Thank you all for coming. It's great to see you all. Welcome to our 2022 Investor Day. Before Luke kicks off today's sessions, I thought I'd take you through the agenda briefly. The first part of the morning, we'll talk about what we think our key differentiators are. After you hear from Lara Carty, our Global Head of Talent, we'll walk across the floor over to our trading desk for the following three sessions. After the break, we'll be back in this room, where you'll hear about why we think we're well-positioned for growth in the future. Antoine will then touch on shareholder value, and we'll move on to Q&A at 12:00 P.M. We'd ask you to save any questions until then, please, in the interest of staying to time. With that, hand over to Luke.
Right. Look, I think one of the bits, we've had a very strong five to six year run, and today is not about telling you how differently we're gonna do things in the next five or six years. It's about the fact that, you know, everybody's waiting for the one number they can write in a report saying, "Luke said the five year plan was X or Y." It's not about that. It's about why we believe that what we've been doing the last five years is sustainable for the next five and beyond that. There is a process about what we do, and we think it really works. There's no new strategy coming out. This is more of the same, but the point about it is you can extrapolate it out.
The point about what we do is that clients have an almost insatiable appetite for alpha. We'll talk through the numbers as we go through the day, but they need added value. They need alpha in their portfolios. We manufacture alpha, and we manufacture it at scale, and we have processes to do that. That means that we can compound and generate more alpha. The more alpha we generate, the more fees we can earn, the more money we can run. It's really about more alpha is more fees, which falls through to the bottom line for shareholders. It's about demonstrating we think that is an ongoing opportunity. It's not about one fund. It's not about one strategy. It's not about we need markets to do X or Y.
You know, we think this process works essentially irrespective of individual markets, irrespective of the broad trend for markets. The way the P&L comes through varies, but the model works, whatever markets you throw at it, the business model. You know, we've made a conscious thing of never being dependent on a single strategy, a single client, a single anything else. It is about a diversified platform every which way. Right. With that, we'll kick off. This is stuff reasonably you know. As you're gonna hear a lot as we go through in the course of the day, we have a big bet on technology in the firm, right? We doubled up on it 10 years ago with the acquisition of Numeric, and we have invested continuously through the process. The firm today is large for liquid alternatives, differentiated.
We have a broad client base. You'll see diversification comes through all the way in the course of the day. It's about how do we deliver this alpha at scale, right? Last year, GBP 6.6 billion of net alpha, GBP 8 billion of gross alpha. That is the thing we manufacture. That's the thing that clients value highly. One could debate the right or wrong percentage of the alpha that we keep, clients keep over time, notwithstanding comments about fee pressure in the industry. The reality is, over time, the percentage of alpha clients are happy to pay is relatively consistent. Individual clients vary a bit up and a bit down, but it's relatively consistent.
If we keep generating the alpha and we have a process for doing that and doing it at scale, one clever thing that makes $1 million is not that exciting. Our clients need it in scale, so $8 billion of gross alpha, that's delivering it in scale, and there's no reason that can't go to 10, 12, pick a number anywhere else as you go on. The demand is very, very high. It's there because if you look at expected returns from beta, classic betas, bonds and equities, you can't get far away from some, you know, in terms of real returns, which is what clients want, you can't get far away from real returns in an economy. Less cost doesn't get you anywhere near the sort of real returns that clients want.
They're sort of 3%-5% behind their target if they look at long-term betas. In any one year, you could get excited, so they need added value. That's what we deliver. One of the key routes to that is the liquid alternatives because, you know, liquid alternatives, it's high alpha strategies and so on and so forth. It's about the alpha. That's what the clients want to help them get to their strategies and, you know, we are one of the biggest manufacturers of that in the world, and that's sort of unique in that the only people we compete with in manufacturing alpha are private businesses. As I talked, it's not about a single strategy. Often people want to ask about this fund, that fund. There are lots of different funds. I don't know.
I think we count 75+, but honestly, you could cut it down into hundreds of individual models. You know, 100+ teams, so on and so forth. The thing that is unique about the firm is we have lots of different forms of generating alpha. All of them are either quant, so driven by technology, or they're empowered by technology, right? Everywhere we use technology to give us an edge to improve what we're doing, which is a sustainable edge. We keep pushing through that, and we have a unique way of combining them so that the client gets what they actually want. That's the solution stuff. It makes life difficult to analyze from an analyst point of view, sorry, but it's really, really valuable from the point of delivering with clients, and we'll talk about that later on as we get through the day.
You will hear all day about how we use technology. We use it to deliver quant strategies. We use it to leverage what the capability of our discretionary PMs. We use it for all the analysts in the room to read all of your analyst reports and parse them. You might see a bit of that demonstrated later on as we go through. It's always hard to put individual numbers down on technology 'cause it depends how much technology you understand, what's the cool thing, what's the not cool thing. The course of the day we'll try and bring it alive. It requires continuous investment, but also it's a compounding effect. We've been investing in technology across the firm for 30 years, aggressively for the last 10 years. The spend every year compounds to keep you further away from the competition.
It's hard for people to catch up, one way or another. The second bit of it is everything goes through a single platform. We use the technology so that once you make a decision to buy something or sell something, that decision might be made by a human, it might be made by a trending model, it might be made by a stock-picking model, it doesn't matter. You make a decision to buy or sell something. After that, everything goes through a central infrastructure. We do everything aside from alpha generation, which is about creativity, everything else we do once, so that we do it really well. Through that, you get the power of really being able to invest in your execution, in your sales team, in whatever, but you also get the leverage for shareholders of doing it once.
You know, we have a much more powerful sales force than you could possibly have with 1 or 2 funds, but we don't have the equivalent of the sales force you'd have if you have a 100 times how many people you'd need to do it. There is real scale benefits that come through by putting it all through the same infrastructure, and it goes for everything after the decision to buy or sell, right? That is all empowered by technology. Without that, you couldn't possibly do it. It's been a really strong five to six years. You know, you'll see over the course of the day, we think the actual demand for the alpha we produce, and therefore the, I mean, it gets delivered through products or solutions. It gets delivered by people investing money, but the thing they're investing in is that alpha.
The demand from investors out there is both permanent and huge. Bigger than, I mean, sort of, it's not quite unlimited, but it's bigger than you could possibly need to produce. The demand is there. We have a demonstrated ability to generate that alpha year upon year, and to grow our capability of generating that alpha in a variety of ways. You know, you have to have the relationships with people to make sure that they'll actually, you know they can buy the stuff from you, and we've really worked on those relationships. All of that uses technology to make it work, but really importantly, people. It's very easy in these things to sort of ignore the people 'cause it's harder to talk about, but the talent here and making the talent work together is super important.
That's why going up first, we have Lara, who's gonna talk about talent.
Good morning, everyone. My name is Lara Carty. I am the global head of talent at Man Group, and I have the enviable job this morning of trying to describe to you all why our culture and talent are differentiated via the use of PowerPoint. I'm not sure if anyone has ever taken on that pursuit before, but I'd argue it's almost impossible. I'll caveat up front by saying the real presentation today is going to be you being here for this event. Not only will you get to see a lot of my colleagues, but you'll also get to see us in our natural habitat as you do a brief tour of the trading floor and move to our demo area.
I am anticipating a slight amount of cynicism about this talk, and that really is okay, because for the last 17 years I've been asked on a daily basis, "What is your job and what do they pay you to do?" I feel like I'm pretty well equipped for any questions that you might have in this area. My hope is that for the next 10 minutes, I'll move you from cynicism to at least curiosity, and if that curiosity leads to questions or intrigue about anything we're doing in the space of talent, I'd be very happy to answer those in the Q&A or over lunch with you today. I am an occupational psychologist by trade. I've worked in financial services for my entire career, and I've supported thousands of people to perform better in their roles.
When I said yes to the job at Man Group, I complacently thought, "How different could it be?" Oh, how wrong I was. There's something quite unique that happens when you combine 1,500 people, where 40% of the workforce are tech and quants, where they sit shoulder to shoulder with discretionary investors, and the second most common language is Python. I quickly learned that it is in fact possible to have in-office banter entirely through the medium of math formula. I also learned in my first week at the firm that it's possible to sit in silence for a full 2 minutes when you ask one of your AHL colleagues a question. That was about 5 seconds, so you might get a sense of just the levels of awkwardness I experienced in my first week at the firm.
Six years on, I can't unfortunately contribute to the math jokes, but I have learned to embrace that there is a real value that comes from people actually thinking before they speak. I believe I have one of the best jobs in the world. I get to hang out with uber nerds every day, who are really intrigued by solving incredibly complex problems, and I get to be a very small part of helping them to be exceptional at what they do each day. I am sure you have heard the phrase that people join a firm and leave a manager. At Man Group, we strive for something a little bit different from that. We strive for experts being managed by experts.
Of the tens of thousands of applications, CVs, or profiles we consider each year, we have the luxury of being incredibly enforcing incredibly high standards in our hiring process. You simply cannot work at Man unless you are at least striving to be world-class at whatever your area of technical expertise is. When you bring talent of that caliber into a firm, you really need to be laser-focused on how you support them. Five years ago, we set up an in-house coaching provision. That means that at any one time, up to 20% of our top performers are supported by one-to-one coaching. We've developed our own in-house data science and coding curriculums.
In addition, when you are at a point of saying, "I'm done with this challenge, and I'm ready to move on to the next thing," it's incredibly safe to do so here. In fact, it's encouraged. You see that culture really reflected not only in our retention rates, our internal moves, but also our employee engagement scores. I'm sure by now that many of you are familiar with the Oxford-Man Institute and OMI, and these are ways that we share some of our internal expertise with you. Let me lift the lid a little on one of our secret sources, which is how we learn from each other each day. When you combine the profile of people we hire with a culture that truly embraces collaboration and openness, you get something quite special. It leads to an incredible learning environment.
Minds at Man is a great showcase of the diversity of thought and experience of many of our colleagues. I don't know about you as you read through the titles, but I really struggle to see past the Boston Red Sox example, and I've accosted Greg Bond on many an occasion to really try to understand what it is that makes him want to work at Man Group versus being the general manager of a Major League Baseball team. You know, that feels like quite a big decision for someone to make.
At any one time, about a third of our workforce are engaged in mentoring, and we have individuals like Rory Powell, Andrew Swan, Charles Long, Nick Judge, Henry Dixon, who are just some of our discretionary investors, who have been generous enough to dissect their investment philosophy, their investment process, and their lessons learned, which are sometimes painful along the way, to other investment professionals across the firm. This is something that's truly unique about our willingness to share what we do in order to ultimately improve our investment capabilities. There is no shying away from the fact that our persistent quest for alpha creates a demanding work environment. Pressure's not necessarily a bad thing. In fact, it often leads to outperformance, and you frequently see that in other elite performing environments. We do have a carefully curated evidence-based wellbeing strategy.
We know that pressure becomes unhealthy when an individual doesn't have a sense of control and when they do not have access to the right support. Our agile working model enables to a large degree of control, each individual to select where they work, how they work, and when they work. In addition, we have a dedicated team of talent consultants and an extensive wellbeing offering that is there to support any individual in the firm. In 2021 alone, over 90% of our workforce elected to engage in talent support. Having worked in financial services my entire career, I can tell you that's pretty rare. We were incredibly proud that we had managed to develop an offering that supported employees in that way.
We know that representation matters, that it's easy for biases to creep in, and that if you are being complacent, that you are leaving a lot on the bench if you don't allow people to bring their whole self to work. I'd argue that creating a representative workforce is one of the most complex talent challenges our industry faces, and we are applying the same intellectual curiosity, the same optimization approaches, and the same personal touch that you will see in our quest for alpha, our development of elegant software, or the interactions you have with our sales force. We do believe that to truly solve this challenge, you have to develop a really intricate understanding of potential. You have to source talent from different pools, and importantly, you have to have a culture that allows you to grow your own.
That is what we are committed to, and we are optimistic that we therefore will not be having the same conversation about diversity a decade from now. We work in an industry that affords us many privileges, and it would be easy to become consumed by that. Not only do we believe we have a moral obligation to invest in our communities, but we also know it makes us better investors. At the heart of great investment management is unbiased, non-ego-driven perspective and decision-making. We have found many opportunities to engage with our communities that delivers this mutual benefit. It's why we don't only lay claim to a very long-standing engagement with our communities, but we are high conviction that this will be an embedded part of our culture as we continue.
Hello everybody. I'm Mark Jones. For those who haven't met me before, I'm the Deputy CEO here. I'm gonna talk a little bit around the technology within the business, and try and do three things for you. Firstly, give you a bit of a demonstration of what quant strategy set-up actually looks like. So Slavi, as I understand, is gonna run you through a demo, both to give you a bit of a tangible feel, but also to give you a sense of the benefit of doing this on our platform and the speed and efficiency that you can see as a result. Then a little bit of how we fit into the wider technology ecosystem. Then lastly, and I think perhaps most importantly for a lot of you, why that matters.
You'll have heard us talk a lot about the technology within the business, but I'm gonna show you some of the examples of how it really drives better outcomes for both the clients in the first instance, but then also ultimately for shareholders. With which, let me hand over to Slavi just to run through that demo, and then I'll come back and talk through the second piece.
Thank you, Mark. Good morning to everyone. I'm one of those people where Python is my first language rather than English. I speak to computers all day long, so I might not be particularly polished. What I thought I would do today is give you one concrete example as to how our tech platform has transformed alpha research. It will be fairly hands-on in that we'll type in some Python code, and we're going to ultimately test an alpha idea functionality. What's our alpha idea going to be? It's going to be that if during an earnings call management and analysts are quickly positive, then the stock is going to outperform, and vice versa if they're negative, right? Going to underperform.
