So this post is about Portfolio Visualizer. This site is going to be of more interest to the efficient markets adherent/quantitative types among us. I think, however, it has a lot of interest and valuable tools even if you’re a Warren Buffett wannabe stock picker.
Portfolio Visualizer has six different areas of tools: Portfolio Backtesting, Factor Analysis, Asset Correlations, Monte Carlo Simulation, Portfolio Optimization, and Timing Models.
The Portfolio Backtesting tool is really neat. You can set up allocations to various asset classes and compare how they have done throughout history. There are some limited data sets, so make sure to pay attention, as the result screen will tell you if the data has been truncated (it runs from 1972 for a lot of assets). Here’s a spoiler, U.S. stocks have smoked pretty much everything else, but as you can see from the charts generated as part of the output, most of the recent outperformance (versus value stocks and foreign stocks for example) is due to the recent bull gallop. I personally think that is likely to revert to the prior experience, but who really knows.
As another example of this cool feature, you can look back see what happened to 10 year treasuries during the stagflation of the 70s and early 80s. Interestingly to me, the nominal returns for 10 year notes weren’t all that bad. 1987 posted a (2.87%) loss which was about the worst figure. I think the real injuries to capital were due to after inflation (“real”) returns. The nominal returns in 1994, 1999 and 2009 were all worse than this historically bad period for bonds. So lots of neat features and you can reassure yourself that value stocks used to crush the market and maybe remind yourself to think in real terms rather than nominal.
Moving on, the second feature I wanted to highlight is the Factor Analysis option. So this offers a number of choices to examine statistical regressions with “factors” identified by academics (and practitioners) as being relevant to explaining historical performance of securities (for example, was your stock very exposed to the old Fama-French value factor of low price-to-book).
Some examples of factors are Value, Momentum, Quality, and Size. You can see the factor regressions, which basically attempts to explain to what past performance may be attributable. You can choose to use the Fama-French data set (and the number of factors to regress against), or AQR data/factors (such as the quality factor).
A discussion of all of the factors, the various ways one can attempt to measure them, and their validity goes beyond the scope of this post, but if you are uninitiated and curious, I recommend checking out AQR.com, MebFaber.com, and AlphaArchitect.com.
For our purposes, you can enter any equity mutual fund or ETF (at least US funds that I’ve tried ) and see what the regression tells you about the historical exposure to the various factors (and the overall market or “beta”). For example, did VBR or IWD offer more value exposure (and over what periods)? Is MOAT, QUAL, or VIG more exposed to the AQR quality factor? Anyways, it is pretty neat.
The Asset Correlation tool, is pretty much what the name says. You can enter various assets (as represented by funds) and see what the correlations have looked like in the past. For example, you can see that EDV (an extended duration treasury ETF from Vanguard) has had a more strongly negative correlation with VTI (total stock market) than AGG. So maybe you want to use EDV rather than AGG, and allocate more to stocks for the same portfolio volatility. I think if the data went back far enough, however, you could also see that correlations between bonds and stocks were positive during some periods, so maybe that is not a foolproof idea.
There is also a Monte Carlo Simulation (basically statistical techinque that relies on computational algorithms to model outcomes…hey, what do you want? I’m a lawyer not a mathlete). Thanks to @Ritholz‘s MIB podcast, I do know that Bill Sharpe doesn’t like that term. There is also a tool to model the efficient frontier for assets under the CAPM and some other stuff that I have never used. If Harry Markowitz is going to be allocating based on Puerto Rico rebuilding, I think I will ignore that stuff (see Barron’s, The Father of Portfolio Theory Bets on Rebuilding ($)).
I do, however, think the “Timing Models” section will hold some fascination for many among us. You can look at what a trend-following strategy using a simple moving average in various asset classes has done over time. I have some of my portfolio allocated to a strategy based on that technique (with some tweaks made for valuation). Meb Faber and Alpha Architect have written a lot about trend following in equities.
Another interesting timing model that you can play with in this tool uses valuations. It has historically performed quite well. It basically allocates only 40% to stocks when the Shiller CAPE is 22 or above, 60% to stocks when the CAPE is between 14 – 21, and 80% to stocks when the CAPE is below 14 (the remainder during all periods is allocated to bonds). This strategy is pretty neat and has done well over time, outperforming a traditional 60-40 portfolio with lower volatility and drawdowns since 1985. You can also play around with the Dual Momentum timing model and a few others.
To sum up, this is a really fantastic resource that has a number of useful free features. I even understand you can use it to run some bespoke backtesting of your own strategies for a fee. I encourage you to check it out!