Does Long Term Buy & Hold really work?

Investment Analysis Clubs Saul’s Investing Discussions
1

A little while ago TMF’s Morgan Housel sent one of famous missives to some TMF subscribers titled “Here are all the smart-sounding, tempting excuses to sell.”

He Tweeted the headline and included a graphic of the stock market index from 1950 to today: https://twitter.com/tmfhousel/status/759007162009923584 I didn’t like that chart, but didn’t really think about it too much. Lots of LTBH defense here on TMF.

However, just recently Austin Frank countered with a very well reasoned argument “Buy and Hold is a Fairy Tale”: https://medium.com/@austinfrank/buy-and-hold-is-a-fairy-tale… That’s a public site that you can read without registration.

Mr. Frank showed that for a full half of the complete time period Housel describes the market was flat. There were two periods, one 20 years and one 13 years, where the market was flat. We’re talking 1962-1982 and late 1999-2013. In the last 67 years the gains mostly came from only 3 periods: 1949-1967 and 1982-2000, with some from 2009 to today.

Frank makes a really interesting argument about when you were born being indicative of your success as an investor:

"Worst of all for buy and hold, it is heavily dependent on when you were born. If 21–25 is the prime age to start, being born in ~1929 put you in position to capture the 100x returns since 1950. And wouldn’t you know, Warren Buffett was born in 1930, Jack Bogle in 1929, George Soros 1930, Charlie Munger 1924, Icahn 1936, Julian Robertson 1932?—?sense a pattern? One of the best opportunities to ever invest in the market perfectly coincided with these all-time greats’ early 20s, making for phenomenal timing.

It runs counter to most investing conventional wisdom, but timing is the most important part of investing. And while we all love to preach about how patience is a virtue, you can’t hold it against those of us who aren’t giddy about the prospect of dead money for 10–20 year periods. 20 years is a quarter of a lifetime! There must be a better way.

Mr. Frank’s argument was so compelling that Housel felt he had to address it in a subsequent TMF letter, and he also included most of that response on Medium, where you can all read it. Housel’s defense was, in my view, completely lacking, changing the argument that LTBH of stocks is great to most people’s only other choice was bonds, which was worse. That’s nonsense, of course, but worse is that it avoids defending what was his original argument - that since LTBH was great in the past it’ll be great in the future. Since Housel has to admit that LTBH wasn’t so great half of the past 67 years, his argument that it’ll be great for the next set of decades falls completely flat - at least in my view.

Whether you believe in LTBH or not, I strongly recommend reading Austin Frank’s post and Housel’s response.

21 Likes
2

What I have found is that despite too many to remember awesome long term buy and sell investments I’ve had, that in the end, I may have done just as well, if not better, by buying and holding just a few quality names that I have owned and flipped over the years.

As example, I bought Amazon in 1999. Everyone called it “over valued” and sure enough it was at $30 billion marketcap. ISRG. Didn’t buy it at 30, 40, 60, $200, but during the housing crash it fell to $90, now $400.

Buying it and holding it, despite always being “overvalued”, through all the ups and downs of the market, just like with Amazon, was a better strategy.

ARMH same sort of thing.

The problem of course kicks in when you get the inevitable losers, or you don’t sell once winners that lose their way (such as Yahoo! or Ebay!).

Examples are for example purposes and not exhaustive of course.

In regard to discontinuous technologies and business models (as discussed below) these you can see over the last decades, one after another as a systematic thing, and very fascinating to follow and identify.

If you are going to buy the market, then your returns will not be so extraordinary. But if you are buying a few of those rare quality names, and you sell when discontinuous technology or business models start to assert themselves (even further reducing the downside to some of the losers), then buying and holding long-term has, and I believe will, produce much better returns over time, with far less stress and effort, than jumping in and out based upon market events, and fretting over 95% of the news that is just noise and is meant to create churn.

Tinker

9 Likes
3

What’s even sillier about Housel’s LTBH argument is that his metric isn’t a long term buy and hold. It’s the S&P 500, which of course changes composition over time. All of these indices (heck, just look at the name “Dow Jones Industrial Average”), “sell” and “buy” companies over time: they just do so according to a formula.

So what we’re seeing is that a formula for buying and selling has up periods and down periods. Since it’s a formula, however, you have no way to know if you’re getting in at the start of a up period or the start of a down period or something in between.

