Needles in haystacks
A stockpicker’s nightmare
With stock market gains being driven by only a handful of companies, Ian McIntosh ran 1,000 simulated portfolios at random to see just how difficult an environment it has been for active managers.
Harry Markowitz, the father of modern portfolio theory, is credited with saying that the only free lunch in investing is diversification.
Diversification is an extremely old concept, older than the Old Testament. King Solomon – the wisest and richest investor of the ancient text – wisely balanced his investments across commodities, agricultural land and infrastructure.
Yet the benefits of diversification might be lost on the new generation of investors; big winners in equity markets have gotten bigger and simply riding the wave has been an extremely successful strategy. This is one of the best years for momentum investing on record.1 So why diversify away your portfolio returns when all you need to own is Nvidia and bitcoin?
Diversification has become more elusive in a world where owners of broad market indexes are taking on concentrated risk that resemble the risk profile of highly active portfolios.
In the S&P 500, three stocks (Nvidia, Apple, Microsoft) now account for around 20% of the index weight and more than one quarter of the risk (as measured by the Barra US Long Term Risk). This is in an index with 497 other holdings. The situation is more extreme in popular indexes like the Russell 1000 Growth that barely pass regulator definitions of diversification such as the 5/10/40 rule of UCITS funds in the EU.
For most of history, active global equity portfolio managers didn’t need to pay too much attention to what they didn’t own.
For most of history, active global equity portfolio managers didn’t need to pay too much attention to what they didn’t own in the MSCI World Index. If you built a diversified portfolio with 50 holdings out of the 2000 or so names in the index, your top active risk contributors were likely large overweights. With the rise of the Magnificent 7, this has changed. Errors of omission are now costly. Not owning the two largest index weight names this year wiped out the alpha target of most active managers.
In this environment, the best class of long only active equity strategies have been ’index huggers’ – systematic strategies where small tilts and deviations from the benchmark have meant that errors of omissions are limited. Low ’active share’, once derided, has helped to limit the damage. An index hugger must own a large index weight like Nvidia. Are these strategies well diversified? By definition, they are only as diversified as the index they embrace.
The concept of active share became popularized after being coined in the 2009 paper How Active Is your Fund Manager? A new Measure that Predicts Performance (Martihn Cremers, Antti Petajisto). Regulators started to pay attention to active funds that had low active share, low tracking error to benchmark, while charging active fees – with 50% being a commonly used threshold that is a warning of low activeness.
To illustrate how a concentrated benchmark can present a challenge for an active portfolio manager, consider the following thought experiment. Imagine 1000 portfolios each with 50 holdings, chosen at random from the S&P 500, with the portfolios built to be equal active weighted, meaning the weight of each holding relative to the benchmark weight is the same. This weighting approach is selected as a simple proxy for the approach of the average active portfolio manager; larger index weighted names tend to have higher portfolio weights as it is the active weight that drives relative performance. Each of the 1000 portfolios is ‘buy and hold’, meaning the holdings are not traded for the duration of the evaluation period.
How many of the 1000 portfolios would be expected to outperform the S&P 500 in a given year? For a year chosen at random, the intuitive answer is 50% – a coin flip. How many of the 1000 portfolios would have outperformed the S&P 500 from 1 July 2023 to 30 June 2024?2 Instinctively you will likely appreciate the number is low, given this was a period dominated by the ascendancy of Nvidia, Apple, and a narrow band of megacaps. How low, though? There is a 10% chance the portfolio is overweight Nvidia, so that might seem to be a reasonable floor. Run the simulation, and the answer is approximately three out of the 1000 portfolios. The median simulated portfolio underperformed the index by a whopping -12.2% (before fees).
Of course, an active portfolio manager is paid to perform better than a random one. But the result illustrates the challenge active managers had during this period as they are in effect searching for the active portfolio in the distribution of possible portfolios that is going to outperform. This search equates to looking for a needle in a haystack.
How could an investor improve on the simulated results? By adding constraints, such a sector constraint (e.g., to limit the underweight to technology), or factor constraints, (e.g., to limit the underweight to the size factor)? Or perhaps by holding large index weight positions – albeit at an underweight – to mitigate the risk of omission. In other words, to start hugging the benchmark more, or in more technical speak to reduce the active share of the portfolio – the measure of ‘activeness’ in the portfolio.
Is this extreme result in favor of the cap weighted index expected to be repeated in future? Will the search for the outperforming active portfolio always be this challenging? Let’s put this result in historical context (see Figure 1). Over the past 25 years, the average hit rate has indeed been about 50%, roughly in line with intuition. The last 10 years have been challenging for active managers, as the chart shows.
Figure 1: How unusual is this outcome?
Figure 1: How unusual is this outcome?
% of simulated portfolios outperforming the S&P 500
The last time the S&P 500 was extremely hard to beat, with less than 80 out of 1000 simulated portfolios outperforming, was back in 1998 in the lead up to the tech bubble. 1999, the last full year before the tech bubble burst, was also challenging with 200 out of 1000. However, this was followed by six years where the outcome flipped; where an active manager was best to stay away from the inflated index weights that were deflating post bubble. In an opportunity rich environment like this, the search for an outperforming active portfolio was easier.3
Is this extreme result in favor of the cap weighted index expected to be repeated in future? Will the search for the outperforming active portfolio always be this challenging?
There are some stark differences between the market today and the tech bubble. The largest index names of today are highly free cash-flow generative, unlike their tech bubble counterparts. However, the dominance of cap weighted indexes is not pre-ordained, and like every regime that came before, this one shall end. We believe in setting up for a multi-year period when hugging broad-based cap weighted benchmarks that merely offer the illusion of diversification could be foolish.
The craft of correlations
The craft of correlations
Make an inquiry
Fill in an inquiry form and leave your details – we’ll be back in touch.
Introducing our leadership team
Meet the members of the team responsible for UBS Asset Management’s strategic direction.