Subscribe

Listen to the show using this podcast player:

Or on YouTube:

In this episode of the Ghost Stories podcast, The Finance Ghost sits down with Reza Fakie, portfolio co-manager of the Old Mutual Investment Group Global Managed Alpha Fund.

The fund has delivered consistent outperformance against its benchmark since inception, but the real story is how it does it. Reza takes us inside the world of quantitative investing, explaining how academic research, factor investing and disciplined portfolio construction come together in a systematic process designed to remove emotion from investment decisions.

From identifying overlooked opportunities around the world to navigating the AI boom and managing risk in a concentrated global market, this is a fascinating look at how a modern quantitative fund is built and managed.

In this episode, we cover:

• How multi-factor investing works in practice
• The factors that drive stock selection and portfolio construction
• Managing risk while seeking consistent alpha
• Why the fund looks beyond the biggest global tech names
• Finding overlooked opportunities in emerging markets
• How quantitative investing helps remove emotion from decision-making
• The growing role of AI in investment research and portfolio management

Old Mutual Investment Group (Pty) Ltd is an authorised financial services provider, FSP 604. The contents of this podcast and, to the extent applicable, the comments by presenters do not constitute advice as defined in FAIS. Although due care has been taken in recording this podcast, Old Mutual Investment Group does not warrant the accuracy of the information contained herein and therefore does not accept any liability in respect of any loss you may suffer as a result of your reliance thereon. Past performance is not necessarily a guide to future investment performance. For more information, visit www.oldmutualinvest.com/institutional

Transcript:

The Finance Ghost: Welcome to this episode of the Ghost Stories podcast. It is the second in what I suppose could be described as a mini-series with the team from Old Mutual Investment Group.

It’s a really cool opportunity to speak to professional fund managers and portfolio managers who are out there, I think, “living the dream” for many, if I’m honest, and actually doing this really great thing where they are managing money on behalf of others. 

And the whole idea behind these podcasts is to get a better understanding of how they actually go about doing this.

So, in this one we will be looking at the Old Mutual Investment Group Global Managed Alpha Fund. 

It’s been around since 2017, so it’s coming up on a decade old now.

Since inception, and just based on the latest fact sheet, they’ve outperformed the MSCI All Country World index by around 200 basis points. I think that’s quite impressive. 

Obviously, past performance is no indication of future performance. The usual disclaimers apply. Go and speak to a financial advisor. Do your own research as well.

But to help you do your own research, this podcast is going to be a really good look at how this fund actually works, and how it has managed to perform like this over the past decade.

To help us understand, portfolio co-manager Reza Fakie is here to take us through it.

Reza, are you tired of people pointing at you and making the “This is my quant” joke? Because a big part of what you do in this fund, of course, is very quantitative in nature as opposed to qualitative, right? 

Reza Fakie: Yes. Hi Ghost, thank you for having me on.

So, just to set it straight, I’ve never won a Maths Olympiad and I do speak English, so we can get started with that.

The Finance Ghost: Ah, there we go. So, you’re scoping out being a quant, I see. Getting that out the way early.

Reza Fakie: So surprisingly enough, I started in actuarial science, which isn’t really considered an investment field for many. But while doing actuarial science, I discovered finance and never looked back. 

And quantitative finance, as sitting in actuarial science, is actually a combination of three different fields that co-join together and become a very interesting combination. 

So, you’ve got that finance background, understanding how the world works and what’s happening in the real world. You’ve got that maths and stats background, where you understand exactly how to measure things, how to understand things, how to understand the mechanics behind things. 

And then, maybe surprising for some: programming! Because once you have a great idea that you’ve come up on the finance side, you’ve tested it on the maths and stats side, you need to actually use it to invest. And that’s where programming becomes very important. And using MATLAB, Python, R, there’s many languages out there; but being able to deploy a solution is actually important in a quantitative space.

The Finance Ghost: I love how that started with, “I’m not a maths genius.” Also, “I studied actuarial science,” – which is basically probably the most mathematical thing possible. There are levels to this game, as the Gen Zs like to say. 