If you bought the actual stocks in the S&P 500 in 1950 at the ratios of the index at that time and literally held them until today (moving money to cash only as companies were liquidated), what would your returns be? Probably pretty lousy. This is the true LTBH, and my guess is that it’s a historically lousy way to invest.

3 Likes
4

It’s quite misleading to say that market was flat between 1999 and 2013. It sounds like nobody would have made any gains in that period when in reality you only wouldn’t make gains if you bought into the S&P in 1999 - if you bought in a year later you would have made a profit.

4 Likes
5

if you bought in a year later you would have made a profit.

Mathematically true, of course.

But in real life, how many people knew that 1999 was a bubble AND acted on it? How many people decided in 1999 not to invest and did in mid 2000?

I do think that the better way to look at history it is to pick some investing time horizon and then do a rolling comparison year by year (or quarter by quarter). Then you can see how many of the rolling 20 year horizons made money versus not.

However, the SPX is simply a formula based buy and sell investment strategy. Hence, I agree with those that use it, or something like it, as a metric to compare their individual investing performance. Because if you can’t predictable beat the S&P 500, you should save your TMF subscription fees and just buy the SPX.

3 Likes
6

Frank’s argument reminds me of the climate deniers starting their charts at 1998 (a peak hot year), then saying global warming is in a hiatus, flat. It’s a cherry-picked start date to ignore the overall trend.

Back to the S&P: Of course if you only invested at the peak of 1999’s bubble, you are not going to have had good returns. But who invested at that time and only that time? Invest in 1998 or 2000, better yet an even wider set of years, dollar cost averaging in, and your returns will follow the long term trend, which is up.

There are peaks and troughs all along the chart; moral is, don’t invest only at the peaks!

Graham

1 Like
7

It’s a cherry-picked start date to ignore the overall trend.

It wouldn’t be hard to do a rolling 20 years analysis to see how you end up.

I’ll come back to two of my own points:

  1. If everyone knew what a company would do in terms of profits and growth in the future, then everyone would know how to price that company’s stock today. Those stocks that many people agree will do well have high multiples: Amazon, for instance. So, while Amazon will almost certainly do really well, everyone knows Amazon will do really well, and so the stock is priced like Amazon is going to do really well, and so you’re not going to make a lot of money investing in Amazon today unless it does even better than other people think it will. And so, is your insight into Amazon’s future better than theirs?

  2. The S&P 500 is a formulaic buy and sell strategy (OK, not strictly speaking, like the Russell 1000/2000/3000 are) that was and is not designed as investment strategy nor vehicle. That the S&P 500 does so well compared to most analysts shows just how hard it is to predict the future.

I’m toying with the idea of a modified Russell 2000 index. As you probably know, the Russell 2000 are companies #1001 to #3000 in terms of market capitalization. There is no curation, just a strict formula. What some people like about it is that it includes “mid-cap” companies that are growing and will eventually be Russell 1000 companies - so you get to profit from their growth. However, it also includes companies on their way down: ex Russell 1000 companies on their way out. So, perhaps adding some kind of filter that only includes (aka “buys”) companies going higher in the index over some set of years and removes (aka “sells”) companies that are going lower in the index over that same set of years.

How hard would it be to construct that formula?

2 Likes
8

Whether you believe in LTBH or not, I strongly recommend reading Austin Frank’s post and Housel’s response.

For readers interested in this subject, I’d also strongly recommend reading recent e-books by William J. Bernstein. He presents (among other things) the results of his extensive research into how the beginning date for investing has a monumental effect on returns at retirement age. It is pretty dramatic.

DT

9

Whenever I see an article like this, I always head over to https://dqydj.com/sp-500-return-calculator/.

When you factor in dividends, the market had an average annualized return of 7.746% between 1962 and 1982, and 4.387% between 1999 and 2013(I used January of the first year to December of the second year).

4.387% is pretty poor relatively, but we’re really cherry picking our dates here.

3 Likes
10

And here’s a chart for historical dividend yield of the S&P 500: http://www.multpl.com/s-p-500-dividend-yield/

So, to answer the question: yes, historically buy and hold does work… except under certain cherry picked years when you back out dividends (because the earnings that companies literally pay out directly to shareholders apparently shouldn’t count).

