Studies of Long Term Portfolios and Retirement Withdrawal Rate Suggestions

When you retire with a large corpus, how do you go about withdrawing from that at a steady rate to sustain your lifestyle, while also letting the corpus grow? Here's what academia has to say on this.

This analysis gives you a basic idea about the historical and current perspectives of institutions which manage money on a large scale. And because the world economy is increasingly getting unified (as compared to past), these ideas will be applicable on a principle basis. Hence, this idea is important as far as the basic strategy is concerned.

However, how it applies tactically in individual cases and in an Indian context will be different.

eg. many people think that since we can get 8-9% tax-free and guaranteed on govt securities (govt backing is as close to guarantee as one can get), we need not think about all this 3.5-4.0% withdrawal rates. But they forget the basic difference between a nominal return of 8-9% versus a inflation-adjusted real return. After 10 years, the same 8-9% nominal return will not be sufficient AT ALL in most cases. A lot of pension plans also show you similar things and make the income look big but it is not so.

It really is a tough thing when you realise in these examples that even the best of the institutions with tremendous money power (and knowledge) cannot have a very high sustainable rate.

The various endowment reports studied:

I. Nobel Prize Committee Financial management Link

The real return (above inflation) expected from the fund is 3.5% (while they plan to use 3.0% at least yearly). They have considered an allocation of 55% (+/- 10) to equities, 20% (+/-10) to fixed income and 25%(+/-10) to alternative investments (including hedge funds and real estate).

II. Yale Endowment Portfolio Management Link

They expect a real return of 4.5%. The main difference from above is that they indulge in lot of alterative assets and active management strategies (with alpha). A good case in point was in 2008 period, when they went too far in their strategy and were having a negative 1.5-2% in cash, and suddenly during the crisis, they were in a severe bind. So now they have corrected that part and keep a decent amount of cash with them.

III. Norwegian govt sovereign fund Analysis Link They expect a return of 4% and unlike the endowment funds, they do not put anything into alternative asset classes. They follow simple 60:40 rule with regular rebalancing and following cost-effective strategies (read passive investing).

Many of the online calculators have used 4% as a reasonable withdrawal rate for 30 year periods using Monte-carlo calculators. However, based on various rolling data periods of actual market returns for US equity markets, a 4% withdrawal rate will not sustain for 30 year periods in 30-40% cases (whatever be the asset allocation). At 5% withdrawal rate, the failure rate over 30 year periods increases to 40-55%. While at 3.5%, it falls to <10% unless one is in 100% bonds or 100% equities.

Why Monte-carlo is not the right thing? because the random number generation in those simulations is normal distribution curve related (while the actual returns are never like those).

Why US equity market data? Since that data is long and reasonably good and since it is the most important economy in the increasingly globalised world, it mirrors capital markets to a large extent.