Just to give you a few examples of what we might mean when we say positive or negative sentiment, let's say here's a sentence from Bank of America's call transcript. "The first quarter of 2007 demonstrated continued strong demand." We consider that to be positive. Or for instance is, "Last year development margin declined," we consider that to be negative. Ultimately, we want to build a sentiment model that's going to go through all the sentences in an earnings call and then give the call a score. We wanna go long stocks that have sentiment and short sentiment as well. Now, the sentiment model we're going to use for this demo is going to come from the most influential academic paper in academic finance.
It was published around 11 years ago by Professors Loughran and McDonald, and it since then has been cited thousands of times. I should say that the sentiment model that we actually built and the trade we did at [Monarch] significantly more sophisticated than this. But of course, because they're proprietary, we can't do a demo with them. But for the purposes, I would say for demoing the tech platform, any model will do. The dataset we're going to work with has about 800,000 call transcripts for a total of about 306 million sentences, which just to give you some context, is about 10,000 times the size of a decent book.
Just to state the obvious, me as a human, if I had to do this exercise myself of going through all these steps, I'd have no chance even if I took no rest, didn't sleep, didn't eat, and it took me 5 seconds to go and analyze each sentence, it would take me about 50 years to accomplish this task. Only versus.
Would you be replacing yourself then?
I'm gonna do this kind of live demo-ish. Even on the quant side, we said, okay, 50 years maybe on the discretionary side. Even on the quant side, it might take months to do something like that. Why? Because there's just loads to do, right? We first have to get the data. We need to parse it, meaning we need to get rid of all the bits of it that we're not gonna use for sentiment analysis. We need to build our sentiment model, which could take days to weeks. We need to run it on all these hundreds of millions of sentences. We then need to work out, okay, which document or which asset does each document correspond to.
We need to build our properly risk managed portfolio, and then we need to evaluate it. Okay. Each of these steps can take weeks to sometimes months, sometimes days if we're fortunate. To just give you an example, even getting the data, we need to work out, okay, what are all the data vendors out there? We need to sign up contracts with them, find the good one, right? Figure out what sort of format they use to store their data. We need to put it onto our platform. We need to parse it, which is a fairly fiddly, but a messy task of textual analysis.
You'd need to build your sentiment model, and then building a sentiment model could be anything from just, I don't know, a couple of days for a simple model, all the way to months for anything smarter. Then even running it on a dataset of 306 million sentences is computationally quite expensive. All in all, even in a good quant place, something like that could take months. What we're going to try to do today is to code it up from scratch on our tech platform in about 15 minutes.
The first thing we're going to do is connect to our natural language processing research platform, which is just a part of our broader tech platform that makes working with text as easy as working with time series, which is what quants are naturally used to doing. We've connected, and then what we're going to do is just get the earnings call transcripts dataset. You see it's as easy as just writing a line of code called Get Docs, and we say that we want the raw version of the earnings call transcripts. Raw meaning we'll get the transcripts the way they come from the vendor. The really cool thing about having a platform is that I could just swap transcripts, for instance, for news and then go get all the news.
I could swap it for, say, filings to get corporate filings in the U.S., or swap it for analysts to get the analyst reports, right? Then everything else that I'm gonna type in, I could just run on a different dataset to get the view of a different market participant. Okay. For today, we're going to focus on just earnings call transcripts. We're going to go back to just leaving that get docs there to be transcripts. That's it. Now we have our transcripts dataset. Let's look at it. The first thing we do as quants with any dataset is to just have a play with the data, have a look at it, see what's in there.
What we're going to do is just get a random earnings call transcript just to see what's in the data that comes from the vendor, how good is it, how useful it is, for our research. Sometimes what you could do is actually limited by what's in the data. Here with this doc equals raw first print doc data, we just get a random earnings call transcript, one of the 800,000. To Lara's point about nerds, this is now proper nerd nirvana in this particular case, because it's a really good data vendor where the data is really highly annotated. We can see who the participants are in the call. We can see that the management discussion section is annotated independently from the Q&A.
You can see that the questions are tagged separately from the answers. You can see who the speaker is for every paragraph or every sentence that has been uttered, which is great because it now means that if I wanted to build a sentiment on just the questions, just the answers, just the management discussion section, the sentiment of the CFO, sentiment of the CEO, we could do that quite easily. But for the purposes of this demo, I want to run the most parsimonious model possible, which is going to be the sentiment of the entire document. If I do something like that, having all these annotations is actually something that I want to get rid of, and this is going to be the task of parsing.
Now I need to start with documents that look like this and ideally get rid of all the annotations, all the punctuation, all the special characters, numbers, and all that. Ultimately, I want to be left with just, the words, in there. Okay. This is the task of parsing, which could be quite messy, can take quite a while, lots to get right. For instance, like there's things to get rid of, simplify, and so on. This again could take quite a while. On our platform, we have the functionality that does, common tasks like that. By just doing parse XML, parse words of doc dot data, I'm now done. This document is being parsed, and I can again, have a look at what it looks like.
This is exactly the same document where it's been all the annotations are gone, numbers are gone, special characters are gone, punctuation is gone, and then I'm just left with the document becomes a list of sentences, and each sentence just becomes a list of nicely simplified, plain English, lowercase words, which is perfect for now scoring with some simple sentiment model like the one we're gonna use for the demo. Now I need to build my sentiment model. Again, how do I do this? Well, I'm done. Now I have the Loughran-McDonald sentiment model in one line with this function called Load LM for the Loughran-McDonald model. I can even poke around and see what's in there. Here I've just printed all the words that that model considers to be positive.
Things like strong, superior, valuable, win, worthy, and, you know, surpass are according to that particular sentiment model considered to be positive. Sentences that contain them are going to get nice positive scores. I can do the same with the negative list. I can see that, I don't know, restructure, reluctant, insufficiency, deter, serious are considered to be negative. I can play around with it a little bit more and see that, just see how many positive and negative words exist in this dictionary. I can see that 354 words are positive according to that model, and 2,353 are considered to be negative. Okay. I have my sentiment model now. I have my documents.
All I need to do is marry the two now and run my sentiment model at scale on these 306 million sentences, right? Which normally have a lot of compute, a lot of waiting because obviously you have to process now a pretty chunky amount of text. How do we do this using our tech platform? Well, we just say we want to score the transcripts words, and we want to score them with a function called Score Sentiment that just comes from our platform. Then we give it the positive words that come from the Loughran-McDonald model and the negative words that come from the Loughran-McDonald model. Now this starts to churn. What's going on? I think that's kind of my favorite part.
Now the 306 million sentences are broken down into 143 chunks. Yeah. They could be broken down into many, many more. I'll explain how we decided on that. Now 143 workers start up on our compute cluster, and 143 is a very small proportion of the compute cluster, by the way. It's a cluster that all of these people here would share. Essentially, that means that while I'm calculating this, everyone can basically carry on with their work without even noticing that I'm doing this.
Each one of the 143 workers in parallel starts to churn through this proportion of the 306 million sentences that's given to it, and ultimately gives me the sentiment of each document. Okay. Now for a simple kind of sentiment model that just counts words like this one, 143 is more than enough such that the whole exercise can finish in a couple of minutes. Now, for our own internal sentiment models that we actually trade, they're significantly more sophisticated, so they take more time to compute, so then we actually run many more things in parallel, such that again, we can get this performance of sentiment being scored pretty quickly.
For a simple model, even something like 140-150 is more than enough. I think if you have this iteration loop where it takes you a couple of minutes to score your entire dataset, it's already good, so you don't need to optimize it much further. Couple of more things to say. What I'm just going to do after this calculation is finished is just call a function called print analytics, which is gonna tell me, okay, how many documents were read by the machine in this exercise and how many sentences were read in this exercise. If all is right with the world, we should see something like 114,000 documents and a little bit above 306 million sentences.
The other thing that's pretty cool is if I wanted to actually run one of our actual sentiment models, I really don't need to change very much. I just need to change that score sentiment function that I'm passing to score with, like, one of our internal ones, and everything else carries on as normal. So kind of the power of the platform where I can just try now a different sentiment model on a different text dataset by just changing two words, right? That's it. Maybe requiring more compute because the actual sentiment models that we trade are more computationally demanding because they're significantly more sophisticated than counting words from a certain publicly available dictionary. Okay.
With that, what we're going to see when this is done, we're gonna see the maybe just to just to see the analytics first to confirm that I didn't lie to you, and we did indeed kind of go through roughly the numbers I was talking about. You can see we processed a bit more than 814,000 documents and a bit more like half a million more than 306 million sentences. You can see the output of this function is like super simple.
If you think about it as a table, we know that now for a particular document with a certain unique identifier, it had 300 sentences in total, 71 were positive, 31 were negative for a net sentiment score of 0.13. The sentiment score of a transcript can go from -1 if every possible sentence is negative to +1 if it's every possible sentence is positive. Okay. I've done that now. I've got my sentiment score. What's left to do is to actually go and do my backtest and see whether there's any value in this.
Now, normally there's another kind of, not particularly enjoyable step, which is trying to work out which particular security that does this document ID map to, and when can I trade it and so on, which again, if you have a good platform, takes one line. So now I'm gonna attach the AHL mappings to this earnings call transcript dataset, and you can see that the sentiment column stayed the same few sentences, same few columns, but now each document is linked to a particular unique AHL identifier, which is the security we're gonna trade off of this document. I get the GICS hierarchy at this point in time, and then the available date time, which is like a conservative estimate of when is it that actually.
The day on which I could actually go and trade this. Okay. Now we're onto the final bit of actually seeing if our idea works, which is going to be to do the backtest, and we're gonna call this function called create predictor. We're gonna use the sentiment scores we just calculated. We say that we want to do our backtest in the U.S. We say that we want to run a backtest at a single stock level rather than at a sector level or at a market level, and then we call this p.construct. Okay. My second favorite thing to talk about. Here what goes on in just this one line of code is we go from our sentiment score to essentially a properly built portfolio. So what does it mean?
Well, it's good to have your scores, but still you now need to work out how to build a portfolio that makes the best of these scores, that is properly risk managed. Darrel's going to talk to you later on a lot more about risk, but it's kind of embedded into constructing the portfolio. We need to load things like our tradable universe, the vol of each stock such that we can make sure that we allocate risk proportionally to volatility. We need to load our risk model such that we can make sure that we're neutral to a variety of different factors as we build our portfolio, so we're not just capturing something that we don't want to capture, and so on and so forth.
We load all the data that is needed here from scratch, from our you know backend systems. We ultimately what this function is going to do is make us a portfolio. I can think of it as a table where each row is a day, let's say, 'cause this is gonna be just a simple daily backtest, and each column is an asset and each vector basically each day if I take a slice, a horizontal slice, tells me how much of my book do I want to allocate to this particular asset, kind of a long or short.
That now would allow me to just plug it into some simple analytics to construct my P&L, to calculate things like my risk, my Sharpe ratio, and you'll see a few kind of basic analytics that we're going to show. In practice, we generate quite rich and detailed analytics off of backtests, but for the purposes of this demo, we'll just look at a few. To evaluate a portfolio, I just need to do this p.evaluate, which is gonna take the portfolio that I just built and calculate all these metrics that I'm interested in. I'm gonna do this p.plot_account_curve to get something a little bit visual, to see, you know, what's the P&L, how many assets I'm trading through time, and some very simple numbers, I guess. Portfolio construction is done.
Evaluation is gonna take me another couple of seconds just to get my P&L. Then I get my cumulative returns on the left of this backtest. Very importantly, you'll see that the account curve starts in about 2004 and ends at the end of 2017, which is common for every research project that we do, that some chunk of history would be hidden away from the researcher until the very end of the project when everyone signed it off, where we look at it out-of-sample only once, to make sure that we don't overfit. Here on this in-sample period, we can have a look at our performance characteristics for this simple backtest. You can see that the Sharpe is about 1.3, which is pretty good.
We can see that the risk is 10% a posteriori, which is also what we would have put in, a priori, which is good, and we managed to TargetR isk the way we want. The holding period is about 90 days, fairly slow and a decent capacity. On the right side there, you can see the number of stocks that are being traded in the cross-section each day. You can see on the same kind of x-axis is time, and on the y-axis is this number of stocks traded, which is just, of course, a function of data availability and the size of our investable universe.
You'll see it oscillates from about 1,200 to 1,500, which makes sense 'cause this is a backtest on the 1,500 most liquid stocks in the U.S. All right, that's it from me. We managed to kind of go from a simple basic idea, through crunching through hundreds of millions of sentences to ultimately getting some basic portfolio analytics in, roughly, 15 minutes.
Thanks, Slavi. Quant research at speed. Hopefully, that gave some of you who've never seen quant research a bit of a feel for what it actually means. It should be clear that that is the very efficient version doing it on this platform. That is not what the everyday experience of it would be. Part of the power of the platform is making people like Slavi as productive as possible so that they can get through that research process as quickly as possible. Turning back to the sort of second point of how do we fit into the sort of wider tech ecosystem. I'm touching here just on a few points of where we're working with people externally, where we're doing things ourselves, and indeed, where other people are using things that we've built.
Sort of starting at the far side, there's clearly things that we don't want to do ourselves, so we're getting data from external data vendors. That's their expertise. Ours is around the processing of it. There's 550 odd datasets brought in since 2018. Our focus is on the speed and effectiveness of reviewing them, not actually sourcing them. On the infrastructure side and just touching on cloud in particular 'cause I think it's an interesting example, you've got lots of people clearly in the wider world moving to the cloud. There's examples within our business where we do see benefit from it. It's scalable, and it comes with a bunch of services out the box, which are sometimes useful. In the core of the business, we think that our infrastructure is actually a big competitive advantage.
