Monday, April 6, 2020

It’s Really Not Rocket Science

The latest book in the not-rocket-science series from Tom Bradley at Steadyhand Investment Funds is out (get the free PDF of It's Really Not Rocket Science).  Like the previous two books (It’s Not Rocket Science and It’s Still Not Rocket Science), it’s a collection of Bradley’s excellent articles on various aspects of investing.  Unlike many investment professionals who seek to make investing seem complex and mysterious, Bradley’s writing is clear.

Disclaimer: I did not receive any compensation from Steadyhand to write this review other than a free copy of the book.

Here are a few parts of the book that struck me as notable:

Nobody can predict the market, but “I see professionals who regularly defy all logic and evidence by explaining the unexplainable and predicting the unpredictable.”  I’ve told many people that nobody can predict the market. What baffles me is that many nod their heads in agreement and immediately ask “Where do you think the market is going?”  Investors need to learn to stop asking, and stop listening to the answers if others ask.

“I’ve never seen anyone, professional or amateur, get [market timing] right consistently enough to make it pay.”  Again, many people will nod in agreement and immediately ask what others think is going to happen in the markets. Contradictions abound.

“The toughest decision in investing” is deciding when to get back into the market after you’ve jumped out.  This applies even if you were lucky enough to jump out before a market drop. As stocks start to rise again, you’ll be “under a lot of stress after missing out on years of good returns.”  You’ll be “heavily invested in [a] negative view,” and you’ll be “waiting for something that will never occur — the perfect time to buy.” Advice from Bob Hager: “Picking the bottom of the market is virtually impossible.”

On hedge funds, “As it turns out, the managers have done much better than the clients.”  Buying a hedge fund because it makes you feel special can be an expensive ego boost.

Bradley’s article on where returns come from is excellent.  The most important factor is how the markets perform. Then it’s your chosen mix of stocks, bonds, and cash.  Third is the costs you pay. “Being an old stock analyst, it kills me to say this, but security selection, whether it’s done by a professional manager or yourself, comes in a distant fourth on the list.”  Finally the wildcard is investor behaviour, which could slot in anywhere in the hierarchy of importance for your returns. With discipline, behaviour hurts you very little. With impulsive changes of strategy, your behaviour could be the most important determinant of your dismal returns.

Whether you’re a DIY investor or work with an advisor, Bradley’s essays are worth reading.


  1. I used to believe that market timing was a waste of time, but I'm not as certain about that anymore.
    William Bernstein is an advocate of overbalancing, which frankly is valuation based market timing. Jason Hsu and Rob Arnott also support overbalancing. When you read Jonathan Clements, he mentions that he sometimes deviates modestly from market cap weighting based on valuations. There's academic support for this (the Stambaugh model).
    The following is from Ben Carlson:
    "In recent years, we have made just a slight overallocation to international's something that we think over the next three, five, seven years can potentially work, because there are lower valuations and higher yields overseas in places like foreign-developed markets and emerging markets. But we're definitely not smart enough to figure out the correct time to make a huge push over there. And definitely, it's one of the things where we don't like to think in terms of extremes. So, we're not going to make a huge shift completely out of one asset class and all into another one, because that's just not the way that a prudent, diversified portfolio works. But it's one of those things where around the edges where we might over-rebalance a little bit, but again, with the correct expectations in mind that this isn't something that we're banking on working in six months, it's a more of a multiyear process."
    The following is from Larry Swedroe:
    “I would add this comment on valuations. Say you decide to be 60 US and 40 international. And now you observe historically very wide differences in valuations, as we have today. And are willing to live with the risks and the tracking error. I would not advise straying far from your 60/40, perhaps moving 5 or at most 10% to international. Not recommending one do it, just advising if considering it, keeping in mind that the lower valuations reflect market's view of higher risk, not a free lunch. And then might want to consider lowering beta exposure as well to cut the tail risks down. That would fall under my sin a little if you will.”
    The following is my implementation of overbalancing. Within stock subasset classes (Canadian, emerging markets, EAFE, US), I will deviate modestly from market cap weighting based on valuations. I have a 20/10 rule. I'll deviate from market cap weighting up to 20% on a relative basis limited by a change of no more than 10% on an absolute basis. For example, US is about 57% of world market cap, and has closest to the highest CAPE of any national stock market. A decrease in 20% in asset allocation would result in an allocation of 45.6%; however this would be limited to 47% by the 10% absolute limit.