Let’s dig into then some of the details actually around how this fund really works. So we’ll spend a few minutes just understanding the underpinning of this thing. And the concept of a multi-factor model is very important here. So, this is something that people may have heard of, they may not necessarily understand what it actually is.

So perhaps just as a starting point, you can walk us through some of the buckets that you use and then an overview of what factor investing actually looks like.

Reza Fakie: For us, a factor (and there’s a lot of material being written about this, you’ll see it in the news, and it all seems very complicated, very mathematical) but what a factor actually just is, is some characteristic of a company or a share, which we believe has some future predictive power. It’s going to tell us how the stock is going to do over the next month or the next year. 

And if we actually go back in time to the ‘50s and ‘60s, we start with people actually identifying this first factor, which is our CAPM model, which many of your listeners would be aware of, which was just saying that how risky a stock is, has some component of where its returns are coming from. 

And then we move a bit forward, we move to the early 90s, and we start with probably the first multi-factor model out there, the Fama and French model. And that again is starting with saying, well, we’re identifying that there are certain components or certain characteristics of a company that have some predictive power. 

And there they found that small companies outperform larger companies. They found that cheap companies, where they measure that by book to value, are outperforming more expensive companies. 

And again, they have that beta component in there.

They’ve reviewed that and they added the five-factor model where they’ve included some quality-type factors, such as profitability. And then a year later you have Jegadeesh and Titman coming out with momentum. 

So, all through time there are these people identifying actually there’s these characteristics of either the stocks or the company itself that is saying that actually, it makes a difference to how it’s going to perform in the future. 

Now, we’ve looked at this, and we’ve identified two families of factors or two groupings of factors, being your fundamental factors and your technical factors. 

On the fundamental factor side (your listeners would be aware of this listening to more fundamental managers or other managers saying, “Well, I’m looking for a cheap company,” which is that value family), there’s the “I’m looking for a well-run company,” which for us is quality. 

And then you have, “Actually, I’m looking for a company that’s growing, it’s growing the earnings, growing the expectations of earnings and expected to grow into the future”, which is your growth factor. 

Those are the fundamental factors that we look at.

But being a systematic manager, we can identify what anomalies in this market tend to persist.

One that everyone knows about is momentum. So, winners keep winning. And we’ve identified this happening time and time again. 

The next one, maybe surprisingly, is actually a reverse of what was initially assumed. It’s been found that lower-volatility stocks actually outperform their higher-volatility peers. So, low volatility is actually what we look for in that factor. 

And lastly, again from that original three factor model, there’s still that persistence of small companies outperforming larger cap companies. 

And those together are these factors that we look at, these six families, three in each of these two groupings.

Now for us, over the long term, all of these factors tend to outperform. But the reality as an investor is, we’re looking at our portfolio every day, every week, every month. We see that yes, these have long-term payoffs, but the reality of it is actually a much more volatile, shorter-term performance. 

We’ve identified (and this is how we look at factors) these factors at work over the long term. And then say, “Well, how do we invest in these today?”

And there’s actually three characteristics of factors that we look at to make our decisions.

The first is that all factors are cyclical. If you hear someone investing in value or quality or momentum, those are going to perform very well for certain periods of time. But they also can have underperforming periods. So, there’s this natural cyclicality. 

Unfortunately, a lot of the time in the market, you’ll hear about the death of value two, three years ago, where value was underperforming for a long time. But again, it’s part of its natural cyclicality. And what we’ve seen the last two years is value coming back significantly well.

So again, there’s this natural cyclicality of factors outperforming and underperforming. And you have to be aware of that at each point in time.

The second characteristic is actually asynchronicity, which is a mouthful, but it just means that not all of these factors move at exactly the same point in time. 

When value is doing well, quality can do poorly and momentum could be doing well or poorly at that same point in time. And this allows us to invest across a multitude of factors, and actually benefit purely from diversified across factors, and not just focusing on one or two. 