For our core investment research processes, some of the ones Slavi was just talking about, we think we're eight times faster and eight times cheaper than doing it in the cloud. That piece of thinking of the infrastructure as a source of competitive advantage rather than as some commoditized things of just boxes in the background, again, is an important part of the tech culture here. It's not just about the software development, it's also about the infrastructure and the platform. The last two points are just a feel for how some industry leaders and then some bottom-up people who work within technology see us. Clearly, you know about HUB partnering with either tech leaders or asset management leaders, so PIMCO, Markit, Microsoft, to build out something which we think can transform asset managers' operations platforms.
We are there because of the IP that we bring around the platform and how to do that effectively and build it effectively. Then lastly, we also give technology back to the open source community, so we release code back so that others can use it. There are a few examples of things we've built at the top. GitHub is a site where engineers can go and effectively vote for software that they found useful. ArcticDB, just as an example there, has been downloaded about 1 million times lifetime to date.
That's thousands upon thousands of engineers, technologists, researchers around the world seeing the value of what we've built and wanting to use in their day-to-day work, including actually some big financial institutions who are using the open source version in the background. We obviously use the enterprise version with lots of extra features, but just to give a feel of that external community seeing the value of what we've built. Now just turning to, you know, why does this matter? So why does this deliver better outcomes for firstly clients and then ultimately shareholders? There's three strands I just want to touch on. One is our ability to process information. There is an absolutely enormous amount of information that flows past your eyes every single day. It's essentially imperceptible for us individually. There's just far too much of it.
Most of you will have something like a Bloomberg Launchpad set up or some equivalent screen. You've got flight prices flashing and so on. You can only look at some subsection of the information that's relevant. You're not looking at it all the time because you're here or you're somewhere else doing something. You can't see the interrelationships with its own history very well. You can't see the interrelationships between them well. Technology processes can do that, and it allows you to process and perceive information that honestly, it's just beyond humans' ability to do at this stage. The second strand I want to touch on is taking some of the techniques we've honed in some big liquid markets that we've been doing for three decades and the ability to apply them to new areas and continue to generate strong alpha, strong outperformance.
Lastly, touching back on the platform, this drives our ability to customize and therefore grow faster and also to grow profitably. There's significant operating leverage you've seen in the business. Just to run through a few examples. This is about TargetRisk, so one of our faster-growing strategies over the last three or four years within the total return segment. What you can see here is its gross exposure over time from 2014 to roughly today, and then each of the different asset classes that it trades. It's trading the most liquid asset classes in the world, so equities, bonds, credit, inflation-linked securities, and you can see the risk going up and down really quite significantly over time.
One of the signals that's making those decisions is looking at intraday information within those asset classes and the interrelationship between them, which again, the information is out there, it's just very hard to process effectively. That is one of the skill sets within the business. I think the other thing that's important is we have the confidence to act off the back of those things. You can see very significant moves up and down in the risk profile. Off the back of those sorts of signals, we really put our money behind that research, and it delivers. On the right, what's the outcome of this? This is for clients if you'd been in since day one, first percentile risk-adjusted performance, first percentile absolute performance, and really handy outperformance of a 60/40 portfolio over that time.
This isn't sort of small scale alpha on single securities. This is big outperformance in the largest and most liquid asset classes in the world. That's an example on the total return side. Turning now to taking some of the techniques and applying them into a broader set of markets. This is AHL Evolution. Taking some classic trend following signals and applying them in a wider set of markets. You can see the growth in the markets that we're applying it to over time. Started out about 100, knocking on the door of 400 today. Every year, we're looking to add new markets that we can apply it to or taking out markets where we think it's less effective. This is on the absolute return side, clearly. Again, first percentile risk-adjusted performance.
Slightly annoyingly, it's fifth on the absolute performance. That's three or four funds running double the risk level as that we do. Again, it's essentially a category killer within that space, so very, very significant outperformance of the SG Trend Index as a sort of benchmark, but also very significant outperformance of equities since inception. It's taking core techniques we've honed in originally sort of FX, equity indices, interest rate markets, and applying them to a much wider set of markets. While alpha does decay, there's still a lot of benefit from taking that expertise into areas where it's novel and where most of the participants aren't aware of or aren't capable of using those techniques. Again, very significant client benefits from doing so. Then lastly, this isn't just about alternative products.
Here's an example from the long-only side. This is Numeric Emerging Markets. Again, taking core models that were initially built on the U.S. equity markets and applying them to emerging markets. Now, emerging markets is obviously an enormous asset class. Some of those markets, it's crazy that it still gets called emerging, given they're some of the biggest economies in the world. Those techniques still work, you know, more than a decade after initially applying them. Again, in big liquid markets, there is still a very significant edge to taking those techniques across different areas. Again, first percentile risk-adjusted, first percentile absolute performance, and very, very significant outperformance of the index over time. Across those three, that is one of our flagship total return, absolute return, long-only products.
It is very, very significant outperformance of the peers or the benchmarks, and it's first percentile. Normally, when you see people stand up at these, they talk about quartiles. We're talking about percentiles because we really think this gives us a huge performance edge, really benefits the clients, which clearly ultimately drives the business. Turning lastly to the platform side, and some of that efficiency that we think we get. So you've got an estimate here of us compared to the wider industry. This is trading volume per person here compared to what we estimate for the wider industry. We think we are 17x as automated or able to process 17x more per person. That's just a measure to give you a feel for the automation and the efficiency of the platform.
I should add that, you know, when we're looking at businesses to buy, we're doing due diligence on managers, we actually often see numbers significantly higher than this. We see businesses that we think we can basically do their whole trading volume in an afternoon without noticing, whereas for them it's their entire platform for the whole year. This is our estimate at the group level. Why does that matter? Two things which I think are very important for shareholders. One is the automation of that platform drives operating leverage. The growth that you've seen over the past five years in the business have also seen significant margin improvement, so 20%-30% on the management fee PBT margin. We've been able to grow and grow very profitably.
Because the platform is automated, we don't need to add significant headcount to cope with that increase in volume. Then secondly, it's also allowed us to access certain types of growth that you wouldn't be able to otherwise, and in particular, the ability to customize business for clients. When a client comes to talk to us about something that they want to do that's bespoke, the answer is basically always yes. The reason is, when we go back and huddle internally, people aren't going, "That's incredibly difficult. I'll need three extra people. I don't know how I'll manage that." It's, "Let me adjust some parameters within the system to allow us to do it." We can write customized business very, very profitably.
We can always figure out a way to do that business with clients, and it's been a very big source of growth over that period again, so going from GBP 40 billion to just over GBP 95 billion and a big outsize source of the growth for the business as a whole. Just before handing over to Rob on the trading side, hopefully that's given you a feel, a bit more of a tangible feel for what it actually is that we do in quant research, but then most importantly, why this is a real competitive advantage for us and why we have a very significant lead within this space. With which, let me hand over to Rob on the trading side.
Thank you, Mark. Good morning. It's my pleasure today to give you an overview of Man's execution. I'll be highlighting the scale and breadth of our trading, how we leverage that to get the best liquidity for our funds, and how that leads to increased capacity and alpha. We start on the alpha journey. As Slavi says, we get our gross alpha, and then we've got some explicit and some implicit costs. Our explicit cost, things like our commissions and brokerage. Implicit costs, our cost of trading, our slippage, and then our prime brokerage and finance. Within that, the biggest cost is that implicit. The role of the trading is to ensure that we keep these costs to the minimum, as I say, and the one that we focus most on is that slippage cost.
In terms of giving you a size of what we're trading. We have over 11,000 instruments, 800 different markets. These can range anything from S&P futures to dry freight shipping routes. The team is made up of execution traders, quant researchers, and technologists, and that team is 56 globally. We're across four trading HUBs, the primary being Hong Kong, London, Boston, and then we have a small presence in Shanghai as well. We're currently 24 hours, five days a week, and depending on how crypto goes, that may get 24/7 one day. Hopefully not soon. In terms of giving you a size of our volume, we actually saw a big increase in trading volume, total number of GBP 7.4 trillion. But you can see spread across the different asset classes, you know, GBP 300 billion, GBP 80 billion of equities, GBP 3 trillion FX.
This makes us a significant player in the market, and therefore, we're able to leverage that scale to negotiate the best liquidity, the tightest spreads, the lowest or some of the lowest commission rates to get that cost down and to keep. I say, it's all about retaining the alpha. I will switch to the live trading system. Okay. This is the live trading system we're now looking at. There's a whole load of different indicators going through on here that we can monitor what's going on while we trade automatically. In the top center, this is just giving us an overview as we trade different markets across the global exchanges. See a few orders popping through, where we're buying some FTSE futures, selling some Australian dollars, buying some German five-year Bund futures.
All these orders are going through and are being traded live now. The top right is giving us an idea of our cost of trading and the volumes. What we'll see, the yellow line is our normal profile for this part of the day. The blue line is the volume we've done today. So far from start of day Asia, our volume today has been slightly higher than normal. The red is our slippage cost. We can see actually it's been so far today a relatively low level of slippage. Volatility's been fairly low in the market, and so we're watching that live, having an idea of how our trading is going as the day progresses. Also, we also monitor all our trading post-trade as well.
Part of the 56 team are focused on analytics, and we look at our post-trade costs. We look if it's in line with our modeling or are our costs in line for each of the markets we trade. What we'll see down the bottom is some of the current orders that we're going through, we can click on a live order. Okay, and I'll see you here. This is where we have made a decision to do a trade, and we're looking at the trade here. The trade has been routed through to our in-house algo, and it's being worked away. The order is to sell 17 Italian government bond futures. I can see, started that order at 9:17.
What we're seeing here is red is the offer, blue is the bid, white is where we place an order, X is a cancel, and then we place an order. What we're seeing is that the algo's looking at the order stack depth of market, trying to decide whether it thinks it will get traded or not. It's broken the order down into smaller pieces, and it's working a way to get that order filled. What we can see so far, we've got, at this point, only one fill that's, and this order's got until roughly probably about an hour or so to complete the order. What we're seeing here is we've got one blue passive fill whereby we've placed an order on the offer, and we've been lifted.
What our research says is by being passive in our execution reduces our costs over time. Oops. Set the top there. If I just select a more active order that we've already completed on the day over here, just to highlight one where we've got a little bit more activity. Okay. This will just give you a clearer picture of when we're placing orders and how we're breaking those orders down. Again, white was place an order, blue was the fill, cancel, moved up. We then reassess, decide to move back down again, and we can see, I see, those blue fills which are passive orders behind. Now, obviously, at times, markets will move against us, and we will aggress to fill that order. But where we can, we will be as patient as possible to get that order complete.
In terms of giving us a headline, this is just giving us an overview of all the orders currently being placed. Orders moving from left to right is their completion fills, and then we can see if we're buying and it's getting cheaper, the order comes down. If it's getting more expensive, it's moving up. It just gives us an overview of being able to see where all those orders are live trading in the market at the moment. Is there some we should be more concerned about? Is there perhaps some that the human traders should be looking to interact and take over? In terms of FX, we trade that slightly differently. Oops. Mouse. So over here, we're looking at the FX. So the FX trades slightly differently. We have up to 12 providers, banks streaming live prices through to us in various sizes.
What we'll see here, I can select a single provider. We're looking at the streaming for cable in GBP 5 million. I've selected one provider. I can see currently this is their bid. This is their offer. What we're seeing is, as we add more and more counterparties, the inside bid-offer spread is getting tighter and tighter. What we actually have in-house, we aggregate these streams, and essentially, when we've made a decision to trade and it gets moved to our execution engine, it's monitoring these live streams. We have a new in-house algo we've built which trades against these. It'll just hit the best stream when it decides to trade. Now, highlighting the sterling here, wrong one. We have currently 57 currency pairs where we're doing this.
There's a few blanks on the screen, dollar Argentina with dollar Chile, which don't open till Latin American time. Also right now, see we've got dollar rubles, because currently banks don't stream that pricing at the moment. This gives us a really rich data source, and as you see, we love collecting data, analyzing data. What we're looking at here, this just goes back to the first Friday of the month. It's the day in the U.S. when non-farm payrolls are released, traditionally a pretty volatile day around the release of the series. What we're seeing is, say, still have sterling in GBP 5 million. These are the streams coming from the banks. What we've got is all the banks listed. This is their width of pricing.
What we're interested in is the dotted blue line because that is the inside bid-offer spread, the spread that we're crossing when we trade. What we'll see is, as we approach 1:30 when the data's released, the banks widen their bid-offer slightly. Obviously, over the data, it goes significantly wider, and then actually volatility stays in there, and as the afternoon progressed, it comes back down again. What you'll notice is the dotted blue line, while widening, doesn't widen anywhere near as wide as the spreads go. What we see is that even during times of short-term volatility, because of that competition and competitive nature, we actually see only a small change in our trading costs. Now, clearly, for this amount of trading and volume, automation is key. This is just looking over the last few years.
We do have very high automation in FX, but have continued to push that forward along with our futures. Actually, a big driver in the last few years has been the big increase in automation of cash equities, where now all of our three major asset classes are running in the 90% of automation. I've gone back to our live order blotter. These are monitoring those various orders that we have. This is where a decision's been made to trade. That trade has been routed through to the electronic blotter. We have a whole series of currencies, commodities, 10-year French government bonds, two-year Italian, Aussie dollar. What we're seeing here is the volumes being traded, how much is remaining. As we said, we break the orders down into small pieces and are very passive in the market.
