Asset Management and how Data Science Geeks can disrupt it
March 20, 2017
Spoiler Alert: This is not about payment apps! It’s a little bit about Robots, but not Robo Advisors. Whoever had to get rich writing them, pretty much has. Tough luck .
I read about a boring merger of fund managers in the UK recently. Standard Life and Aberdeen Asset Management are supposedly merging to stave off competition from low cost, passive fund managers. That’s what triggered my thought process.
These days we have seen a flurry of disruption in the asset management business. Passive investing, made possible by has dramatically brought down the cost of asset management. Robo advisors like Wealthfront, which diversify assets across index funds (and their own direct indexing as well) have brought down the cost of advisory too.
The typical management fee charged by fidelity on a mutual fund is 0.8% to 1% of the capital (https://www.fidelity.com/fund-screener/evaluator.shtml), while Wealthfront typically charges only 0.25% of the assets (http://cashcowcouple.com/service-reviews/wealthfront-review/).
Statistics reveal that active fund managers by far are lousy at beating the benchmark index (like S&P500). It’s should rightly be so if the market is efficient, the US equity market. So, does it make sense pay active managers when passive investment in index funds is cheaper and performs better too? Apparently, it doesn’t. But does that mean active managers should cease to exist? Hell no. Active managers can survive and thrive if they can –
a) Lower their management fees to the levels charged by passive managers
b) Show that they can beat passive investing
c) Make absolute returns. i.e.
That’s because there is nothing inherently great about passive investing, except
Do note that the operative word above is “OR”. What if we change it to a tighter “AND”? Is it possible at all for a fund manager to be active, low cost as well as better than passive and can earn returns even when the benchmark loses? Is that anymore realistic than finding the holy grail? Let’s handle the issues one at a time.
What does the cost structure of an asset manager look like? The costs of fund management are mainly the following –
a) The cost of raising capital (sales and marketing cost)
Investment products are not the simplest of things (or should they be?) in the world and need consultative selling.
b) The people costs and overheads (technology etc.) in actively managing money.
Fund management involves hiring extremely smart people pick stocks/assets, rotate them and empowering these people with tools and technologies and more people!
c) Transaction costs and other overheads
Brokerages, transaction taxes etc.
Let’s see how these can be reduced.
The cost of raising capital (as a percentage of capital raised) is a function of the brand name / track record of the fund manager. While there are a zillion performance metrics, track record can broadly measured based on 2 parameters – Returns produced and how (in)consistent they are. The ratio of average returns (in excess of risk free returns) to their variability is called Sharpe ratio. Passive funds have lousy Sharpe ratios (less than 1). If a fund manager has an investing style that produces a Sharpe of more than 1 over time, it becomes easier to attract capital while not spending a whole lot on incentivizing financial intermediaries (https://www.youtube.com/watch?v=gvZSpET11ZY) to sell it to the investor. Distributing digitally (while simplifying products) helps reduce costs too.
The only way to reduce people costs is to replace people with machines. If people are good at something and they have a method to it, chances are machines can be programmed to do what these smart people do. The machines can also be programmed to live like self learning animals. They can be monitored to see if they are goofing up and if so why. This is doable as long as the fund management job is not something driven just by intuition. i.e. If there is a method to the madness. Paying people to help create algorithms is a way more scalable thing than people for their intuition.
Not many humans can do much about transaction costs. The most that can be done is to minimize transaction slippage (makes sense only if transactions are frequent) by writing/buying execution algorithms (different from fund management algorithms).
We talked about the cost part. But is it possible to beat passive investing at all, given that the market is efficient. The simple truth is – Market efficiency is not a binary. It has (50?) shades of grey. The market is efficient in terms of disseminating information companies and economies. Market prices subsume all the available information in the short term. But if you expand the time horizon, you will realize that the market does misprice assets. Valuations, historical prices, fundamentals etc. may be useful in estimating (using time series models, machine learning etc.) what the prices should be and detecting mispricing.
Apparently, it does look a lot more realistic than finding the holy grail? Yes and no. And remember there are a lot of irrational investors (with excessive fear or excessive greed than what data warrants) out there, both individual and institutional. If you can figure out the nuances of their irrational behavior, you can position yourself to be their counterparty when they do behave that way. There is indeed a method to their madness, from what I’ve observed. What are these specifically? Being greedy when everyone else is fearful and diversifying your greed are probably the most robust and time tested philosophies. Should you also be fearful when everyone else is greedy? That’s not as important. But it helps to have some dry powder when the next opportunity (blood on the streets) presents itself. How can we quantify these things - Fear, greed etc? A simple indicator of fear is Vix (volatility index) and a simple indicator of freed is valuation (especially of the benchmark index). But what qualifies as excessive? That’s more quantitative. For starters, you can plot the probability distribution of any indicator and use it to determine outliers. When you find an outlier, chances are, you have found a buying /selling opportunity too.
Now, let’s try and address the elephant in the room. Is possible to make money when the benchmark loses? This is generally hard the long only way. You have to be long short to accomplish this. That is, you have to have the ability to go long or go short on any asset that’s publicly traded. But if you do this, you become a hedge fund. Is that really bad? No. But chances are retail investors will be off limits. You may (or may not) become a successful hedge fund. But you will not be a disruptive venture, unless you can disrupt retail. Implementing the short side by going long on inverse ETFs might be a useful thing to make a long short fund palatable to the retail investors. Thankfully things like SH (inverse S&P500 ETF) do exist.
Hang on a minute. Does it really make sense to risk your career and savings to pursue this holy grail? That’s up to your risk appetite. You would be well advised to remember that VCs hardly understand this space (despite being in the asset management business themselves)! Whoever said investors are rational! However, it is possible to play in the markets without risking personal money. For example, a Boston based ‘crowd sourced hedge fund” called Quantopian. They would be willing to risk capital on your ideas if they pass their performance criteria. They would even teach you some basics if you are uninitiated to the world of trading and investing - https://www.quantopian.com/lectures. Knowing a dash of machine learning always helps - https://www.coursera.org/learn/machine-learning .
By the way, I and you are not the only junkies trying to crack investing problems. (https://www.bloomberg.com/news/articles/2017-02-06/silicon-valley-hedge-fund-takes-on-wall-street-with-ai-trader). Would love to see if they can grow beyond being a hedge fund and disrupt the retail investor segment someday. That said, it’s not easy to disrupt the retail investor market without an age old brand name. A stodgy financial institution can mobilize a lot more money than an innovator, even in this day and age. Also, India is still some distance away from sophistication in financial markets. So, as an innovator it might be easier (faster) to crack international markets than it would be to make a difference to the financial lives of fellow Indians.
But why do I call it a trillion dollar opportunity, that too with a prefix multi? Here's some stats on the AUM of alternative asset managers (not open to retail really)! Imagine what the AUM could be if you can disrupt retail?