Research

Publication

"Liberalizing Home-Based Business.” (with Sumit Agarwal, Tien Foo Sing, and Jian Zhang), Management Science, forthcoming

Abstract: The COVID-19 pandemic has unleashed an unprecedented disruption in commuting and strengthened the interest in creating and running home-based business. We evaluate a large-scale reform, Home Office Scheme, in Singapore that allows for business creation at one’s residential property and study whether the option of home-based entrepreneurship spurs entrepreneurial activities. Difference-in-difference estimate shows that the reform leads to a significantly higher level of business creation and the newly created firms in response to the reform have a higher survival rate. The effect is more pronounced for low-income, female individuals and industries with high startup capital, implying that financial constraint and non-pecuniary benefits are likely to be the main mechanisms driving the effect. The reform also encourages entrepreneurs to become serial entrepreneurs, and they open a larger business with similar survival rate for their second firm. Overall, our findings suggest that the program effectively attracts more entry into self-employment without significantly lowering the average quality of the pool.

“Liquidity Constraints, Consumption, and Debt Repayment:  Evidence from Macroprudential Policy in Turkey.” (with Sumit Agarwal, Muris Hadzic, and Yildiray Yildirim), Review of Financial Studies, Volume 36, Issue 10, October 2023, Pages 3953–3998. 

 Abstract: Using account-level credit card data from a large Turkish bank, we study the impact of a unique credit card policy that increases minimum payment on consumption and debt repayment. We show that the policy reduces credit card spending and debt, boosts existing debt repayment, and reduces credit card delinquency. The credit card debt of affected consumers falls on average by 50% two years into the policy implementation. An increase in minimum payment has a stronger effect than a decrease of similar magnitude. We build a benchmark life-cycle model with soft liquidity constraint to explain the reduction in credit card spending.

"Inflation Expectations of Households and the Upgrading Channel."  (with Sumit Agarwal, and Yeow Hwee Chua), Journal of Monetary Economics, Volume 128, May 2022, Pages 124-138.

Abstract: Motivated by the evidence that households purchased better quality goods across time in Singapore, we use a survey experiment to study the relationship between consumption upgrading and households’ beliefs about inflation. We find that providing price information on better-quality products will lead to higher inflation expectations. The effects on inflation expectations are smaller when price information on both higher- and lower-quality products is made available, suggesting that product replacement increases inflation expectations. Additional tests show that our results are not driven by mixing price with quality or numeracy. Our findings highlight the relationship between product replacement, product variety, and inflation expectations.

“Financial Illiteracy and Pension Contributions: A Field Experiment on Compound Interest in China.”  Review of Financial Studies, Volume 33, Issue 2, February 2020, Pages 916–949

Abstract: We design a field experiment to study the relationship between neglect of compound interest and pension contributions in rural China. We randomly assigned some households to a financial education treatment, emphasizing the concept of compound interest. This treatment increased the pension contribution by roughly 40%. To pinpoint mechanisms, we elicited financial literacy after the intervention, and added a third group in which we explain the pension benefit in general. We find that the neglect of compound interest is correlated with low contributions to the pension plans in the control group, and that financial education about compound interest does help households partially correct their erroneous understanding of compound interest. Moreover, explaining compound interest increases their ability to translate benefits into their own situations.

“Do Real Estate Agents Have Information Advantages in Housing Markets?” (with Sumit Agarwal, Jia He, and Tien Foo Sing), Journal of Financial Economics, Volume 134, Issue 3, December 2019, Pages 715-735

Abstract: Real estate agents play an important intermediary role in housing markets. We use a merged transaction dataset that identifies houses purchased by registered real estate agents and other buyers in Singapore to empirically test the hypothesis that real estate agents use information advantages to buy houses at bargain prices. We estimate that agents bought their own houses at prices that are 2.54% lower than comparable houses bought by other buyers. Agent-buyers have more information advantages in less informative environments, high ability agent-buyers have even more information advantages, and agent-buyers are more likely to buy houses from agent-sellers. These results support information asymmetries in the housing market. We further study the sources of information advantage. We find that agent discounts are likely from both “cherry picking” and bargaining power. In our main results, bargaining power contributes more to the agent discounts. Agent-buyers have information advantage over non-agent buyers since they have a larger choice set from which to choose, allowing them to “cherry pick” cheaper houses. Agent-buyers have strong bargaining power since they have information advantages regarding the purchase price of houses.