    At extremes of valuation, I think you can do more than then 20/10 rule. Howard Marks is an advocate of this.
    Jack Bogle said "“After nearly 50 years in this business, I do not know of anybody who has done it successfully and consistently. I don’t even know anybody who knows anybody who has done it successfully and consistently.” But he himself decreased his stock allocation around 2000.
    Jonathan Clements increased his stock allocation in 2009. Jonathan Clements recently wrote that if the stock markets went down further, he would increase his stock allocation.

    The following is from Larry Swedroe:
    "I agree that market timing at EXTREMES is a good idea, but only if you are prepared to
    a) be wrong for long time (I sold all growth in 98 and looked bad for 2 years, then looked great)
    b) don't confuse strategy and outcome
    c) stay disciplined (the hard part is when do you get back).
    so for most, especially if have well diversified by asset classes/factors, simply rebalancing is likely to prove better.
    and I've only made a few trades over the past 20 years including getting out of growth in 98 and out of REITS a few years ago. Valuations do matter, and they matter a lot."

    1. Anonymous: Thanks for the interesting comment. I think very thoughtful people might be able to apply these ideas with some success, but I have some warnings.

      Any kind of market timing success depends on others who fail at market timing. Even rebalancing success depends on taking advantage of someone else who fails at market timing. So, we should take it as a warning that the dollar-weighted average market timer fails somewhat (if we treat rebalancers as non-market timers).

      Another warning concerns subtle data mining. To make any kind of call on the state of the market, we use CAPE or relative valuations between countries or some other type of indicator of what the market "should" look like. When we examine the past, we know the distribution of our favourite indicator. Then when we simulate a strategy, we get the benefit of knowing that this indicator will swing up and down in a known range. However, going forward, we can't know what "normal" will look like. Maybe CAPE will permanently shift to a higher range, or the U.S. share of global markets will shrink to half its current level. (These aren't predictions, just examples.) If something like this happens, it hurts rebalancers somewhat, and it hurts over-rebalancers more.

    2. Anonymous: Another thought: Over-balancing is tempting, but I wouldn't recommend going beyond fairly subtle over-balancing. Maybe shifting assets levels within a range. Those who make massive shifts end up with massive gains or losses. We can't recover from blowing up.

    3. I’m advocating valuation based market timing. Such timing between stocks and bonds is almost always a negative sum game. The reason is that there is almost always an equity risk premium. Valuation based market timing between stock subasset classes is a zero sum game.
      However, there are those who select securities based on factors. An example would be those who tilt to small and value. Security selection is also a zero sum game.
      Those who tilt to small and value might say that the increased expected return associated with such tilts are due to increased risk. But I’ve heard the same argument made for valuation based market timing.
      One can also engage in valuation based market timing not to increase expected return, but to decrease risk.

      Figure 8 in the above link shows the correlation between CAPE and drawdown risk: the higher the CAPE, the greater the drawdown risk.
      Valuation based market timing is a low return high dispersion activity, so I would agree that one should do it modestly.
      But I don’t necessarily agree that at extremes of valuation, it should be done modestly.
      Look at Fig. 5 in the above link. Once CAPE greater than 30 in a country that is of at least medium size, returns are poor. Similarly, once CAPE is less than 15 in a country of at least medium size, returns are reasonably good using contemporary data. Your chance of losing money is very small. CAPE works less well in smaller less diversified countries.
      IMO, the data supports more than modest market timing at extremes of valuation.
      You make the “shifting goalpost” argument, which is reasonable. At least in the US, CAPE has increased with time. That’s very relevant to market timing between asset classes, such as stocks and bonds. But for market timing within an asset class, such as between stock subasset classes, that argument isn’t as relevant.

