AI, IA and AM: Robots for Asset Managers

Everybody has heard about Artificial Intelligence, or A.I. However for asset managers I.A., or Intelligence Augmentation, is far more important.

A.I. has long been part of the public consciousness, more due to Sci-Fi than reality. Mechanised intelligent entities appear in fiction starting around the beginning of the 20th century. Generally these machines learn as they go and eventually compete with or supplant the human brain, becoming the base story-line of many a Sci-Fi work.

In 2015 we are still a long long way from that, as you know if you’ve ever tried to use Siri on your iPhone.

Why we humans are the difference between A.I. and I.A.

A.I. is a real red-herring when it comes to business. Certainly within Asset Management there are plenty of algorithms and systems and data flying around, but there is no real A.I.

Instead, pretty much all commercially useful technology in the knowledge industries is better classed as I.A.

That’s because with I.A., human inputs and human skill remain vital. A computer merely augments them, performing calculations and other tasks that the human brain can’t (at least not in a reasonable amount of time). The I.A. computer amplifies our human skill.

I.A. doesn’t replace the human, we still require a skilled person, but it makes an extra contribution to the human’s decision making. I.A. helps people get the most of their own skills, using their own accumulated knowledge, experience, learning and intuition. 

Look again at the picture above: the guy with the robot arms is augmenting his capabilities in executing a physical task. He is still directing that task, and neither he nor the machine can do it alone: this is the physical analogy to the I.A. computer helping a Fund Manager in his/her decision making, such as when creating a portfolio of assets for their investors.

 There’s really nothing new about I.A.

I.A. isn’t a new concept. It’s pretty much been the foundation of business innovation since the industrial revolution.

Take a process => Isolate the true human skills => Augment the ability to produce through technology.

In businesses that don’t require vast amounts of brain power – the industrial revolution started by automating aspects of textile manufacturing - human skills can be mostly replaced, or more accurately re-directed.

When I say this I mean that an assembly line still requires human skill but that skill is now found in designing the product for easy-assembly and designing the process of assembly rather than doing the work of assembly.

But in the knowledge industries - and Asset Management is the purest commercial application of knowledge for profit - human skills are an integral part of the production process. They can’t be replaced. But they can be augmented.

And the way to do this is to design the (decision-making) process so that you benefit most from what humans are good at, and you automate most where they fall short and where computers can help.

I.A. is a major differentiator in A.M.

This may seem obvious, but historically the Asset Management industry has been slow to adopt I.A.

As I’ve explained previously in this blog, in A.M. innovation can be tough and personalities still drive decision making.

But there are very real advantages for asset managers willing to invest in I.A.

In fact, I observe a definite correlation between I.A. awareness and A.M. success:   one difference between those large AM businesses that have succeeded and those that remain small is that the larger businesses are open to I.A. They have used I.A. to augment the skill sets of their staff. That is one of the reasons that they perform better, and grow faster.

 So what role does I.A. play in your business?


Art vs Science, have we progressed since 1959?

In 1959, the British scientist and novelist C.P. Snow gave a lecture on how art and science separated people. It caused a huge furor which has continued to this day.

Snow argued that people who sat on one side of the arts/science divide didn’t fully appreciate the methodologies of people on the other.

And, because of the British tertiary-education system’s emphasis on humanities (especially Latin and Greek), this meant most ‘educated’ people simply didn’t understand science - and therefore they failed to use the many benefits of scientific process in their day-to-day lives. 

Snow obviously never worked in modern finance, which is stacked with scientifically-educated (often very highly-scientifically-educated) people. But, if he were looking at the industry, he may well notice that few people apply their scientific training to their work.

Instead, they tend to treat what they do as more of an art - a cottage industry where intuition, force of personality and individual brilliance hold sway.

Of course industrial process engineering has taken hold of back-office operations, settlements and cash management in most areas of finance, but in many places, Front-Office decision making remains more intuitive than scientific.

Intuition has its limitations

I think that we should be treating much more of what we do in Finance as a science, not an art.

After all, finance requires all the same elements as any other scientific pursuit: a process, data, analysis and inputs.

There will always be room for good old gut feel. It’s gut feel that may lead us to hypotheses that we can then test in a scientific manner.

But gut feel on it’s own will not lead to consistent long-term success.  We need to be scientific about it.

