Looking at their "Aggressive" portfolio (80% stocks/20% bonds, which is similar to my target asset allocation) you have a Twenty Years Annualized Return of 12.3% with a Thirty Three-Year Model Annualized Standard Deviation 11.8%. The "Very Aggressive" portfolio (100% stocks) has a Twenty Years Annualized Return of 13.7%, but the Thirty Three-Year Model Annualized Standard Deviation shoots up to 14.6%.
This shows that, as can be expected from modeling of the efficient frontier of a portfolio composed mainly of stock and bonds, the optimum return-risk outcome is achieved from a portfolio comprised mostly of stocks, but with some bonds included. The mix within the stock component is usually around 60% domestic:40% foreign, although in various ten-year periods you would have done better with the opposite ratio (so a 50:50 split may be a good bet).
Moving from the "Aggressive" to "Very Aggressive" asset mix boosted returns by 11.38%, but the "risk" (variability of returns, as measured by the Standard Deviation) increased by 23.73%.
For this reason, if you are seeking higher returns over long time periods, it seems a better strategy to use gearing of an "Agressive" portfolio, rather than moving to a "Very Agressive" portfolio.
Taking the Twenty Years Annualized Return of the "Very Conservative" portfolio (100% bonds) as a proxy for the interest rate cost of gearing (via margin loans or a real-estate backed investment loan such as a HELOC), one can make a rough estimate of the Twenty Years Annualized Return and Thirty Three-Year Model Annualized Standard Deviation that would result from a 100% geared (50% LVR) "Aggressive" portfolio:
Return Std Devn
Ungeared "Aggressive" 12.3% 11.8%
Estimated Cost of Loan 5.9% 2.4%
Estimated 100% geared 18.7% 23.6%
Estimated 22% geared 13.7% 14.4%
Ungeared "Very Aggres." 13.7% 14.6%
Using gearing could therefore increase your average returns by 52.03% at the cost of increasing standard deviation by 100%. This is somewhat better than shifting your asset allocation from "Agressive" to "Very Aggressive". However, the absolute "risk" has increased 100% compared to 23.73%, so the strategy of making use of 100% gearing ratios should probably be called "Hyper Aggressive". A more modest use of gearing (say, 22%) would produce similar average return as a "Very Aggressive" asset allocation, but with a slightly lower standard deviation.
It was interesting to see that the returns for the 100% geared "Aggressive" portfolio are very similar to the long-term increase in value of my own investment portfolio. When I started out I didn't use gearing and had a more conservative asset allocation, but this was offset by the relatively large impact my savings had at that stage. These days my savings have a more modest impact on my overall increase.
One final note, when using gearing the cost of funds (interest rate and any annual fees) can have a major impact on the long-term performance of this strategy, so it is worth shopping around.