Lastly, how we actually decide which factors to invest in. So, we’ve got this pool of factors, and we wanted to be determined whether we want to be overweight or underweight any of these factors at any point in time.

What actually allows us to make that decision is the short- to medium-term trend in factors.

So, we look at each of these factors, and what tends to happen is when a factor starts to trend positively, it generally tends to continue that positive momentum. Similarly, when a factor starts to underperform, it also continues that downward momentum. 

And that trend looking over the last year, actually tells us, right now we should be overweight these factors and underweight those factors. And that ultimately informs us on a high level, on how we should be positioned within the market.

The Finance Ghost: So much cool stuff coming through there. Thank you very much. That really does give a strong indication of how this thing actually works. 

And I think people hear terminology like “quantitative versus qualitative investing”, they hear words like “algorithms”; I mean “algo trading”, which is obviously not what this is.  

But people hear these kind of terms, and maybe as part of answering the next question, I can ask you to just help us understand exactly which umbrella this fits into. 

You’ve already touched on it, which is how the model is maintained, how it’s built, how some of the backend academic-type research has informed the way this thing is actually put together, the amount of back-testing. 

But ultimately a lot of it also comes down to managing human emotion, right?

And perhaps my… not perfect definition, but the one way I would think about a more quantitative fund, there’s a lower probability of emotion coming into it because it’s very model-driven, as opposed to something where there’s more in the way of judgment calls, right? 

Reza Fakie: Correct. How we view the world is that we’ve developed this model to tell us how to invest in factors. But ultimately, any model, being quantitative in nature, you trust your model, you understand your model. But you know, every model has downfalls, it has risks. 

And for us, this takes us to the next step, which is actually portfolio construction. Which is where we say, “Actually, what is this model really good at, and where can we utilise that in our portfolio? But also where is this model really weak, and how do we prevent or mitigate those risks?” 

And that is what we spend a lot of time applying our mind to. Where we say, well, we like choosing factors. This model is really good at deciding how we should be positioned. 

But ultimately, we don’t want to be taking any single stock risk. That is, we don’t want to have a large active tilt relative to our benchmark. So, our benchmark is the MSCI All Country World Index, which has both EM and DM in it.

When we look at this benchmark, this is our guiding light. This is what we want to generate alpha against.

When we look at it, we say, well, this model can be really good at producing alpha by choosing factors. But given the universe and this benchmark that it’s being invested against, there are certain components of it that we want to mitigate. So, we don’t take a large active tilt, we won’t go more than plus-minus 1% relative to our benchmark, as well as on country and sector. 

So, we’re not going to take a huge punt on the US and say we’re going to go only into the US or a significant overweight into US, or underweight. We cap that around 3%, and the same with sectors. 

What lets me sleep at night is that I trust the model. This model has been continuously tested and back tested. But we’ve put the safeguards in place on the portfolio construction side as well to ensure that the model is actually controlled for what it’s good at and what it’s poor at. 

When we run this process, we never ever make any changes. We never override it, saying, actually, I want more Nvidia or I want less Samsung. We allow the model to do what it does best, and we trust in our portfolio construction to actually mitigate those risks. 

We never change the outcome. If we believe something can be improved, it’s always backed by research. So, we’ll go back and say, well, is the portfolio behaving as we expected? Is it from portfolio construction? Is it from the model? Are there any improvements we can do to either improve either or both components? 

And it’s a research-to-improve mindset, rather than “I don’t like this outcome, let me change the outcome”.

The Finance Ghost: Yeah, that’s great. Another wonderful set of insights there. And I think you’ve spoken so well to some more of the design elements around things like tracking error, etc. and how you just actually build this thing. 

So perhaps just one more question then, before we actually get into some case studies in the fund, which I think is where it gets really interesting. 

Just the costs of churn. It sounds like this is the kind of fund that might be making changes (and you can confirm whether this is correct or not). Would you say that the churn in this fund is perhaps higher than some other models and how do you actually manage that?

