> If one bought and held the Dow from 1921 to present, the total return would be 500,000% (5,000x)!
If we can use hindsight in our trades, I'd recommend using the numbers 60, 54, 36, 24 and 7, on last week's Powerball lottery that paid out $211mm. That's a 100,000,000x return!!
The Dow was designed for easy calculation before computers. And it's pretty correlated to the S&P 500 anyway.
In theory, price weighting is a poor way to index, but look at it from the other side of the coin - if you choose 30 stocks, from all sectors, how likely is it that you'll get a different return from the market even if you try? Not very.
It's only relatively recently that it became popular to never split and let stock prices grow without limit, too. If the prices are mostly in a small range, then the index is similar to an equal weighted one.
So version it in a repo e.g. on Github, and ensure continuity in the DJIA index from one version to the next on the date the version is committed. We have technology, we should use it.
I could add a clause that says if a share splits its weight gets multiplied accordingly, and that would have the effect of (a) DJIA stays continuous now (b) DJIA stays continuous through next split. It doesn't disrupt anything now, and prevents future inadvertent disruptions.
Anyone that is insisting we shouldn't change it is stuck in a backward age. Honestly I don't understand why there are so many no-sayers on HN. We should be building the future, not making excuses.
The index you're describing has already been created by Charles Schwab as SCHV in 2009 (and SPCH in 2005 for the S&P, and ...). There are already new indexes -- lots of them. DJIA is useful because it's an index calculated using the same metric for over a hundred years, so you can track performance for reporting and public perception.
> so you can track performance for reporting and public perception
Well you can't track performance if splits wreak havoc on it. Garbage in, garbage out. So __at least__ fix that by adjusting weight when splits happen. THEN you can track performance for reporting.
Sounds like a simple thing to solve if they just let me make a pull request for the function getDJIAValue() or wherever the hell it lives. Based on price is fine -- when splits happen just adjust the weight by the split factor. Are they really that incompetent at coding?
So change the specification. Fire the person who hasn't changed it already and schmoozing with beer in a Manhattan high rise, and put an engineer in charge.
The s&p 500 is worse In my opinion. They make it sound like it is the top 500 public companies in the us. Which is not the case. They have group that picks them based on a bunch dumb rules. I like the total stock market index's. If you have people picking what goes into the index they why bother just buy active managed index funds.
There's a published methodology on how the S&P 500 works. It's quite predictable, as this is a desirable feature on any index. While it's true that there's a human index committee that meets, it's mostly for tie breakers. They strive for diversity and typically lean on precedence to make their picks.
Also, almost nobody thinks the S&P5 is worse than the Dow. I'm not sure how you can back that up.
The S&P 500 isn't ideal, but it's much better than the Dow. Both indices are a selection of companies, but the Dow is smaller which leaves much more room for judgment calls. Additionally, that the Dow is weighted by share price is completely ridiculous.
(If you actually want to index, though, a total market fund makes much more sense)
Yea true. But most people I think know the Dow is kinda of BS handpicked list. Where the S&P 500 pretends it is the top public 500 companies in the USA. Just always kinda bother me how they advertise it. Most of them will become useless anyways. Firms are going to be using in house index so they don't have to pay the fees. Fidelity is doing this with there free index funds.
The simple premise that S&P500 weights its components based on market cap already puts it lightyears ahead of DJIA.
Also including 500 big companies is much more reasonable than including just 30.
---
People lately criticize S&P committee for the rule that a company needs to report 4 consecutive profitable quarters in a row to be included. Thanks to this they've missed the boat on Tesla and they had to eventually include it as the 8th largest component.
Someone on this forum joked that they should amend the rule to say: A company needs to report 4 consecutive profitable quarters in a row and the CEO name must not rhyme with melon tusk.
There are funds that are more diversified than funds tracking S&P (e.g. $VOO tracks S&P 500). $VTI should represent the total US market. $VWRL should represent the total worldwide market.
Not sure about “better to own” though - that depends on your risk profile. Of course both of them underperformed S&P 500 in the recent history.
Dow kicks out underperformers and brings on good ones from time to time. That ensures only the good ones remain. I wonder how those changes affect performance.
The Dow is arbitrary. It happened to do well, if you bought a Japanese index in 1929 you’d do well for 60 years and then make losses which would require another 30 years to recover from.
You're measuring conveniently from the top of the epic bubble in the late eighties. But yes the US has been a better long term bet than Japan. That doesn't mean it's "arbitrary".
I’m not measuring from anywhere, I’m describing what happens if you invested over the same time period as the original comment with the same strategy but a different index.
Nor did I say the US is arbitrary, I said the Dow was, which it is given it’s fairly illogical weighting rules and arbitrary size limits. But the US is also somewhat arbitrary, it happens to have been the best performing economy over a period in which it won a World War and Cold War, and became the largest power in the world. To say it is the obvious choice now is to make a massively uncertain bet. I’m sure in 1929 the obvious choice was Britain who at the time controlled the largest empire ever, rather than the US, yet a British index would not have performed nearly as well.
Your 30 year comment is measured precisely from the top of the epic Japanese stock bubble of the late eighties. Whether you are aware of that or not. Measure from the bottom shortly after, and it's a very different story.
