Research

We document widespread use of personal financial advice among retail investors. Individuals seek competent and trusted sources for financial advice among their family and friends. Investors who provide advice to family and friends are positively selected and emphasize the reputational costs of giving risky financial advice. While previous studies have shown that advice shared on social media promotes active trading, we show that personal financial advice encourages investing in funds over single stocks. Our evidence complements the existing literature on financial advice in online social networks by highlighting differences in incentives and outcomes of advice to close personal connections.

How does the design of debt repayment schedules affect household borrowing? To answer this question, we exploit a Swedish policy reform that eliminated interest-only mortgages for loan-to-value ratios above 50%. We document substantial bunching at the threshold, leading to 5% lower borrowing. Wealthy borrowers drive the results, challenging credit constraints as the primary explanation. We develop a model to evaluate the mechanisms driving household behavior and find that much of the effect comes from households experiencing ongoing flow disutility to amortization payments. Our results indicate that mortgage contracts with low initial payments substantially increase household borrowing and lifetime interest costs.

What are the aggregate and distributional consequences of the relationship between an individual's social network and financial decisions? Motivated by several well-documented facts about the influence of social connections on financial decisions, we build and calibrate a model of stock market participation with a social network that emphasizes the interplay between connectivity and network structure. Since connections to informed agents help spread information, there is a pivotal role for factors that determine sorting among agents. An increase in the average number of connections raises the average participation rate, mostly due to richer agents. A higher degree of sorting benefits richer agents by creating clusters where information spreads more efficiently. We show empirical evidence consistent with the importance of connectivity and sorting. We discuss several new avenues for future research into the aggregate impact of peer effects in finance.

We study how mortgage innovation can cause a housing boom even within a robust regulatory framework and strictly enforced recourse borrowing. Specifically, we find that the 2003 introduction of interest-only (IO) mortgages in Denmark ignited a housing boom that increased house prices 36 percent. In line with IO loans lowering debt-service payments and relaxing payment-to-income constraints, results show higher IO loan uptake and house price growth in areas with greater ex-ante benefits of such mortgages. Overall, our results are relevant for the many countries where IO loans play a sizable role in mortgage finance.

We use household-level data to analyze how the introduction of interest-only (IO) mortgages in Denmark affected consumption expenditure and borrowing. Using an ex ante measure of exposure to the IO mortgage reform motivated by mortgage-payment and leverage constraints, we show households more likely to use an IO mortgage to relax their mortgage-payment constraint increased consumption following the reform. This increase in consumption is financed by borrowing at the time of refinancing and by borrowers with lower pre-reform leverage and higher needs for liquidity. We find even larger post-reform consumption growth for the leverage-constrained homeowners through house-price growth stimulated by the reform.

More than 20 million Americans are affiliated with multi-level marketing firms (MLMs), but there is little empirical evidence on who participates in this controversial part of today's labor market. We link data on 350,000 individuals cited in a Federal Trade Commission settlement with one of the largest MLMs to detailed county-level information. We find that the share of refund claimants is greater in areas with higher median income and where women are absent from the labor market, suggesting value in flexible work. However, check amount, a proxy for losses, is correlated with higher inequality and lower social capital, suggesting that the pitfalls accrue to vulnerable groups.

We study the 2000s Danish legalization of interest-only (IO) loans, a mortgage market innovation aimed at increasing affordability and homeownership rates for cash-strapped buyers by substantially reducing first-year payments. Our results show that IO mortgages rapidly became popular in regions with higher house prices pre-treatment as well as with both the targeted cash-strapped borrowers and mortgage applicants spanning the wealth and income distributions. Just three years after policy implementation, IO mortgages constituted half of all outstanding Danish mortgages. The introduction of IO loans thus led to an increase in housing turnover and transactions. However, as these increases were dispersed across the distribution of homeowners, they did not lead to any changes in homeownership rates for the targeted group or allow any underserved buyer groups to enter the market. Broadly, our policy analysis documents how reforms targeting housing affordability and inequality can be exploited by the wider population, limiting intended effects.

We show that high-income buyers earn higher capital gains on housing using detailed transaction data from Denmark. Geographic location statistically accounts for nearly all the difference, with little role for aggregate market timing, property type, or other buyer characteristics. This finding is consistent with income-based sorting, whereby higher-income households systematically sort into locations with persistently higher price growth. We test whether credit conditions shape access to locations with higher house-price growth and find no detectable change in buyer composition by income rank around major credit expansions and contractions.

Macroprudential policies are a key policy tool for financial regulators, but concerns persist that these policies restrict access to homeownership. I examine this concern using cross-country data on homeownership for 28 countries. I find little evidence that macroprudential policies reduce homeownership rates in aggregate or for select groups such as low-income households. The estimates are precise enough to rule out large negative effects of macroprudential policies on homeownership rates. The null effects are consistent with models where credit shocks primarily affect prices rather than quantities. My results alleviate concerns that macroprudential policies systematically exclude certain households from ownership, but also indicate that relaxing such policies is unlikely to increase access to homeownership.

We examine episodes in which investors mistakenly trade similarly named stocks to identify short-term, non-fundamental pricing errors. Using a long time series and a broad cross-section of US equities, we systematically document the frequency, magnitude, and market response to these "confusion events". These incidents generate pronounced abnormal returns and wider effective spreads, consistent with transient mispricing. By exploiting confusion events as exogenous, we provide new causal evidence that such dislocations primarily arise from investor misperception rather than inventory frictions. Importantly, the two channels are not mutually exclusive but represent orthogonal sources of inefficiency — belief-driven demand shocks versus dealer balance-sheet constraints. We demonstrate that, under normal conditions, the misperception channel dominates, whereas under systemic stress, inventory constraints become binding. This duality bridges the behavioral and structural perspectives on market inefficiency, underscoring that belief errors remain a pervasive driver of transient dislocations, even in deep, liquid, and technologically advanced markets.