Reza Fakie: We actually don’t see as high a churn as you might expect. Generally, within this fund we’re looking at about 100% one-way turnover within a year. And that may seem a lot to fundamental standards where they may churn a lot less, but you have the potential (given a quantitative strategy) to actually churn significantly more. And how we actually control that is how we look at our trading every single month. 

Now if you think about what I’ve covered so far, when we enter a new month, we have this model that we’ve updated and tells us, well, what are the best stocks to invest in given our model views? And we have our existing portfolio.

So, if you think about just manually (and again we’re obviously applying an optimiser to get to this result) you’d say, “Well, what is the best thing in my model that it now really likes that I don’t own?” And I’d buy that.

“What is the worst thing that I do own now in my model that I want to get rid of?” And if you buy and sell, you “net neutral” that trade, it will give you some sort of gains. Let’s say that’s 1% expected gain.

If you repeat that and say, well, what is the next best thing I don’t own and what is the next worst thing I do own? And do that same trade again, you’re expecting to generate some positive alpha. 

But the reality of what happens when you do this through each level of turnover is that you get a “turnover frontier”. And what you end up seeing very quickly is that when the model hasn’t changed significantly – and again, due to the short-medium term persistence, a lot of the time over one month, two months, three months, we’re not seeing a significant shift in the model. 

It depends on what Trump decides to wake up in the morning and say and tweet about. But that just creates this added volatility level in the market. If we take an example of what happened over last year, when the trade impact happened, when Liberation Day happened, that actually caused a significant shift in the market. 

Because people were actually changing their minds on how they should invest, should they go more value, should they go less quality. That actually impacted the model. And we saw that change happening in the model, and in our portfolio. Whereas if you look at the Iran war, this Iran war just adds a lot of inflationary pressures, but it never actually changed the investor’s mind. 

So over that period, our model would have changed quite a bit. Over this period, our model has actually almost ignored the Iran war, and we’ve benefited from effectively ignoring the noise. 

Coming back to my example, what then happens is you see this frontier, and if you think of any sort of efficient frontier, there’s a point where your additional gain kind of flattens off. So, what we look to do is maximise that marginal gain. 

We’re only trading as much as we’re getting more signal from our model into our portfolio. But we won’t go just trade everything we can because that ultimately just introduces costs into your portfolio, which then detracts from performance.

Coming into each month, we dynamically assess what is the optimal amount to trade to actually maximise the signal within our portfolio without just trading for the sake of trading.

The Finance Ghost: It just shows you how many judgments calls there still are in something like this, right? As much as it is very model based, quite correctly so, there still needs to be, dare I say it, that “human in the loop – and of course, that concept is key to all the debates around AI at the moment, which is driving global markets. 

And maybe that’s the perfect opportunity for us to now jump into some of the case studies.

Because unsurprisingly, if I look at your fact sheet, a number of the big tech names that I would expect to see in a global fund with momentum as one of the factors are there, and that makes sense to me. 

But the weightings do look quite different to what you might see if you go and actually buy just the broad index. 

Let’s start then with the big tech names. How does your fund treat these stocks? Why are they important? And am I right that the weightings do look somewhat different to what you’ll find in the index?

Reza Fakie: That’s correct. So again, our starting point is always this model and what it likes in this market and what it dislikes.

But why you’d see those big tech names comes down to a portfolio construction process. As I’ve mentioned, we don’t take a large active tilt, say plus-minus maximum 1% relative to these names. 

And as we’ve seen over the last few years, we’ve seen this growth in mega-tech companies with significant weightings, like Nvidia, Microsoft, Tesla. Micron just became a trillion-dollar company the other day. You’re seeing this growth in these large companies. They’re taking up more and more of the index.

Now, taking significant risk against not holding these ultimately leads to a poorer outcome in our process. So, we limit our active tilts around these stocks and focus on holding the factors themselves and generating, or rather harvesting, from the factors themselves. 

