Maximizing Risk-Adjusted Returns on Crypto Investments
After the hack of Mt.Gox, I had lost interest in Bitcoin & Cryptocurrencies until the spring of 2016 when I got a chance to take a deeper dive into the world of Blockchain and Distributed Ledger Technologies. Fast forward to the present-day, and I’ve worked on some exciting projects using Hyperledger Fabric, Ethereum, and Corda. While I was at it, I happened to stumble upon Coinbase. Trading accounts on exchanges followed rather quickly. Before I understood where this was all going, I was partaking in ICOs & IBOs — it was all moving at a breathtaking pace.
Veni, Vidi, VC.
I’m a sucker for adrenaline and that’s unmistakably counterproductive for the crypto-trading game. I decided to catch a breather during my vacation in December 2017 and take a fresh look at how the cryptocurrency markets are playing out.
Until I figured it out, I decided I would just HODL.
In the meanwhile, Robo-advisors had fascinated me sufficiently that I finally found time to explore Wealthfront.
If you have not checked it out yet, you sure should. You are not very likely to regret it and to sweeten the deal, you will also get your first $15,000K managed for free. (But please DYOR).
As I saw it, automatic investing using AI was all fine and dandy, but there was a different itch I had to scratch — I had to know more about how all of these worked under the hood. I couldn’t let it be a black box that magically arrives at the best investment strategy for my risk appetite. So, one night, I decided to finally peruse the whitepaper on the Wealthfront investment methodology. I thus got introduced to Markowitz’s Modern Portfolio Theory. One thing led to another, and I ended up implementing MPT (however limited in efficiency it may be) for optimization of a crypto portfolio.
The idea behind MPT in English is this:
You want to invest some of your disposable money in the crypto market, but you know zilch about the state of the crypto markets. Your friend from high school, who has been smoking some funny stuff lately, strongly recommends you buy some $BTC. The nerdiest guy on the block, who has been buying mining racks and shoving it all down in the basement of his mother’s house, strongly believes your best bet is $ETH. And then, there’s the guy who thinks that the Big Brother is watching all of us all the time and even holds back from using Tor due to the fear of the egotistical giraffe thinks you must be buying nothing but $XMR and $XVG. As if these friends weren’t enough you have listened to experts who feed off of their collective intelligence to predict how Ripple,$XRP, is going to create waves in 2018.
Well, that’s too much to process. So, let’s break it down.
The relationship between risk and return is not something you discovered while sipping your coffee this morning. If you demand higher returns, you must take greater risks. If you know what your goals in terms of returns are, then half the battle is won because all you have to do now is find a way to minimize the risks for that target return. In other words, you want to make enough money to buy that Lambo, but you want to do it in a way that minimizes the chance of making you look like, well, the guy who lost money in the crypto market.
So, here’s what we do: We select a bunch of coins as potential candidates for investment. By representing the historical RoI of each coin as a random variable, we estimate two parameters — covariance and mean return, for the variable. For a given risk level, our goal is to maximize the return, by selecting the optimal combination of coins. We find that by solving an optimization problem.
The optimal allocation of investments dollars over 14 large cap crypto-assets (market cap > USD 5B), based on price data from last 180 days gathered from exchanges is below:
The plot for finding the efficient frontier:
In the part-2 of this series, I will get into the details of the implementation.