It seems like this follow up paper clarifies the data's vision a lot more. Notable changes from the previous version discussed in a sister thread here:
- There is no more emphasis on price-only-inflation-adjusted returns. Good riddance: getting rid of dividends makes no sense and is borderline intellectually dishonest just to make the point.
- He no longer argues stocks don't work for the long run, just that bonds were as good in the past. This is a lower bar to meet as bonds in the past, especially corporate bonds as he's included, are actually quite risky!
- Finally there is some argument to be made that bonds are better investments when monitoring technology is poor -- since insiders can steal equityholders' wealth. But the 20th century invented good accounting, auditing, etc to reduce that and drive up equity returns.
Moreover the more modern approach is share repurchases, which are largely not subject to the tax drag and use the same money that was used historically for dividends.
You should be taking money out at your chosen rate, not depending on how those companies choose to allocate money between dividends vs. buybacks vs. cash piles vs. reinvestment. So treating dividends as reinvested by default makes sense to me.
If you plan to withdraw 4% per year, so you preserve your wealth indefinitely, you're more than 2 percentage points short when the dividend yield is 1.71% [1]
If you want to live solely from dividends, you'll need more than double the capital.
If you want to die with zero [2], it's impossible.
I'd much rather invest in a dividend-accumulating index fund and sell as I please.
It may be hard to imagine, but dividend yields were not always this low [1]. Investopedia has it usually something healthy over 4% up until 1990s it seems. Over that 1926- time frame, dividends are said to have contributed 32% of the total return of S&P 500 [2].
I think parent is saying that you can't die with zero if you plan to live off dividends. (Because you need to keep owning the stock throwing off the dividends.)
In most other developed countries a) healthcare is funded by the government (to a first approximation).
b) end-of-life healthcare expenditure is considerably lower outside the US.
OK sure; same's true of bonds or CDs or any other asset type though no? The point is that "invest in equities with dividend reinvestment until retirement and then buy a life annuity" is a totally viable strategy. That's basically the way pension saving in the UK works historically, for example.
Right. Life annuities may or may not be a good deal. But that's certainly the main way to not be essentially forced to pass on assets. (Modulo real estate you own and are living in.) And, as you say, defined benefit pensions basically work the same way--although, in the US, current ones are fairly uncommon outside public sector--although a lot of people still have them from years past.
Just to add. Bonds and CDs do have durations though so you can essentially do your own actuarial calculations to a certain degree.
In the UK this is actually the typical pattern for defined-contribution schemes; I haven't looked recently but it used to be the case that you were legally required to buy an annuity with 75% of your tax-deferred retirement savings.
Interesting. I'm not sure how common annuities outside of defined benefit pensions are in the US. My impression is not very.
I've noticed them mostly in the form of charitable trusts (which can offer the benefit of basically shielding large asset gains from taxation). But it doesn't seem to be a widely-used investment strategy in general. Maybe it's more common if someone doesn't have an interest in passing down any money.
Once you start selling off your assets, the """returns""" are worse, but equally so no matter what you invested in. It's better to leave that math out of the situation and look at the returns of the actual assets by themselves. Which includes reinvesting.
If you really want to factor in the sell-off, then every dollar of dividend means one less dollar of sold stock. If dividends go higher than withdrawals for a year, then you need to buy more stock to compensate. So the math comes out the same. What you don't do is ignore dividends, or let excess dividends pile up in cash form. Which the original paper apparently did.
Why would you exclude part of the total return on an investment? It'd be like ignoring the principal value of a bond because you expect to live on the coupon. Cashflows are cashflows.
There are enough “cash cow” securities that maintain a same / similar share price by distributing heavily for this to make sense. The price wouldn’t show the whole story and the cash could go much further over 100 years than just sitting in a bank account.
I don’t know many people that spend 100 years in retirement.
- There is no more emphasis on price-only-inflation-adjusted returns. Good riddance: getting rid of dividends makes no sense and is borderline intellectually dishonest just to make the point.
- He no longer argues stocks don't work for the long run, just that bonds were as good in the past. This is a lower bar to meet as bonds in the past, especially corporate bonds as he's included, are actually quite risky!
- Finally there is some argument to be made that bonds are better investments when monitoring technology is poor -- since insiders can steal equityholders' wealth. But the 20th century invented good accounting, auditing, etc to reduce that and drive up equity returns.