The Four Pillars of Investing
On the Grannan family’s recommendation, I recently read William Bernstein’s The Four Pillars of Investing. It’s really good, and I see no reason not to recommend it to more or less everyone.
In fact, I wish I had read it as soon as it came out. It’s quite straightforward, and while I need to read it again to go over some of the math in it, close mathematical study isn’t really necessary to get quite a lot of value from reading it.
It is US-centric and aimed at Americans, but I think that it could be adapted for use outside of the US without much difficulty—most of that difficulty would probably be in relation to tax codes, and in researching alternatives to some of the US-based services he recommends.
A core message of the book is that you should be taking care of your financial future by investing. This is something I have absolutely not been doing, and therefore really need to do some catching up on. (Brian has been telling me this for years, and he’s been right.) At the risk of hypocrisy: if you’re not investing your money now, you should start, immediately, if you have any intent of having a comfortable retirement. Reading The Four Pillars of Investing would not be a bad way to figure out how to go about investing.
Bernstein emphasized that you shouldn’t trust stockbrokers, an argument that appears strongly backed by the facts. He also argues against mutual funds and for index funds, which makes a lot of sense to me, because an index fund is (by definition) going to track the market, while a mutual fund will either do better or worse—and the odds are that it will do worse over time.
One of the reasons it’s likely to do worse over time is another key point in the book: you can’t beat the market, and neither can anyone else. Not in the long run, anyway. There are always people who get lucky. But planning on getting lucky as a way to fund your retirement seems rather dodgy to me.
(Naturally, there are also one-offs where people beat markets, but these are usually not the stock market per se but very clever exploitation and arbitrage of other, limited, markets. If you’re doing that kind of thing, maybe you don’t need the book, but on the other hand, it wouldn’t hurt to put whatever you earn through that arbitrage into an index fund…)
Bernstein also convincingly makes the case that high risk and high returns go together, and that it’s more or less impossible to (legally!) guarantee low risk and high return.
He also puts forward the claim that your investments should be boring. The money you’re putting towards retirement should have a long-term strategy that’s unrelated to fads and fashions in the finance world. It should be something that you let quietly accumulate over time, and not something that provides excitement, or that provides ammunition for bragging to your friends and neighbors. Finally, in order for it to work, you have to think in the long term, and the long term is generally a lot less exciting than the short term…
Again, I highly recommend it. Even if you don’t have a lot of money to invest, it’s got good tips in it, and will help you on the road to finding a good way to invest whatever you’ve got.
10 May 2007 at 21:17
I would like to read this book, because I’ve heard views of this kind from various sources and have always been skeptical. While I agree that up to half of index funds do not perform as well as index funds (hence the popularity of ETFs), this does not mean that there is no value to hiring competent advisors to help you pick which companies to invest in. In my personal experience as an investor since 1993, I’ve always been happy with my mutual funds (all 4- and 5-star funds using the Morningstar system). In fact, in every year since then, these funds have outperformed the corresponding index. This is no surprise, as this is the definition of 4- and 5- star funds. Moreover, this is not merely blind luck— the additional value of active management can be accurately measured.
16 May 2007 at 11:37
You should definitely read the book, and then apply your skepticism directly.
I must admit to some skepticism about your mutual funds having outperformed the market for fourteen straight years! Is that their entire system, or have you been optimizing it in some manner?
17 May 2007 at 14:30
My mutual funds haven’t outperformed the market in every one of the last fourteen years, but they have done substantially better than most mutual funds in most of those years. Hence their status as 5 star funds. The Morningstar system examines performance compared to the appropriate indexes over 1, 3 and 5 year period, factoring in the added cost of active management. Consider, for example, the Janus Contrarian Fund or the T. Rowe Price Capital Appreciation Fund
which have performed admirably well in both the short and long term. As they always say, past returns are no guarantee of future performance, but a fund such as this has a long track record of success, but surely these guys are doing something right. Thus, I don’t mind giving up <1% of my earnings in exchange for the services of these fund managers.
I propose a wager. You and I can each invest a modest sum of money in the instrument of our choice… let’s say $500. You choose your favorite index fund and I will choose my favorite mutual fund. After a reasonable period (let’s say 3 years) we will calculate the difference in value between the two investments. Then, the person with the lower-valued investment will pay the other person the difference in value between the two. In other words, I bet double my money that my strategy is better than buying index funds.
Lev
20 May 2007 at 15:58
I’m planning to start my own hedge fund. Its going to use quantum physics to predict stock price changes. I’m going to pull a random equation from my friend’s lecture notes (he studies theoretical physics) and use that to pick stocks. I bet it will do as well as all the others.
Seriously though, mutual funds are a load of rubbish. Excluding the fact that they’re insured (hedge funds need not be, that is the major difference), you are essentially paying someone else to randomly invest your money on your behalf. You may as well just randomly invest it yourself.
Index funds are less prone to risk, but of course, the market as a whole can still decrease in value overnight, and its impossible to predict.
20 May 2007 at 16:07
“a fund such as this has a long track record of success, but surely these guys are doing something right”
Consider this example, lifted from Taleb’s “Fooled by Randomness”:
Take 10,000 companies who return barely the risk free rate. They engage in all forms of volatile business. At the end of the first year, we will have 5,000 “star” companies showing an increase in profits, and 5,000 “dogs”. After three years, we will have 1250 “stars”.
