Fair warning this is not a tweet ….I repeat this is not a tweet. Grab a coffee and settle in for a few minutes.
Having found some time on my hands I decided to document some thoughts on the state of play in the asset management industry specifically how it applies to trading. In my humble and for anyone who knows me, obviously very biased opinion execution trading has become an integral part of any asset management business. Conversely, although still seen as a cost it is in fact on the front line in protecting client’s assets in a world where returns have dwindled and squeezing every penny out of a trade has become more scientific and important. MIFID in my view while has been an administrative nightmare and logistical challenge has delivered a rapid change of approach to how trading is conducted. A much more scientific and professional approach has been adopted by both the buy and sell side and the reduction of related trading costs has been immense.
To some of us who have been around for more than a couple of years this is a big change and the temptation is to shake your head and tut tut at the state of play and lack of human interaction. To some it has been a de-skilling of dare I say it an art form developed over many years of long battles fought with order books, brokers and “the market” in general. “Back in the war” a trader who was committed enough had to live a pretty high stress life and really immerse yourself in order book dynamics – there was an art to sniffing out whether you are the aggressor or is the bogeyman on the other side the aggressor and then adjust accordingly. That’s all trading in the short term is – who wants it more. The much vaunted “more buyers than sellers” has always raised a laugh and while obviously it’s silly it takes away from the important point of who is the “more motivated buyers than motivated sellers” or vice –versa. Judging the panic or lack of panic in the opposing order is still a thing of beauty when executed correctly.
As with most industries you either adapt or die.
Now we have large trading volumes spread across the equity market mostly conducted by algorithms. The human element is gone or so most would say. But let’s not forget what an algorithm at its basest level is – a set of instructions designed by a human to process a task. We get lost in words sometimes and has been the case very much recently where big time money managers with great track records blame their poor performance on the “algos” like they are a living breathing thing. They are not – they are a set of instructions designed by a human. Understand that humans perspective and you will likely have a better clue of their algorithm and what it’s doing and why. Most of the brokers I have had relationships will attest to my very apparent reaction when they attempt to explain what their algo does and why it’s the best. A cloud comes over the room and I just want to know what the simple basis for what this algo does is. Is it trying to be aggressive or is it trying to be passive? Once I know that I know how to use it correctly, that is if I evn need to use it.
The “electronification” of the trading landscape and use of algorithms for executions has been vast. Why might this be the case? A variety of reasons come to mind:
Cost cutting: Asset managers and banks believe that using algorithms for doing a once human task is a cost saver. It’s hard to argue with this point on the face of it. In almost all industries labour has always been one of the largest costs of production and almost always been replaced by machinery where possible. The finance industry was certainly no different and probably was one of the largest examples of largesse when it came to pay. I believe in paying for performance so in that respect I would argue that a lot of it was fair ….often mind boggling numbers but fair if they performed. Before I became a trader the desk that I serviced had somewhere close to 25-30 traders covering just large and mid-cap Pan Europe & UK, that’s not even counting the sales staff. That same desk probably has most likely less than ten now on a good day, at least 2-3 of which are likely to be grads. My most recent experience saw me conduct trading for over $200bln notional a year all on my lonesome and barely broke a sweat. So fair enough you have reduced headcount and most likely increased efficiency. However the costs don’t just go away I’m afraid. There is a huge tech spend associated with all this electronic trading. Data, its storage and use does not come for free either. While this is decimating the mid-tier brokerage industry the large or bulge bracket banks have responded in my view brilliantly to this changing landscape and invested heavily in their trading architecture, some more so than others but that an argument for a different day. It’s fair to say that there are 9 banks worthy of taking on large electronic trading volumes – five of which will be the eventual “winners”.
