Skip to main content

Preliminary Agenda

The two-day conference will feature a keynote address and eight sessions of researchers presenting their work with members of CFPB’s Office of Research serving as discussants. The conference will be held in person at CFPB headquarters at 1700 G Street NW, Washington DC.

In person registration is open until midnight December 11th. Livestream registration is now open.

Thursday, December 15, 2022

Check-in and breakfast (1 hour)

Opening remarks (15 minutes)

Buy now, pay later credit: User characteristics and effects on spending patterns

  • Presenting author: Emily Williams (Harvard Business School)

Firms offering “buy now, pay later” (BNPL) point-of-sale installment loans with minimal underwriting and low interest have captured a growing fraction of the market for short-term unsecured consumer credit. We provide a detailed look into the US BNPL market by constructing a large panel of BNPL users from transaction-level data. We document characteristics of users and usage patterns and use BNPL roll-out to provide new insights into consumer responses to unsecured credit access. BNPL access increases both total spending levels and the retail share in total spending, with magnitudes too large for standard intertemporal and static substitution effects to explain. These findings hold for consumers with and without inferred liquidity constraints. Our findings are more consistent with a “liquidity flypaper effect” where additional retail liquidity through BNPL “sticks where it hits”, than a standard lifecycle model with liquidity constraints. (Joint with Justin Katz and Marco Di Maggio)

Buy Now Pay (Pain?) Later

  • Presenting author: Ben Lourie (University of California, Irvine)

“Buy Now Pay Later” (BNPL) is a largely unregulated FinTech innovation that provides consumers with easy access to credit for specific retail purchases. The BNPL market is projected to reach $1 trillion by 2025, but we know little about the effects of BNPL on consumers’ financial wellbeing. Using banking data for 10.6 million U.S. consumers, we find that first-time BNPL users experience rapid increases in overdraft charges and credit card interest and fees, as compared to non-users. An instrumental variable exploiting consumers’ pre-BNPL shopping habits increases the credibility of BNPL having a causal negative effect on users’ financial wellbeing. Our results contribute to the academic literature by expanding our understanding of a major development in household finance, and indicate that regulators should take seriously the concern that BNPL could have significant negative welfare implications. (Joint with Ed deHaan, Jungbae Kim, and Chenqi Zhu)

Breakout (15 minutes)

Is there crowd-out in mortgage refinance?

  • Presenting author: Nick Frazier (FDIC)

We examine whether supply-side capacity constraints contribute to the welldocumented “failure to refinance” among certain borrowers who would benefit financially from doing so. Using comprehensive loan-level data from the National Mortgage Database (NMDB), we show that, conditional on the potential financial benefits of refinancing and other observables, borrowers with low loan balances, low incomes, or low credit scores have substantially lower mortgage prepayment rates during refinance booms compared to periods of lower refinancing activity. In contrast, borrowers with high loan balances, high incomes, or high credit scores have higher prepayment rates when markets are operating at or near full capacity. These patterns hold after controlling for a rich set of observable characteristics and among borrowers that are likely able to qualify for a conventional refinance loan, indicating that some borrowers are crowded out from refinancing when capacity constraints bind. Overall our findings suggest that in addition to demand-side explanations for differences in refinancing highlighted in previous literature, supply-side factors also play an important role. (Joint with Ryan Goodstein)

Frictions in Mortgage Refinancing

  • Presenting author: Gaston Illanes (Northwestern University)

Households often fail to refinance mortgages, foregoing substantial reductions in interest payments. We implement a large-scale experiment on the mortgage borrowers in a Chilean bank to study whether information interventions affect refinancing behavior. We randomly assign borrowers to receive treatments that target distinct information frictions, and track their search and refinancing behavior over time. We find that the most effective intervention—giving information about gains from refinancing and details about how to refinance—increases search for refinancing offers by 20%, with significantly larger effects for those with the largest gains from refinancing. We develop and estimate an equilibrium model of refinancing to evaluate alternative policies that affect refinancing frictions, as well as its implications for market power and the passthrough of monetary policy. (Joint with Vivek Bhattacharya, José Ignacio Cuesta, Ana Maria Montoya, and Raimundo Undurraga)

Closing Costs, Refinancing, and Inefficiencies in the Mortgage Market

  • Presenting author: David Zhang (Rice University)

I use a structural model to quantify the cross-subsidization in the US mortgage market due to heterogeneous borrower refinancing tendencies. Actively refinancing borrowers gain up to 3% of their loan amount relative to non-refinancing borrowers in expectation. In equilibrium, the presence of borrowers with high refinancing inertia reduces mortgage interest rates particularly on lower upfront closing cost mortgages which have more valuable refinancing options. As a result, actively refinancing borrowers refinance excessively relative to a perfect information, no cross-subsidization benchmark, an effect that accounts for around 28% of all US refinancing and generates significant deadweight losses due to administrative resource costs. Policies that reduce disparities in refinancing behavior can increase efficiency in the market while also having attractive distributional properties.

