By Pablo Azar and Sergey Gorbunov
There is a seemingly impossible tradeoff between privacy and systemic stability in the financial industry. Financial institutions are reasonably reluctant to reveal their portfolio positions and trading partners. These are important business secrets, which if revealed out into the open would grind the financial industry to a halt. On the other hand, regulators need access to financial institutions’ information in order to ensure that there is a properly functioning market, and that the collapse of one institution—as happened with Lehman Brothers in 2008—cannot bring down the entire financial system.
In the aftermath of the 2008 financial crisis, regulations were put in place to ensure that financial institutions had to reveal their information to government entities and register their trades with trade repositories. The Dodd-Frank Act of 2010 requires significantly important financial institutions (SIFIs) to report detailed information about their counterparties, leveraged positions and overall risk exposures. This information allows the Federal Reserve to perform stress tests, which simulate the effects of severe recessions on the financial system. The Dodd-Frank Act also imposed the requirement that financial derivatives be cleared through central counterparties, which are required to report their risk exposures to regulators.
These reforms were instrumental in mitigating the excesses that led to the financial crisis, but they threaten to introduce new types of systemic risk in the financial system arising from centralization. For example, trade repositories and central counterparties may become the targets of state-sponsored hackers, who may want to leak sensitive financial information.
The tradeoff between privacy and stability seems unavoidable, but it does not have to be. For example, Abbe, Khandani and Lo propose a new secret sharing protocol to compute correlations and covariances---which form the basic building blocks of systemic risk measures. They also show how a secret sharing protocol for sums can be used to report the aggregate amount of real estate loans made by Bank of America, Wells Fargo and JP Morgan without revealing individual banks’ positions.
Beyond correlations and sums, more sophisticated cryptographic protocols have been developed to compute arbitrary functions. Together with Oded Goldreich and Avi Widgerson, Algorand’s founder Silvio Micali invented the first protocol for secure multi-party computation that can support arbitrary functions. That is, if a group of n people, each holding an input x1, wants to evaluate a function y = f(x1,...,xn), they can participate in a cryptographic protocol where each of them only learns the output y without learning anything else about others’ inputs. Regulators can use these kinds of cryptographic protocols to compute arbitrary risk models on banks’ balance sheets without revealing the details of individual deals.
At Algorand, we have the expertise to implement the tools that the financial industry needs. Silvio has assembled a world-class team that makes state-of-the-art cryptography available to everyone. This team has already improved blockchain scalability with the use of Verifiable Random Functions (VRF), standardization of Boneh-Lynn-Shacham aggregate signatures, and design of new Pixel-signatures for consensus. With expertise in both economics and cryptography, the team at Algorand is excited to continue developing cryptographic innovations that can enable new markets in the borderless economy.