Terms that are important to know before you read further:
1. Asset Preservation: A conservative investment strategy where the primary goal is to prevent loss and preserve capital.
2. Genetic Algorithm: An adaptive algorithm which draws inspiration from Charles Darwin’s theory of natural selection. Commonly used to generate solutions for optimization problems, the algo combines random search queries with historical data to provide the best solution for the given query. Here’s a very geek-y explanation of the same.
3. Value-at-Risk: Also known as VaR, this is an important part of asset allocation because it helps advisors determine how much a given portfolio’s value can decline over the years, as a result of changing market prices or rates (Hendricks, 1996). More here.
4. Drawdown: Markets have their ups and downs, and this affects your investments as well. A drawdown is the difference between a peak and the subsequent low of a specific investment. Drawdowns are calculated only once the investment recovers. Let’s say a trading account has $10,000 in it, and then the funds drop to $9,000 before recovering and moving back above the 10k mark. We would say that the trading account witnessed a 10% drawdown. Detailed explanation here.
5. Downside Volatility or Downside Risk: Wiki says it best – “Downside risk is the financial risk associated with losses. That is, it is the risk of the actual return being below the expected return, or the uncertainty about the magnitude of that difference.”
If there is anything that 2019 has taught us, it is that a single tweet from the right (or wrong, depending on perspective) person has the capacity to ruin a bull run at the market.
Volatility has been the norm for much of this year, and in volatile times investors often start to worry about preserving the assets they have in hand. Capital preservation is a term used in the investment industry to describe a couple of specific financial objectives. Preservation is protecting the absolute monetary value of an asset as measured in nominal currency.
Now, if preserving assets were the only criterion one could easily deposit money into say a term deposit or an RD (Recurring Deposit). Herein lies the clincher. No one wants to have their money just lying around doing nothing. When most people talk of asset preservation they do expect a certain amount of monetary increase, or return; all the while keeping the underlying asset intact.
Enter, Kristal.AI’s Asset Preservation Algorithm
On a scale of nothing to dream-like returns, our patented algorithm is handmade to give you returns that are better than – or at the very least, at par with – a term deposit, while minimizing risk.
Now, here’s something we all need to understand. Risk in the investment world is a well-defined entity. Return, on the other hand, can be pretty subjective. For some, a good ‘return’ correlates with getting maximum profits out of their investments. For many others, a ‘return’ is equivalent to being able to achieve a certain financial or personal milestones like repaying their home loan in the next 5 years. An asset preservation model would work best in such a scenario.
Asset Preservation Based on Modern Portfolio Theory
Most modern portfolio optimization methods are extensions of the mean-variance approach proposed by Markowitz which allocates assets based on a risk-return computation. Kristal’s asset preservation algorithm is an extension of this core approach which identifies portfolios with asset preservation as the underlying investor objective.
Given that an exhaustive search for the best portfolio is not feasible, we identify the optimal portfolio using GAIAA. In case you haven’t met her, GAIAA is our home-brewed Genetic Algorithm-based Iterative Asset Allocation model. A true fan of Darwin’s evolutionary theory, GAIAA uses a biology-inspired methodology to select the best portfolio design for every individual’s needs from a pool of thousands of permutations. At each stage, the curated strategies are tested against user-defined objectives, and only the ‘fittest’ strategies are chosen to finally evolve into a risk-limiting model which optimises asset preservation, while maximising user-expected returns.
Inside the Mind of an AI
AI feeds on data. To make our asset preservation algorithm intelligent, we use historical data and simulations of real and improvised stock market scenarios so that the algo can understand market movements and patterns.
Let’s say an advisor has a pool 100+ assets to choose from and create a client portfolio. We’re in effect asking the advisor to curate a subset of these assets (typically <10) which are conservative in nature because asset preservation, while providing returns higher than FD and capable of beating inflation rates. Add to this, the caveat that these chosen assets also have to be less susceptible to market fluctuations – both predicted and unforeseen.
Allowing the AI to learn from simulated scenarios enables it to generate risk-return profiles that are stable over a range of actual stock movement. To this framework, we introduce a customer.
As you can see from the flowchart above, these are the steps we follow to create a single asset preservation portfolio:
- Customer enters their details.
- These details are used to do an initial clustering. For the purpose of this piece, let’s assume we have a pool of around 200 assets. The initial clustering will use the details given by the user (such as their risk profile, their desired investment amount and investment horizon, and other constraints) to reduce the pool from 200 to say, 50 assets which are better-suited for asset preservation.
- In the second step, this curated pool is further reduced using a bootstrap algo, and VaR calculations.
- For the final layer, we optimise. Kristal’s in-house Investment Committee keeps its eyes on the market and prepares recommendations and advice accordingly. These IC views act as thresholds for our genetic algo’s workings, along with fitness functions and other parameters defined by GAIAA.
- The end result is a personalised, hand-picked list of about 5 assets which form the final client portfolio.
But, why trust an algorithm?
Industry reports suggest global algorithmic trading market size is expected to grow from $11.1 billion in 2019 to $18.8 billion by 2024, expanding at a compound annual growth rate (CAGR) of 11.1 per cent. Algorithms build on an expert’s understanding; while reducing bias. We take literature on model building (over 60 years of academia), simulate scenarios with both high and low return; with special emphasis of periods of low return/downturn (using bootstrap). But this is not all. Our IC looks at the results, and confirms thresholds for the algo which will give reasonable returns.
To keep our algorithm robust and on-point with market trends, we actively track recommendations and the algo’s output so that expert-based modifications can be made based on expected market fluctuations. This makes our algo unbiased to short-term fluctuations.
And, To Sum Up…
We have talked about how we have modelled our asset preservation algorithm. One of the limitations of any AI-based system is that it cannot predict the future. An algo learns from the past and uses repeated patterns to suggest the best course in the present tense, but it cannot preempt a Trade War or geopolitical tensions which affect global finances.
Since asset preservation strategies usually have investment horizons longer than 2 years, it is important to have some inkling of the future. This is where our IC steps in (the ‘I’ in our ‘AI’ as we like to call them). Our man-machine allows the IC to rate assets on an on-going basis, change their weightage on client portfolios and also modify and update an asset’s expected returns so that the algorithm’s recommendations are always real-time.
As I have pointed out earlier, algorithms feed on data. The more our algo munches on numbers and user profiles, the stronger it gets. So, while I hope you have enjoyed this introduction to our asset preservation algorithm’s inner workings, I would be happier to know from you, dear reader, what you expect from your term deposits, your long-term investments, and your AI-based investment helpers.
The materials and data contained herein are for information only and shall in no event be construed as an offer to purchase or sell or the solicitation of an offer to purchase or sell any securities in any jurisdiction. Kristal Advisors does not make any representation, undertaking, warranty or guarantee as to the update, completeness, correctness, reliability or accuracy of the materials and data herein. All opinions, forecasts or estimation expressed herein are subject to change without prior notice. Kristal Advisors and its affiliates accept no liability or responsibility whatsoever for any direct or consequential loss and/or damages arising out of or in relation to any use of opinions, forecasts, materials and data contained herein or otherwise arising in connection therewith.