Market Valuation and Asset Allocation Decisions
I am often asked how market valuation plays into asset allocation decisions. Should an investor consider reducing equity exposure when the market is trading at a high price relative to earnings? My standard answer is no, but in some situations, yes.
Figure 1 illustrates the valuation method that I prefer to use for the stock market. It’s called Price-to-Peak Earnings (P/PeakE). This method smoothes over big earning drops that can occur during recessions.
Figure 1: Historic S&P 500 P/PeakE (Jan. 1871 to Jul. 2013)
Source: Online Data: Robert Shiller, P/PeakE calculated by the author.
The average P/PeakE shifted higher during the 1950s as corporations began reducing dividend yields and reallocated cash to expansion and stock buybacks. The average P/PeakE had been about 10.0 times PeakE for 80 years before shifting to 15.0 since 1950. Currently, the ratio is at 17.4 times PeakE. This places stock valuations slightly higher than normal, but not excessive.
Investors who plan to work full time for more than 10 years years shouldn’t be worried about market valuation at this time. Even if a person put money into stocks back in April 1999 at peak market valuation, they would have earned respectable 4.4% annualized return through June 2013. Even in the worst of times, over the long-term, stocks have generated the best returns for investors.
Figure 2 highlights the amount of time it took for the S&P 500 to earn back losses during two big downturns, March 2000 to October 2006 and October 2007 to March 2012. In both cases, the market recovered within 10 years. The chart shows total return, which includes dividend reinvestment.
Figure 2: Recovery period for the S&P 500 total return index
There are three factors to be considered when considering an asset allocation decision based on stock valuation. They are cash flow needs, time horizon and the valuation itself. Cash flow need is the amount of money to be withdrawn from portfolio at some future time horizon.
If a person is within 10 years of retirement, and they plan to withdraw 3% or more from their portfolio each year to pay living expenses, then an asset allocation change may be in order when the market P/PeakE is above its historic average as it is today. This decision depends entirely on whether the person has saved enough to make this change, or expects to have enough saved for retirement based on the lower expected return from a higher fixed income allocation. My rule of thumb is 20 times living expenses that will not be covered by Social Security, pensions, annuities, rents or other income sources.
Even if an investor is approaching retirement and has enough to live off, there may not be a need to make an asset allocation change. If there is no need to withdraw money from a portfolio during retirement because other sources of income are available, or if the amount of money to be withdrawn is low, perhaps equal to or less than the current 2% dividend yield on stocks, then there is no reason to rush an allocation change when the market is trading at slightly above average.
Here is where the level of valuation makes a difference. In 1999, market valuation was over 30 P/PeakE, more than twice its average. It didn’t matter how close an investor was to retirement or how much they were going to withdraw in retirement, if they had enough saved, or would have enough though savings and a earning lower return from more fixed income, then I recommended they change. The only thing worse than not having enough to retire, is having enough and then losing it.
Market valuation does play into asset allocation decisions at times, but only when a person is closing in on retirement and is considering an allocation change anyway. There are three factors involved in this decision: cash flow needs, time horizon and market valuation itself.
On the rare occasion when stocks reach an extreme valuation, it may be prudent for all pre-retirees who are close to their accumulation goal to consider an asset allocation change even though they weren’t planning to do one.