 We’ve come a long way since Kitty Hawk

In fact, I think we should see the scientific process for building a decent investment portfolio as no different to the process for designing an airplane. We should be trying to get the best out of our decisions based on analysing all the data - and using every tool at our disposal to help us do this.

Sure, you can try to figure it all out as you go along. That’s exactly what the Wright brothers did at Kitty Hawk back in 1903. It was perseverance through trial and error that got them off the ground. It was a huge achievement, and flight has come a long way since then.

In 2015, you’d never dream of designing an aircraft without Computer Aided Design, wind tunnels and Finite Element Analysis. These technologies have made airplanes faster, more efficient and infinitely safer. Failing to use them when building a modern aircraft would make the aircraft unsellable, it would also indicate a reckless indifference to the economics as well as the lives of pilots and passengers.

Consider a small change like the vertical 'wingtips' that arrived on commercial jets in the 1980s. These improved fuel efficiency by around 3% and are directly a result of designers gaining access to improved computational modelling.

There is still a lot of skill in being an aircraft designer: the software doesn't do it all by a long way. You need training, a deep understanding of your goals, and experience. And indeed there will be a huge difference between the designs created by one person or another using the same software.

But the software still ensures that the plane will fly, not break- up under stress and be optimally engineered for cost. All problems that the Wright brothers solved through trial and error.

Think laboratory, not studio

We should approach portfolio construction in the same way.

Instead of treating the Portfolio Manager’s office as an artist’s studio, let’s see it as an industrial setting where we use a robust, repeatable process to inform all our decisions.

Let us build an Investment Portfolio that benefits from a PM's skill, in a robust way and that uses science so that it also will fly, not break-up under stress and be optimally engineered for cost!



Robo-Advisors, 2/3rds of the way there…

It’s 2018, Robo-cop doesn’t yet exist, but Robo-Advisors are very much ‘in’. And it’s not hard to see why.

Businesses like US-based Wealthfront and UK-based Nutmeg promise to build anyone a portfolio that suits their risk profile without any human involvement at all. Pretty much all you need is an email address and a little cash.

On top of that they’re intuitive and easy to use. They have transparent, low fee structures. And they have great web design and marketing too.

Robo-Advisors harness two of the three great powers of modern computing: Networks and Databases.

Networks give their sales efforts huge economies of scale, letting them access many clients without having to knock on doors or have a retail presence. Databases give them the power to manipulate a lot of data, which they use to allocate your money and then keep track of it.

Combine these two great powers and Robo-Advisor companies are able to open and manage a multitude of small accounts - in the process giving small-time investors the feeling that they are getting the kind of individual treatment once reserved for those with access to a private banker.

 Things look great when the bulls are charging

A Robo-Advisor doesn’t pick stocks. It constructs a portfolio of ETFs from a restricted menu, choosing weightings relating to your risk appetite. They don’t pretend there’s any active management or asset selection (i.e. ‘Alpha’): the ETFs they invest in are mostly those that have done well recently.

The Robo-Advisor builds a portfolio of ‘Smart Beta’. A portfolio constructed without any active skill added to it.

Right now that looks great. After six years of QE, when stock market price action has been going all one-way. So any historical simulation of what a ‘Smart Beta’ portfolio would have done looks fantastic.

The S&P Index for the past 6 years, one way traffic!

And that track record is a very powerful marketing tool. Almost all investors buy on the basis of the graphically displayed track-record.

But I’m not sure that how you want to look at the next few years.

Think Smart Alpha not Smart Beta

Robo-Advisors don’t rely on skill at all. That’s partly why they can be so low fee. 

But now I think many investors are going to be better off using someone who can apply skill and experience right down to the asset-level decision making. 

When it gets difficult, who do you want on your team?

And if Investors can access that ‘Alpha-Skill’, they can create their own Smart Alpha portfolios. They can combine this skill with their own risk tolerance by using the third of the three great powers of modern computing...

The final 1/3rd: Computational Analysis

Computers started off as Analytical Engines (1940’s): it is only more recently that they became databases (c 1960) and now networks (1990s).

And Robo-Advisors do very little analytical computing.

But you can. Today in Finance you can access analytical engines that do real maths and aid decision-making. So if you want to move beyond Robo-Advisors and into the realm of Smart Alpha, then you can do it.

