Bank insolvency, how to avoid the subsequent riskNature quantum simulation!
Since the Silicon Valley Bank debacle festered, some analysts are predicting more pain for the sector as concerns spread about the banking industry's hidden risks and its vulnerability to rising currency costs. "There could be a massacre next week because ...... short sellers are out there and they will attack every bank, especially the smaller ones."
The failure of one financial institution can cause a domino effect of more bank failures, leading to systemic risk. As we are now witnessing with the Silicon Valley banking saga and the 2008 financial crisis, this risk can threaten the entire global economy - and the global economy is an intimate and complex network of financial institutions.
Just now, a team of researchers from New York University and the University of Toronto just recently demonstrated, according to the journal Nature, that quantum computers play a role in developing effective strategies to mitigate systemic risk.
Optimizing financial networks to reduce risk
The research team reports that one way to mitigate risk is to optimize the connections between institutions in a financial network, such as aligning loans, holdings and other liabilities that connect banks. However, this task is challenging due to the complexity of financial systems and the potential for nonlinear and discontinuous losses.
This time, the researchers offer some ideas about this complexity. The total number of decision variables grows in four steps with the size of the financial network. In general, the complexity of optimization problems grows dramatically with the number of decision variables, which in turn grows quadratically with the size of the network. Thus, the problem can become intractable when the size of the interbank network is large, which is often the case in real-life banking networks. For example, there are currently about 5,000 commercial banks in North America, according to the FDIC quarterly reports of the Federal Deposit Insurance Corporation.
Examples of interbank network visualization
A cascading simulation visualization of an interbank network with 100 banks, where banks are shown as orange nodes, failed banks are shown as black nodes, and failed banks will cause further losses to other banks through cross-holdings are shown as red edges. The red edge can be interpreted as a pathway for crisis contagion or loss propagation.
New two-stage optimization algorithm
To address this challenge, the researchers developed a two-stage optimization algorithm.
Example of an interbank network using two-stage optimization. The left panel visualizes the circular layout of a randomly generated interbank network of 100 banks; the right panel shows the network after a two-stage optimization using Algorithm 2. Note that the change in cross-holdings (cross-holdings) is not as drastic as before.
In the first stage, the financial network is partitioned into highly interconnected bank modules: this approach improves scalability because it is much easier to optimize smaller bank modules than to optimize the entire network. To partition the network, the researchers developed new algorithms for classical and quantum partitioning of Directed Graphs and Weighted Graphs.
In the second phase, the researchers developed a new method to solve mixed-integer linear programming problems with constraints in the context of systematic risk. This approach was used to optimize the connections within each module of the bank while ensuring that the entire system remains stable.
The results of the Average cascade simulation (ACS) show the number of bank failures relative to the increased perturbed assets (β) for the stochastic network, the one-stage optimization network and the two-stage optimization network.
Quantum simulation to improve solution efficiency
According to the researchers, who used a variety of D-Wave quantum devices and relied on quantum annealing methods, quantum computers are particularly promising for solving these types of problems. The team added that these devices offer significant advantages in terms of computational power and efficiency in solving complex optimization problems.
Comparing quantum and classical partitioning for interbank networks
A time comparison of single-stage optimization, classical and quantum two-stage optimization for an interbank network of 50 banks. The two-stage optimization with quantum partitioning takes much less time.
The team investigated single- and two-stage solutions to these problems, as well as a mixture of classical and quantum approaches. Finally, the experimental results indicated that the two-stage quantum partitioning performed the best.
Finally, the original article concludes, "The two-stage quantum partitioning algorithm outperforms the classical approach. It beats the two-stage classical partitioning because it can alleviate the cascade problem more efficiently. The implementation is also more realistic: real-world results are consistent with the synthetic results, and we demonstrate that the two-stage quantum algorithm is more resilient to financial shocks and delays the cascade failure phase transition (CFPT) and reduces the time complexity under systematic risk."
Reference links：
[1]https://thequantuminsider.com/2023/03/11/timely-research-quantum-computers-could-help-reduce-cascading-financial-crashes/
[2]https://www.nature.com/articles/s41598-023-30710-z#Sec25
