The quote “What gets measured gets managed” is attributed to the celebrated American management consultant Peter Drucker. By quantifying tasks or goals individuals can better understand their performance, set clear objectives and track progress. In this article, we will explore the implications of that statement in the context of investment risk measurement of retirement funds in South Africa.
Regulation 28 of the Pension Fund Act requires that retirement funds don’t put all their eggs in one basket. Investments have to be spread across several asset classes, geographical locations, industries and durations. There are limits on individual securities and companies, meant to avoid concentrating risks. Investment policy statements for many retirement funds often explicitly embrace diversification.
Investopedia defines diversification as “owning a wide variety of investments with different characteristics to reduce volatility”. The rationale behind diversification is that, on average, it yields lower risk without sacrificing higher long-term returns.
Diversification eliminates the specific risk attached to a single asset and the portfolio is exposed mainly to systematic risk.
Systematic risk
Systematic risk is also known as market risk or undiversifiable risk. This risk is unavoidable because it arises from factors such as inflation, recessions, war, interest rates, exchange rates, natural disasters and other macroeconomic events that impact the market as a whole.
The opposite of systematic risk is idiosyncratic risk which is the risk inherent in an asset due to specific qualities of that particular asset. Idiosyncratic risk is eliminated in a diversified portfolio because it is not correlated across different assets. Markets do not reward idiosyncratic risk because it can be eliminated. It is thus reasonable to eliminate it from an investment portfolio.
Systematic risk is correlated across different assets, hence it accumulates as different assets are added to an investment portfolio. Markets reward systematic risk because it cannot be avoided. The only way of reducing systematic risk is to sacrifice returns through hedging.
Retirement funds are diversified hence idiosyncratic risk is small. The focus should thus be on systematic risk. We will thus focus on Beta and R-squared in this article which measure systematic risk. Measuring systematic risk means it will eventually get the attention it deserves and gets managed properly.
Beta
Beta shows the volatility of an investment portfolio compared to the volatility of the entire market. Beta effectively describes the movements of a portfolio as it corresponds to swings in the market. It is used in the capital asset pricing model (CAPM) which describes the relationship between systematic risk and expected return for assets.
Mathematically, Beta is calculated as the covariance of a portfolio with respect to the market as a whole divided by the variance of the market as a whole. The covariance of a portfolio measures how changes in the portfolio’s returns are related to changes in the market’s returns. The variance of the market measures how far the market’s returns deviate from their average value.
A Beta of 1 means the portfolio moves in line with the market. A beta of less than 1 means the portfolio is less volatile than the market. On the other hand, a beta of more than 1 means the portfolio is more volatile than the market. A negative Beta means the portfolio actually moves in the opposite direction as the market. It thus goes up when the rest of the market is going down!
R-squared
R-squared is a measure of the percentage of an investment portfolio’s performance that can be explained by the market as a whole. Mathematically, R-squared measures the strength of the correlation between the portfolio and the market. It is reported as a number between 0% and 100%. An R-squared of zero means the portfolio is not correlated with the market at all. An R-squared of 100% means the performance of the portfolio exactly matches the benchmark.
While Beta tells you how volatile a portfolio is relative to the market, R-squared tells you how precisely you can determine the performance of the portfolio using the beta of that portfolio. If R-squared is high, e.g. above 90%, then more than 90% of the portfolio performance can be explained using the Beta of the portfolio. Thus predictions on portfolio performance can only be credible if R-squared is sufficiently high. A fund with an R-squared higher than 70% is considered one with a good R-squared.
Relative risk measures
Both Beta and R-squared are relative risk measures because they do not tell us the absolute risk in the portfolio but they inform us on the risk relative to a benchmark. The benchmark typically used as a proxy for the market is a broad market index such the JSE All Share Index in SA and S&P 500 in the US.
The higher the relative risk of an investment, the higher the possible return. Relative risk is a driver of performance in the long term. Relative risk measures are inherently more stable than absolute risk measures. This is because absolute risk measures go through cycles but relative risk measures are ratios hence maintain stability as long as the numerator and denominator are moving in the same direction.
Impact of two-component arrangement
The long awaited two-component arrangement came into effect on 1 September 2024. It ushered in access of retirement savings for members without having to leave employment via a savings component. It also brought in a retirement component which cannot be accessed by the members prior to retirement.
While most attention has been focussed on the savings component, the retirement component is the pillar on which the future of the retirement industry rests. The retirement component will start with a balance of zero on 1 September but due to forced preservation it will inexorably accumulate over the coming years. Decades from now it will have the lion’s share of retirement fund investments. This will have an impact on the investment risk profile of funds.
The savings component can be accessed anytime by members hence there is a need to hold liquid assets. The retirement component cannot be accessed as a cash lumpsum even at retirement. Liquidity risks are thus less of a concern for the retirement component, and this will impact funds going forward.
The forced preservation of the retirement component is meant to prevent depleting their savings as has been past practice when members change jobs. As the retirement component becomes a significant portion of funds then this will increase the investment horizon of funds. There is thus a greater need to focus more on relative risk measures that drive long-term performance.
Practical considerations
STEFI can be used as a proxy for the riskless rate. This is done by subtracting STEFI’s returns from the returns for the fund and the market index before doing the statistical regression to calculate Beta. STEFI is thus allocated a Beta of zero in this approach. This avoids the assumption that the risk-free rate is constant. The estimation of Beta thus captures the dynamic that risk-free rates change in line with inflation and risk appetite.
Capped SWIX can be used as a proxy for the market portfolio and thus allocated a Beta of 1. Capped SWIX applies the caps on weights as required by Regulation 28 of the Pension Funds Act. Capped SWIX is more volatile than most funds and also outperforms in the long term hence Beta for funds will be normalised to be less than 1.
Let’s look at an example. The following figures were obtained for a fund for the 10-year period 1 September 2014 to 31 August 2024.

- The R-squared is 78% which means that the relationship is statistically significant. 78% of the movement in the fund is explained by movements in SWIX and STEFI.
- Beta is 53% which means the volatility of the fund is 53% relative to SWIX. This figure can be compared to other funds to rank the fund’s riskiness. The figure can also be compared over time to monitor the fund’s risk.
- The Alpha of 1% is the portion of the fund’s return that is not explained by the movements in STEFI and SWIX. It thus represents the excess annualised return that the investment manager has managed to earn after adjusting for risk. If Alpha was negative, then it would have meant the investment manager didn’t manage to add value after adjusting for risk. The table below shows the returns for a fund with a negative alpha. The investment manager for this fund has destroyed value by taking on active risk. The fund can thus improve performance restricting the investment manager to not take active risks.

- The fund in question was invested in a balanced fund hence the numbers above were reasonable. The fund was not invested entirely in equities but had a significant portion invested in cash and bonds. The portfolio was actively managed and the asset manager was allowed to tactically deviate from the strategic asset allocation. This explains why Beta and R-squared were not 100% and consequently why there was alpha in its performance.