I Turned ChatGPT Into My Robo-Quant Advisor
Take an inside look at how ChatGPT can be used to change the world of quantitative investing.
We’ve all been enamored by the success and capabilities of ChatGPT. But there hasn’t been much talk about how it can be successfully used in the investment world. The model is incredibly smart, so I wanted to give it a real challenge.
I wanted to see if I could use the system to not only offer quantitative and systematic investment advice, but to actually beat the market.
Let’s see how that went:
Testing
Prompt: “I am an entrepreneur that recently sold their company for $5 million dollars. I want to invest this, and my goal is to generate monthly income that can cover my living expenses while still receiving the benefits of capital appreciation. I want to have no more than 25% of the $5 million in the same asset class.
How exactly should I structure my portfolio? Provide the most optimal portfolio based on my goals. Try to provide specific security names, for example, “SPY” for the SPDR S&P 500 ETF. Be as quantitative and systematic as possible, try optimizing for areas like beta, volatility, yield, and correlation. For each suggestion, provide an explanation.”
Response: “To construct an optimal portfolio based on your goals and constraints, we can utilize Modern Portfolio Theory (MPT), which aims to maximize return for a given level of risk. We’ll use quantitative metrics like beta, volatility, yield, and correlation to select securities and allocate your $5 million investment.
Here’s a suggested portfolio allocation with explanations:
- US Equities (25%): $1,250,000
- SPDR S&P 500 ETF (SPY): 50% ($625,000)
- Provides broad exposure to the US equity market
- Moderate beta and volatility for capital appreciation
- Invesco S&P 500 Low Volatility ETF (SPLV): 50% ($625,000)
- Targets the least volatile stocks within the S&P 500
- Low beta and volatility help reduce overall portfolio risk
2. International Equities (25%): $1,250,000
- iShares MSCI ACWI ex U.S. ETF (ACWX): 50% ($625,000)
- Offers diversified exposure to international equities
- iShares MSCI EAFE Minimum Volatility ETF (EFAV): 50% ($625,000)
- Focuses on developed market equities with lower volatility
3. Fixed Income (25%): $1,250,000
- iShares iBoxx $ Investment Grade Corporate Bond ETF (LQD): 50% ($625,000)
- Provides exposure to US investment-grade corporate bonds
- Moderate yield for income generation
- Negative/low correlation with equities for diversification
- iShares U.S. Treasury Bond ETF (GOVT): 50% ($625,000)
- Offers exposure to a range of US Treasury bonds
- Lower yield than corporate bonds but adds safety to the portfolio
4. Real Estate Investment Trusts (REITs) (15%): $750,000
- Vanguard Real Estate ETF (VNQ): 100% ($750,000)
- Diversified exposure to US REITs
- Moderate-to-high yield for income generation
- Moderate correlation with equities for diversification
5. Dividend Stocks (10%): $500,000
- ProShares S&P 500 Dividend Aristocrats ETF (NOBL): 100% ($500,000)
- Targets S&P 500 companies with a track record of increasing dividends
- Moderate-to-high yield for income generation
This allocation provides a balance between income generation and capital appreciation potential while maintaining diversification across asset classes. The selection of specific ETFs and securities optimizes for low correlation and volatility, providing a smoother return profile. Regularly review and rebalance your portfolio to ensure it continues to align with your investment goals and risk tolerance.”
Let’s see how that portfolio performed historically:
Definitely not bad for a language model.
Use-Case
Investment advice can be expensive, which is why it is an industry ripe for disruption. This is where advanced AI models like ChatGPT come in handy.
One of the primary benefits of using it as a “robo-advisor” is it also solves a lot of anxieties about asking “stupid questions” to an advisor. Depending on how recent you need the information to be, you can essentially just flood the system with a stream of all of your investment questions, and you can get the data to back it up; all in seconds.
Not just that, but it also provides investors with a level of objectivity that is difficult to achieve with human advisors. Because it is not influenced by personal biases or emotions, it can analyze data and make recommendations based solely on the numbers. This can help investors make more rational and informed investment decisions, which can potentially lead to better outcomes.
Now, obviously, it isn’t the best idea to buy a stock because a model said so. But it objectively provides a rather robust foundation that you can later build off.
If this article piqued your interest, you’d likely enjoy some of my other posts just like this one:
- Exploiting The Volatility of Volatility
- Arbitraging The Silicon Valley Banking Crisis
- It’s Showtime: Our Sports Betting Algorithm is Live and Profitable! [Code + Data Included]
Happy trading! :)