Consolidation of assets: While there were always mega pension or mutual funds back in the day – the growth in AUM of a concentrated bunch of asset managers has been mind boggling. We all know the names of the people I’m talking about – and they exist in hedge fund, ETF and pension fund space. While you will hear many complaining about how regulation has killed the industry most of the people at these mega funds must be crying out for “more more more” as the barriers to entry have increased significantly in the asset management industry. Had it not been for electronification of the industry all those traders I serviced back in the day would have merely moved from the sell side to the buy side and shifted that cost base from one to the other. Adapt or die, AUM of a large magnitude correlates pretty strongly to an enormous amount of continuous trading.
Success breeds imitators: All traders biggest problems over the last ten years on a day to day basis could often have been reduced down to “what the hell are the HFT crowd doing to me.” HFT or ELP’s whatever you want to call them nowadays have received a pretty bad rap over the years and while I don’t condone predatory behaviour or illegal trading practises I do wonder what everyone is complaining about. These guys are just doing what everyone has always done in trading. As referenced above they are sniffing out who is a more motivated buyer or seller and reacting accordingly. It just so happens they can do it a million times a day and not be tired. So they are exploiting an advantage over human traders. I really struggle to see the objection to them (again assuming it’s all legal practises). These guys made a killing and if sharpe ratio in its normal form applied to them (which it sort of doesn’t) it must be mouth wateringly large. Now that the light was shone upon them and the haters hated enough it’s a bit more difficult for them. Also the brokers have responded to try and design their algorithms to temper the worst behaviour from them. However the point being is that the rapid and frightening success of these strategies led most people to the forced conclusion of “if you cant beat them join em”. While most electronic trading is not even comparable to what HFT does the concept of using algorithms to execute being more efficient has stuck.
Style shift: Factor rotation, Risk parity and target vol for the most part require a large amount of continuous trading. While returns have not been on fire over the last few years this would suggest that you also need to ensure your cost base is low and nothing eats into a trading strategy more than inefficient trading costs. A typical long short equity portfolio that employs leverage has to contend with large financing costs at their friendly PB and execution commissions and that’s before they even think about what slippage they will experience between the signal to trade and the trade execution. While financing costs are usually the most expensive you also cant afford to pay the 15-25bps per trade that brokers used to garner on the hi-touch side you need to pay the much lower costs on the low touch side. This makes more sense when you look at the typical trading which is usually never big in any one name but rather made up of lots of little trades across a very diversified universe. What I’m trying to say here is that these type of strategies trade a lot but do not need nor can afford a hi-touch service.
ETF’s: Writing this the day after Jack Bogle passed away I couldn’t not mention the staggering growth of AUM in ETF’s over the last ten years. Jack Bogle has done amazing things for the average investor and truly made it easier and cheaper than ever for Main Street to participate in the market at the right price. The mostly low management fees of these products and fervent focus on tracking error requires a large and continuous amount of trading. Again none of which would be possible under the old hi-touch model.
So these are some of the reasons why we got here. Now that we are here what do we do about it.
If we were in a world where returns were buoyant and everyone was crushing it and making money – I wouldn’t be writing this post because no one would care. We aren’t and haven’t been for some time. When returns diminish costs have to come down also. Never waste a crisis and all that. The work that is going on now under the hood in most asset managers to reduce costs will in the future pay dividends as long as discipline is kept as returns are unlikely to stay low for long. A good execution desk is worth its weight in gold to an asset manager and the end investors. Don’t throw the baby out with the bathwater. Execution desks have always been sniffed at by “traders” as they are thought to just blindly execute like a monkey. I’m sure there are examples of this but in my experience it’s a skill like any other. It’s the same skills as portfolio managers possess just the time frame is different. PM’s must filter through the constant barrage of noise to identify an opportunity for a trade or investment. It’s the same for an execution desk only the one thing you don’t have is choice. You are handed a trade or bunch of trades that you did not devise of often in stocks or whatever asset you have never traded and you have to think quick and fast about what is the right approach to take with this trade given the incomplete information you have.
As mentioned high trading costs are cancer to a trading strategy. They eat into returns that we have already established are on the low side with a ravenous appetite. So many strategies on paper have an equity curve that goes from bottom left to top right in a nice trending fashion until you factor in trading costs.
