Friday 12 February 2010

The 4% Retirement Withdrawal Rule Gets Potshots from Nobel Laureate

Came across The 4% Rule—At What Price? by Jason S. Scott, William F. Sharpe, and John G. Watson (SSW) on the A Loonie Saved blog (who posted some interesting comments on the paper). Sharpe won the Nobel Laureate for Economics in 1990 so it behooves us to pay attention, especially when he deals, along with his co-authors who, of course, should share the credit equally, with the critically important topic of how to manage one's finances and investments in retirement.

The 4% rule states that a retiree can withdraw 4% of portfolio value, adjusted upwards for inflation each year, based on the portfolio at the start of retirement. This provides a constant real spending amount. The portfolio is assumed to be a mix of about 40% bonds and 60% equity. Based on the past history of the US market going back as far as the late 1920s, various researchers have found that no portfolio would ever have run out of money in 30 years or less. Who can argue with that?

SSW have harsh words for the 4% rule, calling it wasteful, inefficient, fundamentally flawed. Their main criticism: "A retiree using a 4% rule faces spending shortfalls when risky investments underperform, may accumulate wasted surpluses when they outperform, and in any case, could likely purchase exactly the same spending distributions more cheaply." And the cause they say is the bad strategy of trying to match volatile investments with a desired fixed income stream. They claim that even highly diversified low volatility portfolios have some risk of running out early in a 30 year retirement horizon - a 3.9% risk for a portfolio with 9% standard deviation. They also say that portfolios waste10-20% of the initial funds generating un-necessary surpluses and another 2-4% spending money the wrong way. They ask, why bother with the 4% rule when a risk-free inflation adjusted bond (RRB) can guarantee 4.46% for 30 years?

My comments
The logic is fine as one would expect, but are the assumptions complete and correct and do practical realities affect the bottom line? To some degree I think they do.
  1. Life expectancy - Will you live exactly 30 years? Do you want to plan retirement based on that certainty? At age 65, there is a 6% chance that one of a couple will still be alive 30 years later. After 30 years, the real return bond is done, there is no money left with 100% certainty but a portfolio will survive (... most likely? [there being arguments about whether 4% is sufficient to guarantee perpetual life, according to market history, or possibly less than that using Monte Carlo simulation]). If you die before 30 years, then you also have a "wasteful surplus" using an RRB. The waste label applied to a portfolio compared to an RRB only is true in the context of a retirement spending plan which lasts exactly 30 years.
  2. Surplus = Legacy - maybe the implicit spending on surplus isn't so bad since many people want to leave money to the next generation. Any money left over is not therefore wasted. With the RRB method, one would need to decide at retirement how much bequest money to leave and set that aside, which would of course mean a reduction in net funds available to live off, i.e. a lower withdrawal rate
  3. Returns - SSW assume a 2% return on RRBs. At one time that may have been true and it may happen again, but right now Canadian Fixed Income is showing yields on Canada RRBs ranging from 1.36% to 1.60%. The little chart below shows how withdrawal/income rates change according to yield and length of retirement. So, if you pick 1.4% as the current base and 35 years to be safe(r) the withdrawal rate is 3.63%. The 4.46% RRB withdrawal rate that SSW cite doesn't seem so available any more.
  4. Transaction costs - will also affect net returns; to be realistic, one would have to get real data on rebalancing costs, purchase and sale of securities and I am not sure how it would work out in a stock/bond portfolio vs RRBs. Maybe it would not even be possible to buy a complete ladder of RRB strips for the time required. The longest maturity Canada RRB is 2036, only 26 years away. What would you do, put 4-5 years worth of money in a money market fund with about 0% effective yield?
  5. Taxes - it doesn't matter in a tax-deferred account since all returns are treated as income and only when withdrawn, but in a taxable account the favourable tax rates on capital gains and dividends that would come from a portfolio mean better after tax money to spend. RRBs in Canada in a taxable account get taxed on the interest (or imputed interest) and the inflation adjustment component and that happens every year so net returns are reduced quite a bit.
Bottom line:
There is much sense to SSW's basic idea that one should match the nature of asset (income generating) cash flows with liability (living expense) cash flows. Essential living expenses in retirement need to be supported by an equally solid and steady income stream. RRBs and government CPP and OAS/GIS fit into the highly certain category. Non-essential / luxury retirement spending is quite flexible and controllable and that's the bit a portfolio can fund, whether it's according to the 4% rule or other types like 4 or 5% of year-end portfolio value, or year-to-year floor and ceiling limits.

1 comment:

Patrick said...

Thanks for the link! Your analysis here touches on all the "real world" shortcomings of the SSW paper.

Wikinvest Wire

Economic Calendar

 Powered by Forex Pros - The Forex Trading Portal.