JohnB wrote:Say your portfolio returns 3% dividends, and 1% capital gain over inflation, averaged over a decade. You could just take the dividends, but you'd be getting richer and be short of income, which seems odd.
And it's very obviously not a good idea to do your financial planning that way if you're going to be short of income - because the capital growth is
not reliable, even over a period as long as 10 years. For instance, imagine that the portfolio concerned holds a FTSE 100 tracker. Some selected FTSE 100 values, for April 9th or a close trading day if it fell on a weekend (April 9th was the day I wrote this bit of the post - it's been sitting around unfinished for a while! And the 2019 value is an intraday one, while the others are at the close):
1994: 3120.8
1999: 6472.8
2004: 4489.7
2009: 3983.7
2014: 6635.6
2019: 7455.4
Those indicate that the annualised capital growth rates over the 10 years covered are:
1994-2004: +3.7%
1999-2009: -4.7%
2004-2014: +4.0%
2009-2019: +6.5%
What's it going to be over the 10 years to 2029? Based on that data, I have no real idea... And while one can find all sorts of predictions about such things if one looks for them, I very much doubt that those who make them have any real idea either!
Even looking at the couple of 20-year periods those figures cover, the annualised rate for 1994-2014 is +3.8% and that for 1999-2019 is 0.7%, so pretty significantly different from each other. That's up in the rough area of the typical lengths of time people live in retirement, so while one might get a better estimate of the long-term capital growth rate by looking at yet longer periods, that isn't very relevant to the fundamental unreliability of capital growth for most
practical planning about living off one's investments.
Dividends aren't all that reliable either, of course. My HYP's dividend income reduced by about a third between the 2008/9 and 2009/10 tax years, tjh290633's reduced by about a half. That was a very unusual year's change, but there
is a risk of something like that happening again. And unless you plan to be able to live on quite a lot
less than the portfolio's dividend income and/or have a good-sized reserve of cash to supplement a shortfall in dividend income, you
have to have at least a contingency plan to sell part of the portfolio if the dividend income is doing badly enough.
There is a major drawback to all such plans, which is that the stockmarket tends to be low when dividend income is low and high when it is high. So the times that you're especially short of income and so particularly need to sell shares to make up the shortfall are also the times when you can't get all that much more for them, and so will have to sell a particularly large number of them. And if your plan also then involves buying at times that you have dividend income surplus to what you planned on, to restore the dividend income to its expected level, share prices are likely to be high at the time, so that you'll get relatively few shares. I.e. such plans have a somewhat hidden Buy High, Sell Low aspect to them -
not a desirable feature of an investment plan!
But having such a feature is basically unavoidable, at least on a contingency basis if things go sufficiently badly for sufficiently long, so all one can do is try to make it as small a feature as possible. In particular, plans that
only involve selling shares on an if-things-go-sufficiently-badly contingency basis, rather than routinely, fairly obviously have a smaller component of it. And that's basically where planning to use only the dividend income of a HYP to live on comes into it - and in fact, especially for someone who is retiring early, I would recommend planning to use
less than the dividend income of a HYP to live on, and use the surplus income to cushion dividend cuts as necessary, to build up a good-sized cash reserve, to grow the HYP to be yet bigger so that you can withstand an even worse period of dividend cuts, and/or for 'bonus' spending that you can cut back on if and when necessary. The exact mixture of those uses depends too much on personal circumstances and preferences for me to be able to say which any particular HYPer should choose, but give
serious consideration to the consequences of the high unreliability of capital gains,
and to those of the lesser but still very significant unreliability of dividends when doing the planning. They mean that any future projections of one's financial future should
not be based on average cases, but on a scenario which is highly likely to be exceeded.
Also, that principle needs to be extended to one's expected remaining lifespan. In particular, suppose that you're 60 when you want to retire and you check up on your expected remaining lifespan and get an answer of 20 years (in the right ballpark, but I haven't checked the exact figure). So you decide to go for a capital drawdown scheme that aims to reduce your portfolio to a low level (as close to zero as possible) in 20 years' time. Knowing the unreliability of all returns from equities, you instead put your capital into bank deposits, gilts or other investments with highly reliable returns. A good plan? Well, no - not unless you want to take a roughly 50% chance of running out of money at age 80 because you happen to be one of the roughly half of people who turn out to live longer than statistically expected. Maybe that's a risk you're happy to take, but if I were to use a capital drawdown plan, I would want to do my planning on the basis of the age that I'm at least 90% likely to die before reaching (which is probably somewhere in the 90-95 range) - and I'd prefer 99% or even 99.9%! That pushes the number of years a capital drawdown plan needs to cater for up quite a lot, which means that if one uses such a plan, one will probably actually die with quite a lot of capital still intact. But basically, because of the uncertainty of the age one will live to, one can trade off the risks of dying with unspent capital and living destitute & very old off against each other: one can reduce either a long way, but only at the expense of increasing the other significantly. And I know which of those risks I'd prefer to take!
