Until 25 years before retirement, Vanguard Target Retirement Funds all have an identical asset allocation of 63% U.S. stocks, 27% international stocks, 8% U.S. bonds, and 2% international bonds. I roughly follow this allocation because it seems reasonable. But I wonder, does anybody know how they came up with these numbers? Did they run historical simulations to determine them?
vanguard target asset allocation led me to this page on the Bogleheads wiki which has detailed breakdowns of the Target Retirement funds; that page in turn has a link to this Vanguard PDF which goes into a good level of detail on the construction of these funds' portfolios. I excerpt:
(To the question of why so much weight in equities:)
In our view, two important considerations justify an expectation of an equity risk premium. The first is the historical record: In the past, and in many countries, stock market investors have been rewarded with such a premium.
Historically, bond returns have lagged equity returns by about 5–6 percentage points, annualized—amounting to an enormous return differential in most circumstances over longer time periods. Consequently, retirement savers investing only in “safe” assets must dramatically increase their savings rates to compensate for the lower expected returns those investments offer.
The second strategic principle underlying our glidepath construction—that younger investors are better able to withstand risk—recognizes that an individual’s total net worth consists of both their current financial holdings and their future work earnings. For younger individuals, the majority of their ultimate retirement wealth is in the form of what they will earn in the future, or their “human capital.” Therefore, a large commitment to stocks in a younger person’s portfolio may be appropriate to balance and diversify risk exposure to work-related earnings
(To the question of how the exact allocations were decided:)
As part of the process of evaluating and identifying an appropriate glide path given this theoretical framework, we ran various financial simulations using the Vanguard Capital Markets Model. We examined different risk-reward scenarios and the potential implications of different glide paths and TDF approaches.
The PDF is highly readable, I would say, and includes references to quant articles, for those that like that sort of thing.
While the Vanguard paper is good, it doesn't do a very good job of explaining precisely why each level of stocks or bonds was optimal. If you'd like to read a transparent and quantitative explanation of when and why a a glide path is optimal, I'd suggest the following paper:
(Full disclosure - I'm the author).
The answer is that the optimal risk level for any given holding period depends upon a combination of:
- The average expected returns over that period
- The potential for losses over that period
Using these two factors, you construct a risk-averse decision model which chooses the risk level with the best expected average outcome, where it looks only at the median and lower percentile outcomes. This produces an average which is specifically robust to downside risk. The result will look something like this:
The exact results will depend on the expected risk and return of the portfolio, and the degree of risk aversion specified.
The result is specifically valid for the case where you liquidate all of the portfolio at a specific point in time. For retirement, the glide path needs to be extended to take into account the fact that the portfolio will be liquidated gradually over time, and dynamically take into account the longevity risk of the individual.
I can't say precisely why Vanguard's path is how it is.