This is an article written for press by Patrick Cairns, The Investor’s Guild
Tracking an index might seem like a simple concept, but for the team at Gryphon Asset Management it is something they are constantly trying to improve.
“It’s very specialised skill set,” says portfolio manager Casparus Treurnicht. “We have been tracking for more than 25 years and we are still improving our process on a daily basis. We don’t just plug a black box into every index and push a button.”
Replicating the return of an index in the most efficient way possible is an iterative process for Gryphon. And, importantly, every index needs to be approached differently.
“You have to look at each benchmark it on its own and decide how you are going to track it,” Treurnicht explains. “For example, the ALSI and the MSCI World, both of which we track, are completely different.”
Fellow portfolio manager Ruan Goosen elaborates: “The ALSI is now around 120 counters, and it’s easier to track a smaller index, especially because the Top 40 constitutes 85% of the market capitalisation,” Goosen says. “So it’s quite easy to maintain.
“But there are over 1,300 stocks in the MSCI world, and they’re not just in a single location. They’re listed on different stock exchanges, in different time zones, in different currencies.”
For most of its history, the Gryphon Global Equity Fund, which tracks the MSCI World Index, was managed through a geographic building block approach. The fund was very small, and Gryphon was able to replicate the index by holding a selection of ETFs to track the benchmark return.
This naturally resulted in some tracking error, but also had to be reconsidered in light of regulations. As a CIS fund, the portfolio was limited to holding 80% of its portfolio in other collective investment schemes. Gryphon had to find a way to meet the regulatory requirements and improve performance with the remaining 20%.
“It’s not practical to hold all 1,300 shares in the index, so we have to use some sort of sampling,” Goosen says. “Initially, we transitioned to an interim step which was buying individual stocks based on index weights.
“The tracking error did get tighter, but we felt we could do better because there were still days where the stocks that we held moved differently to the index. That led us to look for a more sophisticated, statistical approach.”
This involved taking each of the nine sectors in the index, and bucketing the stocks in that sector based on their correlations. In this way, each sector was split into 15 groupings.
“For example, Meta and Apple sit in the same bucket. So, without taking a view on Meta or Apple per se, we can see they tend to move together so we don’t need to hold both of them,” Goosen explains.
“The implication of that is that in a bucket of ten stocks you can just own one and that will mimic the daily moves of the entire bucket. That shrinks the population of stocks you need to invest in.”
Gryphon also calculated that it’s not necessary to hold a representative of every bucket. Some can be left out, given their weighting in the index.
“For example, Apple makes up about 5% of the MSCI World so you want to track the bucket that it’s in,” Goosen says. “But there may be a different bucket that is made up of Japanese names that make up only 0.05% of the benchmark. We would probably choose not to pick a stock in that bucket.
“This is also about cost efficiency because we would have to pay custodial fees for a Japanese bank account, trade in a different time zone and so on.”
In this way, Gryphon is able to hold only 50 direct equities. Since implementing this sampling approach, the tracking error of the fund has reduced to zero and even turned positive over certain periods.
Goosen adds that Gryphon evaluates each bucket on a monthly basis as correlations between stocks will move.
“It’s not static,” he says. “We have to evaluate them constantly.”
This approach to the MSCI World is quite different to how Gryphon tracks the ALSI. While the firm also uses a sampling approach locally, it does so purely on the basis of liquidity, not correlations.
“We look at the three-month liquidity of every share in the index based on value traded, and at the moment any counter that will result in flow of less than R300,000 we consider not worth holding,” Treurnicht explains. “That quant is constantly evolving.
“We also look at black holes, which would be shares that appear to have traded on average quite well, but experienced days where liquidity disappeared. We strip out those as well.
“What you’re left with is shares with sustainable liquidity. We look at every sector and track it by using the most liquid counters.”
Since tracking the movements in an index is effectively about following trades of stocks in that index, it’s not necessary to hold every counter to replicate the performance.
“What we also tend to find is that when certain stocks don’t have liquidity they also diminish in size,” Treurnicht adds. “So, you don’t sit with the problem that you need trade something with no liquidity because your process has already removed it.
“The flipside of that is an example like Capitec many years ago. It was still relatively small when we noticed that it was starting to pick up in liquidity and so we needed to include it in our portfolio. The process makes sure that the moment a counter gets enough liquidity it is included and so we will have exposure to the stocks that end up moving the index.”