I think we highlighted that our annual trading volume in FX was about GBP 3 trillion. It's really highlighting how we break that order down to try and be as passive as we can to leave a minimum footprint in the market. Here we can see this is our method of execution. [AX and AEA], these are our two internally developed algos for futures and FX that were built by our execution research team. I can see we've got the FTSE is going through on a bank algo on a TWAP. We, you know, use machine learning to decide which is the best route to trade.
Where we have capabilities in both camps, where we talked about for FX, futures and equities, we have a machine learning algorithm that will decide whether to route the trade to the human trader or whether to route to the electronic trader, and I'll demonstrate how that works. Just click through this one. What I'll do is I'll just pause this midway through and just explain what's happened. This is how the experiment works, and this is a fast-forward of the 10-year treasury note future. What we're seeing is we have a trade comes out, and it has a choice to start with. It's a 50-50 choice. It either goes, on this occasion, to the trading desk, we use a whole series of algos that are provided, or it goes to our electronic trader.
What happens is, over time, we start to see a distribution of the trades it's doing. If I let it run through a bit further. What it does is records the volume and trading. What we're seeing here is, here's the slippage density. What we're seeing is the slippage for the blue is around 0 and the distribution around there, whereas the red is around, essentially what it's worked out to be $20 per lot. What we're seeing is, we stop going from our 50/50, and this is the probability of which route we take, and we start routing more to the blue.
We continue to route more to the blue with the occasional route to the reds to continue to keep a foot in that camp to check something hasn't significantly changed. What we're seeing down here is, by choosing the right route, we will continue to see our slippage saving. We've got this essentially continued machine learning going on of which is the cheapest route and looking to continue doing that. If it would play through, you'll continue to see that distribution building, the saving cost building up, and the how those distribute. This line will stay solid.
We currently go to the extreme we do is 90/10, so that way we get enough information still from the more, I guess, on this occasion, more expensive to make sure that costs are in line. I guess to finish up and give you a sort of overview of where this leads us. Last year, we traded over 7,500 different equities. You can see that I tried to highlight the scale. That scale essentially gives us diversification, which gives us an improved product set for our clients. Over here, we're looking at systematic equities. Over the last four years, we've seen growth in our in our AUM there, but actually the blue is our efficiency, our slippage cost.
What we've seen a 50% reduction over that time in the equities. Essentially this gives us additional capacity. The capacity means that we can grow further. Then over to the right, what we're looking at here is over the last four years, we've roughly doubled our AUM, and what we've seen a 25% reduction in unit cost. This means that we can retain more alpha. Hopefully that's given you a sort of live idea of the, I guess, the insights we have into our system, how we're monitoring our trading costs and live trading as we go. I'll pass over to Darrel.
Good morning. My name's Darrel Yawitch, and I'm the Chief Risk Officer for Investments at Man Group. Risk management at Man is an incredibly important skill, and today I'll explain to you why we think that differentiates us as a business. We have an established risk management process that is an integral part of the investment management process and the engines. Risk management at Man strives to be both the traditional second line of defense, but to be proactive and to influence the investment outcome because we believe that better risk-managed funds are better funds. That's better for clients. That's better for shareholders. Practically, what does that mean? The first point I'd make that differentiates us is risk management is an integral part and embedded within the investment engines. This happens both on the discretionary side of our business and within the systematic side.
The daily, weekly conversations that we have with investment professionals and the researchers, risk management are very much a part of that. This happens also on the systematic side of the business. You heard Slavi's demonstration earlier, where risk management is just built in as one of the considerations when constructing a portfolio. Allow me to tease out maybe two points on the systematic side that I think help to make my life a little bit easier and to sleep easier at night, and there's two things that we embed within our trend following strategies. The first is inbuilt diversification. Okay, diversification is an extremely powerful risk management tool.
When you build a system, not only are we putting maximum position sizes that you can take in any one equity for, as per the previous example, but we give a lot of thought to how you should appropriately diversify that portfolio using a number of different techniques that makes for a better diversified portfolio. It's a very powerful control. Second control within our trend following strategies and some other strategies is volatility scaling. What this means is we size our positions inversely proportional to the volatility. When volatility increases, like the environment we're in at the moment, that means we reduce the size of our positions, and when volatility comes down, we increase the size of our positions. What that means is you track your intended risk exposure more closely over time because you keep scaling towards that.
The second point I'd make is that we have a very open and collaborative culture, and that encourages many beneficial things from an asset management point of view. The first is it avoids groupthink. When you combine an open culture in an environment where you don't have an in-house view, we don't have an in-house view at Man, you get lots of different opinions, and that forces portfolio managers to validate their own decisions and to avoid the groupthink of everybody thinking the same way. The other thing which we saw as we trade, as Rob showed, thousands of instruments across hundreds of strategies. So if we want to know what's going on in the market, one way would be to go externally and see what's happening. But actually, we trade many, many strategies.
At any one point in time, we have a fairly good idea of how different strategies are performing, then because we're trading them. The third thing is, as we've heard quite a bit this morning, we invest to stay at the cutting edge, and there's really two parts within risk management where this applies. One, the obvious one, is within technology. Okay, it's a real pleasure to work in an organization where we have access on a daily basis to cutting-edge tools, technology, systems, and we very much utilize that. Risk management as a skill is no different to many other skills. It requires that constant investment in time, money, technology, and risk management within Man. We very much do that. The other way that we do it as well is through publishing cutting-edge research in industry-leading journals.
We believe that doing research will keep our mind sharp on different risk issues, and they change. They change through time. It helps us to use our intellectual curiosity to solve the complex problems that Laura mentioned earlier. Maybe to make this a little bit more tangible, let me talk you through one recent example of how risk management within Man helped us in the case of the Russian invasion of Ukraine. What we have here on the screen is a plot through time of the exposure we had in four different categories of funds: the absolute return, discretionary long only, systematic long only, and total return funds through time, starting in November 2021 and ending just after sanctions were introduced.
The first and obvious thing to notice is we had a fair amount of exposure back in November, and we pretty much had almost no exposure by the time that the war broke out. How does this happen? I'll come back to those three things just by way of illustration. In an open and collaborative environment, the first time we spoke about Russia was April 2021. That was the first time there was an apparent buildup of troops near the Ukrainian border. Nothing happened for a few months. It was one of the many risk issues that are on our radar that we talk about at any point in time. It gained more attention in November when there was a further buildup of troops, and I think this was the largest buildup of troops since the Crimean War.
At this point in time, we start to see some interesting things happening. We have inbuilt risk management within our systematic side of our business, and the first thing that moved was not only a systematic risk reduction in some of our trend following programs, but they actually wanted to go short. They were the first ones to notice actually there's something going on. We're gonna position ourselves short in Russian, in Russian assets. We continued to discuss this at our risk committees, and what we then did was we started to set up some shock scenarios to say what would happen if Putin was to actually invade Ukraine. At this point in time, it seemed quite unlikely. One of our favorite risk management techniques is to do what we call a historical replay.
We pick different stressed environments over time, and sort of like some people collect stamps, sort of collect these stress scenarios as they happen through time. One we had collected was the Crimean invasion. We collected things like Trump gets elected, Brexit. There's a whole range. There's at least 35 or 40 of these. We started to run the Crimean invasion stress scenario again to see how exposed we were to different assets. While we were doing that, one of the discretionary portfolio managers said, based on everything that was happening, his own assessment was to reduce his exposure to Russia. He too, at 0.3, went short, and he said he has a very negative view. He's gonna go short. We get to this period that you could call war or not war.
Running through January and February was will he invade, won't he invade? Will diplomacy work? Won't diplomacy work? We continued to run our stress tests using the technology infrastructure that we have, and this culminates. The week of the 20th of February was a very interesting week. It was the end of the Beijing Olympics. Not surprisingly, Putin waited until after that to actually invade Ukraine. One of the things we started doing was saying, "Should we start to take action?" The next thing that happened in our systematic side of our business, some of our signals had been a strong signal one way. They were starting to converge and starting to turn signals the other way. So, for example, some of them wanted to short the ruble.
At this point in time, we took a step back and we said, "Does this make sense? Are we not entering an environment where it's gonna become much more difficult to trade these assets? We're very uncertain about the liquidity, and if sanctions are introduced, you won't be able to trade these things at all." What we did in conversation with the systematic side of the business was we said, "Now that our positions are pretty small, it's as good a time as you're gonna find to actually close those positions out." We may land up giving up a bit of alpha on those individual markets, but we trade hundreds of markets across a very diversified portfolio. One market is not really what the trading strategy is about.
Let's rather avoid a problem and close out while positions are small and costs are low, which is what we did, which was the Tuesday before they had Defender of the Fatherland Day. That was the Wednesday. The big green line you see here are our systematic long-only benchmark funds. These are funds that are benchmarked versus an emerging market index, where you're required to hold a certain amount of emerging market exposure. Normally what you do is you might go a little bit underweight versus that exposure, which is what we decided with the portfolio managers to do on that Tuesday. As things started developing Wednesday, Thursday, we actually took the decision, we actually wanna cut completely out the positions because there's a real risk you won't be able to trade Russian assets at all.
On Thursday, Russia invaded, and Friday, sanctions were introduced, by which time we're pretty much outside of the position. I think the example there is to illustrate that open culture, that good use of technology, and the integral part that risk management plays within the process to avoid a problem which otherwise could have been much bigger. I'll spend just a brief time now doing a quick demo of some of the tools we used, some of the shiny tools that we used to analyze that. The first one is a tool that we call Spotlight. I'm doing a very simple example here. What I'm showing you here is the ten-day Russia-Ukraine escalation from March 2014.
This is a historical replay that we started to run on one portfolio. I've just picked one. We could run this, and we do run this every day across all our funds in many different stresses. I'm showing exposure, and I'm showing just the stress. This is when we started running the stress. As I said, we started running it in December. You could see how I'm using the same example, where we were positively exposed from a stress point of view, and then we changed positions and actually went negatively exposed. This is how much money would that portfolio have made or lost if we were to see the same 10 days we saw in the lead-up to the Crimean invasion.
Now, one of the powers of this tool is whenever you run the scenario, you say, "Is it right? Do we have a data error? Where does it come from?" It allows us to drill through in any way into the underlying portfolios. For example, you might wanna look by country. The obvious one was, in this case, we were looking for Russia. Here is a sort. You can sort by different columns, or you can sort by your stresses. Then you can say, well, if I see that actually, I mean, I'll pick the United States. If my biggest exposure positively is coming from the United States, which type of asset class, for example, is that coming from?
I can drill through there and sort it again and see the biggest positive exposure is futures on a commodity. I could drill in again and find exactly which asset it is. As I'm doing that, it is. I'll just plot it over here. It is showing through time how those have actually evolved and changed. You can see at any point in time for any of these analytics which strategy it comes from, which industry, which country, which sector, which instrument, and how that's evolved through time. The tool enables us to very quickly analyze and make decisions on that stress test information, and this is a tool that we've used during that process. The second tool I wanted to share with you is what does that portfolio look like today.
These tools all are part of what we call our proprietary risk portal, and this is one of the tools we use, which is called Prism. Now, Prism is live risk and live P&L. What you're looking at today is today's P&L. Okay, I came in at 7 just to check, well, 8:00 A.M. just to check this was all up and running 'cause it's a live demo. It was last calculated at 6:46 A.M. It recalculates periodically through the day. What we're looking at here is our frontier fund, and I have just selected five fields. There's normally more than five. The daily P&L is the first column split up by the different sectors that we trade. Today, we're up three basis points, and four of those are coming from commodities.
Over the last week, I could change week to month at the top, the. You can pick month to date. But the equity curve here saying over the last week, we're actually down nine basis points. Whilst we've made money in commodities, we've lost money in fixed income. How has our risk been changing? This column is showing our forecast annual volatility. So at the portfolio level, it's 12.3%. This fund should track around 15%, so it's slightly below its target, and most of it is coming in the commodities area. You can see again, I can do the usual sort on any column. What are our exposures? What is our interest rate sensitivity? What is our credit sensitivity?
Again, you can interact with this data and you say, "Well, maybe I wanna see that by strategy." This particular fund trades about eight strategies. When I click the button there, it recalculated. You see a few things. The next recalc was actually at 8:45 A.M. The risk information is as at 8:45 A.M. It's recalculated the P&L. It's recalculated the risk as well. It's calculating on the fly and allows us to drill in. The purpose of this demonstration, and this is a tool at any point in time, if we wanted to know where's our risk, where's our P&L, where's our exposure, how has it changed, this is a tool that enables us to do that.
Just to wrap up, I've really tried to illustrate in the brief 50 minutes we've had that risk management is a skill that differentiates us as a business in three ways. 1, because it's actually an integral part of the process. It's not an afterthought. Better risk-managed funds are better funds. 2, we have a very open and collaborative culture where risk management thrives within that culture, but it extends to all parts of the business, whether it's on the PM side, the research side. We continually invest in our technology platform and in our thinking to stay at the cutting edge. Thanks for listening, and I think now we go back to the previous room for coffee.
Great. Thank you. Good morning. I hope you had a good session so far. My name is Steven Desmyter. I co-head sales and marketing at Man Group. This morning in my session I'll discuss frankly, just three topics. First of all, who are our clients? Second, what is our sales approach and sales process? Then importantly, what kind of growth potential do we see going forward? Who are our clients today? As you can see, from the slide, we are in every market globally that has large and sophisticated pools of capital. We're obviously deepest in EMEA. That's our home market. We've got good penetration there, but really importantly, we've had good growth really for quite a few many years now across every region equally.