Breakout (30 minutes)

Credit When You Need It

  • Presenting author: Ben Keys (The Wharton School of the University of Pennsylvania)

We estimate the causal effect of emergency credit provision on the household balance sheet. To do so, we link application data from the U.S. Federal Disaster Loan program, which provides loans to households that have uninsured damages from a federally-declared natural disaster, to a panel of credit records before and after the shock. We exploit discontinuous variation in loan approval rules, which led those with debt-to-income ratios below 40% to be differentially likely to be approved. Using an instrumented difference-in-differences research design, we find that credit provision at the time of a shock can significantly reduce the worst outcomes, reducing the likelihood of loan delinquency by 33% and the likelihood of bankruptcy by 2 percentage points. Credit provision in a time of crisis has real consumption effects in the form of additional car purchases even 3 years after loan receipt. Our findings suggest that well-timed liquidity provided to households in acute need can have substantial and persistent positive effects. (Joint with Ben Collier, Dan Hartley, and Xian Ng)

The Impact of Social Insurance on Household Debt

  • Presenting author: Sasha Indarte (The Wharton School of the University of Pennsylvania)

This paper investigates how the expansion of social insurance affects households’ accumulation of debt. Insurance can reduce reliance on debt by lessening the financial impact of adverse events like illness and job loss. But it can also weaken the motive to self-insure through savings, and households’ improved financial resilience can increase access to credit. Using data on 10 million borrowers and a quasi-experimental research design, we estimate the causal effect of expanded insurance on household debt, exploiting ZIP code-level heterogeneity in exposure to the staggered expansions of one of the largest US social insurance programs: Medicaid. We find that a one percentage point increase in a ZIP code’s Medicaid-eligible population increases credit card borrowing by 0.74%. Decomposing this effect in a model of household borrowing, we show that increased credit supply in response to households’ improved financial resilience fully accounts for this rise in borrowing and contributed 33% of the net welfare gains of expanding Medicaid. (Joint with Gideon Bornstein)

Pay-As-You-Go Insurance: Experimental Evidence on Consumer Demand and Behavior

  • Presenting author: Raymond Kluender (Harvard Business School)

Pay-as-you-go contracts reduce minimum purchase requirements which may increase market participation. We randomize the introduction and price(s) of a novel pay-as-you-go contract to the California auto insurance market where, despite a universal mandate, 17 percent of drivers are uninsured. The pay-as-you-go contract increases insurance take-up by 10.6 percentage points (87%) and days with insurance available by 4.5 days over the 3-month experiment (26%). Demand is price-sensitive and coverage increases are smaller at higher prices. The pay-as-you-go contract increases insurance coverage in part by relaxing liquidity requirements: purchase behavior for more than half of drivers is consistent with a cost of credit in excess of payday lending rates and 19 percent of enrolled drivers have a purchase rejected for insufficient funds. Insurance coverage converges between the traditional and pay-as-you-go contract over time. I discuss potential explanations and their implications for similar financial technologies (e.g., buy-now-pay-later, earned wage access) and the uninsured driver problem.

Breakout (30 minutes)

Attention Constraints and Financial Inclusion

  • Presenting author: Jiacui Li (University of Utah David Eccles School of Business)

We show that attention constraints of decision makers function as barriers to financial inclusion. Using administrative data on retail loan screening processes, we find that loan officers exert less effort reviewing applicants from unattractive social or economic backgrounds and reject them more frequently than justified by credit quality. More importantly, when quasi-random workload variations tighten officer attention constraints, unattractive applicants receive even worse treatment—review-time halves and approval rates drop by approximately 40%—while attractive applicants are not affected. Our findings suggest that financial technologies that reduce information-processing costs may promote more balanced financial access. (Joint with Bo Huang, Tse-Chun Lin, Mingzhu Tai, and Yiyuan Zhou)

