Be your own wealth manager: How algorithms build your investment future
When it comes to your finances, things should be simple and straight-forward.
However, ask most investors how their portfolios were constructed and they’ll tell you they have no idea.
After this post, you will be able to ask your (robo?) advisor the questions that led to the construction of your portfolio. Or at least you will have a better grasp on how quantitative portfolios are made. If you don’t want to go through all this trouble, you can always talk to Richard and have him make one for you.
Let me tell you how dark magic works.
First, we have asset classes like Stocks, Commodities, Energy, Real Estate, Bonds etc. Each asset class has dozens of ETFs to select.
Now this is where the fun begins.
An algorithm creates thousands of portfolios using every possible combination of these asset classes and their ETFs. Sounds simple but there are some restrictions to tackle.
For example: a portfolio might need to feature a specific percentage of REITs because the client might own a home, and thus already have a large part of their assets in Real Estate.
At AgentRisk, one of the restrictions we regularly encounter is that some ETFs need to feature a certain level of option liquidity in order for our Machine Learning Strategies to be applied.
Once these are identified, we end up with around 500k unique portfolios. And we are not done yet.
Next, another algorithm ranks the portfolios and finds the best in terms of fees, liquidity and any other parameter we might choose to factor in (like which asset is historically low or high in case we want to apply a mean reversion strategy).
At this point we put the portfolios on the Efficient Frontier to identify the best ones considering your risk tolerance (using A.I. to figure out your real risk profile is another post but in the meantime read here).
The results are then back-tested to see how each portfolio behaved during past market turmoils.
The portfolios that performed better are saved and the one that best suits your financial goals and profile is finally applied to your account.
The interesting thing is that we follow this process in “real time” which means that if the market moves and the best allocation for your portfolio changes, we recalculate everything, making sure your portfolio is again “the one”.
Circle back to where we started and ask most investors when exactly their portfolio was created, or worse, when was it last updated? They have no idea.
So, there you have it. All you need in order to create a portfolio is to choose your asset classes and ETFs, mix and match them, rank them, apply your restrictions, calculate the Sharpé (or Sortino) ratio, fit them on the Efficient Frontier and tada! Your own quantitative portfolio!
P.S. If we get more than 1,000 claps, I promise to ask our CTO to open-source a small version of our portfolio creator, so you can experiment with having fun building your own robo-advisor.
P.P.S. If you have mutual funds instead of ETFs, have you ever asked yourself why their fees are so much higher? Food for thought.