Now, our client base today is largely institutional, so well over 80%, and that has been the case for almost a decade now, so it's been quite a while. Now, importantly, we do have a large number of sophisticated intermediaries with whom we have, you know, strong relationships. I actually think there's quite a good deal of growth opportunities. You know, Eric and I see good growth opportunities there as well, especially in U.S. wealth and European wealth. Then lastly, it's been a focus for a while, and it's gonna remain a focus to be really the number one liquid alternatives and solution provider industry, especially with a focus on quant. Now, the institutional focus I mentioned has really shaped our client base.
I hope you can see all the way to the back, but on the left-hand side, that just shows all our clients split across AUM. We have over 600 institutional clients today, and you can see there's you know, very broad variety of them. On the right-hand side, you see the same but for intermediaries, and as I mentioned, when I say intermediaries, I'm really talking about private banks, wealth managers, and other distributors into retail. Importantly, we really cover third-party distribution, so we don't sell directly to the end customer. Now, Luke mentions this often to us internally. Obviously, these aren't just nameless corporations. We manage assets for well over 100 million pension beneficiaries, as well as, of course, some of the largest global wealth funds and banks.
Now, clearly growth is gonna come from both sides, but we don't really expect the shape to change materially, and we really see a lot of upside even with the large existing investors that we have. We're only a small share of their external wallet, but I'll get more into that a bit later. Now, what is our sales approach? Maybe most importantly, we have one sales force for the whole company, so we don't have any specific sales force focused on a subset or sub-strategies. It's really important to us that every investor has one client point of contact, and that person is then the quarterback, so to speak, for the whole organization. In practice, we have over 240 people in sales and marketing, and that does not include what we call CPMs or client portfolio management. These are effectively product specialists.
Now, these product specialists are crucial. As you can imagine, we have a generous sales model, and we really lean on them for technical expertise as well as, you know, many other people in the firm. The approach here is the following. We like to have our salespeople be as close as possible to the clients, and I'll get more into detail on that later, but it means we have 16 local offices. Often these are just sales reps, but the idea is the salespeople are senior. They spend time with the clients and then in the kinda headquarters, the local regional headquarters, so New York, London, and Hong Kong, we have support staff including client services, marketing, and so forth, and that allows for scale, of course.
Now, to give you the idea of the typical profile of our senior salespeople, these are people with a lot of tenure at the firm. Many of our regional sales heads have been at the firm 15, 20 years at Man and, of course, many more years in the industry, and that's really the typical profile. It's not an exception. That gives quite a few advantages. It means that you have great stability with the investor base. The top investors that you'll see examples of also with Eric are complex. They have many points of contact. We often have to cover 15, 20, sometimes 30, 40 people at those organizations. Being an expert of the client is really their first job. Then, of course, they lean on the rest of the organization for technical support.
Even those people tend to have high technical expertise. Now, it doesn't mean we don't bring in new talent. We definitely hire a lot of juniors in to bring them up through the ranks, as well as hiring senior people where we see good growth opportunities. Now, I won't go into too much detail on this, but sales really is a thorough and repeatable process. There are kinda four elements to it, all of which happen continuously, but just on a high level, first of all, salespeople obviously engage with the clients, understand their needs and desires. Once those are understood, we bring in the investment engines, and here we obviously present the strategies that we think will fit the needs. Once we have them engaged, we look for seamless execution, and then maybe the fourth step is the most important one.
It's about client services. Here, the focus is on cross-selling and upselling, as well as client retention. I think the best analogy is if you've seen some of the presentations this morning, I hope you've seen there's a lot of rigor applied to technology, to investment approach, and we really have the same high standards that we wanna apply to our sales process. Now, what is our sales or growth potential? This is maybe one of my favorite data to look at when we look at what we've achieved the last couple of years, as well as having a good insight of what the future may bring. Hopefully, you can see this clear in the back. Colors are. We like blue at Man, as you know.
On the left-hand side, it shows the growth sales per year for the last couple of years. At the bottom side here. What you see is more business that we've done with existing clients, and a significant amount of our flows over the last couple of years and going forward has come from existing clients. The bottom line is effectively top-ups. These are increasing in existing mandates, existing strategies with existing clients. The second part of the bar shows new mandates with existing clients, and only the smaller part, I should say, on the top is mandates new business with new clients. Importantly, not to have too much, you know, too much detail to it, but the reality is we can see easily these assets double for existing clients in the future.
The reason is we're on average only a small percentage of their wallet. Even for the very largest clients, it's really low single-digit percentage of their external allocations. Now, as a result, on the right what you can see, and this is a result of focusing on large institutions like whom we can do multiple pieces of business with, you'll see that we've not just increased the number of strategies, but also the average ticket size that we have from them. Hope you can see this on the bottom on the right, you see the top 50. In 2014 we did on average just under 4 strategies per client. Today it's just under 5.
This is important, the average ticket size has gone up by 50% and as a result, the total AUM of our top 50 clients has effectively doubled over that period. Now, net sales is a key target for us, for everybody in sales, for everybody at the firm, and as a result of the efforts, obviously not just of sales, but of the whole organization, they have trended upward nicely. Again, at the top you can see the gray shaded area, you see that trend up nicely, upwards there. In the bottom the bottom chart, what it shows is effectively our net flows globally against the industry. We've consistently outperformed the industry on an asset-weighted basis. Again, hopefully showing the quality of the investment strategies and our overall process.
This is a relatively simplistic point, but it's an important one, and it's hopefully a straightforward one. This is showing the gap that you'll have between traditional assets and the required returns, and in this case the example is the 130 U.S. state retirement schemes. If you just take U.S. Treasuries and U.S. equities on the left-hand side and apply 60/40 portfolio and put that against the required liabilities, you easily have a gap anywhere between 3%-6%. Now this is, I hope, common knowledge, but it's a very important one. Obviously it's clear that traditional assets won't do it. You'll have to have really creative solutions. Now, creative solutions, this is where we try to come in, and our solutions obviously require a good breadth of products.
Our offering has expanded more and more. Whenever I speak to clients, and I know it's the case for many of my colleagues, it's clear we are more relevant than ever. I've been at the company for two decades. As many of the people in the room have also long tenure, and you can just tell the conversations you have are more productive, they're more interactive on a much more broader variety of assets. I think it's a two-way situation. It's not just from our end, also the investor base. This is a trend I've seen for quite a few years now. The investor base are looking for partners. They want fewer counterparties. They want people that have a larger menu with whom they can do a broader variety of things with.
This is a focus for the firm. I believe Luke's gonna touch on that, in his session later as well, and then Eric will also explain to you how we actually deliver these solutions for our clients. This is kind of the key thought process that we have in our sales approach. We probably cover about 14,000. I looked at it yesterday. We have 14,000 accounts that we cover at the firm, a bit more than 14,000. 94 different countries. The reality is, the key part is that we disproportionately focus on large existing and prospective investors that we believe we can do, you know, large and multiple things with. Now the approach that we use is really one of segmentation. Let me run you through that. The bottom 65, these are 65 what we call core accounts.
These are existing clients. We have at least $250 million with them, at least 2 strategies. In practice it's many, many strategies more, and actually quite a few are more assets than $250 million. Often it's multiple billions. These 65 accounts are not just key in terms of servicing, but also key in terms of prospective opportunities. We also have very large pipelines with them, and as you'll see in the examples later, whenever we have new strategies or new capabilities that come on board, either new hires or new quant strategies that we develop, there is a high interest from them. They have a lot of interest to see what we do and wanna do more with us. Importantly, like I said, our share of wallet is, frankly immaterial at this point, especially as our offering grows. You have upgrades.
180 names currently. These are, again, existing clients. They either don't have quite $250 million with us, when in practice it's often way more, but they only have one strategy. Again, here the real opportunity is to add strategies with them. To me, maybe the most important part is what we call targets. We currently have 120 names. That's a selection. That's a subjective list that we have put together. These are not clients, these are prospects. The best way I think of it is we treat them as if they were clients. These are really long-standing relationships. These are relationships we're on the cusp of doing business with. Often, we've covered them for many, many years. We might have done business with them in the past.
Just for context, we probably expect to do some business with them in the next year or two. It sounds straightforward, but in practice it's a lot of allocation of resources and efforts on names that you actually don't have short-term opportunity with. They've given me some examples. We are on the academic advisory board of some of these investors, even though they're not clients. We invite them to, you know, share IP with. We spend time at their offices to help them with their asset liability management. You might have heard of our kind of flagship conference that we have once a year in Oxford. That's a key event for us. It's a day and a half. We again share IP, go through a lot of different research opportunities there.
Probably only two-thirds of the attendees there are existing investors, even though it's very exclusive and the investor base really sends CIOs or very senior people. Easily a third of those attendees for many years now have been prospects. Of course, they then can mingle with our existing client base. The great thing for me I find is that you can then easily track, you know, have those efforts paid off. We've been able to track, we've been running it for eight years now, eight, nine years, I believe. We've been able to track clearly to see the engagement increase and improve and eventually get them obviously into the other cohorts of upgrades and core. Importantly, that universe can expand.
There literally are thousands, if not tens of thousands, of potential names that we could end up in this kinda target focus list that we have, especially as our, you know, solutions and/or offering expands. With that, I will hand over to Eric, who will talk to you more about offering these solutions to our clients. Thank you.
Thanks, Steve. Good to see everyone. Thanks for your time. I am guessing I'm not gonna be the first person that's ever talked to you about solutions in asset management. It seems to be one of the most overused kind of buzzwords that's out there. We've got a decent number of reasons I think that we've got credibility here, and I'll take you through those over the next 15 minutes or so. What's critical is that this really underpins our growth strategy as Steve's just described it. As I say, the client strategy's really built on developing the deep relationships with sophisticated investors, both in the institutional but also on the third party side. We're kind of, our DNA is linked to this because we bring partnership between Man and our clients.
In a way, the market's coming to us for a number of reasons. Firstly, as Steve said, a number of our clients are wanting to do more with fewer providers. They're looking for deep benches of strategies and other partnerships that people like us can bring to them. Secondly, we were debating this. Are clients' problems becoming harder or they're coming to us for more of them? I actually think that the problem's the same. Clients are realizing that some of the stuff that's simple, they can internalize, and then some of the stuff that's more complicated, they want partnerships with.
For firms like us that have high alpha and that have a broad swath of strategies that we can offer, the investors are coming to talk to us about how we might help the complex part of their asset allocation. As we think about it, just a point of definition, solutions to us means anything that's tailored to a specific client, which may be a very bespoke investment list for a simple long only mandate, or it might be something that's much more complex on the hedge fund solution side. As it says here, from a client's perspective, I think we can offer a number of things. We've got investment quality that comes from a breadth of different strategies and different mandates. We can offer structural flexibility, so cash efficiency, tailored fees, performance fee netting, structuring for specific jurisdictions.
You're sort of limited by your imagination sometimes when you come to Man for that. We offer true partnerships. As Steve said, we offer the Oxford Conference for people to come and talk to us about IP and about non-investment things. We offer education sessions. We offer tailored modeling for them. We offer portfolio construction. All of that needs sophisticated engagement between our team on the sales side and sophisticated people on the client side. What you'll see at the bottom here is the good news is it's not just one way, so Man gets a lot out of these relationships as well. Firstly, there's a number of strategies that we develop that might be too niche or too specific for us to be able to offer as a standalone product.
By packaging them together in these solutions for investors where they may fit better with a broader swath of strategies, we're able to give clients access to things that are built both on the quant side and on the discretionary side of our business. Some of our best ideas come from our clients, so we've got hundreds of people that sit in investments coming up with new ideas all of the time. Every once in a while, a client will come to us and say, "We think this is a good idea." We'll build a solution around that for that specific client, and then potentially be able to go on and scale that and turn that into a flagship product that we offer more broadly.
Finally, actually, Steve showed this on the slide before, the solutions that are built on partnerships where investors have some ability to tailor, tinker, make the asset allocation fit them, lead to much, much stickier assets. We see the redemption rate on those dropping significantly, which all benefits the net sales objective of the firm. What's our edge? Why do we think that we're good at this? To be a credible solutions provider, I think you need the following, right? To be able to offer the investment alpha piece, you need the broad range of products and strategies, both on long only and hedge fund across multi-assets. We can credibly go and talk to investors about helping across many different parts of their portfolio.
To be able to offer the structural alpha piece, you need non-investment people to be able to engage with those clients. These aren't just operational people that sit in the background somewhere. These are people that are front and center to our client offering, being able to work with them to figure out jurisdiction and tax and structuring and portfolio construction, all the non-investment pieces that really add value into those clients. Broader than that, you need this cultural DNA across the firm to want to sit and listen and not just flog products into the client base. Finally, on the right-hand side, and I think you've heard about this before, all of this sits on a central operating platform underpinned by technology and data. There's a single version of the truth. Clients see the same data that we see.
It allows us then to bring lots of different strategies across content engines into one place without having to replicate things across multiple different systems. It gives us scale and benefit. The critical bit is that we continue to invest in all of these different pieces as well. Going to a client today and saying, "Look, we've got all these different things," is great, but the problems in two years' or three years' time is gonna be different. We're gonna need to continue evolving and innovating to be able to offer all of these different pieces. I thought it might be interesting to try and illuminate this with a couple of examples. The first, and to give you context here, this isn't all just about U.S. and European pensions. The first example's with an Asian sophisticated third-party distribution platform.