You can build a portfolio taking the skilled input of a stock-picker. Computer software such as Sherpa Funds Technology’s ORS can combine it with your own specific risk tolerance to produce your own Smart Alpha Portfolio.

You can use the analytical capabilities of computers, rather than just the networks and databases that the Robo-Advisors use, and make it work for you. 


Asset Management and the struggle with Innovation

I saw the cartoon above on LinkedIn a while back, and I’ve been thinking about what’s going through the minds of the team dragging the box: not least because they remind me of how it sometimes is when you work in finance.

 They know what they’re doing is tough. But they’ve got a delivery to make. And they know what they’re doing will mean that the box is delivered on time, even if it takes some effort.

 Then the innovator comes along with a new idea that will make the box go faster. It’s obvious to him that it will. But the guys pulling the box are sceptical.

 Here come the questions….

 “How does it work”, they ask?

 “Will the circular 'wheel' fit the axle? Will it break? Do we know how to fix it? Will the supplier be there to help?

 If we decide to fit that wheel will we have to stop what for a while to fit it? That’s going to make us lose time. We’ll fall behind in our goals.

 After all, we have stakeholders, bosses, employees and clients, you know. These people depend on the box being at the goal at a certain time. We can’t put that in jeopardy.

 So I’m afraid we’re going to have to pass.”

 How Other Industries get around this

Innovative (i.e. successful) manufacturing businesses often change processes, and do it in a controlled way. 

  • When a new process looks like it will increase efficiency they set up a small parallel line to test it.
  • If it does, they build up the capacity of the parallel line while gradually running down the original line.

A key difference with Asset Managers  that manufacturers' capital comes from shareholders who expect investment in R&D and know that it involves some level of risk: if an innovation does work it can dramatically impact on their bottom line in what is a very low-margin business.

 Finance Industry Conservatism

 I think there are three reasons finance is much more conservative about process innovation (although not necessarily product innovation!): 

  1. There is very little emphasis on front-office process in the first place. (This is something I’ve written about in the past [Process makes Profits]). The Industry's focus on individual decision making clouds our judgement on what humans are good at and what’s best left to the machines.
  2. Although the total ROE in financial services is low, the differential margin attributed to individuals' decision making is huge. So there’s limited pressure to make decisions more efficiently, and a reluctance to question the way those decisions are made.
  3. We’re in thrall to the mystique of Fiduciary Duty and use it to justify the status quo. Testing an innovative idea on real money (someone else’s real money) can easily be misinterpreted. Shareholders in other Industries expect their money to be spent on R&D. Investors in Funds are not so keen on this.

 So it’s not surprising that innovation is hard, especially in established businesses.

 What that means for Asset Managers, Innovators and Investors

 As someone who wants to show you the benefits of potentially game-changing technology, my message is NOT just that our clients - Asset Managers - must be more accepting.

But first that the Innovators and the Investors (the end clients of the Asset Manager), must have some dialogue. Unless these clients understand that we are all working for them, they will be nervous, and that nervousness will spread through the chain.

And secondly the Innovator has to show an implementation plan that does not cause disruption to the Asset Manager.

And only finally that the Asset Manager can benefit from controlled process change, and should look more at Innovations like these I wrote about previously.

That’s also why after we demo our software, SherpaORS, we run parallel portfolios for our clients. You get to experience the immediate and real difference it makes to your portfolios with no impact on your daily routine.


Image result for the perfect portfolio

To build the perfect portfolio, answer these 4 questions by Richard Waddington 

It’s fair to say that good Portfolio Managers earn their keep. Building an investment portfolio is a rigorous process that involves research, knowledge and diligence, as well as good instincts and strong self-discipline.

To do this well, I think that you need to answer four simple questions.

Here they are.

1. What am I trying to achieve?

The first question you should always ask is: “What am I trying to achieve?. This goes for whether you are making the investment decisions yourself or you’re giving your money to a third party Fund Manager. If you are the Fund Manager you'll need to ask "What is my investor trying to achieve?".

It might sound obvious, but you would be surprised how often the answer to this question is taken for granted or even overlooked altogether.