If you compare those Mag 7 or those big stocks relative to the benchmark weights, you’ll find we’re slightly underweight six of those seven. The only one that we overweight now is Alphabet. 

It’s holding these stocks because they’re quite large in a benchmark but actually taking active tilts away from them to generate alpha from the factors themselves. And we’ll see that, for example, the model doesn’t necessarily dislike some of these large Mag 7s that we’re slightly underweight, but rather it’s found better opportunities elsewhere. 

For example, we like SK Hynix and Micron. If you look at the benchmark itself, it is holding a large weight in these large tech stocks, as the index itself has become more concentrated and these large stocks have seen significant growth over the last few years.

While we’re underweight a lot of these Mag 7 stocks, it’s not to a large degree, and it’s not that necessarily the model “dislikes” some of these stocks that we are underweight. It’s just that it’s found better opportunities elsewhere in the market. It’s always a balancing act when we’re optimising. 

It’s the difference between how we maximize the potential alpha for the risk we are taking, but also mitigating taking active tilts where there isn’t necessarily that benefit to be had.

We are taking those active tilts around the benchmark to maximise that factor return. But it’s mitigating that overall risk. And as you mentioned, we employ a fairly strict tracking error of 2% to 3%, because within that, we believe there’s sufficient opportunities to meet our performance goal.

The Finance Ghost: It’s very much about finding alpha at the margins, right? That’s really what this is about. It’s not, for example, a hedge fund which might do something wildly different to what the benchmark might be. 

Where the benchmark almost becomes like, “Well, you could have invested in this”. But actually, the things are so different that there’s almost no comparability left at all.

Whereas what this is basically saying is there’s going to be a lot of clever stuff applied here. It’s going to be different, but it’s not going to be wildly different, right?

Reza Fakie: Correct. And as our motto is “Champion the Unseen”, we’re looking for those opportunities. Many people find it surprising.

So, looking at MSCI ACWI, going a little bit deeper, 90% of it is in developed markets. Only 10% is weighted in emerging markets. But actually, by number of constituents, it’s split 50/50. 

Half the universe is in emerging markets and there’s a massive amount of opportunity available in that. So, it’s finding those opportunities that maybe may not be apparent, may be overlooked given their size, but given what our model is telling us, this is attractive. Even though they are a smaller company. 

That is what ultimately gives us confidence that we can invest in these stocks. They are aligned with our model. We expect them to outperform. One of the examples we have is we’ve been invested in Samsung, SK Hynix and Micron since late last year, and we know that the big story for this year, starting from January, has been this massive ramp-up in performance. They’ve done over 100%, some of them over 200% year-to-date. 

The reality is when we looked back at that point in time, there were these factor characteristics that we really liked. We liked high-beta stocks, and they were definitely high-beta stocks. We like the momentum component of them. And what may be surprising to some, especially around value, is that these were actually very good value stocks. 

You think of this large run-up and like, well actually how can they be value stocks if they’ve seen such a large run-up? Well, they actually had really good earnings because there was this significant push in demand in their product for this AI build-out that’s happening. 

All of a sudden, everyone needed memory, especially the big AI scalers, Meta, Amazon, and that significantly pushed up their margins, significantly pushed up their demand and they saw that revenue come in as earnings. Relative to their share price at the time, they were seeing a significant growth in earnings relative to share price initially, which actually made it very attractive on a value basis. 

Yes, some of that value basis has declined somewhat, given the continued share price increase. But again, we’re not seeing it as a detractor. We still don’t see these stocks as expensive stocks, relative to the rest of the universe. 

It’s this combination of factor views that ultimately allows us to have this confidence. It’s not just one factor telling us to invest in a stock. It’s the combination of a series of factors, and they are well aligned with our overall factor views to actually say well, this is something that should persist into the future.

The Finance Ghost: Of course this leads to the obvious next question, Reza, which is what do you do first in the morning? Brush your teeth or check the South Korean market? Because it sounds like it might not be the teeth, huh?

