Investment returns are not fixed or reliable. Goldswell's website pages show how variable our returns and our benchmark returns have been. After years of doing monthly updates of investment statistics, it is clear to Goldswell that the investment return changes filter through to change all investment statistics (mean, volatility, value at risk, sharpe ratio, information ratio, correlation etc). None of them are fixed and sometimes the changes are dramatic.
Not only do investment returns change but generally a distribution of investment returns does not conform to the Gaussian normal probability distribution with much confidence. This can be confirmed using the Jarque-Bera test on a series of investment returns. It is Goldswell's experience that generally an investment return distribution will fail this test at a significant confidence level. Generally a distribution of investment returns has higher kurtosis than the normal distribution i.e. generally an investment return distribution has a higher peak and fatter tails than expected with a normal distribution.
This means applying normal distribution mathematics to calculate a forecast range for future investment returns at a particular confidence level (e.g. next year’s investment return at a 95% confidence level) is flawed and we should lack confidence in the answer. Especially we should be careful about the fat tail at the negative extreme where the normal distribution will understate the risk. Suffering a large investment loss is more likely than the normal distribution would tell us. The same lack of confidence goes for all investment statistics which build on from the normal distribution mean and variance (i.e. value at risk, sharpe ratio, information ratio, correlation etc). If the normal distribution doesn't apply, making calculations as though it does and confidently stating expectations is a bit silly.
As if to hammer home the point, it seems there are many instances of professional investors and financial institutions that have got into trouble from having too much confidence in their investment mathematics.
Apparently the fractal distribution is a much better fit for investment returns but working out the parameters for this distribution is problematic. When somebody figures out some better mathematics to deal with investment returns or a user-friendly way to deal with the fractal distribution we will use it.
Meanwhile... a portfolio or investment style's recent performance reflects how it was suited to recent conditions. These conditions may or may not continue, or may have changed already. Periods of out- performance are often followed by periods of under-performance.
Except for Goldswell's Market Timing fund, Goldswell's funds consist of a spread of investments of a particular style. Industry allocations and individual investments are selected to give the funds better overall "value" investment fundamentals than their benchmark indexes. Also we overweight towards smaller companies. Our choices will be right or wrong to varying degrees. On a long term portfolio-wide basis, we believe we have a rationale to outperform our benchmarks. But there are no guarantees. Certainly, on a month by month basis, we may outperform our benchmarks or not, and often the returns will be negative.
Goldswell's calculations won't exactly match those of the various information services as we use slightly different methodology, interim results updates and sometimes make estimates. However we are consistent with our methods across all our funds and index calculations, making our calculations directly comparable with each other. Be careful with information on this website. You may disagree with our calculation methodology and sometimes we make mistakes.