Authored by: Erez Katz
With the enormous government-infused liquidity and an interest rate close to zero, the stock market remains the natural choice for investors to park their dollars. At the same time, many are weary of the dire state of businesses heavily impacted by the pandemic and the mandatory lockdown. Naturally, this environment presents active investors with unique stock picking opportunities geared to profit handsomely.
By merely inspecting last week’s earnings reports from the largest financial institutions, a common theme emerges. Buried deep in otherwise disappointing earnings results, profits from investment activity have consistently beat projections, some by meaningful margins.
As it stands now, even with the rebound of the past two weeks, the S&P is still down -12.25% for the year. It makes one wonder what professional investors are actually doing to generate such handsome market relative returns?
Indeed, many AI-based algorithmic investors are killing it and there is no doubt that we will see many more funds transition away from the traditional buy and hold index funds into smart stock-picking alternatives powered exclusively by big data and AI. We are rapidly moving into a trading environment not vulnerable to human emotions and knee jerk reactions. An environment where excess returns are a pure function of the novelty of data and the machine-learning algorithms that mobilize it for alpha or market relative returns.
In today’s blog I want to demonstrate one example of a long/short market neutral portfolio that is completely free of human discretion. In other words, stocks selection, rebalancing schedule, allocation guidelines, and exit criteria are all determined algorithmically based on data science, statistics, and AI.
Tech Hedge Market Neutral — An Actively Managed Long/Short Cash Neutral Portfolio
Tech-Hedge is a long/short market neutral portfolio. specifically, it’s a cash neutral strategy that identifies large cap winners and hedges them with equal cash allocated towards an array of short only positions.
A long/short portfolio is neither a new concept, nor is it a novel idea. What makes this approach unique and wildly successful is the way the core long positions are selected and how the hedger is able to subsequently identify the right mix of short positions geared to not only protect the core holdings from market selloff, but also to generate net returns under various market regimes.
The core long holdings are selected based on an assortment of uncorrelated data sets that identify which set of constituents present the highest likelihood of near term market outperformance. Lucena’s AI engine is able to dynamically construct the aforesaid multi-factor stock screeners and incorporate independent data sets into a single cohesive model. Just as an analyst conducts top down or bottom up research, our AI engine constructs models based on corporate earnings data, 10K/10Q filings, insider transactions, news feeds, social media sentiment, analysts’ consensus, fundamentals, technicals, sector level and macro level data, and ultimately additional proprietary data set. It is now evident that a well formed AI engine can build such modes more efficiently, accurately and effectively compared to a human analyst. In addition, just like a human analyst, the algorithm perpetually assesses the model’s efficacy, and as needed, it reconstructs new models to conform to new market regimes.
Once the core holdings are identified, Lucena’s proprietary hedging technology incorporates a unique pattern-matching technology geared to identify a set of constituents that exhibit inverse characteristics relative to the core holdings. By deploying our backtest simulation engine, we were able to empirically demonstrate that when combining the long and short holdings into a single portfolio, it consistently exhibited smooth, low volatility returns that accrue over time.
Image 1: Tech-hedge backtest, 2013 through 2/10/2020. The orange line represents our cash-neutral portfolio against SP 500 (benchmark). The backtest simulaties a 1X leverage, generating a Sharpe ratio of 1.20, with average annualized returns of 20.5%. Click here for a live report.
Past Performance is not indicative of future returns.
Since most backtests could be prone to biases and overfitting, we started to test the concept of the above backtest on a perpetual paper traded portfolio. To ensure authenticity, we adhered to strictly publishing the trades we were about to execute at 7:30AM, well before the market opens. The chart below represents the paper traded portfolio results from 2/11/2020 to today, 4/18/2020 (up 32.44% in 2 months).
Image 2: Tech Hedge paper traded portfolio. February 11 to April 18, 2020, boasting returns of 32.44% with a Sharpe ratio of 5.28 and a much muted max drawdown of 4.88 (Orange line). The portfolio is measured against two benchmarks:
- Low beta market neutral ETF (BTAL) denoted in blue.
- SPY, denoted in green.
It is also Important to note that the portfolio’s max drawdown of 4.88% is vastly muted in comparison to the S&P max drawdown of 33.68% for the same period.
Past performance is not indicative of future returns.
If you wish to follow the daily performance of this model portfolio in the future, feel free to bookmark this link.
Recent extreme volatility and protracted drawdown have caught many investors by surprise. The buy and hold concept that has worked so well in the past 10 years has yielded massive losses to ordinary investors. Conversely, sophisticated hedge funds were able to exploit such market displacement and profit handsomely by employing data science and sophisticated machine learning algorithms. Lucena’s partnership with leading data providers combined with our advanced machine learning and data science capabilities uniquely position us as a technology partner for investment professionals looking to embrace this new reality and regain their competitive edge. If you have any questions and/or you would like to explore our capabilities for your firm, please feel free to contact me directly.
Questions about algorithmic trading? Drop them below or contact us.