Monday, January 14, 2019

Skating Where the Puck Was

Diversifying your portfolio reduces volatility and improves compound average returns. Portfolio theorists dream of finding risky asset classes with low correlation to known risky asset classes. However, author William J. Bernstein argues that correlations between asset classes have been rising. He explains why this is so in his book Skating Where the Puck Was: The Correlation Game in a Flat World, the second book of his four part Investing for Adults series.

One of the side effects of rising correlations is that everything tends to crash at once. We are moving toward “a cohort of nearly identically behaving asset class drones.” “When everyone owns the same set of risky asset classes, the correlations among them will trend inevitably toward 1.0.”

Bernstein asks “Will things really get that bad?” The surprising answer is “We are, in fact, already there; further, it’s always been that bad. Yes, international REITs were a wonderful diversifying asset, but the ordinary globally oriented investor, and even most extraordinary ones, did not have access to them before 2000. Ditto commodities before 1990 or, for that matter, foreign stocks before 1975.” “Diversification opportunities available to ordinary investors were never as good as they appeared to be in hindsight.”

There is some good news, though. “Short-term and long-term risk are indeed two very different things.” Bernstein gives data showing that monthly returns show high correlation, but 5-year returns show much lower correlation. So, your asset classes tend to crash together, but if you wait long enough, their returns tend to drift apart and give some diversification benefit.

In the past, early investors in new asset classes benefited from diversification, but those who pile in later are “skating where the puck was.” It’s necessarily the case that as many investors buy into an asset class, its correlations with other asset classes rise.

My takeaway from this book is that I can’t avoid market crashes that affect most of my portfolio. I’m best off to live on safe liquid assets and wait out any market declines in the rest of my portfolio.


  1. Milevsky argues that "living on safe liquid assets and wait out market declines" could leave you in a very scary situation if the bear market is a very long one. You could end up with no safe assets left and having to sell equities to live on. Rare but possible.

    Also, you need rules to decide when to replenish the safe bucket from the risky bucket... not that easy...

    1. @Garth: What Milevsky says is true, but we have to look at alternatives. I've read Milevsky's books related to using various types of annuities. However, when I look into such annuities, the terms and implied returns aren't as good as the ones in Milevsky's examples. I don't know if I'm looking in the wrong place (Canada?), but the products I've found don't work nearly as well as he says.

      Here's my attempt at fixed rules for replenishing the safe bucket:

      This is what I'm using myself.

    2. Garth, Nick Murray, in his excellent book "Simple Wealth, Inevitable Wealth", offers some suggestions. 100% equity portfolio during accumulation. At retirement have set aside 2 years of cash (3 years "if you must"), continue 4.5% annual withdrawals from equities until the market is down 20% from the previous all time high, Then switch to the cash reserve. Switch back to the equity portfolio when the market returns to the 20% below the previous peak level. Replenish the cash reserve when the market reaches a new high. Certainly not without risk in a prolonged down turn, and ought to be combined with some spending cut back rules.

  2. Michael...your opening statement is questionable.
    "Diversifying your portfolio reduces volatility and improves compound average returns. "

    Tren Griffin has a good overview on why diversification is less than ideal. If you're indexing, good for you -- diversification makes sense. If you're not indexing, the diversification has many flaws.



    2. @Anonymous: The word "diversification" gets used in different ways. If we restrict it to owning multiple assets with the same expected return (and not fully correlated), then you get the benefits I described. If you diversify into assets with much lower expected returns, such as bonds, then you get lower volatility at the cost of lower expected returns.

      Stock pickers usually believe they can discern stocks with high expected returns from those with lower expected returns. If they're right, then diversifying into poor stocks doesn't help. However, 90% of professional stock pickers (and 99.9% of non-pros) are wrong about their abilities. These people do get benefit from diversification even though they don't believe it.

    3. Michael, I think that Larry Swedroe has some data that indicates that the percentage of professional stock pickers that are wrong about their abilities is now up to 98%.

  3. " If you diversify into assets with much lower expected returns, such as bonds, then you get lower volatility at the cost of lower expected returns."

    Michael, have you read Bernstein's "Intelligent Asset Allocator" or "Four Pillars of Investing"? I believe in both of those books he shows the risk and return of various portfolio mixes based on historical data. He shows that introducing a small amount of bonds can both increase returns and reduce risk.

    I do wonder how much of that result is impacted by the multi-decade decline in rates from the early 80's until the date he wrote those books. I'm not convinced this would remain true in today's exceptionally low-rate environment. It is probably true though that a small amount of bonds would reduce risk and lower returns by less than the weighted average of expected returns.

    I personally do not hold any bonds or fixed income, but over the next 1-2 decades, I expect I will mix in 10-20% of bonds.

    1. @Returns Reaper: I've read The Four Pillars -- my review here:

      I've also seen many analyses and examples of portfolio optimization. The math isn't too difficult, and it shows that an asset can have somewhat lower returns than other assets and still help boost portfolio returns. But the expected return gap can't be too wide. You're right that during recent decades, booming bonds (fueled by dropping interest rates) have had returns high enough to help boost portfolio returns. But there was no point when it was clear that this would be true. Averaged over the full history of return data, owning bonds lowered volatility and reduced returns.

      I don't hold any bonds either. Now that I'm retired, I have some GICs and a high-interest savings account that together hold about 5 years of my family's spending. The rest is stocks.