Problems with the “X best days” argument

If you’ve ever asked your investment manager about whether you should stay in cash or get invested, you’ll likely have been presented with some variation of the following chart:

Source: JP Morgan

The chart is from the excellent JP Morgan Guide to the Markets, but almost all managers have some variation in their marketing decks.

It’s supposed to show how important it is to get your cash invested, and keep your cash invested, so you don’t miss out on the few days where a good proportion of your portfolio’s gains are made. The more you delay investing, the more likely you are to miss out on the X best days. It’s often accompanied with the slogan, “time in the markets beats timing the markets”, suggesting that it’s more important to get cash invested and hold for the long run, rather than to try and time the market’s fluctuations.

The finding that a relatively small number of days account for a large proportion of the market’s gains has been confirmed by research from the University of Michigan’s Professor H. Nejat Seyhun.   He showed that just 90 days in 30 years accounted for 95% of all the market’s gains. Going back even further, 96% of market gains in the last 40 years (between 1963 and 2004) occurred during only 90 trading days (0.85% of 10,573 days).

So it’s true that a small number of days have provided a large proportion of the gains, and it sounds like a convincing rationale for buying-and-holding.

But the argument comes with a few problems.

Missing the X worst days

If you stay in cash, you may well miss the X best days in the market, but it works the other way too – you may also miss the X worst days. Missing the worst days has a big impact on returns, just as the X best days do, but missing the worst days massively boosts a portfolio’s returns, rather than harming them. Missing the X worst days is great for a portfolio, and if they can be avoided, then it means that staying in cash rather than investing would’ve been the right decision.  

The best and worst days are two sides to the same coin, but only one side is ever presented. Nobody ever shows the “X worst days” chart.

The University of Michigan research cited above illustrates this “worst days” side, which is much more rarely seen. They showed that between 1963 and 2004, the index gained at a geometric average annual rate of 10.84%, for a cumulative return on $1.00 of $73.99 over 42 years. If the best 90 trading days, or 0.85% of the 10,573 trading days, are set aside, the annualized return tumbles to 3.20% and the cumulative gain falls to $2.70. With the 90 worst days out, the annual return rises to 19.57% and the cumulative gain to $1,693.68.

Astute readers will have noticed that the effects aren’t symmetrical. That $1 turns into $73.99 fully invested, falls to $2.70 through missing the 90 best days, but rises to $1,693.68 by missing the 90 worst days. That’s a big difference on the downside by missing the best 90 days, but an absolutely massive difference on the upside by avoiding the worst 90 days.

This effect comes about because the largest daily market losses have been larger than the market’s largest daily gains (known as ‘negative skew’). For example, the largest daily percentage gain for the S&P 500 was +16% (15/03/1933), but the largest daily loss was -20% (19/10/1987).

So by missing both the best days in the market, but also the worst days in the market, you’d actually do much better than the index. Missing the best days has a big impact on the downside, but missing the worst days has a much bigger impact on the upside.

Research from Michael Batnick of Ritholtz Wealth Management demonstrates this point, by showing what would’ve happened if you missed the best 25 days of the S&P’s returns, but also missed the worst 25 days:

Source: Michael Batnick

By looking at both sides of the coin, and thinking about the cost of possibly missing out on the market’s best days versus the benefit of possibly missing market’s worst days shows how this argument for staying invested falls apart.

The “X best days” argument is generally used to convince investors to get invested as soon as possible, to prevent them from withdrawing cash, and to demonstrate that market timing is inferior to a buy-and-hold approach.

But by looking at the X worst days alongside the X best days, the “X best days” argument starts doing the exact opposite of what it’s supposed to do. By looking at both the best days and the worst days together, it actually starts opening the door to market timing arguments. The message changes from “you should get invested and stay invested, otherwise you’ll miss out on big gains” to “get invested, and we’ll try to avoid the losses for you so you’ll outperform the index”.

That doesn’t sound very buy-and-hold to me.