You are right about the arbitrary rules of the dow. If you're going to do index investing, I would much rather the S&P 500.
Well yes. We are talking in the context of buying an index in the 20s and holding it until the present day. I split the Japanese index’s performance into two periods, but I’m not sure what your point is because I’m commenting on the entire 90+ year timeframe. Buying straight after the crash would obviously have a different outcome, but that seems like you’re cherrypicking a particular date rather than me.
I shouldn't have to spell it out so much for you. You have no return for thirty years only if you cherry pick the top of that bubble in the late eighties. Start anywhere else and there is a positive return (ie what you said is false unless you cherry pick the start). You can do the same cherry picking in the US market and people frequently do. To look at long term average returns properly picking fair start and end dates is important. And then remember, we won't see growth over the next 100 years like the last hundred, so expect much lower returns. But the same is true for other assets as well. Bonds have negative real yields right now.
But I never said cherrypicked the 80s, I merely described what happened then. The entire point of this comment thread is that buying an index in the 20s is a strategy highly dependent on the index one chooses, with certain indices suffering for decades. I never said that one would be down overall, merely that anyone holding over such a long timeframe may need to wait decades to recover from certain losses (so in fact the opposite of losing money). The whole point of index investing is buying and holding, so anyone in the Nikkei for the long time had to hold through the bubble in the 80s.
If you pick the top or bottom of the market as a starting point you can tell a very misleading story about returns over time. Starting and ending points matter a lot.
I didn’t though. That wasn’t my starting point, the 20s was. It’s just the 80s are when things start going wrong. The Nikkei declined over years, hitting just over 8k in 2003, whereas post bubble it went from mid-30s to mid-20s. The bubble wasn’t even the worst of it.
"and then make losses which would require another 30 years to recover from."
Make losses starting from that high in the late eighties. Why choose that date to start from. Start a decade earlier and it doesn't look so grim. Look at the average return since 1929 and it's good as well. Pick a better date than 1929 and get a fairer picture.
Are you missing the ‘and then’? It’s a compound sentence, I was just stating the point at which the losses occur. I didn’t ‘choose’ that date, that’s just when things happened. The start point was the 20s as is the point of this entire comment thread.
It's 30 years to recover losses from the high. That sounds bad, but how long to recover losses if you don't count the ridiculous paper gains during the bubble? The point is you're cherry picking to give that statistic.
You gave a statistic that is meaningful only in the context of picking the worst possible date for comparison. Surely you can understand that.
I chose no dates. I described what happens if you used the same dates as the original comment at the root of this thread. The American indices recovered from the 2008 crash in a couple of years. Japan’s problems are far worse than a single stock market bubble in the 80s.
I see two options (or, 3 options, depending on how you count it?): Either we interpret this taking probability to refer to some objective probability distribution, or one takes it to refer to a subjective probability distribution.
In the first case, uh, one can either talk about the empirical distribution, and like, because the downside is limited ("all the money one invested in it"), if one puts p-value-ish confidence intervals on it (I don't know the right way to talk about this) on it, and, assuming we treat the behavior at different times in history as comparable, or like, if we assume that the current moment is randomly sampled from the time in which the stock market exists, or something like that, uh, I imagine that it would be possible to say something like "If the stock market across time-as-a-whole and like, selecting when you buy in and cash out at random, with like, some bound on how far apart those two are, had a negative expectation value, then we would see behavior like this with probability less than p" ? I don't know what the value of p would be, but, I suspect it would be fairly small, at least, for some ways of formulating the statement.
Or, one could take a subjective probability distribution view of, uh, what the unknown objective distribution is, and so the statement that it has a positive expected value is just a statement that one assigns a high probability to it having a positive expected value.
Or, one could just take a subjective probability distribution view of like, how it will behave, and interpret the statement as subjectively assigning positive expected value of investing in the stock market.
I think? This seems to make sense to me, but, I've not like, read much about philosophy of probability or whatever, and also I could be missing (or wrong about) some of the math.
But, in any of these 2 (or 3) cases, it doesn't make sense to me to say that it is "impossible to say" whether "the stock market in not a game with net negative expected value".
The underlying factors that drive and produce the outcomes of these two systems makes me think your comparison is a false equivalence.
I guess there's an element of randomness to everything and any company could go under overnight, but in general I think you can rely on hindsight in terms of performance. The higher you get up the ladder and aggregate, the more stable I feel that is.
If you're just comparing the methodology of extrapolating expected returns from hindsight I would agree with you.
I guess what I'm saying is using hindsight to pick a direction seems logical. And doing it from a broad perspective seems more stable (using hindsight to predict if a company will go up vs using hindsight to estimate if a sector ETF will go up). However, I agree that I don't think you can estimate how much it will go in that direction.
There's probably a fallacy in my logic there somewhere. But the results in the real world are good. If those returns are all just luck, well at least I had fun and made some money.
Based on the example of USA Inc? Over this time period what about Germany Inc? Or China Inc? Had Hitler used his artillery to wipe out the British Expeditionary Force retreating at Normandy, UK Inc investors would have lost their shirts.
If we can use hindsight in our trades, I'd recommend using the numbers 60, 54, 36, 24 and 7, on last week's Powerball lottery that paid out $211mm. That's a 100,000,000x return!!