The difference between active & passive

• Passive investments tend to show higher risks than expected.
• The tracking error only partially captures the magnitude of active risks.
• The active share provides a forecast-independent measure for the ex ante assessment of active risks.
   

Passive with active risks?

The importance of passive investments in asset allocation has grown massively over the past ten years. While passive invest­ments initially consisted mainly of replicating concepts based on global indices such as the MSCI World or S&P 500®, typically in the form of ETFs, there are now a variety of extensions:

• Smart beta indices that attempt to collect factor premiums based on systematic rules
• Indices that consider various sustainability criteria
• Individual criteria, often found in the institutional segment, such as exclusion lists or minimum requirements with regard to ESG

Despite many of these products being classified as passive or semi-passive, the sobering reality reveals that the inherent risks, in the form of active deviations from the base universe, likely have a more significant impact than anticipated. In this article, we explore when an investment can no longer be considered passive. We also demonstrate that the commonly used tracking error metric is insufficient for classification and assessing risk.

Analysis using the example of Dividend Global

For the following analysis, we construct five portfolios that em­phasize the dividend factor based on the global equity market (base universe). Portfolio 1 aims to achieve a dividend yield ten per cent higher than the base universe. Each subsequent portfolio increases the dividend target by an additional ten per cent, with Portfolio 5 having a dividend yield 50 per cent higher than the global equity market. All portfolios are constructed holistically. This means that, except for the dividend factor, which is empha­sized to varying degrees depending on the portfolio, all factor, sector, and country effects are neutralized compared to the global equity market. This ensures that the increasing level of activity in the portfolio (increasing the dividend yield) does not introduce additional risks beyond the dividend factor. 

Table 1 presents descriptive data for the global equity market and the five portfolios. The 1,510 companies have an average current dividend yield, the target metric in this case, of 2.14 per cent.

Tab. 1: Descriptive statistics of selected portfolios

With each increase in dividend yield, the tracking error progress­sively rises, and the number of positions decreases as stocks with lower dividend yields are underweighted to achieve the higher dividend level. While a dividend yield of 3.22 per cent may not seem high, especially compared to the European market, it rep­resents a significant increase in the payout level from a global perspective and is accompanied by a tracking error of 1.4 per cent.

The challenge with tracking error is that it only provides limited insight into maximum relative risks. Furthermore, factors such as the interest rate shift in 2022 can lead to a significant future increase in tracking error. The substantial rise in interest rates has fundamentally altered market structures and correlation relationships, resulting in an increase in tracking error. Therefore, this metric is particularly limited in its ability to assess whether an investment can still be considered purely passive, especially ex ante.

Figure 1 illustrates the active share for the five dividend portfolios alongside the relative maximum drawdown. While a purely ex post view is possible in the first case, as with the tracking error, it nevertheless reflects the resulting active underperformance risk and often serves as the basis for investor sell decisions.

Fig. 1: Active share and realtive drawdown

As evident, in the case of the portfolio with a ten per cent in­crease in dividend yield and an associated tracking error of 0.74 per cent, there is only a relative maximum drawdown of -2.0 per cent. However, aiming to construct a portfolio with a dividend yield 50 per cent higher than the base universe results in a rela­tive drawdown of approximately -10 per cent. Although the tracking error of 1.4 per cent may suggest low activity compared to active strategies, the risk of underperformance, around ten per cent, exceeds what can be classified as passive.

Assenagon Equity Framework

The active share can be used for the final categorisation of in­vestments into passive and non-passive. It reflects the sum of open active risks (underweighting and overweighting) compared to the benchmark. Additionally, the active share offers the ad­vantage over both the tracking error and relative maximum drawdown, providing an ex ante and forecast-free assessment of an investment's active risks without susceptibility to estimation errors.

Fig. 2: Relative performance compared to the global equity market
(January 2000 – December 2023)

For investors

We believe an investment can no longer be considered passive if the active share exceeds 20 per cent. This observation holds true across different factor styles or basic universes, irrespective of whether the portfolio was constructed holistically or undesirable side effects are present. Holistic construction ensures the active share is optimally utilized for the target metric (in this case, dividend yield) without introducing additional risks in terms of tracking error or relative maximum drawdown.

Figure 2 illustrates that purely passive portfolios can be opti­mised by slightly emphasizing factors, such as dividend yield in this example (e.g., 10 to 20 per cent), without compromising their passive nature. However, as evident from the performance of the Dividend +50 per cent portfolio, there are limits to such optimisa­tion: despite a moderate tracking error, there is a relative perfor­mance with active risk characteristics.

P.S. Stay tuned for the upcoming issue, where we delve into the efficient construction of index-linked investments.