Reza Fakie: Yes, I generally do check what’s happened around the world. South Korean market opens at around 2am our time, depending on daylight savings. So, a lot has happened by the time we’ve woken up. 

And it’s maybe just being a global portfolio manager that many people think that no, you just care about the US; the US is 60% of your benchmark. But actually, it’s where you have your active tilts, right? It’s where you are invested in. And we’ve invested across Thailand, Hong Kong, Korea, India. 

There’s a lot that has happened by the time I switch my desktop on at 8:30 in the morning. A lot of the trading has already happened. And it’s more to just understand, well, what has actually happened? 

But again, being a systematic investor, I’m not putting my finger on the trigger every day and saying we need to change something. It’s about understanding what is happening out there in the world. How is it influencing your portfolio? What is likely to change into the future? 

If you start seeing a trend emerging from a certain sell-off or a certain bull run, you know that when you get to the next model run that that is going to ultimately influence what your next month’s portfolio is going to look like. So, it’s a good idea to understand exactly all the moving pieces.

The Finance Ghost: Yeah, absolutely. And well done on those trades obviously because those are the positions you wanted to be in this year. But of course, as we speak, lots and lots of question marks around AI stocks and especially I think those top-of-the-value-chain type names. Your memory stocks, etc. 

It’s the shovel in the gold rush, of course. And we saw some interesting news recently from Meta selling “excess compute”, which I think are two words that gave the market a little bit of a skrik.

Everything has been about a supply crunch. “What is this excess compute that you are speaking of, Mr Zuckerberg?”

And look, no one knows obviously, we’re all just trying to do our best to figure it out and try and guess what’s going on and make educated guesses around what’s going on. 

But in terms of your approach, and the model and the cyclicality that is inherent in a number of these stocks, and making difficult judgment calls like, “Are the memory stocks still cyclical or are they actually enjoying a structural underpin now?” 

How do you handle that in a multi-factor model? What are you thinking about at the moment as markets look increasingly hot, let’s be honest, around some of these stocks?

Reza Fakie: There are two components to it. The one is, the model will identify what is a good value stock. So, for example, SK Hynix and Micron, where those stocks became really good value stocks, and now they’ve declined in value. 

So, through time, as these stocks outperform/underperform, as they release their quarterly results, the picture of a stock and what it’s exposed to and whether it’s a good value or good growth or good quality stock, that transforms through time. 

On the other hand, what moves a lot quicker is actually our model itself, determining, do you want to be in momentum right now? Do you want to be in high-beta stocks? Do you want to be in quality? 

Those two components are moving through time. And at the moment we’re seeing, while there’s these sell-offs that happen for a few days, there’s these structural changes that could be happening. At the moment, we’re still seeing that the same factors are playing through. Momentum is still playing through quite well. Value is still playing through quite well. 

We’re starting to see a bit of a correction on that. But one month is not a correction, right? You need to see a significant trend change for it to change your mind. It’s not always a good idea to just pull the trigger quickly and see, “Okay, something is changing. It looks like it’s changing. I want to get ahead of it”. 

The reality is you don’t know, at that point in time. You need to actually sit back and say what is measurable, what is actually investable is a trend. Monthly signals aren’t a trend. So how is this month influencing the longer-term signal? Is it shifting it back? 

What we generally see, and like I mentioned, our model doesn’t generally change one month, two months. When volatility starts coming off, when momentum starts coming off, you’ll start seeing that pullback in our model as well, until a point where it’s pulled back far enough to go underweight. 

But at the same time, in those stocks that are now performing well or underperforming, we might see, for example, a stock starts to underperform, but the momentum theme itself might continue. 

All it means is that that stock itself isn’t a good momentum stock. But there are other stocks that have now come up and have now bolstered this momentum theme further. And we still believe in momentum, but it’s just not those same stocks anymore. Hope that clears that up a bit.

The Finance Ghost: Yeah, it makes sense. Thank you for being willing to share this stuff. Obviously, you’re sharing ultimately your proprietary approach publicly, so you can’t send us a screenshot of the model. But it certainly helps to just understand more of how you think.