In that market, they are attuned to investing in multi-asset, and they're attuned to investing in specific products. They need unit trusts in the market that they sit in. We've been working with the client and talking about what Man does in the multi-asset side, and they didn't want another me-too product to go into the market. They asked us to build some combination. I'm sure you're all aware of TargetRisk, which is our risk parity strategy. Trying to do something that mitigates the drawdowns there. We worked with them and built a solution that combines TargetRisk and Trend, two things that we're pretty good at.
The final stipulation, which is music to our ears, was they came to us and said, "This needs to be able to grow significantly because we're pretty good at distribution, and we don't want to have to stop this at some point in the future." We were quite happy to be able to solve that problem. That gives you the investment and the structural pieces. From the partnership side of things, we have a team out in that market that went and spent a lot of time with their distributors, with the financial advisors that they're selling into, to help educate them, because it's a fairly complex product versus 60/40, and make sure that they can go and tell the story into the markets that they're in. That's all been pretty successful.
We launched that three years ago, and the product range is up to about $4 billion now. What we've done is launched 3 products initially, and another 3 have just come to market recently. We feel pretty good about that engagement and growing in that market. The second, it's a bit blurry there, is a North American pension fund. Right at the start, December 2015, actually, we went and spoke to them about AHL. They were looking at kind of crisis risk solutions and trying to figure out how to protect drawdowns in the long-only portfolio. We traveled a long way to go and see them pitch AHL, and they decided not to do anything at that point in time.
The sales guy feeling pretty sore about that, sat in the public session of the board and heard the trustees talking about the concerns that they had about investing in hedge funds and embarking in that crisis risk piece, because particularly around transparency and around control of assets. Salesperson's ears prick up, and that led to eight months of discussion with the team at FRM around how to build a hedge fund allocation and how to do diligence on managers and how structurally to set things up to mitigate those issues around control and transparency. That led to an engagement in August 2016 to help them build a hedge fund portfolio. You can see over time, that's both steadily grown, and the engagement has increased as we've gone through.
What's pretty cool is in May last year, after six years initially of pitching them AHL, we managed to get them convinced to invest in AHL. It became not just a third-party manager platform that we built for them, but they're also investing and looking at more and more internal projects as well. This gives you the sense that we need high EQ and high IQ salespeople sitting in the firm to be able to sit in those trustees meetings and actually listen to what they're saying and translate it into a way that we can help these clients. It's not an immediate solution. It's not flogging a product on day one. It's much more, how do we build a long-term partnership with them to turn that into a big relationship for them and a big relationship for the firm.
If we take a look at the solutions business today, about two-thirds of our assets are in solutions. Like I say, that spans from customized mandates on the long-only side, which might have a specific restricted list, all the way to the hedge fund solutions. To be clear, though, we're only able to offer that because of that single operating platform. If we didn't have that and weren't able to bring everything into one place, operationally, we wouldn't be able to deliver these solutions to our clients. The technology and the platform that sits under these solutions is critical to us being able to go and engage clients on pretty much everything that we do across the firm. The institutional solutions are the really complex hedge fund pieces that we put together and allow clients to tailor portfolios to suit their investment needs.
That's growing quickly. We expect it to continue to grow, which again, is quite important for us because that's a part of an investor's portfolio, unlike their traditional long-only stuff, that's quite difficult to replicate in-house for those more sophisticated investors. We're spending a lot of time thinking about how we can partner on the more complex side of their business. And then as I mentioned before, not to belabor the point, the redemption rate on those institutional solutions is very low. You'd expect that because clients feel that they're in the room, building these things with us, so it's not just judged on performance. If I have a crack at summarizing Steve and my sessions.
For a number of different reasons, hopefully they come through, the target investors that we're going after, big sophisticated institutions and third-party platforms are tending to do more with fewer. We're well-placed to benefit from that because of the breadth of strategies, the people that we have working for us, and the operating platform that we have. While we're not trying to disintermediate our investors, there's a number of things they can't do. They tell us they can't do them, and so we're spending all of our time really thinking about that part of the market. Things that focus on alternatives, on execution capabilities, on risk management, and on technology. They're bits that legitimately we can go and say we have an edge on, and the client really shouldn't be bothering thinking about those parts of things because they're gonna get a better solution coming with us.
We can't rest easy. We've got to keep expanding both strategies but also the structural piece as well, and you'll have heard throughout all of these sessions. I think the investment's there to do that. That gives us, on the sales side, a pretty interesting toolbox to be able to go to those investors with. Finally, if we leave you with one message, it's really that we have a strong growth potential with our existing clients. We're not just reliant on new clients or new channels or new targets. We often say that our existing clients are our best prospects and we can grow into those. That, again, is only realized by the quality of the platform that we have and the breadth of the strategies that we have. That was a session on clients and solutions.
A number of these solutions now are coming with an ESG flavor, and a lot of our investors are caring more and more about ESG and about impact. As Robyn will take you through now, that's a place of great importance actually to the firm and growing day by day. Thank you.
Morning, everyone. The aim of this section is to talk about how Man thinks about responsible investing and why we believe we're well-placed to navigate the challenges and the sheer complexity of the area. It is complex, and it is multidimensional. The data is inconsistent, it's incomplete, it's conflicting. There's not a single taxonomy. We have countries and governments and regulators all approaching this in a different way. Clients have very different views on this and are at a very different stage of their journey, depending upon where they are.
Even the separation that happens sort of weirdly between the environmental and the social and the governance piece is in and itself belies the significant intersection, the overlap of those particular factors. All to say, there's not a single interpretation. There is certainly not a single way to approach it. Against this backdrop, we see, as Eric was saying, we're seeing clients talking to us about this, and we're seeing flows into these strategies. We are here to help clients navigate it. It's very simple that in many ways, we are just here to listen, to understand, and deploy what we do best, which is quant and data and tech and thought leadership and solution driving. That navigation is not about being evangelical. It's not about saying this is the right thing to do from Man's perspective.
It's about finding solutions for our clients and their needs. Scorecard. Let me give you a scorecard, at least. At the end of 2021, you can see that circa a third of our AUM has ESG factors demonstrably integrated into the investment process. We converted or launched 20 funds since the beginning of 2021, and we will continue to introduce more strategies in this space. As you're probably aware, we are also signatories to the Net Zero Asset Managers initiative, which commits us to achieving net zero emissions within our investment portfolios by 2050, and we'll be publishing targets of our journey to to that number, and our targets to 2030, I think, in the next couple of weeks or so. Lastly, it's picking on another piece about the highest standards and our involvement in creating and engaging with the market.
We are engaged with, among others, the UNPRI, and we have an A+ rating in that space. Let's talk about the challenge of data. I'm not sure you can see that very well. It's okay? It's quite pale. We believe our greatest strengths where we have depth of experience and capability is in understanding datasets, which is critical given just the vast explosion of this data hitting the market. This is moving from, I guess, the broad ESG rankings into more specialized ESG datasets. As data sources expand, it's important to try and connect the dots between what vendors actually have to say. To do this, there is a lot that needs to happen below that waterline, and it did not escape me that I am using an iceberg in talking about responsible investing.
That was not as deliberate as it might have seemed. Below the waterline, you need to have a robust data platform, which in and of itself requires a tech stack and a mapping of data to companies. You need to understand what the data is actually representing. You need to be able to investigate data integrity, and you need to be able to look ahead. Of course, you need to also link the data sources to each other. The goal is to do all of that below the waterline so that you can in fact create alpha and identify the risk signals for above the waterline. It's not only a data story, it's an expertise story. It's about not only bringing that data in-house, but being able to interpret it and find the signals through it.
I mentioned that slightly weird conflation of ES and G, and that can make it difficult to understand and interpret the data you're getting through the door. The chart on the bottom, which I hope you can see, tries to demonstrate the inconsistency of ratings between agencies across four different issuers. I'll take Tesla, 'cause it's easy to do, I guess, which the MSCI rates as very highly for environment. It does so because its view is that Tesla's products will positively impact environmental change. Seems sensible. Whereas the FTSE rates Tesla very low for environment, based on factors such as the environmental impact of mining, of the core minerals needed for battery use, which we know all mining techniques have an enormous environmental impact.
We're looking to mine more data sets to enrich the analysis, to aim to integrate these signals seamlessly, and whether that's things like patents, pending for certain technologies, or whether it's Glassdoor employee rankings, it's about bringing all of the data sets together, joining them up, and then being able to interpret them with expertise. We are thoroughly able to do just that. We analyze data. It's what we've been doing for many, many years, and we create intelligent, responsible investing solutions across our entire business, and that's through our depth of knowledge, experience, and quant techniques. I wanted to explain just how organizationally we're structured. I talked a little bit about data, but the data science capability that we have is completely centralized, and that means that when we're onboarding ESG data sets, we onboard them once, and we onboard them for the entire firm.
That allows us to both distribute that data in raw format or indeed in a curated format. That allows every single part of our business to look at that data and to decide how it can inform and benefit the investment process, and indeed, how we might be developing new solutions for clients. We also established, in the middle of the slide, a dedicated responsible investing team last year. We want to deploy our experience and the expertise we are building across the entire organization to leverage and enable the creation of product and solutions.
The thematic research piece of that, which I'll talk a little bit on later, is where we're taking traditional research as we sit embedded across each one of our businesses, and we think about supplementing that with the expertise needed in this new space, something very different to that we've seen before. Of course, this plays to another core strength of the firm, which is collaboration. We do, and we fundamentally believe that the strength of this firm is built on the back of the entire organization working together. When you're listening to Eric and Steve, you're hearing about how the firm comes together to drive solutions. Even more was this the case in responsible investing. Then lastly, we use our expertise and significant tech and data science resource in creating reporting solutions. These are proprietary in-house RI technologies.
We use them internally, but we're also demonstrating and showing them externally to our clients, and they are desperate for this type of tech, and it doesn't exist very easily. It allows them to think about the impact of their portfolios more broadly and beyond that of just the Man Group investments. Eric talked about the edge. I'll do it again. Where's our edge? Why do we think we got an edge here? We have a circle, and of course, I'm going to now mess it up and talk about it in different orders, but go with me. Our quant capabilities, combined with that central expertise I've talked about, and then combined with common infrastructure, provides us with a unique position to innovate.
We're dynamic in our ability, top right, to integrate RI across investment strategies, and we partner it with huge investment and a robust, strong governance structure. Bottom left, our strength in deep fundamental analysis allows us to put nuanced ESG into perspective. It helps us give ESG data meaning in our investment processes. We're also active stewards of our clients' capital, and we aim to vote 100% of all proposals in AGMs and interact with management to try and affect change on a one-on-one basis. As I talked about before, we have focused on growing a strong climate expertise, and that expertise is there to measure portfolio risk in advance of that of climate and in prediction of what the financial impact of climate change will be. I'll talk about that a little bit further in example.
That's irrespective of whether we're dealing in alternatives, in multi-manager solutions, or in private markets. Let me try and bring this to life a bit. I'll do my best. St. James's Place was repositioning its $18 billion global equity fund, and it wanted to pivot it to be ESG and climate aligned. After an exhausting, and exhaustive RFP process lasting more than a year, certainly more, probably closer to two, SJP selected three new managers, of which Man Group is one. They wanted to align the mandate with the Paris Agreement and materially reduce its carbon intensity to below 50% of the benchmark.
In its new form, the strategy seeks to insulate the portfolio from risks of climate change, capitalize on the alpha opportunities from transition to cleaner energy, and to maintain positive exposure to ESG and fundamental investing, ESG fundamental investing concepts. In order to meet that mandate, Man developed bespoke, cutting-edge proprietary models. We also developed, I wish there was a nattier way of saying this, but there isn't, the Man AI Climate Change Computation System, which we neatly called MACCS, but I had to actually spell it out here. That is a climate projection system that aggregates existing climate models to build a best-in-class higher resolution model. That model enables us to incorporate various climate change probabilities and then forecast temperature change and other impacts, climate impacts, within extremely localized areas, somewhere around 10 sq km , that sort of small area.
All of the climate modeling is bias adjusted and back tested to 40 years of data. The goal is to determine how a company's physical presence in a particular region might be affected by climate change, and that helps us measure the physical risks of climate change on companies at a really, really granular level. Now, this is testament to all of that stuff I've been talking about. It was a long sales process, it was engagement with client, it deployed what we do well in tech and quant and data modeling. It is one of the largest ESG funds in the world. One of the other reasons that SJP in particular engaged with us was 'cause they were interested by the work we also do on engagement and stewardship. Let me talk you through a little bit in these slides.
The numbers speak for themselves probably, but it would not be possible without quant and data and tech to literally vote 7,409 at meetings, or to think about 7,200 resolutions, or to engage with 384 companies, pulling some numbers off there. What do we focus on? Environmental reporting, board tenure, executive comp, audit tenure, and shareholder protections. We do so in this centralized stewardship function that we have in the RI team, because we are able to provide information and recommendations across all of our strategies into the firm. We also are listening into and through the discretionary businesses that we have to see where their preferences and what their recommendations are through their fundamental analysis too.
In other words, our stewardship function is driven through data, but has a top-down and a bottom-up view of how we need to engage, and that engagement is very much in the format of trying to persuade and join together. Which is why that little yellow button over there on the left-hand side is a very live current example of us alongside HSBC and Amundi co-filing three resolutions proposing that J-POWER have a more detailed business plan and remuneration policies connected with meeting their goal of 1.5% in line with the Paris Agreement. Very live, happening right this minute. Last piece of the puzzle. I talked about how we partner our clients. We've heard from Steve and Eric that commitment, and how we develop solutions alongside them.