21 May 2007 at 03:45
Lev: that sounds like a reasonable idea, we can discuss implementation details when I get back.
Niall: I’m inclined to agree with that assessment, not just because of The Four Pillars but also due to chicanery I read about in Infectious Greed, where investment banks create funds, designed to offload risk, that are sold to mutual fund managers because they’re quite likely to do well in the first two years and badly afterwards, but by the time the risk hits, the mutual fund manager has gone somewhere else at a higher salary due to the success from the first couple of years.
21 May 2007 at 22:14
Niall— I would be happy to extend my wager to you as well.
Lev
22 May 2007 at 16:16
Lev – I have no particular interest in your wager as it proves nothing. If I took you up on it I would simply be gambling. I don’t particularly mind gambling – in fact I am already gambling my own money in the stock market.
My point is to be wary of Wall Street and don’t fool yourself into a false sense of security. Mutual funds gamble on your behalf.
22 May 2007 at 16:35
To further illustrate my point, I used Monte Carlo methods to write the following simulator in Python:
import random total_funds = int(input("How many funds at the start: ")) years = int(input("How many years running: ")) end_funds = total_funds for i in range(0, years): for j in range(0, end_funds): r = random.randint(0,1) if not r: end_funds -= 1 print "After %s years and %s total funds, completely randomly, %s funds have made a profit each of those years consecutively" %(years, total_funds, end_funds)Sample runs:
$ python mutual-fund.py
How many funds at the start: 1000
How many years running: 3
After 3 years and 1000 total funds, completely randomly, 129 funds have made a profit each of those years consecutively
$ python mutual-fund.py
How many funds at the start: 1000
How many years running: 5
After 5 years and 1000 total funds, completely randomly, 35 funds have made a profit each of those years consecutively
$ python mutual-fund.py
How many funds at the start: 1000
How many years running: 7
After 7 years and 1000 total funds, completely randomly, 5 funds have made a profit each of those years consecutively
23 May 2007 at 11:24
From: http://en.wikipedia.org/wiki/Mutual_fund
“Historically, only a small percentage of actively managed mutual funds, over long periods of time, have returned as much, or more than comparable index mutual funds. This, of course, is a criticism of one type of mutual fund over another.”
And
“Certain empirical evidence seems to illustrate that mutual funds do not beat the market and actively managed mutual funds under-perform other broad-based portfolios with similar
characteristics. One study found that nearly 1,500 U.S. mutual funds under-performed the market in approximately half of the years between 1962 and 1992.[7] Moreover, funds that performed well in the past are not able to beat the market again in the future (shown by Jensen, 1968;
Grimblatt and Titman, 1989.”
29 May 2007 at 15:52
Niall— your Monte Carlo simulation (gambling? ha ha!) is a nice piece of code. However, I’m not sure that we’re on the same page here. I don’t dispute that there is a more or less normal distribution of mutual fund performance around the mean market return. But this doesn’t take into account that some funds consistently perform better than average. This is the small percentage of funds that Harry has mentioned. The Morningstar rating method attempts to identify those funds using a set of criteria concerning long term risk and return. Here’s a link to a document describing the methodology . The Morningstar rating system measured accurately models volatility as well as risk-to-return. Thus, while not all mutual funds with a high annual return earn a high Morningstar rating, all mutual funds with a high rating have a higher than expected long term return.
Thus, while it may be true that the typical actively managed mutual fund doesn’t do better than a comparable index fund, empirical evidence does demonstrate that some of them do. Of course, you can’t be certain that any one well-performing fund has benefited from active management, or, as in Niall’s Monte Carlo simulation, simply lucky. In Modern Portfolio Theory, the this uncertainty can be quantified in the alpha coefficient
a measure of risk-adjusted excess return compared to an index.
If the process of picking winners is totally random, a mutual fund picked five years ago on the basis of its high Morningstar rating is today equally likely to under- or outperform the market. Moreoever, it’s not enough to merely beat the index to earn a high rating; the Morningstar method weighs the margin of out-performance as well as the volatility from year to year. In your Monte Carlo simulation, all it takes for a fund to score as “successful” is for it do infinitesimally better than the index. If you were to set boundary conditions such as “1%, 2% or 5% better than the index per year,” I’m certain that the number of funds that make the cut after 3, 5 and 7 years will drop considerably. In a totally random system, there is high year-to-year volatility in the performance of any given mutual fund relative the market as a whole.
The corollary to all of this is that there exist mutual funds that consistently perform worse than the market and active management is contributing to their poor performance. These would be the funds with a Morningstar rating of 1.
29 May 2007 at 16:59
As a followup, take a look at this article summarizing the performance of funds selected as standouts by Morningstar’s analysts. They claim that 70% of the mutual funds they recommended have outperformed the market over the last three years; that value is 65% for the last five years. This translates to a 9.56% return compared to 7.76% return for sticking with an index fund. Of course, Niall would say that this is just dumb luck and that they might as well be pulling fund ticker symbols out of a hat. But this sort of difference on return compared to index funds has been the case with Morningstar for years. Luck or analysis? Only your hairdresser knows.
21 Jun 2007 at 08:04
Well, I rather liked the book myself, in fact it’s my favourite investment book.
Mike