As we start to maybe bring this to a close. Let’s talk about some of the smaller names in the fund. As you quite rightly pointed out earlier, Old Mutual Investment Group is busy “championing the unseen” at the moment. 

And that means just putting the spotlight on some of the areas of these funds, etc. that people might not know are there, and might find very interesting.  You’ve given us some quite big names that I think people will know. They were unseen; I’m not sure they are now; but as you say, that’s how momentum works. 

But some of the other smaller names in the fund, maybe a couple of examples, and at what weighting they tend to come in. Because I think, as you said earlier, it’s interesting the split between developed and emerging markets in terms of overall exposure, but the number of names in each of those portfolios was a particularly interesting insight.

Reza Fakie: In terms of smaller companies, maybe going back to the MSCI ACWI Index, there’s roughly 2,500 Maybe surprisingly, there are as many US as Chinese stocks. So even though the US is 60% of the index, it’s got 600 stocks. China actually has around 600 stocks as well, and it’s only around 4% of the index. 

So, there’s a lot of opportunity within China and Chinese stocks, especially around the same AI build-out. So, the two examples I have are Zhongji Innolight and Eoptolink. I’ve probably completely butchered their pronunciation…

The Finance Ghost: …I mean, I’ve never heard of them. So, Reza, on the money there, championing the unseen. I’ve never heard of those names. You’re going to have to let me know for the transcript how to spell them (laughs). That’s how unseen they are. Fantastic. Carry on.

Reza Fakie: They’re both optical companies, so involved in the AI build-out. So Zhongji Innolight produces optical receivers; Eoptolink produces optical modules. 

And again, as AI grows, the data centre components. Yes, there’s that massive Nvidia chip. Maybe not in the Chinese servers for now, but there’s these massive chips, there’s these optical providers, there’s different components. 

The two components that are probably focused on the most in the market is the chip itself, made by Nvidia and AMD; or now at the moment, the memory producers being SK Hynix, Micron and Samsung. But the reality is there are thousands of other components that actually go into building this AI build-out. And these are two of those companies. 

And again, they were identified quite early on by our model based on their factor exposures and given their size. So, these will probably be less than 5bps. Well, they will both be less than 5bps in the benchmark. And we generally take an active tilt of around 30bps to 50bps initially, depending on how well the stocks are liked, and how much risk they contribute. 

Because there’s always a payoff, right? Between a stock that is really liked, versus how much risk it contributes to your portfolio.

As an example, just going back the last month, Micron is still really liked in our model, but given that it’s run significantly, it generates a lot of risk, it’s actually been pulled back in our portfolio construction because of its significant contribution to risk. 

So similarly, when it comes to these smaller stocks, we know that including a very large active tilt will significantly increase the risk of the portfolio. But we’re trying to maximise this gain across the entire portfolio. 

These are two small stocks it’s identified. It won’t put them at significant overweights, maybe 30bps to 50bps. But ultimately, we’re looking at those small plays that we can actually generate alpha from. 

The third company I’ll bring across is – I know there’s been a lot of hype around SpaceX lately, so a lot of people are worried that all of these index providers are including it. Is it going to dominate our indices? There’s a brand-new company, we all have our different thoughts around Elon Musk and Tesla, and people were worried: “Are people just going to be forced to buy the stock?”

Well, very early on became obvious to us that yes, by market cap it’s a massive company given its size, but it actually has a very small free float. And the reality is within the MSCI ACWI Index it has come in at less than 10bps. It’s almost small enough to ignore. It’s not this big player everyone thought it would be. 

Yes, it’s still a significant size. Yes, it is a fairly large player in the space. But actually, it’s coming at a smaller size, maybe, than what people expected.

On the other hand, a company that our model did identify a few months back and we’ve been invested in, is Rocket Lab. Very similar to a SpaceX. But actually, it provides end-to-end launch services, spacecraft design, satellite components, flight software.