Do Investors Read the Fine Print? Salient Thinking and Security Design

  • Presenting author: Petra Vokata (The Ohio State University Fisher College of Business)

Banks use financial engineering to distort product headline rates upward by adding value-decreasing conditions. I refer to such engineering as headline distortion. If investors overweight headline rates because they are saliently advertised, normatively irrelevant headline distortion may affect choice. I find results consistent with such salient thinking. Controlling for fair values of 28,000 retail investment products, a one percentage point increase in headline distortion leads to an increase in sales equivalent to a 0.5 pp reduction in fees. To identify a causal effect of headline distortion, I show that shocks to the structuring costs of headline rates affect demand. Yet, they have no effect once headline rates are fixed at the beginning of the offering period. Products with more distorted headline rates charge higher fees and deliver lower returns. Banks distort more when disclosure or market conditions make the products appear less attractive.

Breakout (15 minutes)

Robo-Advice for Borrower Repayment Decisions

  • Presenting author: Jonathan Shaw (Financial Conduct Authority [UK] and Institute for Fiscal Studies)

Poor debt-management skills lower financial security and wealth accumulation. And yet, most households, including the most vulnerable, are left to their own means when making repayment decisions. Because optimal solutions to credit repayment problems depend on neither risk preferences nor beliefs, loan repayment is a natural application for robo-advising. At the same time, the vulnerable households who would benefit most from robo-advising tend to distrust new technologies and override suggestions that do not align with ingrained heuristics, such as matching the minimum payment on a credit card balance. Lower adoption rates by these groups might increase rather than reduce wealth inequalities. To assess these trade-offs, we design and implement an RCT in which robo-advice for borrower repayment decisions is offered to a set of representative UK consumers. On average, the availability of free robo-advice significantly improves loan repayment choices. When asked about their willingness to pay, many subjects report values larger than the monetary benefits of the tool. Non-adopters and overriders report lower trust in algorithms at the end of the experiment. Providing tips alongside robo-advising barely improves subsequent unassisted choices, suggesting the lack of learning from using robo-advice. In fact, learning-by-doing is highest for those who make all choices unassisted. (Joint with Ida Chak, Karen Croxson, Francesco D’Acunto, Jonathan Reuter, and Alberto Rossi)

How Do Borrowers Respond to a Debt Moratorium? Experimental Evidence from Consumer Loans in India

  • Presenting author: Martin Kanz (World Bank)

Debt moratoria that allow borrowers to postpone loan payments are a frequently used tool to soften the impact of economic crises. One concern with such policies is that they might give rise to moral hazard by changing borrower beliefs about credit enforcement and the likelihood of future relief. We partner with a large consumer lender in India to issue randomized debt forbearance offers to a nationwide sample of borrowers. In the experiment, borrowers receive identical forbearance offers that are presented either as an act of generosity by the lender or as the result of government regulation. We find that delinquent borrowers who are offered a debt moratorium by their lender are 4 percentage points (6.9 percent) less likely to default on their loan, while forbearance has no effect on repayment if it is granted by the regulator. Borrowers receiving forbearance offers from their lender are also more likely to do future business with the lender. In a follow-up experiment we find that demand for the lender’s products is 16.3 percentage points higher among customers who were offered repayment flexibility by their lender than among customers who received a moratorium offer presented as an initiative of the regulator. Overall, our results suggest that, rather than generating moral hazard, debt forbearance can improve loan repayment and lead to higher value banking relationships. (Joint with Stefano Fiorin and Joseph Hall)

Happy Hour

Friday, December 16, 2022

Check-in and breakfast (1 hour)

Bank Competition amid Digital Disruption: Implications for Financial Inclusion

  • Presenting author: Jinyuan Zhang (UCLA Anderson School of Management)

This paper studies how banks compete amid digital disruption and the resulting distributional effect across consumers. Digital disruption increases the geographic coverage of banking services, bringing new entrants to local markets. However, as digital customers shift from branches to digital services, banks close branches, and the remaining branching banks gain market power among non-digital customers that rely on branches. Consequently, digital customers benefit from the intensified bank competition at the cost of non-digital customers who pay higher prices for branch services and face the risk of financial exclusion. We provide empirical evidence by exploiting the staggered expansion of 3G networks, instrumented by regional distribution of lightning strike frequency. Using a structural model, we further quantitatively decompose the benefit and costs of digital disruption resulting from banks’ pricing, branching, and entry decisions. The results highlight the role of banks’ endogenous responses to digital disruption in widening the gap in access to banking services. (Joint with Erica Xuewei Jiang and Gloria Yang Yu)