    4. Anonymous: Depending on how far you're planning to push "more than modest market timing", I disagree that the data supporting it is a good reason to proceed. My data-mining argument limits the amount of market timing one should sensibly use.

      These could be famous last words: "But for market timing within an asset class, such as between stock subasset classes, that argument isn’t as relevant." It's not unusual for large businesses in some countries to form powerful oligopolies that undermine smaller competitors. Currently, US small companies flourish. Maybe that won't always be the case. I certainly hope we don't get this kind of stagnation, but who knows? Maybe investors who make big shifts to small cap will get crushed. It's not hard to make up possible secular shifts.

      I've chosen to avoid market timing altogether. I agree that some smart investors might be able to make market timing work. However, big market timing shifts would bring ruin possibilities that I wouldn't want to take on.

  2. I mentioned more than modest market timing at extremes of valuation. What defines more than modest is in the eyes of the beholder. It will be highly dependent on the person and their circumstances. Also, a variable such as CAPE has to be considered in the context of other factors, such as inflation and interest rates.

    The following is what I would consider reasonable for an intermediate level investor with at least a 15 year time horizon, based on present inflation and interest rates. If CAPE less than 15, consider leverage. The leverage used should be such that there could be a subsequent drawdown of 70% without risk of a margin call. If CAPE greater than 30, don't buy and consider selling.

    Your comment regarding secular shifts is well taken. That's why I have a 20/10 rule, which results in modest changes. Also, this market timing is at the level of emerging markets, EAFE, US and Canada. With the important exception of Canada, that decrease the risk of secular shifts having a negative impact on overbalancing. All overbalancing is is rebalancing taken to an extreme.

    About more than modest market timing at extremes of valuation, it looks like we're going to have to agree to disagree. Such extremes of valuation are uncommon. In recent times, they've occurred in 1999-2000,2008-2009 and outside North America at present. I've written a market timing strategy in my IPS, and my goal is to stick to it.

    1. Anonymous: Your plans are better thought out than the vast majority of market timers, so your odds of success are better than average.

      One more remark of warning: Like a turkey that has never seen a thanksgiving, those who look for evidence-based strategies (that rely on studying past market data) risk being harmed by an event that has never happened before.

    2. There is always the risk of the black swan. William Bernstein might say that it's more about the risk of insufficient personal knowledge of market history than the risk of a black swan. Regardless, such risk can be devastating. Long Term Capital Managment, with its Nobel laureates, is as example.

      Whenever you use leverage, black swan risk goes up considerably. Leverage may be the only way for an investor to go bankrupt.

      One could consider other options, when valuations are unusually low. One could modify one's stock/bond allocation or increase overbalancing.

    3. Anonymous: My point is a little different from what people usually mean by "black swan." People call the coronavirus a black swan, and in many ways it is. However, the only aspect of its effect on markets that we haven't seen before is the speed of descent in prices. So, any plan that didn't foresee the possibility of a big drop was not waylaid by a genuine black swan, unless their harm came strictly from the speed of price descent. (However, we have seen smaller drops that have happened very quickly.) In the future, many things will happen that weren't foreseen (and will be called black swans), but their effects on markets will be similar enough to market data that we have seen before. The big risk to data miners is seeing market patterns different enough from any data we've ever seen before.

  3. I have no problem with someone who eschews market timing entirely. Of course, there are those who make the argument that rebalancing is a form of market timing, although I don't agree with that myself.

    Valuation based market timing can be more about risk management than increasing expected returns. In 2000, US CAPE reached around 44. An even more extreme example is the Japanese stock market at the end of 1989. CAPE was around 90, and the Japanese stock market was around 45% of world market cap. In both cases, investors were sheep to the slaughter.

    The use of leverage at low CAPE is more consistent with increasing expected returns than with risk management. But even there, the use of leverage provides the liquidity necessary to buy securities at low price. And low price increases the margin of safety. If you're an accumulating investor, you can use subsequent contributions to pay off the loan used to purchase those cheap securities.