Thursday, May 24, 2012

Michael Kitces on 4% as an income floor

Michael Kitces has an important new blog post about an issue I've been grappling with, "Annuities Versus Safe Withdrawal Rates: Comparing Floor/Upside Approaches"

I currently have some deadlines approaching that limit how much time I can write about this now. But I would like to make a brief post to guide readers to Michael's thoughts, and I've been tentatively planning to return to this theme for my June Advisor Perspectives column. I had been asking Michael to get his thoughts out on his topic to help me avoid confirmation bias, as most of what I've been looking at recently is decidedly against thinking of a 4% withdrawal rate as a safe income floor. And do recall that the international experience with the 4% rule has been no where near as positive as in the United States.

Nonetheless, still weighing heavily on my mind at present is the demonstration by Bill Bengen that 2000 retirees are still on the path to success using the 4% rule.

I did explore this issue in "Lower Future Returns and Safe Withdrawal Rates." Most of what I discuss now is based on the methodology described in more detail at that post.

There I was setting current real bond yields to 1%. Michael Finke asked me about failure rates if current real bond yields are 0%.

Essentially, let me now just provide 3 figures of failure rates under 3 different sets of assumptions. 

The first figure is calibrated to the standard historical data. With the 4% rule, failure is minimized at 6% with a 50% stock allocation. 

The second figure keeps the historical data parameters, except that the average real bond return is reduced from 2.52% to 0% to better reflect current bond yields (which, in turn, are the best predictor of future bond returns). Here, 4% withdrawals and a 50% stock allocation result in a failure rate of 15%.

The problem with the second figure is that the stock return is not adjusted, implying that the equity premium would be even larger than historically, while most pundits suggest that Americans should expect lower equity premiums in the future. In the third figure, real bond yields are 0%, and the historical equity premium is maintained by also reducing stock yields by 2.52%.  Now the failure rate for the 4% rule with 50% stocks is 34%.

A 34% failure rate does imply a 66% success rate. So there is a 2 in 3 chance for 4% inflation-adjusted withdrawals to survive for 30 years. But those are not the types of odds you would seek for an income floor to meet basic needs. And the problem would be worse if investors underperform the return indices used for these calculations either because of account management fees or bad market timing decisions. 

Michael has made compelling arguments, but I think I am still not convinced that it is a good idea to not build a more proper floor to meet essential needs.