So, everyone is focused on SpaceX, but actually we’re seeing a growth in the space exploration business as a whole. And this Rocket Lab company that we’ve identified is actually something that we’ve been invested in and has actually grown nicely. 

Maybe some of it is due to hype from SpaceX, but actually in its own right, it’s aligned very well with our factor views and it’s done really well in our portfolio.

The Finance Ghost: Well done, I really enjoyed that. So, let’s bring it home now. Reza, I’ve got you for a couple more minutes.

AI, we’ve spoken about it a great deal in terms of something you can invest in, but I’m guessing it’s starting to have an impact on how you actually run the fund as well? 

These tools are always interesting to think about. They certainly do have their limitations, but they tend to have some benefits as well. Maybe for the sake of the interest of listeners, give us a couple of minutes on, in this fund, how AI tools are starting to make a difference to your daily life?

Reza Fakie: It’s actually been making a significant difference. From my perspective, when we look at the factors themselves, and the model, we want them to be understandable, right? We don’t want to just generate a black box and hope for the best. Because if you don’t know why something is working, you don’t know when it’s going to stop working or why it will stop working. 

So, when it comes to the actual modelling process, we try and keep our process as transparent and understandable as possible. But in the ways we work – so  I mentioned before, I spend a lot of time programming, and just in terms of using Claude Code and using Claude to generate code for me, it’s been helpful.

I’ve been coding for over 20 years. I started in high school, so maybe revealing some of my age here, but I’ve been coding for many years. I’ve just noticed that I focus less on syntax (on figuring out the perfect amount of code, the perfect way to do something). 

Claude knows exactly how my database is set up, how I usually do my queries for analytics, and I can spend a lot less time worrying about typing out code and a lot more time focused on the analytics side. It allows you to be more productive from that perspective. 

Even in my commentary, I will write out my views of what has contributed performance, all the information that I think is relevant; and I’ll then have Claude review it and say, “Well, this can be more succinct. You are duplicating words”. 

So, you focus on the important parts of your job across the board, not just in portfolio management, but subject matter experts are going to become more important because you have to decide: is Claude telling you the truth? If Claude gives me a bad piece of code, I need to understand, well, what is it doing wrong? I can’t just say this isn’t working, fix it. You need to understand what exactly it’s doing wrong. 

And sometimes it’ll give you a piece of code that works, but the output is wrong. And using your experience (your subject matter experience) you need to understand why this is wrong.

What is it doing wrong, what assumption is it making? 

So, the “Claude is going to take all our jobs” hype is going away and it’s focused more on, as people, we have certain knowledge sets – being stats, maths, marketing, writing. 

Ultimately where Claude will help us is to actually improve our output. It’s not about us not doing any work, it’s about verifying and understanding exactly what is coming out and actually using it to help us make better decisions, ultimately.

The Finance Ghost: Reza, thank you. Really appreciate your time today. Where can people go and actually find out more about this fund and potentially engage with you if they are interested in investing?

Reza Fakie: We have our website, oldmutualinvest.com. That is our main portal. Feel free to contact anyone on the website at the bottom, and within our distribution team.

The Finance Ghost: Excellent. Thank you so much. I am really enjoying getting to know the team on that side and how you guys operate. It has been a lot of fun, and I’m looking forward to the next podcast coming along as well. 

To the listeners, if you enjoyed this, go and check out the fund. Also go back and listen to the previous podcast with Old Mutual Investment Group. That was with Maahir Jakoet, and he runs the Shari’ah-compliant fund. Let me tell you, you’ll get some really cool insights there as well into how Shari’ah -compliant investing can deliver some unexpected performance outcomes, versus what I would call traditional investing. 

But Reza, thank you so much for your time today. You’ve given us a wonderful example of multi-factor investing. It really helped us understand what’s going on there and all the best for the remainder of this year. 

I think we’re in for an interesting time in the markets around some of these tech stocks. I’m sure you’ll do a great job of navigating it, so well done and thank you.

Reza Fakie: Thank you.