Hi Davidrmoran
You asked if the risk-reward curve has a distinctive character such that it attracts special financial attention. The simple answer is No.
The marketplace risk-reward curve rises in a continuous well-behaved manner as the equity fraction increases; higher risk, higher rewards. There are no outstanding features.
Note that I did not answer the title question in the original post; “How Much Of Your Retirement Portfolio Belongs In Bonds?” One size does not fit all; there is no single overarching reply. Each investor has a logical different answer to that question for very disparate logical reasons. The answers lead to the complete spectrum of the equity-bond tradeoff.
One historical standing rule is that younger folks should have a portfolio that heavily favors equity positions, while older folks should be more conservative with a portfolio weighted towards bond products.
Today, some industry experts are challenging that wisdom. In the end, it depends on the individual investor, his wealth, his plans, his risk aversion. One size definitely does not fit all.
To help answer your question, I input the annual returns (AR) data and the cumulative annual growth rate (CAGR) data into a curve fitting program available on the Internet. The program automatically “best fits” the data sets to Linear, Exponential, Power, and Logarithmic equation formats.
This statistical curve fitting was done on the following mathematical website:
http://www.had2know.com/academics/regression-calculator-statistics-best-fit.htmlGoodness of fit values (correlation coefficients) were high for all the tested equations. The Logarithmic form was slightly superior for all cases examined. However, the Linear modeling did an excellent job also. For simplicity, I’ll report the Linear modeling. Here are the equations:
AR = 0.4
52 X SD +6.41 Correlation Coefficient = 0.972
CAGR = 0.369 X SD + 6.71 Correlation Coefficient = 0.9
50
The percentage signs were just ignored in these correlations (use
5 for
5%). You get to choose whatever volatility (Standard Deviation) you find comfortable, and the equations provide an estimate of returns using the historical data sets.
For every unit that you move up the risk curve, estimated AR increases by 0.4
52 units and the CAGR increases by 0.369 units. If the more complex Logarithmic formulation were deployed, a slightly more refined estimate would be predicted that is not constant over the entire range of Standard Deviations.
This submittal might be a little more than folks wanted, but it puts the trends and relationships into a rigorous statistical framework that uses historical data. I hope you find this first-order analysis of some utility.
Best Wishes.