Retirement Income: Natural Yield AND Free Dividend Fallacy
One common approach to managing retirement income is to rely on natural yield from one’s portfolio. The goal of a natural yield approach is to mitigate sequence risk by relying exclusively on dividend/coupon income from the portfolio and so avoid selling down in poor market conditions.
This means that the portfolio (and the income from it) can last almost indefinitely. Dividends also tend to grow each year; they also experience less of a decline than the capital value of a portfolio during extreme market conditions, such as the 2008 financial crisis. This makes a natural yield approach the ‘perfect’ retirement income strategy. Or is it?
The natural yield retirement income approach makes a distinction between income and capital from your portfolio. Under this approach, capital is sacred and must not be touched, or you risk killing the goose that lays the golden eggs. Retirees must be content with living on only the natural income (dividends and interest) from their portfolio.
This is contrary to the SWR framework, which advocates a total return approach and makes no distinction between income and capital. The SWR framework focuses on working out - and drawing - a sustainable level of withdrawal from the portfolio, whether that comes from income or capital growth. As the saying goes: whether it’s rainwater or water from the well, water is water. It’s about drawing a level of income that would survive even the worse case scenarios.
Proponents of the natural yield approach like to pitch it as a superior alternative to the sustainable withdrawal rate framework. This makes no sense, because assuming the same overall asset allocation, it’s difficult to see how drawing only dividends can be superior to the ‘total return’ (a combination of dividend/interest and capital growth) from the same portfolio.
Yet one searches in vain to find any empirical evidence to support the natural yield approach for retirement income planning. Much of the research on natural yield as a retirement income strategy is provided by asset managers (whose livelihood depends on such a strategy) and their research suffers from three significant fatal flaws:
- Much of the research on the natural yield approach is based on limited data, going back no further than the 1990s. This gives a very limited picture of how the approach might fare under extreme market conditions, particularly periods of extreme inflation such as the 1960s.
- They typically focus on yield in % as opposed to actual income, in £/$. What matters to a retiree is how much ends up in their account every month, not the theoretical definition of yield in %.
- Income is seldom adjusted for inflation. An example of this is AIC’s Dividend Heros, a list of 20 Investment Trusts that have increased their cash dividend to investors for at least 20 consecutive years. Leaving aside the inherent hindsight bias and survivorship bias of such research, the failure to account for inflation limits the credibility of such research.
In this research, we want to illustrate how a retiree’s income by adopting the natural yield strategy might actually fare under a wide range of market conditions.
For this research:
- We use long historical monthly data for UK and US Equity, and the corresponding CPI for UK and US between Jan. 1924 and May 2019. Returns are decomposed into dividend and capital growth.
- Unlike much of our research, where we would use a balanced portfolio consisting of equities and bonds, we assume that the retiree invests in a 100% equity portfolio. This is what natural yield proponents tend to advocate. Indeed our previous work on the subject did use a balanced portfolio, so the overriding result of this research remains unchanged by the asset allocation.
- We looked at the monthly inflation-adjusted natural income (or dividend in £/$) for every rolling 30-year period within the dataset. So, we have a total of 792 scenarios, each for UK and US Equity portfolios.
- Given the large data-set, we ranked all the 30-year scenarios that start in January of each year (eg Jan. 1924, Jan. 1925….. Jan. 1989) based on the cumulative income in real terms over the entire 30-year period. We identified the worst, 25th percentile, median, 75th percentile and best scenarios for our discussion.
- We ignore the impact of fees. In practice, fees would reduce both the income and capital since fees are charged as a proportion of outstanding capital on which dividends are paid. Nonetheless, the underlying lesson from the research remains unchanged by the impact of fees.
UK Natural Yield Investor
The chart below shows the monthly inflation-adjusted dividend received by a retiree invested in 100% UK Equity for every 30-year rolling period since 1924.
And below is the portfolio balance, adjusted for inflation, for every rolling 30-year period. This illustrates how the capital value changes over time, in real terms.
The results can be summarised in the table below.
US Natural Yield Investor
To illustrate that this is not just an anomaly with a UK investor, the chart below shows the monthly inflation-adjusted income for a US retiree relying on the dividend of the S&P 500, in all the rolling 30-year periods since 1924.
And the table below summarises the results for a US investor relaying entirely on the dividends from their $1,000,000 portfolio.
The very idea of relying on dividend for stable retirement income is nothing more that a fallacy.
The illusion that relying on natural income mitigates sequence risk in a retirement portfolio is exactly that, an optical illusion.
- Once you take account of what really matters to a retiree (inflation-adjusted income), a natural yield retirement income strategy is grossly inefficient. Income varies on a monthly basis beyond what most retirees can tolerate. Retirees following this strategy must be willing to tolerate significant volatility in their income, including periods of prolonged reduction in real terms.