How Much Do Small Businesses Rely on Personal Credit

  • Presenting author: Julia Fonseca (University of Illinois at Urbana-Champaign)

This paper estimates the degree of substitution between personal and small business credit for U.S. entrepreneurs between 2009 and 2018 using a novel, individuallevel dataset. We identify the effect of business credit supply shocks by exploiting geographic variation in the market share of large banks, which sharply reduced credit supply to small businesses relative to other banks after the 2008 financial crisis. While this contraction decreased total business credit by $13,572 per firm in our sample, we find that entrepreneurs were able to substitute about 68% of this decline with personal credit, driven by mortgages. Entrepreneurs with subprime credit scores, below average income, and high credit utilization do not substitute lost small business credit with personal credit, suggesting that personal financial characteristics of entrepreneurs meaningfully affect the overall access to external finance of small businesses. (Joint with Jialan Wang)

Breakout (15 minutes)

Strategically Staying Small: Regulatory Avoidance and the CRA

  • Presenting author: Jacelly Cespedes (University of Minnesota)

Using the introduction of an asset based two-tiered evaluation scheme in the 1995 CRA reform, we examine the consequences of regulatory avoidance. Banks exploit the attribute-based regulation by strategically slowing asset growth, bunching below the $250M threshold. The regulatory avoidance also produces real effects. Banks near the threshold experience an increase in the rejection rate of LMI loans, while areas they serve experience a decline in county-level small establishment shares and independent innovation. These results highlight a bank’s willingness to take costly actions to avoid regulatory oversight and subsequent credit reduction for individuals whom the CRA is designed to benefit. (Joint with Jordan Nickerson, Carlos Parra)

Bank Access Across America

  • Presenting author: Alexander Zentefis (Yale University)

We use location data from mobile devices to construct a granular measure of bank access throughout the United States. The measure originates from a spatial gravity model and is an expression of a local area’s distance from bank branches and branch characteristics. To overcome methods that protect user privacy in the mobile device data, we estimate the access measure using the Method of Simulated Moments (MSM), and we introduce an econometric method that can handle thousands of fixed effects in MSM routines. We find that residents of both low-income areas and areas with higher Black population shares use branches less, but for different reasons. Residents of poorer areas experience greater access but exhibit lower demand, which leads to less branch use. In contrast, weaker access, rather than lower demand, explains the entire drop off in branch use among residents of block groups with higher Black population shares. (Joint with Jung Sakong)

Breakout (15 minutes)

FinTech Lending with LowTech Pricing

  • Presenting author: Mark Johnson (Brigham Young University)

Fintech lenders are known for the use of alternative data and sophisticated technologies in delivering financial services. However, based on the pricing and performance of over two million unsecured personal fintech loans, pricing appears rather simplistic and does not necessarily correspond to default likelihoods. Pricing is oversensitive to credit score bins, including a substantial interest rate jump of 9.4% around the nonprime cutoff. Pricing is insensitive to known predictors of default (e.g., affordability, location), leading to borrower cross-subsidization. Pricing patterns are influenced by institutional factors, e.g., lack of competition from banks, originators’ incentives, and low demand for funding risky lending. (Joint with Itzhak Ben-David, Jason Lee, and Vincent Yao)

Invisible Primes: Fintech Lending with Alternative Data

  • Presenting author: Dimuthu Ratnadiwakara (Louisiana State University)

We exploit anonymized administrative data provided by a major fintech platform to investigate whether using alternative data to assess borrowers’ creditworthiness results in broader credit access. Comparing actual outcomes of the fintech platform’s model to counterfactual outcomes based on a “traditional model” used for regulatory reporting purposes, we find that the latter would result in a 70% higher probability of being rejected and higher interest rates for those approved. The borrowers most positively affected are the “invisible primes”–borrowers with low credit scores and short credit histories, but also a low propensity to default. We show that funding loans to these borrowers leads to better economic outcomes for the borrowers and higher returns for the fintech platform. (Joint with Marco Di Maggio and Don Carmicheal)

Keynote Speaker: Sendhil Mullainathan

2:15 Adjourn