Our efforts extend well beyond that to promoting and raising awareness of RI, not just within the firm, not just with our clients, but across the industry more broadly. That also includes sitting down with the regulators in multiple jurisdictions and informing them on what we think can work and what the new regulatory regimes can feel like. At Man, I think we believe we thrive on solving complex and difficult problems. I think we can all agree that ESG is definitely one of those. It's not going to get any simpler anytime soon. We are really well-positioned to navigate this. We have much better understanding of these markets and much better capabilities in navigating datasets than many of our competitors on the street. We enjoy rising to the challenge.
We actually enjoy finding the solutions, and that is, I think, what we're well placed to do. With that, I'll hand over to Luke to talk a little bit more about innovation.
Thank you.
All right, cool. Remarkably, we are actually sticking to time. There's a big sign here saying, "Keep on your schedule." Cool. I can witter away a bit, which is my joy. The thing about innovation. What we've talked about through here is the technology platform. We've talked about the need of clients. What we have to do is continue to generate new content, whether it gets delivered in a fund in the traditional way or through a solution as is more typical here, it's all about producing content. The bit of having Slavi go earlier through that, what was at one level simplified but actually not simplified example of how in a quant process you develop a new model.
It's quite hard to show innovation at the model level, but actually that's where it's most important in terms of what we do. When I got here, AHL had, depending who you ask, 2 or 4 models. We still run those models today. We run them on a lot more markets than we used to run it, 'cause that's been one of our forms of innovation, is broadening what we traded on. The alpha in those models, in each individual market has decayed. The alpha in everything decays over time. If you are gonna maintain alpha to deliver to clients, and if you're gonna grow it, which is what we do, you have to keep innovating all the way through. One of the parts is that bit of going to 800+ markets where back then it might have been 50, maybe.
Can't remember what number we used to talk about. Similarly, you know, if we didn't do any new research in AHL or in Numeric, we wouldn't need those 100 plus researchers that are doing the quant research. To maintain the existing stuff is maybe 10% of people's time. The more you invest in the platform, the less time people spend on existing stuff. Fiddling with an existing model is not a good use of a researcher's time. The models all have built into them the process to say if something is changing, something's different. I can't tell you how frequently I get asked the thing about, "but what happens if the market's different?" Well, the truth is markets are never really different.
If the model starts behaving in a way that is not in the distribution of ways we thought it would happen, that is automatically built, the risk reduction followed by going to cash and turning the model off is built into the models automatically. You don't spend lots of time monitoring the existing models. They're monitoring themselves. The research people are developing new models. Today, there are 400 plus models in AHL in active use. All right? There is a lot of organic innovation of individual things. Trend, which was what those original four models were all about, still very important to us. In a period like we've had in the last six-12 months, trend is fantastic, but it's all of the other things which are consciously orthogonal to trend that improve the diversification within AHL.
It's the same at Numeric. We're an order of magnitude more models in operation today. You know, when we first bought Numeric, they were very proud of their technology and all the way through. We kept saying, "Wait till you get here and maybe you'll think about it." Three months in, they went, "Can we rewrite our whole system into the same Python environment that you have for AHL?" Which we did, and it meant they can develop models much, much more quickly. Before they had one standard set applied in different geographies in a totally standard way. Today, to Robyn's example, you can choose any combination of different models, a lot more models coming through. This then gets delivered either through existing content and just having more in it, or through doing new strategies which we hope to become independent funds.
You know, we don't really mind whether people buy things combined or they buy them in separate bits, but in the end, you have to have certain separate strategies in order that you can have the conversation with clients. There's an interesting thing on this chart with the 16 and 4. Last year we hired 4 teams. People you may or may not know out there. Actually, one of them was in pure quant from Weber. We hired 4 teams to do new things. We launched 16 strategies that we seeded. 16 against 4, what's that about? That gives you a sense about the amount of the content production that is really organic, internally produced using existing teams against the stuff that we either hire or buy. We didn't buy anything last year. No, that was not a secret.
You know, that would come in that right-hand column, but we can constantly innovate things internally, and that is what really builds up the volume of content. What that does is let us grow with clients. It lets us generate more alpha, which is what we're getting paid out of. Lets us grow our assets, lets us grow our profitability. It also diversifies what we do. The more different sources of that alpha we have, the more stable the performance is, the more stable the alpha is, and for those of you who care, the more stable the fees are and the performance fees. Nobody here cares about that one.
If that's all about innovation in sort of the existing things we do, we look constantly for where there may be extra white space where we can expand. Somebody asked me earlier. We'll never have two teams doing the same thing. We won't have two, I don't know, whatever value strategies in Japanese equities. We've got one. They're really very good at it. We don't need a second one. We might have quant and discretionary look at the same assets. In fact, we do sort of everywhere. We might have a value person and a growth person, but we won't have two of the same things. We're always then looking for one of the criteria is how do we innovate in new places that we haven't done things before. That can be in asset classes. Thought of when I first did this a few years ago.
I am getting gray hair, I'm told by my wife. When I did this, you know, we talked about the fact that we had long risk asset strategies in our long equity stuff. We obviously had the hedge funds and trend that would do well in a bad environment. We didn't really have the sort of income-y type of strategies. We've done a lot to innovate in credit. I'll come to it in a minute, but credit and real estate is about having more income strategies. It creates extra dimension of diversification in the firm. That's been a key area for us, but we've got further to go. You know, we pushed to do more in the U.S., more in Asia.
Asia would have been a more exciting thing to talk about if it wasn't for the last two years, and you can't get there, but we would like to do more in Asia if we could get there. It's interesting, you know, we are avowed institutional specialists. We talked about, you know, it's 80% plus of our business is institutional, but you saw one of Steve or I can't remember. Somebody's solution they talked about was a retail distribution solution. Actually, SJP's. There were two of the things we've talked about were retail. We're quite happy to expand into that area as long as we do it with the people where it's at the scales we understand, at the quality we understand, and where we can use our, you know, it doesn't get in the way of our institutional business.
As we build out the credit things, the carry things we're doing, it opens up the ability to do more in the insurance space, which has historically been a sort of pointless exercise for us because they only want carry stuff in there. Let's get to a few examples of how we innovate and how you can see it coming through. TargetRisk, we've talked a lot about. We should talk a lot about because it's now nigh on $20 billion of our assets, right? That's from zero, six years ago, from nearly zero, four years ago. The first three years of TargetRisk, we raised literally not a penny, and it used to be a source of great entertainment, unless you were on the receiving end of it, of grief internally about why weren't we raising money?
Once people understood, it's really taken off. The thing of it is, what TargetRisk did was to take the risk management techniques that we understand really well across the firm and then to apply them into a client problem, which is actually a big client problem. The client problem is, how do I manage my asset allocation? It's a pretty simple problem, but honestly, I mean, as I go around talking to large asset owner CIOs, they know it's a problem, and it's the least one they know how to deal with. It is incredible when you talk, and they know they're too slow at making changes. The typical CIO of a large pension plan will tell you it'll take them two to three years to reposition the portfolio if you guarantee there's a recession for the next three years.
Well, if you guarantee there's a recession for three years, they're all max long equities, as you know. Takes them three years to reposition. By the time they get there, it's time to reposition again. It's too late. What they're doing in working with us, the original TargetR isk idea, and then as we've expanded it into various other things, is basically outsourcing the active part of their asset allocation to us. That's the fundamental solution we're solving, but we're bringing it first with a product that builds a track record because otherwise, people don't listen to you. It's nice to say you've got an idea. They say, "Show me the numbers. Show me something that's not a back test." We did that, we did that, we did that, and suddenly people went, "Oh, you really are better than us.
Well, what can you do? That's when we start to do it. Now we are expanding that idea of taking our risk management techniques into more and more different problems that the client has. We use the target sort of whatever. I mean, use that as the phraseology because clearly it's got a nice amount of brand awareness now. A lot of the things we're doing with a target bit at the beginning are really taking other risk management techniques we know and applying it to their problem. The innovation is really about using things we know how to do to solve a specific problem, and there's lots of room for growth in that, which is nice. This is going at another angle. I talked about we're trying to grow some of the things we're doing in income.
Robyn talked about trying to do things in ESG. It's quite hard to do impact things in public markets. We're trying, and we're, you know, that's what the engagement is around. You know, doing things in a housing market is a place where you can have impact, and you can have impact surprisingly quickly because housing is in the housing markets are in a mess in most countries and people need innovative ideas. The first is the stuff we've been doing in the U.K.. Well, I'll use the simple example of where I live. I live in Kent. They built a beautiful new hospital like three miles from us, so if I ever get ill, I've got a very quick run to the hospital.
The small problem is the average house around where we live is a sort of, well, a two-bedroom house will set you back half a million pound. Pretty hard for a nurse, never mind a doctor, frankly, to live in a half a million pound house. They built this beautiful hospital, and they couldn't get anyone to work in it. They need affordable housing. They need community housing. We work with the local authorities, actually not in Tunbridge Wells, just so I'm not conflicted, but we work with the local authorities where they are the people we're leasing the properties to, and then they are leasing them on to nurses and other key workers. You get something which can deliver a nice financial return with a good credit. That's what clients like, but actually does something good as well.
Similarly, in the U.S., we've had this single-family residential business for a while. It's in the bottom right-hand corner. I never know if you're allowed to call it the Southeast in America. They get very confused as to what's Southeast and Midwest and whatever. The bottom right-hand corner, as you look at the map, is the growing housing parts of America. The nice thing down there is you can buy a really nice three, four-bedroom house for $250,000. Hell of a lot better than Kent. That's the part of the country you see a lot of people moving to. The interesting thing is, you know, a lot of the capital wants ESG-related solutions. How do we get. Here's a thing we're working to build, in fact, we think the first Net Zero housing program from scratch.
It's build to rent. We know how to do build to rent. We're working with clients who want to have an ESG, and you can build net zero housing, and you're bringing an interesting opportunity. We do some things in private markets. It's not huge for us, but we never want you to forget it. Then credit, which I talked about a little bit. Here is an area where there is a mixture between hiring and organic. I'm trying to stick near the speaker endpoint, which I can't quite do, but you'll see that what we have over this time period grown what we're doing in credit significantly. You could see we could be a lot bigger in credit. You all know how big the credit markets are. There's lots of room to do more in that.
The interesting thing, the yellow box is quant credit, systematic credit. We think that has a really interesting future. It's something where, you know, we think you could generate some really nice alpha. More and more of the credit markets are available for electronic trading. We're now doing a number of things in electronic trading in credit markets, which lets our skill come through. You can access data even on private companies. That's pure organic. Then you'll see we've added a number of teams. Jean's gone, so we added a couple of very good teams from Schroders, which I can now say, thank you, Schroders, for good training that day. We've been able to bring them in, develop the teams, add resources around it, expand.
We now have a nice platform across credit, and the sort of suggestion on the shading is if we're sort of where the green is and each of the different things can get to the light blue, assuming all things being equal. Now, tough credit markets as we speak today, two interesting bits about that. One is it's a place where the alpha is really coming through. We've got several, I can't remember the numbers, but, like, 4, I think, of those boxes are top percentile. Again, top percentile, not quartile. Top percentile returns in high yield in emerging markets. I can't remember which one or the other. Anyway, in the course of this period, right? We've been delivering the alpha because the information flows, so that is very good for scaling possibility.
The other thing, the balance in the firm means we can, you know. You might say, "Oh, it's a bad time to be growing credit because, you know, bad credit markets, aren't you gonna be stuck with seeding of credit things?" Whatever. The balance of the firm on the other side of the floor in Evolution, we might be not long credit, shall we say, and it's been an extremely good position for us to be short credit, and you generate performance fees off that. The business is naturally hedged. The more we diversify things, the more different parts of the business hedge other parts of the business, which creates shareholder value. Damn. Sorry. Antoine's presentation. Here is. I mean, this is the solution it was getting to, what Eric and Steve.
I know we feel like we go on about this. I guess I partly go on about it because the number of times I get somebody from a bank comes to tell me they've got a solution for us, and what they mean is, "I've got a product." Every one of them, I mean. Sorry, my email gets 20 of these a day. "Have I got the solution for you?" What they mean is, "I've got a product. I've been told to sell it to you." That's not a solution, right? Solution is about understanding the client's problem and blending an answer to meet their problem. It's something that you've seen some different examples. It is pervasive through the culture to listen. Whenever we hire a new salesperson, the interview process is basically, do you listen enough?
We want somebody who listens 90% of the time. Whenever anybody goes out to talk to a client, it's go and listen. Don't sit there talking. Listen to what the issues are, listen to what the problems are, then we can build a solution. We can do something which really gets to adding value, and it's really good for hoovering up content across the firm. You know, Eric, I think, talked about the fact that the solutions bit can use the content that's in blocks that are too small to sell individually, but it also is a great way of selling, you know, spare capacity in particular markets, spare capacity in individual names, different bits that you can push together. It's the solutions, as well as being stickier with clients, means that we can pull in the content.
what is so important is that all the investment side is focused on new content, new dollars of client P&L, right? We're sort of had to get people away from thinking my goal was to have the most perfect track record, the highest Sharpe ratio, whatever, 'cause that's not what our clients care about, right? Clients are not sitting there going, "Tell me the Sortino ratio on this," and whatsoever. They're like, "If I give you know, money, how much return can you generate? How many dollars can you generate that I can use to pay out the retired teacher in Texas, his pension, or the hospital worker in Denmark? How do I, you know, how many dollars of P&L can you get me?" That's what we get all of the research process focused on.
That's what we get the investment teams focused on, and it is a constant innovation process to get there. With that, I think we then try and pull it into why it's good for shareholders.