Returning to HYPs after that excursion into investments with much more reliable returns, the real question is what withdrawal rate it's prudent to use in one's planning. I've found a very useful free tool to use in such planning is
FIRECalc - it's got its limitations for me and for HYP purposes, especially the fact that it's US-based rather than UK-based, doesn't allow one to choose what types of equities one invests in, and doesn't distinguish between dividend income and realised capital gains, but it does have the advantages of being based on real-world data and allowing one to see a good range of outcomes from a scenario rather than just the average outcome. Their default scenario of $750k capital, £30k income (inflation-adjusted) taken for spending (so an initial 4% withdrawal rate), 30 years to go (in the "Start Here" box on the right hand side) is quite instructive: if you click on its "Submit" button, you'll get a set of numerical statistics about the 118 30-year periods in their data. The average final portfolio value (also inflation-adjusted) is $1,395,451, which is about a 2.1% annualised real capital growth rate over 30 years. So just looking at the average outcome suggests that one might reasonably expect to be able to take withdrawals that are initially 4% of one's capital and rise with inflation, and still see one's capital rise in the long term, and one might naïvely decide that a plan based on doing that was pretty obviously on solid foundations.
But when one looks at the statistics of the 118 individual outcomes, 6 of them run out of money in the 30-year period, and looking at the chart as well, one can see that the first of those did so in year 24. It's fairly clear (at least to my eye) that quite a few over half of the lines end up below the ~$1.4m average - the average is pulled up by the fact that a few of the best 30-year periods do very well indeed, with the best line ending at over $4m. And it's also fairly clear that several more of the lines have got so low that they're heading inexorably towards running out of cash - which one can confirm by changing the period to 40 years and rerunning. The number of run-out-of-money periods rises from 6 out of 118 (5.1%) to 16 out of 108 (14.8%) - which at least for me would be a move from a rather uncomfortably high chance of ending up old and destitute to a
very uncomfortable chance of it. Though of course, how relevant that change is does depend on one's age: the chances of living another 40 years are very low indeed for a 65-year-old retiree, but rather higher for a 50-year-old early retiree. (One can combine the figures with longevity data so as to assess the combined effect of equity uncertainties and date-of-death uncertainties somewhat more definitely, which is something I did back in 2010 for
this TMF post, but that takes quite a bit of careful work - more than I'm going to tackle at present!)
The net result of all that is that FIRECalc suggests that a 4% withdrawal rate can easily be uncomfortably high, though whether it actually is will depend a lot on the HYPer's age and preferences about just how far they want to drive down the risk of ending up old and destitute. And I don't think I've ever seen a HYP designed that had a portfolio yield below 4% - individual holding yields below 4%, yes, but not the overall portfolio yield. So in practice, if you want to run a HYP and take an initial 4% income from it, you can easily do so from the dividend income alone - no need for share sales to boost a 3% dividend yield up to the 4% you'd like, and if you were to make the share sales anyway, you'd just have to decide what to do with the surplus cash. And in fact, you're probably going to be getting such cash in the form of surplus dividend income anyway and having to decide what to do with it...
One other point to take into account: the FIRECalc graph and its wide range of outcomes suggest that just taking an inflation-adjusted constant income is ultimately likely to end up in a vicious circle of reduced capital producing reduced returns, meaning more capital has to be taken out with sales and so yet further reduced capital, or in a virtuous circle of increased capital producing increased returns, meaning there's yet more surplus income to reinvest to increase capital yet further. And if you increase the period to e.g. 100 years (as a thought experiment, of course - it's not feasible for individuals in practice!), it becomes apparent that that really is what's going on - virtually no lines in the resulting chart end up near an inflation-adjusted $750k, but instead about a quarter of them have ended up in the vicious circle and have gone strongly negative, while the other three quarters have ended up in the virtuous circle and have inflation-adjusted values far above $750k. The point is that if one sees that one has got into the virtuous circle, one can easily cut back on its dying-with-unspent-capital effect by awarding oneself a higher-than-inflation 'pay rise', but if one sees that one has got into the vicious circle, that's much harder to break out of - and very conceivably impossible to do so, leaving one watching one's capital inexorably draining away and merely able to wonder which of capital and life one is going to run out of first... I.e. the vicious circle is a trap, the virtuous circle isn't, so for me at least, it's much more important to avoid straying into the vicious circle than into the virtuous circle.
So to summarise the main points of the above:
* Future capital growth is very unpredictable, even over periods of 10 years and more. (If one needs to cash in on it or be "short of income", one is actually probably short of capital!)
* Dividends are also rather unpredictable, but not as unpredictable - so it's only prudent to rely on rather less from them than your portfolio can produce. (In particular, one can very safely predict that they won't go negative - which is not the case for capital growth, even over 10 years...)
* Good financial planning for retirement income needs to involve taking all that unpredictability into account by making a serious attempt to look at the range of plausible outcomes, both for one's investments and one's lifespan,
not just look at the average case.
* While some types of equity investment strategies do have a lower dividend yield than a safe withdrawal rate, and so need the dividend income supplementing with sale proceeds if one is going to withdraw at that rate, HYP strategies have significantly higher dividend yields that may well be (and IMHO probably are) above the safe withdrawal rate.
* If one wants to keep the risk of running out of capital before one dies down to a reasonable level, plan for living considerably longer than one's statistically-expected remaining life expectancy. In particular, this affects how long any capital drawdown planning one does should assume one is going to live, making it a lot longer and so making the amount of capital it is reasonably safe to draw down per year a lot smaller. Just how much longer that should be depends on one's age and how big a risk one is willing to take, but think in terms of at least a decade longer.
* Erring on the high side about how much capital one leaves oneself with is easily corrected. Erring on the low side is
not!Gengulphus