Why should it be that way?
The keys to economic growth he identifies are (1) property rights, (2) scientific rationalism, (3) capital markets, and (4) adequate transportation/communication. All of these appeared in sufficient form for prosperous growth several hundred years ago.
There is of course no guarantee of continued growth at same rate as last several hundred years. But given the conditions that have prevailed it has settled at a fairly stable rate as sort of a natural law.
Without two hundred years of unsustainable consumption of fossil fuels, property rights or capital markets wouldn't have given us a fraction of the economic growth that we got. Stock exchanges don't do much for you when 97% of your population are either subsistance peasants, or make hand-crafted tools used by subsistance peasants, and you have to spend 8 hours a day banging rocks together to stay warm and to scare away mountain lions.
#4 is also only possible because of #5.
"Long term" is longer than 20 years.
The 20th century was particularly good and I think the average return was closer to 6 than 3.
The assumptions should be documented there.
However, the assumptions have only 80% in capital gains while this should be much higher for long term investors.
If you’re investing in a 401k then you add about 1-3% because you won’t be taxed each year and those gains are compounded.
And it’s assuming 1% admin fees from 2000+. This is way too high and is closer to .1% starting in the 80s with vanguard index funds.
So if you invest in tax deferred index funds you are looking at 5-6% after taxes and inflation.
But I like the way this matrix displays info and I want to find a version with assumptions for efficient 401k investors.
I mostly agree with you, but this is basically the gambler's fallacy.
If I'm flipping a coin every second for days, and I hit a run of 10 heads in a row, "reversion to the mean" just means that the next 10 flips are likely to be less extreme than the previous 10. It does not mean that I should expect "more tails than usual" for the next flips. We revert towards the mean, not past it.
Unlike in a casino, the returns and value of stocks are loosely coupled to the real economy.
If stock values grow quicker than the economy, then we should expect a correction, because of that loose coupling.
Of course, markets can stay irrational longer than we can stay solvent, so I wouldn't try to time it.
This is incorrect. Prior performance of the market over the span of years has little to no predictive power on future performance of the market.
Your statement is like saying: Because I flipped a coin and got heads 10 times in a row, I'm more likely to get tails in the future.
While it's true you should expect the market to revert to the mean over the coming years, there's no evidence that it will grow 'more slowly than average' in the coming years to 'make up' for the hyper growth in past years. If you were a betting man (and a non-sophisticated investor), you should bet that the future years will grow at exactly the historical mean.
Edit: I did some analysis on historical S&P 500 pricing to validate my intuition.
On average, the monthly growth rate of the S&P in a month following a bear month is -0.39%
On average, the monthly growth rate of the S&P in a month following a bull month is 1.08%
You might argue that it takes longer than 1 month for the market correction to occur, so I've included the script and data set I used here for you to play around with: https://pastebin.com/F78pLUka. You can use any cadence, and will find the same relationship.
Empirically, you cannot time the market, which implies future growth rate is not affected by past growth rate.
For the same inputs (sequence of future cash flows, enterprise value, and weighted average cost of capital [discount rate]), the long-run value will be the same. If the near-term share price value rises more quickly than the long-run intrinsic value model, it is entirely reasonable to assume future growth of share price will moderate, as it must in order to converge on the same long-run value.
I think it's not at all like your example with 10 coin flips in a row.
I think we only value stocks the way described (intrinsic value model) because its a cultural myth to do so. We can also think of stock certificates as baseball cards - of value mostly to other collectors. Sure there's a 'book value' or 'dividend value' behind them, but that's irrelevant most of the time for most stocks.
My assertion is that you cannot time the market at all. As a result, I'd just invest passively and immediately.
For data that supports my assertion, please see the edit of my original comment.
Robert Shiller (Yale finance professor, of Case/Shiller housing index) likes to make the following analogy. Predicting the stock market is basically the opposite of predicting weather. With weather our short-term predictions can be fairly accurate, but our long term predictions are very poor. With the stock market it's the reverse, short-term predictions are worthless, but long term predictions are generally fairly accurate.