In our last post in the series, we looked at some underlying theory around diversification and minimum variance portfolios. We also learned that the key downfall of this theory, when used to pick a diversified portfolio of stocks, is that correlations among stocks change over time. In particular, stocks become highly correlated during broad market sell-offs. (Dynamic correlations really have the worst timing, don’t they?)
We saw however that if we look beyond stocks to a broader set of asset classes, we can find the opportunity to create true diversification that won’t abandon us when we need it the most.
In this post, I’ll illustrate those concepts with a few charts, looking at stock diversification versus asset class diversification, through good times and bad. Ultimately, the powerful difference you’ll see comes with a clear rationale, which we’ll start on here and continue in the next post of the series.
About that free lunch…
Going back to Markowitz, the free lunch opportunity we have lies in finding a portfolio of assets which are uncorrelated in their movements, or even better, inversely correlated in their movements (with one asset tending to rise as the other falls).
Because asset classes are affected by large and often slow moving economic forces, the correlations between asset classes are more likely to be persistent over long periods of time, through different economic conditions. The result is a stable, hedged portfolio that provides strong and consistent risk-adjusted returns.
With that in mind, let’s visualize this difference between a portfolio of stocks and a portfolio of asset classes.
Rising apart and falling together
First let’s look at how different stocks behave in rising and falling markets, and observe some evidence supporting the claim that stocks move up in their own ways but all move down together.
Let’s start by looking at a “steady” economic environment, like that of 2007.
This graph shows six stocks out of the S&P 500 representing stalwarts of different industries: technology (MSFT), airlines (BA), construction (CAT), health care (PFE), energy (XOM), and financials (BAC).
As you can see, under “normal” circumstances, stocks of different industries are not highly correlated. Some may go up while others go down, and they are capable of moving in opposite directions. For example, Pfizer started rising precisely as Caterpillar and Boeing started to fall. Microsoft kicked into high gear as financials started to suffer. The result of these differences is a portfolio that has less overall volatility.
Now let’s look at what these same stocks did during 2008, a year of around 40% losses in the broad market. (Read: Not so steady.)
While the individual names in this basket of stocks had a wide range of performance, they unanimously fell. Not a single instrument was able to hold up this portfolio, and during the panic selling in the fall of 2008, they all fell by 20% or more in a very short period of time.
In those times of panicked market crashes, all stocks tend to participate. As such, the correlation of all stocks approaches 1, or fully correlated. And in the wrong direction! The risk protection we seek in diversification then disappears during such times, and our overall portfolio can exhibit dramatic volatility.
In the meantime…
Now let’s see what some other asset classes were doing in the meantime. Our instruments in this case are all ETFs that track the price of an entire asset class:
- SPY - domestic large cap stocks (i.e. the S&P 500)
- TLT - long term US treasury bonds
- GLD - gold
- LQD - investment grade corporate debt
- TIP - inflation protected short term treasuries
This is definitely not exhaustive of what’s out there in terms of classes, but it’ll do well for our purpose of illustration.
Now to revisit our two economics periods. First, let’s hop into the DeLorean and head back to 2007.
In some ways, this looks similar to our stock portfolio during 2007: different instruments moving along their own paths in a fairly uncorrelated way. There are some key differences though.
For one, there’s less overall variance in the returns. This might be explained due to each ETF itself being an aggregation of other securities. TLT represents a portfolio of long term U.S. treasuries. SPY represents 500 different large cap US equities. When we looked at stocks, we could see more individual variance due to the unique influences on each company which had a profound effect on its specific valuation. Here though we’re observing the valuation of an entire class of assets through the prices of each ETF.
That’s not to say there are no outliers in an asset class portfolio. Clearly gold was a significant outperformer during 2007, no doubt due to the fact that inflation appeared to be picking up during the year. In other years, such as the one we just experienced in 2019, equities can become the major outperformer. Sometimes it’s treasuries, when interest rates are falling or even just expected to fall. But these factors go beyond those faced by individual stocks, like a change in executives at Boeing for instance, and instead represent major trends or changes in broad investor sentiment.
Now let’s move on to 2008, and see how those changing sentiments unfold during the heart of the global financial crisis.
Well that’s a bit of a different story. Investors didn’t just sell out of the US market and stuff the proceeds under the mattress, they moved that money into other assets. In particular, demand for U.S. treasuries, show above via TLT, exploded in Q4 after the collapse of Lehman Brothers signaled major issues in the financial markets, and led to a nearly 30% rise in bond prices with just a couple months. While equities were crashing down, bonds were crashing up.
Partly this effect is because huge amounts of money can’t just sit as riskless cash in some savings account at a bank. Hedge funds and trading shops would quickly and far exceed the $250,000 insured by the federal government (though this insurance does a good job of preventing consumer bank runs).
The money has to go somewhere, and the very next best thing to government insured cash is government debt. Though it’s not quite as liquid as cash, it’s actually a safer place, in relative terms, to park a lot of money. And the rates surely beat those in money market accounts.
Another factor here is the anticipatory nature of the market. When stocks fall and economic conditions worsen, the expectation is that central banks will step in to help stimulate the economy by lowering interest rates. And when interest rates fall, the prices of bonds rise.
Because of this rush into a safe place to park money, combined with an expectation of lower rates, we see this counter effect in bonds during fast falls in the market. As stocks come crashing down, bond prices tend to crash up. A similar effect can happen with gold, as investors look to park cash in a “safe haven” asset, and in anticipation of potential inflation from future monetary policy, though this is just one of many reasons gold prices may rise.
Each asset class has its own influences and interactions with the others. Some assets are affected very similarly by the same economic forces, such as domestic and foreign stocks. In that case, we see a correlation in the returns of those assets. Others act quite oppositely as outlined above, and in those cases we see a negative correlation between them.
Now, to build
In the next post in this All Weather Portfolio series, we’ll talk further about the key elements of change when it comes to economic conditions and outlooks, and lean on the work and wisdom of Ray Dalio to help explain them. Then we’ll use the asset classes discussed in this post to build a very simple and easy-to-maintain portfolio that can hold up well in any weather.