Knowledge that Transforms

We study the impact of alternative data on credit access and borrower outcomes using anonymized data from a major U.S. financial technology (fintech) platform that incorporates education, employment, and other nontraditional variables into its underwriting algorithm. The platform approves 15%–30% of low-credit-score applicants rejected by traditional models and offers them lower rates. It particularly benefits “invisible primes”—borrowers with thin credit files and low credit scores but low default risk. We show that gains for invisible primes are primarily driven by alternative data, while model sophistication yields additional improvements in segments where traditional credit report information is more extensive. Using exogenous variation, we find that expanded credit access improves borrowers’ subsequent financial outcomes. This paper has been accepted by Kay Giesecke for the Virtual Special Issue on Digital Finance.

When a government agency considers tightening a standard on a pollutant, the agency often takes into account the proportion of firms that are able to meet the new standard (what we refer to as the industry’s “voluntary adoption level”) because a higher proportion indicates a more feasible standard. We develop a novel model of regulation in which the probability of a stricter standard being enacted increases with an industry’s voluntary adoption level. In addition, in our model, the benefit of a new green technology is both uncertain and correlated across firms, and firms’ decisions exhibit both strategic substitutability (because the marketing benefit of a new green technology decreases as more firms adopt it) and complementarity (because the stricter standard is more likely to be enforced as more firms adopt it). To analyze such strategic interaction among firms’ decisions under correlated and uncertain payoffs, we use the global game framework recently developed in economics. Our analysis shows that regulation that considers an industry’s voluntary adoption level, compared with regulation that ignores it, can more effectively motivate development of a new green technology. Interestingly, uncertainty in the payoff can, in some situations, help promote development of a new green technology. Finally, we find that more aggressive regulation (a higher probability of enforcing a stricter standard for a given voluntary adoption level) encourages more firms to adopt a green technology once the technology becomes available but may discourage a firm from developing it in the first place when facing intense competition. Therefore, for an industry with intense competition, a government agency should exercise caution about being too aggressive with regulation, which could potentially stifle innovation.

This study contributes to our understanding of what accruals capture and how they relate to the distribution of future returns. It examines the past and future growth components of accruals and shows that, whereas past growth is negatively associated with idiosyncratic skewness, future growth is positively associated with it. In addition, although both past and future growth are negatively associated with future returns, the association is more pronounced for past growth when volatility is lower, but for future growth when volatility is higher. The study also shows that the association between the future growth component and future returns reverses in the long run, whereas the association between past growth and future returns does not.

As a platform improves trust between the two sides of its market to facilitate matching and transactions, it faces an increased risk of disintermediation: with sufficient trust, the two sides may circumvent the platform to avoid the platform’s fees. In this paper, we investigate the relationship between increased trust and disintermediation by leveraging a randomized control trial in an online freelance marketplace. We find that enhanced trust increases the likelihood of high-quality freelancers being hired. However, when the trust level is sufficiently high, it also increases disintermediation, which offsets the revenue gains from the increase in hiring high-quality freelancers. We also identify heterogeneity across clients and freelancers in their tendencies to disintermediate. We discuss strategies that platforms can use to mitigate the tension between trust building and disintermediation.

How should a firm price a new product for which little is known about demand? We propose a simple and practical pricing rule for new products where demand information is limited. The rule is simple: Set price as though the demand curve were linear. Our pricing rule can be used if three conditions hold: the firm can estimate the maximum price it can charge and still expect to sell some units, the firm need not plan in advance the quantity it will sell, and marginal cost is known and constant. We show that if the true demand curve is one of many commonly used demand functions, or even a more complex (randomly generated) function, the firm can expect its profit to be close to what it would earn if it knew the true demand curve. We derive analytical performance bounds for a variety of demand functions, calculate expected profit performance for randomly generated demand curves, and evaluate the welfare implications of our pricing rule. We show that with limited demand information (maximum price and marginal cost), our simple pricing rule can be used for new products while often achieving a near-optimal performance. We also discuss the limitations of our method by identifying cases where our pricing rule does not perform well.

This study examines cross-regional differences in news tone for the same news event and relates it to local stock ownership of individual investors using a combination of detailed investor and media data. The results show that news tone amplifies the overall attention effect and decreases the difference between the trading activities of local and nonlocal investors when locally sourced news is republished across regions. In general, trading activity decreases for given republished news, but it still exists because the same news event garners attention as it diffuses to other regional outlets in a different news tone. Overall, individual investors seem to realize that the news is stale but expect subsequent price reversals following the republished news, thus showing news-contrarian trading behavior. Furthermore, they distinguish good firms from bad firms, indicating that adverse selection is partially responsible for the observed news-contrarian trading behaviors.

This paper uses the 65th anniversary of Management Science as an opportunity to examine how well the journal has performed against its primary goal to promote development and encourage application of a science of management. We also highlight recent trends that are impacting this mission and discuss how recent initiatives address these issues.

We assess the impact of geographic diversification on a bank’s costs of interest-bearing liabilities. We employ a new identification strategy and discover that geographic expansion across U.S. states lowered funding costs. Consistent with expansion facilitating risk diversification, we find that (1) funding costs fall more when banks expand into states whose economies are less correlated with the banks’ state and (2) geographic diversification reduces the costs of uninsured, but not insured, deposits. Consistent with expansion intensifying agency frictions, which puts upward pressures on funding costs, we discover that geographic diversification reduces the costs of interest-bearing liabilities more in better-monitored and better-run banks.

We report the results of an experiment in which a company, Firm Vary, temporarily suspended its sponsored search advertising campaign on Google in randomly selected advertising markets in the United States. By shutting off its ads, Firm Vary lost customers, but only 63% as many as a nonexperimental estimate would have suggested. Following the experiment, Firm Vary merged with its closest competitor, Firm Fixed. Using combined data from both companies, the experiment revealed that spillover effects of Firm Vary’s search advertising on Firm Fixed’s business and its marketing campaigns were surprisingly small, even in the market for Firm Vary’s brand name as a keyword search term, where the two firms were effectively duopsonists.

Information systems generate copious trace data about what individuals do and when they do it. Trace data may affect the resolution of lawsuits by, for example, changing the time needed for legal discovery. Trace data might speed resolution by clarifying what events happened when, or they might slow resolution by generating volumes of new and potentially irrelevant data that must be analyzed. To investigate this, we analyze the effect of electronic medical records (EMRs) on malpractice claim resolution time. Use of EMRs within hospitals at the time of the alleged malpractice is associated with a four-month (12%) reduction in resolution time. Because unresolved malpractice claims impose substantial costs on the entire healthcare system, our finding that EMRs are associated with faster resolution has broad welfare implications. Furthermore, as we increasingly digitize society, the ramifications of trace data on legal outcomes matter beyond the medical context.