of CAPE at which the expected real return would be zero. If CAPE goes any higher, it’s likely that the real return going forward would be negative. Lastly, I would calculate what I call the point of losing shirt. This is the CAPE level at which there is a 67% (or any number above 50%, determined using the confidence interval of the regression) probability of negative returns.
• I would leave the money on the table and have it compound itself as long as the CAPE is below the point of no fun. I would start skimming profits when CAPE is between the point of no fun and the point of no return. When it starts going above the point of no return, I start taking capital (not just profits) off the table. When CAPE hits the point of losing shirt.   

This is a version of value investing which can be used by someone with a penchant for quantitative techniques and absolutely no idea of analyzing what the accounting numbers of companies mean (or don’t mean)!

I don’t have the foggiest idea on how to use some of these principles in the world of stocks / sectors. Will and to see if a learning algorithm can be used. I do have some idea of how to use the principle of compounding in short term trading as opposed to long term investing.  That calls for a dedicated post.

Till then, happy investing;