Why 60/40 May Have To Become 80/20

We recently had a client ask about the prospects for the classic 60/40 portfolio of stocks/bonds given the possibility of higher inflation. It’s a great question, so here’s our take.

#1. 60/40 is popular for a reason: it has a good historical track record of delivering equity-like returns while lessening the risk of serious annual portfolio drawdowns. Here are a few basic statistics, assuming an annual rebalance at the close on December 31 of each year:

  • Since 1928 (first data available), a 60/40 portfolio of the S&P 500 and 10-Year Treasuries has delivered an average annual total return of 9.0% or 78% of the total return for just the S&P 500 (11.5%). After inflation (using annual CPI), this translates to a 5.9% average total return for 60/40, or 70% of the average real returns for the S&P 500 (8.4%).
  • The 60/40 portfolio saw 19 years with negative total returns from 1928 – 2017 (21% of the time). The S&P 500 has been down 24 times over the same period (27% of these years).
  • The worst drawdowns for the 60/40 portfolio since World War II: 1974 (-14.7%) and 2008 (-13.9%). The returns on just the S&P 500 in those years were -25.9% and -36.6%, respectively, or more than twice the losses of the 60/40 portfolio.
  • Because stock and bond returns show no historical return correlations (0.03) over the 1928 – 2017 time frame, the standard deviation of returns for the 60/40 portfolio is lower than the weighted average of the two assets: 12.1% versus a 14.9%.

#2. The wrinkle: stock and bond correlations are not static across time. Over the last 20 years (ending in 2017), the correlation between the two has been -0.66 based on total annual returns. The prior annual low for 10-year historical correlations was in 1967, at negative 0.48.

The reason for this exceptionally negative correlation sits in the data from the last 10 years. From 2000 to 2009, for example, stock and bond returns were -0.86 correlated (and remember that correlations only go to 1.0). For 2008 – 2017, that relationship didn’t weaken much: just to a negative 0.78 correlation.

Worth noting: the prior peak for 10-year stock-bond correlations was back in 1964, at negative 0.64.

#3: Stocks and bonds can – and have – shown positive return correlations throughout history, lessening the ability of a 60/40 portfolio to provide the benefit of diversification. Trailing 10-year return correlations peaked in 1995 at 0.74, for example.

#4. Given that we are ending a period of exceptionally negative stock-bond correlations and face the very real possibility of an extended period of rising inflation, what should a 60/40 portfolio owner consider? A few thoughts:

  • The five years after the prior trough (1964, as noted above) for stock/bond correlations (1965 – 1969) saw 10 Year Treasuries deliver just a 0.06% compounded average total return. Moreover, CPI inflation went from 1.0% in 1964 to 6.2% in 1969, so bond investors saw negative real returns over this period.
  • Equity returns over the 1965 – 1969 period were choppy and substandard. The compounded annual total return for the S&P 500 over the period was just 5.0%.
  • By comparison, the 60/40 portfolio showed a compounded annual return of 3.2% from 1965 – 1969. This lagged inflation, which showed compounded growth of 3.9% over the same period.

#5. Our conclusions:

  • While we doubt US inflation will rise as quickly as in the late 1960s, the 1965 – 1969 period shows the pitfalls of the 60/40 portfolio. Equities may not excel when inflation rises, but they certainly do better than bonds.
  • Long periods of rising inflation recouple stock and bond correlations, and this continues over the crest for the peak of higher prices and into the next down cycle. Ten-year price return correlations peaked in 1994 – 1995, at 0.74, reflecting a long journey through high inflation (early 1980s) through monetary discipline (mid-to-late 1980s) and eventually lower long-term inflation expectations (1990s).
  • The bottom line: we think this analysis has two takeaways.
    First, 60/40 seems too heavily weighted to bonds for this point in the correlation cycle. Whether the optimal mix is 80/20 or 65/35 is a matter of risk preference. We tend towards the latter out of conservatism, but understand it leaves little room for positive real returns if inflation accelerates.

    Second, any blend of bonds and stocks will have lower future returns and higher volatility than the recent past. A lot of our work reveals a similar message. Stock valuations are historically high, and bond yields are still near all-time lows. Neither is the recipe for outsized future returns.

Source: http://pages.stern.nyu.edu/~adamodar/New_Home_Page/datafile/histretSP.html