You should always ask this question without giving any consideration as to how you are going to do it. Otherwise you're likely to get bogged down in the detail of your investment choices, and this will alter your thinking about what you are trying to achieve.

Always remember to be realistic. Don’t come up with fanciful or unachievable goals that are going to cloud your judgement when it comes time to trade. Be clear and certain about what you are trying to achieve.

2. What's my Edge?

Once you’ve worked out what you want to achieve, ask yourself what’s your ‘Edge’ for achieving it. In other words, what is it that separates you from any other investor or portfolio manager?

Here’s a tip for answering this question. There’s no such thing as a free Edge.

Your Edge comes directly from your work. It may take the form of more detailed analysis, better models, better data or anything else that you do particularly well. But it always has to have a grounding in something you do better or know better than anyone else.

If it’s freely available, it’s simply not your Edge.

It's this Edge that will drive your asset decisions and lead you to make a call on which assets to own or ‘go short’: which assets you want in your portfolio.

But you're not finished yet. The assets you have just selected need to be put together in a way that gives you the best chance of achieving your goals (the answer to Q.1) and that still involves answering questions 3 and 4.

3. What does the rest of my world look like?

The third question to ask is what the rest of your world looks like. What other assets do you own? What other risks are you exposed to?

The answer to this question has a big impact on your perfect portfolio. Answering it well means that two people with the same goals, the same edge, and the same asset decisions may end up with two very different looking portfolios.

4. How do I bring all this together?

You now have your answers to these 3 questions:

So how do you use this information to build your portfolio?

This is by far and away the most difficult question to answer. That’s because you must use all the information you gathered in answering Q's 1-3, and do it consistently.

It is very common to let human qualities like emotion or distraction get in the way at this stage, and when that happens, your carefully prepared responses to Q's 1-3 go out of the window.

If you want to be successful then you must perform this task: combining the three answers consistently every time.

And that is why we at Sherpa Funds Technology developed SherpaORS, a tool specifically designed to answer Q4 and produce an optimum risk size for every single trade.


Why Size Matters by Richard Waddington

In my 20 years trading in financial markets I watched my fellow traders both make and lose a lot of money on behalf of their clients. Often the same traders who would pull off a series of wins would be the same ones who soon afterwards suffered a massive loss.

In fact, of the 1,000 or so people I worked or dealt with over the years, I could probably count on one hand the number who had consistent returns.

And I wasn’t one of them.

Maths in Asset Management: Man and Machine

Several difficult conversations with clients made me think about the role that maths could play in aiding consistency and improving the relationship between asset managers and their Investors. So in 2011, I took a two year sabbatical to research this question.

One conclusion I ended up with came as a big surprise to me...

When it comes to asset selection, the human skills of a portfolio manager are very valuable.

My research showed that analysing which assets to invest in was rarely a pure mathematical exercise. Instead it involved characteristics only humans possess - skills such as intuition, relationship building and the sort of complex and un-definable pattern recognition that our brains are very good at.

But there is a caveat.

Trading: the Tale of Two Decisions

The second issue my research revealed is that selecting which stocks (or FX, Rates or Bonds! will go up or down is simply the first part of the money making equation. Just as important is the percentage of a portfolio you allocate to that decision.

In other words, the size of the trade matters way more than many traders and managers give it credit for.

Too Big or Too Small?

If the trade is too small you won’t get rewarded for your skill and insight. But if it’s too large, and you are wrong (there are always times when you will be wrong), you will lose too much. When this happens investors are likely to take away your mandate altogether.

When I witnessed a big loss it was usually because traders got the decision about the size of the trade wrong - not the decision about which asset to select.

That’s because those same human emotions that make us great at asset selection, can hold us back when it comes to knowing how much to allocate. Emotion and the act of buying or selling simply don’t mix.

And where human skills can help in asset selection, human emotion interferes in the sizing decision.

So what is the Optimal Size?

The final - and most important insight - my research showed was that there is always an optimal size for any investment. This can be worked out by properly understanding an investor’s risk profile and then applying advanced maths.

When you get this right, you don't just get more consistent returns. Conversations between fund manager and investor become much easier: there are no surprises. Instead there is a direct link between the risk tolerance of the investor and the portfolio that the Asset Manager constructs on his/her behalf.

And that’s exactly what our product, Optimal Risk Sizing, can do.