What Is The Optimal Asset Diversification Historically? (Statistics, Facts, Backtest)

What is the optimal asset diversification historically? This is debated, but we came across an interesting study that looks at the best historical asset class diversification models over the last 100 years, all based on statistics and backtests. The main takeaway is this:

A study finds that a constant allocation of 50% domestic and 50% international stocks (no bonds!) is optimal for the full lifecycle, including retirement, for an American saver or investor.

Such a strategy produces better results than the traditional 60/40 portfolio, or any portfolio that includes bonds, for that matter. 

Let’s look at the findings, results, statistics, and how the results came about:

Diversification benefits

The study is called Beyond the Status Quo: A Critical Assessment of Lifecycle Investment Advice, and is written by Aizhan Anarkulova, Scott Cederburg, and Micheal O’Doherty.

This is a study that is highly relevant for long-time savers and investors and those saving for retirement. The study looks at what is the optimal asset diversification historically for a US-based investor. The study compares international diversification for stocks and bonds. 

Bonds for diversification and safety

Conventional wisdom generally recommends an increasing allocation to bonds as retirement approaches, aligning with the notion that bonds provide greater stability and are more suitable for individuals with less tolerance for market fluctuations during their retirement years.

Also, Bonds offer correlation benefits in an investment portfolio. But is it good advice?

According to the authors, you risk losing valuable returns by allocating capital into bonds. In the long run, volatility should not matter, thus bonds should be “ignored”. Or at least you’ll expect lower returns by including bonds. 

Asset allocation and retirement planning

Asset allocation advice fundamentally revolves around anticipating the expected returns and correlations of various asset classes. While predicting future asset class behavior is notoriously complex, historical data provides valuable clues about past trends and potential future movements

One typical problem with any asset allocation, is that very little is backtested historically. This study addresses this by using a block bootstrap procedure with 38 developed markets from 1890 to 2019 (see more of the methodology at the end). 

It’s a very impressive study and covers nearly 2 500 years of monthly data and returns. 

Risk management and asset allocation

The study suggests that you should skip bonds and go all in on stocks. 

This doesn’t mean stocks are not risky – quite the contrary – in terms of volatility. 

But bonds are more risky at long-term horizons. Time horizons for most savers, spanning decades, bonds get increasingly correlated with domestic stocks, and offer little downside risk in real terms, while it might offer diversification benefits in the short term.

Bonds vs stock asset diversification

Small asset allocations to bonds, as low as 5 to 15%, result in significant underperformance. 100% stock allocation offers better returns, plain and simple. 

Tha