But really, the author is correct in one sense. The dates you’re born do have significant impact on your overall return. Buffett has said many times that we was fortunate to be born when he was.

I’m not saying buy and hold is the only way to make money in the market, nor am I guaranteeing that it will make money going forward, but there’s no denying it has in the past (and I’d be willing to bet that it will going forward as well).

3 Likes
11

1999 is bubble territory.

Denny Schlesinger

12

Made a ton of money in 1999. Lost a ton in 2000-2001 and made it back in 2003-2006. LTBH does work, as does LTBS, but after
Reviewing all the effort that went into it, unless you have rock solid character, your emotions don’t gyrate with the markets ups and downs, and you never let anything waver your investing convictions, LTBH and sell when time is to sell, is the best system.

Twilio as a current example. It just keeps going up. I see no reason for this valuation. At this point one might say it is time to sell because it is truly a bubble. 20x current year sales. Then again never much hurt MBLY. But perhaps one best for LTBS at current valuation as it is at a true extreme.

On the other hand, if not at an extreme then LTBH has worked best in my experience. And this comes from someone w too many awesome LTBS investments to count. But they were all hard work. And all it takes is just one bad one to set back years of effort.

Sip a beer, LTBH a few great companies when you can get them at temporary FUD
event lows.

Tinker

5 Likes
13

I finally got around to reading Morgan Housel’s full response, and then went back to re-read Austin Frank’s article, and I noticed one other thing. Frank states:

Supposed you had sold in 1962?—?you were spooked out of the market by the Bay of Pigs debacle. Truth is, you didn’t miss out on much; nominal returns were zero for the next two decades.

What if Y2K (or obscene valuations) freaked you out in late 1999? Were you a fool to sell then? Not really. If you sold out in December 1999, you missed out on approximately zero gains in the market all the way up to 2013.

Although he uses the term ‘nominal’ in the first paragraph, I’m not sure if he fully understands it (or, perhaps, he’s trying to mislead his readers). Housel’s original chart (that Frank drew the lines on) was adjusted for inflation. The 70s and early 80s had double digit inflation some years (13.5%(!) in 1980 for example).

So to say, “you didn’t miss out on much” by pulling your money out of the market is flat out wrong. You lost a significant amount of purchasing power. To be honest, we should be praising LTBH for maintaining purchasing power through times of high inflation like that instead of damning it.

35 Likes
14

Mr. Frank showed that for a full half of the complete time period Housel describes the market was flat. There were two periods, one 20 years and one 13 years, where the market was flat. We’re talking 1962-1982 and late 1999-2013. In the last 67 years the gains mostly came from only 3 periods: 1949-1967 and 1982-2000, with some from 2009 to today.

1962-1982 was flat? Well, in the sense that you could draw a horizontal line through those two years on a chart of the DJIA. But the Dow went from about 700 in 1962 to a record high just over 1000 in 1973, then slumped back below 600 in 1976-77, then back to over 1000 in 1982, when the 18-year bull market took off. That’s not my definition of flat. The market was actually much more interesting in those years.

And keep in mind that the DJIA, or the S&P 500, or any index, measures nominal stock prices. It does not measure the return that a shareholder receives. People who held stock from 1962-1982 received good returns in the form of dividends, spinoffs, and increases in wealth from mergers and acquisition transactions that the indices don’t measure. And dividend yields were higher then than they are now.

Bill

7 Likes
15

What if Y2K (or obscene valuations) freaked you out in late 1999? Were you a fool to sell then? Not really. If you sold out in December 1999, you missed out on approximately zero gains in the market all the way up to 2013.

I bought individual stocks for the first time in 1999. I believe my money was cut in half by 2001.

In 2005, I bought a small position in Netflix.

I sold that position in 2014 for more than I had in 1999. I used the money to completely renovate my basement to include adding a full size bathroom in it.

My XIRR since 2005 to date is around 15%.

If this has been a flat market, I’m happy.

Fool on,

mazske

All positions are listed in my profile

5 Likes
16

“I’m toying with the idea of a modified Russell 2000 index.”

I recently discovered that there now is such a thing as a “value index” for large, mid and small caps.

Vanguard’s funds in this area are:
VMVAX (mid)
VSIAX (small)
VVIAX (large)

1 Like