Quantum Finance: How Quantum Computers Will Change The World of Quantitative Finance
“Quants”: An Overview
Quantitative Analysts, or “quants”, use complex algorithms and mathematical models to analyze data and predict outcomes, critical skills in the world of stock trading. Before the 1970s and 80s, successful traders sat down with CEOs and economists to gain exclusive information that, when combined with their intuition, allowed them to hedge educated bets on stocks. As market news became digitized, people with a background in math began to realize that stock markets displayed patterns, and complex models could be used to predict whether a stock was going up or down. They started using computers to analyze vast amounts of data, putting information through algorithms designed by mathematicians and computer scientists that returned instructions on buying or selling a particular stock. As computer technology advanced, algorithms became more complex, and the field of quantitative finance grew. Today, almost every trader, from day traders to massive hedge funds, uses research done by quants to help them decide which stocks to buy and sell.
Where Quantum Computers Come In
The key to “winning” in the trading game is receiving and analyzing information fast. Very fast. Being one of the first to place a trade based on new information means making more money than others who placed the same trade a few seconds later. Although classical computers have become faster over time, they all rely on the same basic principles: performing one computation at a time to arrive at an answer. This puts severe limitations on the amount of information computers can process. Quantum computers, however, can solve in a few hours what could take computers hundreds of years to complete by doing multiple calculations at once.
One example of how quantum computers can revolutionize the world of quantitative finance is their optimization of the Monte Carlo Integration (MCI), an indispensable tool in risk modeling. The Monte Carlo Integration can be visualized by a random game of darts. Picture a wall with an invisible dartboard. A computer randomly throws a dart at the wall and receives information about whether it hit the dartboard or not. The computer will repeat this process thousands of times until it can estimate the location and size of the board based on the number of successful hits obtained. This process can be used for risk modeling by simulating complex events thousands of times and then predicting outcomes by taking random samples of the simulations. The estimation error "e" gives analysts the distance between the true value and the estimated value in order to quantify the accuracy of the algorithm. For classical computers, the time complexity (the average time it takes for a computer to complete an algorithm based on the number of samples) to achieve the estimation error is O(σ²/ε²), where σ is the standard deviation of all samples. Yet, there is a quantum algorithm that can obtain ε with O(σ/ε) time complexity and has a constant probability of success. This means that the quantum algorithm can reduce the time complexity of a critical process in quantum finance while obtaining a more consistent finding of the accuracy of the value obtained by Monte Carlo Integration (Herman et al., 2023). JP Morgan Chase has already begun taking advantage of this finding. A paper co-written by JP Morgan Chase quants and IBM researchers details how quantum computers can price simple options with less error than classical computers using the IBM Q Tokyo quantum device (Stamatopoulos et al., 2020).
How Quantum Computers Work
The explanation for these promising findings is the use of qubits. Most computers use bits to store all of their information. A bit can be either a 1 or a 0, representing “on” or “off.” Using millions of bits, computers can store complex information such as numbers and letters. The limitation of bits, however, is that they only have two states of being, meaning a computer must use multiple processors or time-sharing methods to perform numerous tasks at the same time, whereas a qubit relies on the quantum principle of superposition to be in a combination of the 0 and 1 states simultaneously. The probability amplitude (a complex number in quantum mechanics used to describe a system's state) represents the probability that a measured qubit will fall under the 0 or 1 state. Qubits can use superposition to enable quantum parallelism, which quantum computers use to explore multiple possibilities in parallel, a beneficial function for algorithms that require checking millions of possibilities to find the solution, such as the MCI.
Limitations of Current Quantum Computers
If quantum computers are so efficient, why are they not in the corner of every Wall Street office? Quantum computing is a new technology. In theory, they should outperform contemporary computers when calculating complex problems, but they are still far from being used by quants. For one, quantum computers process information differently than other computers, meaning data must first be translated into quantum information before being processed and then translated back to information we can understand. This adds significant time to any computations done by the quantum computer, the opposite of what quants are looking for. Researchers are heavily invested in solving this problem, known as the “quantum-classical interface problem”, by designing hybrid algorithms that use both quantum and classical computers in order to minimize the amount of information that must be translated to and from quantum languages. Secondly, the number of qubits in quantum computers limits their abilities. Currently, the most advanced quantum computers contain about 1000 qubits. The number of qubits needed to implement quantum algorithms scales linearly with the number of steps that need to be performed, meaning a quantum computer that would be useful to quants would require anywhere from hundreds of thousands to millions of qubits. Finally, quantum computers must be kept in cold conditions for qubits to work properly. How cold? -450 degrees Fahrenheit, a few degrees above absolute zero (Zewe, 2023). They are cooled using large, complex, and expensive refrigerators. When we look at images of quantum computers, what we really see is the cooling unit. The actual quantum computer is a chip stored inside of the cooler. This is perhaps the biggest reason quantum computers are currently used for research and not practical applications. The cost of temperature regulation is a massive investment that provides too little return for profit-maximizing firms. Researchers are working towards two main solutions to this issue: more efficient cooling systems and qubits that can perform at room temperature. While significant progress is being made, these limitations must be overcome before quantum computers can achieve widespread practical use in fields like finance. Until then, their potential remains largely theoretical, as researchers continue to work on improving performance, scalability, and cost-efficiency.
Conclusion
Quantum computers are currently unable to model complex market movements and place trades, but at the rate this new technology is advancing, it may not be long before they are used for much more. IBM plans to build the first 100,000 qubit quantum computer by 2033, which could be a massive step forward in the race to the first million qubit computer. When these computers reach the capabilities necessary to be used by quants, the financial world will change forever. Researchers will be able to code quantum algorithms that act on information faster than ever, and “quant” will take on a whole new meaning: quantum analyst.
Works Cited
“10 Quantum Computing Applications & Examples to Know.” Built In, builtin.com/hardware/quantum-computing-applications. Accessed 22 Sept. 2024.
Adam Zewe | MIT News Office. “A New Way for Quantum Computing Systems to Keep Their Cool.” MIT News | Massachusetts Institute of Technology, news.mit.edu/2023/new-way-quantum-computing-systems-keep-their-cool-0221. Accessed 22 Sept. 2024.
“Charting the Course to 100,000 Qubits.” IBM Quantum Computing Blog, www.ibm.com/quantum/blog/100k-qubit-supercomputer. Accessed 22 Sept. 2024.
Herman, D., Googin, C., Liu, X., Sun, Y., Galda, A., Safro, I., Pistoia, M., & Alexeev, Y. (2023). Quantum computing for finance. Nature Reviews Physics, 5(8), 450–465. https://doi.org/10.1038/s42254-023-00603-1
IonQ Newsroom and media resources. (n.d.). IonQ. https://ionq.com/posts/quantum-computing-monte-carlo-algorithms-and-financial-modeling
Stamatopoulos, N., Egger, D. J., Sun, Y., Zoufal, C., Iten, R., Shen, N., & Woerner, S. (2020). Option Pricing using Quantum Computers. Quantum, 4, 291. https://doi.org/10.22331/q-2020-07-06-291