There are two different classification of trading costs in my opinion – those that can be controlled at the broker level and those that are an in-house problem. I often say to brokers that is a YP or a MP (Your problem or My problem). Too often I feel asset managers have made trading costs a broker problem. Purely focusing on driving down commission costs and declaring victory over trading costs is only half the battle won. Controlling these costs come down to negotiation and having a strong working relationship with a broker. The old combative relationship between brokers and clients are long gone – you are partners whether you like it or not – brokers are the necessary medium you use to achieve the results you want- don’t forget it, work with them and they will happily work with you.
The much bigger impact a trading desk can have is to identify the role they play in driving down the in-house costs (an MP) to a trading strategy and as I keep referring to adapt or die. Your role has become less about waging that war with the market day in-day out and more of a strategist. You need to identify the characteristics of your flow or trading over time and then employ or match your trading strategy to those of the alpha signal or generator.
For very basic examples there are often two distinct styles of investing- let’s say that X PM has a value orientated approach to equities while Y PM has a momentum approach to equities. Is it good enough to simply receive orders from both PM’s and route to a broker algo and work the same way? Clearly not – one style demands an approach that will likely be more aggressive in trading than the other. One style earns its alpha from capturing short term gyrations and trends – so you better make sure you capture as much of that often fleeting trend as you can. A momentum strategy is inherently of the belief that stock ABC is going up/down at a faster rate than stock XYZ so you better get on the train quickly to capture that. Employing an aggressive liquidity seeking type algo for these trades makes sense in theory right? The very widespread adoption of momentum strategies (definitely a post for another day) would suggest that if you don’t aggress someone else will.
Conversely a value type strategy can often be described as being contra or early to the party. So often a good value strategy is picking up unloved stocks or selling ones that have strayed too far from their thought to be intrinsic valuation, either way at the heartbeat of the alpha is usually swinging the opposite way from the crowd currently. Do you stand and buy these names? No, you employ a much less aggressive style and let the stock come to you, it’s more likely than not that there are currently more motivated sellers than buyers in this name.
Obviously this is not going to “work” in 100% of the trades that you do – but show me a strategy that does and I will show you a liar or a fool. What you are attempting to do is win the marathon not the 100 metre dash. If you can control your in-house trading costs and achieve a smooth profile over the long run that is worth more to any strategy than a blind squirrel who finds a nut every now and then.
This is all well and good and a very simplistic approach to take correct? Where it goes wrong is in two ways: first off the PM doesn’t know their own style and you don’t have enough data captured to make an informed opinion on what their style is. Simple remedy to this is to talk it over with PMs and identify a basic plan that sounds like the right approach to take over time. The next part is important …DATA DATA DATA. Yes I know it’s boring and hard to do, methodical tasks are often just that. But you have to collect a variety of nuggets of data every single time you come to market. Without that feedback you are shooting in the dark. The data may show inconsistent results but not just my guess but my experience is if you capture and measure correctly over time the correct approach screams out from the spreadsheet. Also pro tip – don’t need a fancy off the shelf very expensive TCA package – you can build in house in excel very easily depending on your requirements.
Problem number two that many face is do you have the stomach to stick to a strategy over time and do you have the buy in from the business? If you lack A then you will almost certainly lack B and if you have A it doesn’t mean you have B. PM’s, the CIO and the business managers must be involved and behind this approach and take a more active approach in treating the execution desk as vital to the goals of the organisation and the underlying clients. Best ex is not just about broker selection it’s about managing both the YP’s and the MP’s in a pro-active and efficient manner. The more importance and focus put on in-house trading costs the better returns will be in theory –(it won’t correct a failing strategy- it’s not a miracle worker) which is all around just better for business.
The main point here is that a good desk must match their trading style for each order to the characteristics of the alpha signal or generator and over time you will achieve a smooth slippage profile or to put simply you can control and put structure around the in-house problems. You do this by adapting to the new environment and changing how you do things – you have graduated from foot soldier to general, focus on what strategy best suits your organisation and its goals.