Monday, December 20, 2010

Jim Otar on Safe Withdrawal Rates in Retirement

Trying to figure out how much money you can safely withdraw from an investment portfolio each year is challenging. Some use rules of thumb such as 4%, but the real answer must depend on the types of investments and total fees and commissions paid each year. Jim Otar has studied this problem extensively, but has a questionable built-in assumption.

Otar’s book, Unveiling the Retirement Myth, contains a near endless supply of worked examples where Otar checks the likelihood of running out of money in retirement based on real historical rates of return over the last 100+ years.

Readers are told to “ignore any retirement plan that includes a forecast” but implicit in Otar’s analyses is the assumption that the future will look like the past – plus a twist. The author replaces historical dividend returns with roughly the current dividend level: 2%. Otar says that using historical dividend yields “creates an artificially higher degree of outperformance compared to prevailing dividend yields.”

If this is true, then isn’t it also misleading to use historical bond returns when currently bond yields are very low? And isn’t it overly pessimistic to use historical inflation figures when inflation is low right now? This handicapping of stock returns affects almost every worked example in the book.

Some might suggest that it makes sense to handicap stocks somewhat because they are risky. The problem is that the author ends up handicapping stock returns doubly. First Otar removes most of the dividends from stock returns and then he analyzes the volatility of these lowered returns to recommend even lower allocations to stocks in retirement portfolios.

It’s no wonder that Otar recommends “never allocate more than 50% of your assets to equities in any portfolio, ever.” Another consequence of the reduced dividend assumption is the conclusion that the percentages of stocks and bonds in a portfolio doesn’t seem to affect how long it takes to run out of money. Yet another is that it makes various types of annuities including the variable types with guaranteed minimum withdrawals look better than they are.

The approach Otar takes to analyzing portfolio longevity in the first half of the book is very interesting. I’d like to see it done using more reasonable return assumptions. Chapters 21 to 26 about seeking positive alpha are mostly an exercise in data mining. Overall, I’m glad I read the book for some useful ways of thinking about retirement, but I disagree with many of the conclusions.


  1. Interesting. Would you consider doing a follow-up post describing the aspects of his approach that you agree with?

  2. The author seems to be arguing that company executives are incented to pay out less to shareholders in the form of dividends in order to buy back shares and thus increase the value to their stock options, but the author doesn't seem to account for higher stock price appreciation based on stock buy-backs.

    I think the author should arrive at a prediction of future stock market returns by either using historical dividend yields and stock price appreciation, or else assuming a lower dividend yield but a higher stock price appreciation based on stock buy-backs. I think either way will result in a similar prediction of future stock returns.

    If the author wanted to justify lower future stock returns, I think he would have been better served to argue that current price to earnings (P/E) valuations are higher than the average over the last 100+ years, and so we should lower our expectations of any future expansion of the P/E multiple. More likely, we should assume the multiple either stays at the "new normal" or else reverts to the mean; either way, stock returns would be lower than the average over the last 100+ years.

    I'm not sure who to believe frankly. However, it does seem prudent, as the author lays out quite nicely, to assume that retirement starts with a poor returns sequence like 1929, 1966, or 2000.

  3. @Patrick: I'm still mulling some of the other aspects of the book. The problem with the issue about expected asset class returns is that it affects almost everything in the book. Even if his approaches to testing portfolio longevity are excellent, they are undermined by the asset class return assumptions.

    @Blitzer68: I agree that buy-backs should lower dividends but also increase stock prices. I'm not sure that this is Otar's only reason for docking dividends. I'm not even sure that docking dividends is a mistake. However, I don't think it makes sense to dock dividends and simultaneously just use historical bond returns.

  4. I'm glad you looked at Otar's book. Your synopsis gives the gist of the book, saving me from reading the whole thing. I found the introduction to be entertaining, talking about how his boyhood dreams of wealth in cabbages was destroyed by an early frost.

    His engineering training might cause Otar to be overly conservative (huge margin of safety), but I like how he used rigourous analysis of past returns to model portfolio performance.

  5. blitzer68, that comment makes sense to me.