One of the prominent potential applications for federated machine learning is in detecting financial crime risks across multiple institutions which cannot share data with each other due to confidentiality and other regulatory restrictions. This article delves into the recent growth of financial crime in congruence with failing financial crime compliance and monitoring systems. The author describes how privacy enhancing technologies such as federated machine learning could help to overcome information sharing restrictions in relation to financial crime compliance and monitoring.
Using real time anonymized data from private companies, this paper focuses on the ripple effects of a sharp decrease in spending by high-income households on both small businesses and low-income workers.
The blog analyzes survey results from the U.S. Census Bureau that provide the first real-time, national data on housing payments disaggregated by race and ethnicity, income, age, and other household characteristics. Among both renters and homeowners, African-Americans, Latinos, and low-income households were more likely to miss or defer housing payments in May than other groups.
The blog links both prior and forthcoming research on when consumers are likely to stop paying their mortgages due to a “double trigger” from income shocks due to job losses or other events that then cause drops in housing prices. The authors conclude that income support to households and small businesses in the wake of the COVID-19 pandemic has limited disruption to date in the housing market in addition to supporting individual families.
This article examines changes in mortgage lending, including tightening of lending standards and narrowing of product offerings, since the passage of the CARES Act. The authors compare these changes to lenders’ reaction to the 2008 financial crisis and assess how they affect broader economic recovery.
Using online retail search data, this article explores how dramatic shifts in consumer behavior during the quarantine affected algorithms used to manage inventory, sell ads, and screen for fraud. It underscores the role of informed and timely governance – including human intervention – to ensure algorithm performance.
This report tracks consumer spending patterns in the weeks after receiving federal stimulus payments using debit card data for more than 16,000 recipients. Consumers who live paycheck-to-paycheck spent 68% of the funds in the first two weeks, while higher-income consumers and those who generally save a significant portion of their income spent an average of 23% in the same time period.
Using data from the nation’s largest payroll processor from February through May, this paper finds that employment losses were disproportionately concentrated among low-income workers while wage cuts were disproportionately concentrated among workers in the top two deciles of the wage distribution. The percent of workers receiving wage cuts was roughly twice that
reported during the Great Recession.
This article examines the prospects for minority-owned businesses for navigating the economic crisis brought by COVID-19, given that they are less likely to have a financial cushion, are concentrated in communities and industries severely affected by the pandemic, and may face particular challenges participating in relief programs.
The IRS has had trouble getting stimulus money to people quickly because of reliance on a network of tax preparation intermediaries to distribute funds. This article explores whether this infrastructure disproportionately affected low- and moderate-income people who were eligible for relief.