- In the worst case scenario, the retiree starts with a monthly income of £2,500, which then fell to £1,800 a month by the 6th year of retirement, a reduction of over a quarter. This is much lower than the worse case scenario under the SWR framework for a portfolio with the same asset allocation.
- In the 25th percentile scenario (Jan. 1968), income started out at £3,500 a month and fell consistently for the subsequent seven years. It reached a low of £2,400 a month in year seven, a reduction of over 30%!
- Even in the median scenario, where the income started out at a tidy £4,466 a month, it zigged-zagged for the subsequent 15 years, making budgeting hard for the retiree. It reached £3,258 a month in the 16th year of retirement, a reduction of 27% in real income!
- A similar pattern and volatility of income exists for a US investor who depends entirely on dividend from their portfolio.
- The capital-is-sacred philosophy of the natural yield approach exacerbates the problem of sequence risk, as the retiree has little flexibility to draw from capital. Sure, it achieves its objective of never running out, but the monthly income volatility is a significant cost to bear. The result is that the retiree ends up leaving a significant legacy, which is a bit pointless if their income needs weren’t met during their retirement.
- Under the natural yield strategy, the retiree gets the worst of both worlds - volatility of their income AND volatility of capital during their retirement. Compare this to the SWR where the goal is to deliver relatively stable income, while the portfolio fluctuates. Under a SWR framework, reduction in income is a conscious decision made by the retiree (with their adviser). With a natural yield approach, income volatility is forced on the retiree by company boards who make their dividend payment policy without any regard for the investor’s income requirement.
One of the key attributes of the SWR framework is that it aims to deliver relatively stable income, in-spite of the volatile portfolio balance.
For instance, an initial withdrawal of £50,000 a year from a £1m portfolio invested in 100% UK Equity has a success rate of 91% for every rolling 30-year period since 1915. A retiree might start out with this level of withdrawal in the knowledge that, should they encounter poor market conditions (i.e. the other 9% of scenarios), then they will need to make a conscious decision to reduce their withdrawal to avoid running out of money.
Under the SWR, the retiree makes no distinction between dividend or capital growth. They draw from their portfolio based on their income requirement, given due regard to the total return from their portfolio. There may be years where the dividend from the portfolio exceeds the SWR income requirement. The retiree merely banks this. Capital isn’t sacred and this provides valuable flexibility that the natural yield strategy lacks.
The Free Dividend Fallacy
Why do people, particularly investment professionals, think of natural yield as a superior strategy for retirement income?
This has been a puzzle for academic research for many years. In reality, dividend payment is merely a capital allocation decision made by companies. As Nick Murray puts it, ‘dividends—like share repurchases and reinvestment in future growth—represent a capital allocation decision on the part of a company’s management. That is, rational managers are constantly asking themselves what deployment of their earnings may, in their best judgment, produce the optimal return to their shareholders.‘ Accordingly, there is no logical reason to prefer £1 in dividend to selling £1 of shares. (Miller and Modigliani, 1961). But investors are hardly rational.
In their paper, The Dividend Disconnect, Prof Samuel Hartzmark of University of Chicago and his co-author David Solomon offer an explanation for why investors tend to prefer dividends. They document what they call the ‘free dividend fallacy’, a behavioural error where individual and professional investors systematically treat dividend as free money that is disconnected from capital value or share price. Many investors view dividends as an independent source of income, like payment from a bond. The result is that, in a low interest rate environment, demand for dividend paying stocks increases and that invariably reduces expected return.
They concluded that ‘Many individual investors, mutual funds and institutions trade as if dividends and capital gains are disconnected attributes, not fully appreciating that dividends result in price decreases. Behavioural trading patterns (e.g. the disposition effect) are driven by price changes instead of total returns. Investors rarely reinvest dividends, and trade as if they are a separate, stable income stream. Analysts fail to account for the effect of dividends on price, leading to optimistic price forecasts for dividend-paying stocks. Demand for dividends is systematically higher in periods of low interest rates and poor market performance, leading to lower returns for dividend-paying stocks.‘
There are other problems with natural yield investing:
- Generally, dividends are taxed more severely than capital growth (both at corporate and individual level), so a natural yield portfolio is inherently less tax efficient.
- Some of world’s fastest growing companies do not pay dividends, e.g. tech companies like Facebook, Amazon, Alphabet and Netflix. A natural yield portfolio that excludes these companies misses out on potential growth. Since dividend payment is a capital allocation decision, one reason these companies don’t pay dividends is that they can deploy capital better than hand it back to investors. Excluding these companies from your portfolio increases portfolio concentration at the very least. At worse, it reduces potential return in the long term.
To conclude, the natural yield strategy is ill-suited for retirement income. The volatility of income makes budgeting incredibly challenging for retirees. This income volatility is exacerbated by the lack of flexibility that results from viewing capital as sacred. Furthermore, that is no empirical evidence to support why an investor should prefer dividend over capital growth.