Another problem with “X best days” argument

What most people who present the “best days” argument don’t mention is that the market’s worst days often occur very near to its best days.

The table below shows the 20 best days for the S&P 500 and the 20 worst days, in date order. The highlighted cells mark an instance where a top 20 gain or loss occurred within thee weeks of a top 20 move in the opposite direction (green for gains, orange for losses).

Source: S&P Dow Jones

For example, the -12.34% loss on 28/10/1929 was followed 16 days later by an 8.95% gain on 14/11/1929.

2008 was a particularly interesting year. The -7.62% loss on 09/10/2008 was followed by an 11.58% gain 4 days later on 13/10/2008, then a -9.04% loss 2 days later on 15/10/2008, then a 10.79% gain 13 days later on 28/10/2008.   

Of the S&P’s 20 largest daily gains, 10 of them have come within 3 weeks of one of the 20 largest daily losses. Similarly, of the S&P’s 20 largest daily losses, 14 of them have come within 3 weeks of one of the 20 largest daily gains.

The largest market losses occur very close to the largest gains.

This is known as ‘volatility clustering’ – the tendency for large swings in the market to occur near to each other.

And this is relevant for the “X best days” argument. Given that the market’s best days and worst days happen so close to each other, missing the X best days also means likely missing the X worst days. The best days and the worst days happen so close to each other that you’re extremely unlikely to be able to capture the X best days without suffering the effects of the X worst days. And as we’ve seen, missing the X worst days has a much bigger impact than missing the X best.

The best days, but the worst years

Another interesting finding from the data is that many of the largest daily gains came during years where the market was falling, and ended the year negative.

Source: S&P Dow Jones

In fact, 14 out of the 20 largest daily gains came during years where the market ended up negative. Looking at averages, the average return for the calendar year in which one of the top twenty largest daily gains were made, is -10%. This means that it would have been beneficial to miss these large daily gains to avoid the losses the market made for the year.

A bad argument against market timing

My final point relates to how the “X best days” argument is often wrongly used to criticise market timing strategies.

The argument is often used to demonstrate that timing the market is futile, because if you happen to miss one of the few days where the market benefits from the big gains, then your returns will suffer.

What this logic assumes, however, is that the market timing strategy will rotate between being completely in the market or completely out. It’s binary – a 100% equity position or a 100% cash position. The argument is essentially saying that if your timing strategy happens to be in 100% cash when the market makes a big daily gain, then your returns get hurt, because you’re not in equity.

In reality, I know of almost no market timing strategies which rotate between such extremes. Usually there’ll be a core exposure (usually a combination of equities and bonds) making up the majority of the portfolio, which is bought and held, with some satellite market timing around the edges to try and improve the risk/return characteristics of the portfolio. Even those who recommend the more extreme timing strategies recommend it being held alongside a core buy-and-hold portfolio. Portfolios using timing strategies are therefore never 100% out of equities, as the “X best days” argument would have you believe.

Whilst there are some legitimate criticisms of market timing strategies, just are there are some legitimate arguments in favour of using them, using the “X best days” argument isn’t one that should be being used.

Conclusion

It seems a bit disingenuous to use the “X best days” argument to imply that by not getting your cash invested and keeping it invested, you’re likely to miss out on the majority of the market’s gains, given that:

• It ignores the effects of the X worst days, which have a much larger impact than the X best days,

• The majority of those best days occurred during years where the market incurred a large loss for the year, meaning you actually would’ve been better off not investing, and,

• You’re extremely unlikely to be able to benefit from the best days without suffering the effects of the worst days, given that the best days occur within a few weeks of the worst.

Whilst it does make sense to invest rather than wait, which we’ve discussed frequently in this blog, the “missing the X best days” reason isn’t the one you should be paying attention to.

In addition, given that almost no market timing strategies will be rotating between 100% in the market and 100% out of the market, this argument isn’t a valid criticism of market timing either.

Now that we have a better idea of when’s a good time to invest, the next post in the series looks at whether it’s a good idea to invest all at once or drip feed your cash into the market.