Yeah. Still ahead on time.
We're trying hard.
Look at this. So we talked so far about the sources of our competitive advantage and how we use them to, you know, serve our clients, deliver solutions, and grow the business. What I'll do in this session is bring it back to our business model, as you alluded to, and make three points. The first one is we're positioned for growth, through the mix of business that we have, a diversification inherently built in our sources of revenues. We are positioned for growth. The second is, although we invest to maintain the competitive advantage, we also have through, you know, particular benefit of technology, huge operating leverage, built in the system. The third point is that models can deliver and has delivered meaningful shareholder value, as we'll go through.
Let's start with growth. Over the last five years, since the end of 2016, we've grown assets at 13% per annum, and our management fee run rates at 8% per annum, which depending on how you look at it, is, you know, quite a bit ahead of asset management overall. While our flows have been and will remain lumpy in nature, we believe that our positioning around solutions, alternatives, it, you know, exposes us to markets that are likely to grow again at sort of 5%-6% per annum over the cycle. There's, you know, positioning for growth in our makeup. Sometimes we get asked about capacity. Capacity is something that we pay a lot of attention to.
It is, you know, kind of the opposite side of the alpha coin that we talk about. However, given the mix of strategies and teams that we have and all the innovation that we've been through already, we believe we have room to, you know, more than double AUM from where we are. So there is, you know, capacity in our strategies, in our content to meet the client demand. And then finally, you know, where we see an industry where is some level of fee compression, we don't think about management fee margin ex-ante. It's not something we target. At no point in our business model we say, you know, sit down and say, we want to aim for 63 or 64 basis points in aggregate.
Really what we want is profitable growth, and the management fee margin is the output of that profitable growth across the mix of business, and funds and solutions that we provide to our clients. We have, you know, a model that has grown over the last five years that is positioned for growth, and the sort of organization itself is geared to deliver that growth. Another benefit we like to argue of our model, and Luke touched on that, is that it is inherently diversified. What do we mean by diversification?
If you think our clients come to us to provide a sort of alpha content, which, you know, by construction should be uncorrelated with the beta exposure of their portfolios, does mean that this gets reflected in the mix of AUM and therefore revenues, management fees, and performance fee revenues, that the firm is exposed to. As of the end of last year, which is the charts you see on the left-hand side, we had just over 60% of our AUM in alternative, but actually, when you bring that back to management fees, it was closer to 80%. 80% of our management fee is kind of exposed to alternative.
What we represented here on the right is kind of a simulation of what the historical beta, equity beta of our asset mix would have been over the last five years as a point in time. It makes the point that, you know, over that period, as a group, our kind of management fee beta exposure would have been around 0.4. Obviously, the long only, which is more equity-focused, would have had a higher beta. That's the green line you see at the top. The alternative line, which is the blue line at the bottom, is fairly low. On average, it would have been 0.3, but actually in some cases as low as 0, and can kind of ramp up quite meaningfully.
It makes sense, intuitive sense, because, you know, if you think of some of the strategies, for instance, tech trend following, it has a, you know, over the cycle equity beta of zero. So if you deliver that to clients, you should expect to see in our, in our revenue number, in our management fee number, that diversification coming through, as is represented here. The other feature of the business model, which, we think is very important is performance fees. Our stock of performance fee eligible AUM has increased by 66% over the last five years from GBP 36 billion to GBP 60 billion, over that period. Performance fees are a very important component of our business. They are the, kind of the economic manifestation of the alpha that we generate for our clients. It means that we're greater aligned with clients.
It also means that, you know, we can attract and retain talent in an industry that is quite competitive. They are an important part of our business model. To try and illustrate that point, we've represented here on the right-hand side, you know, as befits sort of nerdy quants, the output of a Monte Carlo simulation which looks at 12 months forward expected performance fees based on AUM and the underlying stock of performance fee eligible assets as of the end of last year and the end of 2016. Like the usual disclaimer, this is simulations. This is not a forecast. It can go to zero. It converges, you know, at some point to zero.
Really the two or three more important takeaways are, they can be meaningful, and the median performance fee over a 12-month horizon can be quite meaningful, and it has grown significantly. Five years ago, our median expectation was on $160 million. It's now closer to $350 million. That has grown by over 100% over that period. The second point, which is almost as important, is that the distribution, the shape of the distribution has changed quite materially. That goes to the point that we've made throughout the day of, you know, diversifying the business is expanding the range of solutions and content that we provide to clients.
The effect of which is that the right-hand tail of distribution is kind of fatter than it was in the past, so we're inherently more diversified. That's an important key part of our business model. To summarize the first point, you know, we have a business model that exposed to structural growth, is diversified, intrinsically diversifying therefore, and has an additional revenue stream, performance fees, which is highly valuable to shareholders. We do not rest on our laurels, however. We continue to invest, focusing on people, talent, technology, which form a key part of our competitive advantage. If you look at our sort of cost base, you know, people and technology together probably represent 80% of our annual cost. There's a majority of, you know, the cost infrastructure that we bear.
You can expect us to continue to invest in those key areas, resulting this year in a material increase in cost, partly to support the ESG initiative that we just talked about, partly to support our execution, partly to support the technology. That is one of the key aspects of our business. However, we do that in a framework and with an approach and mindset that pays a lot of attention to operating leverage. In particular, we'd argue that the sort of positioning and the investments we're putting into technology attract a kind of higher operating leverage than some of our peers. We look here at revenues, core profitabilities, and then cash conversion from 2020 to 2021.
What you saw was, you know, in a good year like 2021, when our revenues increased by 59%, our core profits actually increased by, you know, over 130%, over that period. The same would apply, by the way, just if you look at it on the management fee level. Our management fee increased by 20%, but our core management fee profits increased by almost 50%. In addition, you know, again, as you should see and expect in a well-run asset management business, our profits convert to cash, you know, immediately in the year that they're generated. That could lead to significant, you know, cash generation.
Last year alone, we generated GBP 650 million of cash before tax, which is almost equivalent to our core profits for the period. What do we do with that capital and cash that we generate? The template is fairly kind of clear and has been for a while now, and we've demonstrated discipline in applying it over the years. The waterfall priority starts with our dividend, which as of last year is now progressive, and then we look at organic and inorganic opportunities for capital. In an environment where over the sort of short to medium term there is no obvious use for that capital, we will diligently return it by way of share buyback. That's the template which we'll continue to apply.
As I mentioned, in the absence of M&A, our focus has been looking at more organic ways to grow the business, and we've used our balance sheets to do so, meaningfully over the last three to four years. One point which is worth highlighting is we are no longer regulated at the consolidated group level, which means that we don't have a consolidated capital requirement at the group level. We are obviously regulated in several jurisdictions, and those jurisdictions have therefore kind of local entity-level capital requirements, but it's not a requirement that we have at the group level. As a result, we have flexibility to use our balance sheet, or book value, which we proxy as net financial assets, to invest in the business and support organic growth initiatives.
We do that with a very strict kind of risk framework. The idea is that we have a VaR threshold which we target expect not to reach. The effect of which is, at the end of last year, we had, you know, a VaR on a sort of one and 20-year, one year forward definition of around $42 million of exposure on our balance sheet. That supported, you know, over 30 different underlying kind of investments that we're making in the business, which together accounted for around $756 million of kind of notional exposure. Some of it, as you see here, was funded externally through repos and swaps.
That's another point that I want to make, is we intend to continue to be a bit more agile and pragmatic on how we fund this exposure. As I alluded to at the year-end results, you can expect to see us draw, you know, opportunistically on our credit facility if we think that it's gonna be sensible to do it because it is relatively cheap compared to some other sources of financing. The final point to make is this can generate and has generated, you know, not insignificant returns for the group and for shareholders over the last three years, through kind of supporting organically our funds. We've also made $67 million of gains on investment, which get reported together with performance fees.
We have a clear framework to use our balance sheet and net financial assets, supporting the growth, paying attention to risk, and eventually that generates in addition to obviously supporting strategies like TargetR isk, which as Luke mentioned, was seeded for three to four years on balance sheet before really taking off. It creates actual P&L gains for shareholders. Last few slides for me. The effect of what I've described is that we have a business model that can and has generated meaningful return to shareholders by way of dividends and buyback. The last six years, we've generated cumulatively $2 billion of capital returns, roughly split 50/50 between dividend and buyback. Last year alone, we returned $544 million to shareholders.
Over the course of the last six years, again, we've bought back the equivalent 17% of our share count, or rather we bought back more than this, because, you know, we have kind of the dilutive effect of share awards. The share count has reduced by 17%, which means that if you look at the reverse, you know, you have a 21% increase in earnings participation. For $1 of earnings now, as a shareholder, you expose to $0.21 more than you would've been, five to six years ago. In conclusion, we have a business model which has demonstrated the ability to deliver attractive returns over time.
First, our management fee and performance fees have grown meaningfully, and are positioned in segments of the asset management markets that are expected to continue to grow. While we invest in the business, second, we also pay attention to operating leverage. Over the last five years as you know was made by Mark in his technology section, we've seen our management fee margin increase by 10 points, core management fee margin increasing by 10 points. Then we return you know diligently capital, excess capital to shareholders. Over the last six years, we've returned roughly half of a market cap to shareholders, which we'd like to think is an attractive proposition for people in the room. With that, hand over to Luke for final remarks.
Thank you. Thanks, Antoine. I've realized, of course, it's the thing that people have actually behaved and not asked questions. The point of this big thing is we are gonna give you all a chance to ask questions. It never occurred to me people actually would listen to asking questions. Look, you know, I'm. The tone of the day is conscious, and as I was sort of sitting there thinking about concluding remarks, it's, you know, we've slightly rammed through some messages, but honestly, I think it's right. I did a strategy update for the people internally, and it was the same thing.
The firm has had a great run over the last five years, and it's very easy to sit there and go, "Well, that was the best of it." Honestly, the great run has just been what we would consider the firm acting as normal. The opportunity to continue to grow is clearly in front of us. We have to execute well, right? I fully agree we have to execute well. I hope the last five years has demonstrated we know how to execute on the model. There is no doubt that the demand for liquid alternatives is only going up. Post what happened in the first half of the year, and we'll see how long it goes on for. Our ability to make alpha at scale has continued through all of this, and, you know, we've been able to deliver.
Last year was a very strong year of alpha at scale. There's enough public information available before anybody asks me the exact number, but it's on the website. You could see how much alpha we're generating this year. It's 2 points about that. It keeps clients happy. They're getting what they need. It means they come to us to ask for more solutions to their future problems. Anyway. It's great the clients who think about a problem before it hits them. Those are the ones you most like to work with. The reality is most clients think about a problem just after it's smashed them in the face.
Well, for most clients, max long equity, max long private markets, they got smashed in the face in the last six months, and they're trying to work out what do they do with their asset allocation, how do they think about the thing going forward, and they're coming to us for advice about how to do that. We believe we've got the capability to deliver more and more solutions that get them what they need in terms of returns, what they need in terms of ESG, what they need in terms of actual solutions to the problems, and that enables us to deliver on this growth. The demand for the alpha is really strong, and it, I mean, honestly, they're sort of.
Steve went quickly over those numbers, but you know, if you take the fact that simply, just taking anybody's estimates of real growth as opposed to nominal growth over the next 10 years and apply it through to what that implies for equity beta and fixed income beta, you know, all of basically all institutional savings plans and frankly retail needs something to top up their returns. You can call it alpha, you can call it added value, you can call it whatever you want. The reason people like private markets is the belief that it's gonna outperform over the long term, and it certainly has done looking backwards. The reason people like liquid alternatives is because we deliver that alpha in scale.
You can fit it within a portfolio very easily, either on its own or with a beta attached, if that's the way you wanna do it. There are not many people who can deliver that alpha at scale. People can have good ideas, but the ability to turn the idea into a trade at scale that you can get into and get out of without leaving all the money on the table for our lovely friends in the investment banks, and actually more importantly, for Citadel Securities and the other high-frequency traders, is a super important part of delivering to clients, right? We have really invested in that space so that we can not just come up with good ideas, but we turn them into dollars of alpha for the clients.
You know, we have a process to keep generating more of that by continuously investing in research. One of the numbers I realized we didn't pick out of this, but maybe we should have done, there was somewhere there was a thing about the headcount growing, what was it, 24% or something? Remember, over the course of that period, the cost base of the firm hasn't really changed. At least the plan was it to change this year, but I think you can all do FX calculations to know it's harder to spend money at a $1.25 than it is $1.35. The point of it is, it is the continuous efficiency gains we get out of the business has let us keep investing in more people, more ideas, and grow the firm and grow the people.
Where we have to add resources is either to do more research or do more legal contracts to execute our more separate accounts, but we have less people doing operations today than we had five years ago. We have less people doing execution who pick up the phone than we had five years ago because it's all become automated. Through all of these things, we're constantly generating the efficiency that means that more of that gross alpha that we generate gets through. It comes down to the people. We started with Lara talking about talent. Hopefully, you've seen, you know. Some of the people you've seen today are used to giving the presentations, others would never have to talk in public at all. What you've not seen is nobody's been allowed to be a nerd. What we all love here is talking about our subject in endless detail.
If you wanna talk about one bit of the execution model, we've got people who can talk to you about that for 10 hours. If you want to talk to Lara about one bit of our talent program, she can talk for longer than you could stand, in a good way. She's walked in, so I meant that as a compliment. For us, loving what you do, being a nerd about what you do, digging into the detail, wanting to is something we prize above everything else in the culture, and it is what helps us deliver this continuous growth in alpha that lets us deliver for clients, which lets us deliver for shareholders.