“Superstition, Conspicuous Spending, and Housing Markets: Evidence from Singapore.” (with Jia He, Haoming Liu, Tien Foo Sing, and Wei-Kang Wong), Management Science, Volume 66, Issue 2, February 2020, Pages 783-804

Abstract: We study the effect of superstition and conspicuous spending motive on housing demand and price in Singapore. We find that buyers pay less for homes with unlucky addresses and more for homes with lucky addresses. There were fewer housing transactions on inauspicious days of the lunar calendar when people are advised to avoid making major economic decisions. This suggests that superstitious belief still affects economic activities. The demand for lucky addresses is also weaker on these inauspicious days, suggesting that superstitious belief indeed affects the demand for lucky addresses. Moreover, the price premium for lucky address is significantly higher for apartments of larger size or on top floors. Since these two housing features can signal wealth and are highly visible, the larger price premium suggests that conspicuous spending motive also plays a significant role in the Singapore housing market. We also find that informed buyers, even with less superstitious or conspicuous spending motive, might still pay price premiums for lucky addresses. In contrast, uninformed buyers are unlikely to pay a premium for these addresses.

"Information Avoidance and Medical Screening: A Field Experiment in China." (with Yufeng Li, Juanjuan Meng, and Kai Zheng), Management Science,  2021 67:7, 4252-4272 

Abstract: Are high-risk individuals more likely to avoid a disease test because of information avoidance? We conduct a field experiment to investigate this issue. We vary the price of a diabetes test (price treatments) and offer both a diabetes test and a cancer test (disease treatments) after eliciting participants’ subjective beliefs about the disease risk. We find evidence that both low- and high-risk groups avoid testing, and this pattern is more salient when the test price is higher and the disease is more severe. We derive new predictions using the optimal expectation model to explain our empirical findings. 

"High Sex Ratios and Household Portfolio Choice in China."  (with Wenchao Li, Shu Xu, and Junjian Yi),  Journal of Human Resources, forthcoming

Abstract: This paper studies how high sex ratios (more men than women) affect household portfolio choice. Using data from a nationally representative Chinese household finance survey, we find that a one standard deviation increase in the sex ratio would rise the stock market participation rate by 2.7 percentage points or 47.9 percent among families with a son. This implies that rising sex ratios can explain around 10 percent of the significant growth in China's stock market size in recent decades. We rationalize our findings with a reference-dependent framework, in which the reference level is expected marriage expenditure for children.

“Benefits of Relationship Banking: Evidence from Consumer Credit Markets.” (with Sumit Agarwal, Souphala Chomsisengphet, Chunlin Liu, and Nicholas S. Souleles), Journal of Monetary Economics, Volume 96, June 2018, Pages 16-32

Abstract: This paper empirically examines the benefits of relationship banking to banks, in the context of consumer credit markets. Using a unique panel dataset that contains comprehensive information about the relationships between a large bank and its credit card customers, we estimate the effects of relationship banking on customers’ default, attrition, and utilization behavior. We find that relationship accounts exhibit lower probabilities of default and attrition, and have higher utilization rates, than non-relationship accounts, ceteris paribus. Such effects become more pronounced with increases in various measures of the strength of the relationship, such as relationship breadth, depth, and length. Moreover, dynamic information about changes in the behavior of a customer’s other accounts at the same bank, such as changes in checking and savings balances, helps predict and thus monitor the behavior of the credit card account over time. To investigate these mechanisms, we show that the observed results are not due to the selection mechanism. Instead, they are consistent with the private information mechanism: Information the lender has at its disposal on the dynamics of the relationship borrower’s other accounts can be used to mitigate credit risk on the credit card account. These results imply that relationship banking offers significant potential benefits to banks in the retail credit market. 

"Do disaster experience and knowledge affect insurance take-up decisions?" (with Jing Cai), Journal of Development Economics, Volume 124, January 2017, Pages 83-94. 

Abstract: This paper uses a randomized experiment to study the effect of personal experience on weather insurance adoption in rural China. First, we conduct insurance games with a random subset of farmers to estimate the effect of hypothetical experience of disasters on real insurance take-up. We show that playing insurance games improves real insurance purchase by 46%. The effect is not driven by changes in risk attitudes and perceived probability of disasters, or learning of insurance benefits, but is mainly driven by the experience acquired in playing the game. Second, we explicitly reveal the true probability of disasters to randomly selected farmers, and find that reminding farmers of their real experience of disasters also has a strong effect on insurance take-up. Lastly, we explain the large impact of personal experience on insurance purchase by a theory of salience.

Working papers

"Ethnic Investing and the Value of Firms." (with Jonas Hjort and Christopher Yenkey), revise and resubmit, Management Science

Abstract: We study ethnic investing, using transaction data from Kenya’s stock exchange and CEO/board turnover. We show that a given investor invests more in a given firm when the firm is run by coethnics and earns lower risk-adjusted returns on such investments. We then model and empirically test for the aggregate impact of (i) the implied taste- or psychology-driven investor discrimination and (ii) counteracting demand- and supply-side forces. Our estimates imply that listed Kenyan firms could collectively be worth 37 percent more—with minority-run firms benefitting the most—if the neutral proportion of active investors increased from 4.2 to 50 percent.

“Relational Contracts in the Housing Market.” (with Sumit Agarwal and Vincent Yao)

Abstract: This paper examines relational contracts (RCs) in the housing market that exist between lenders and appraisers. We document that 42% of appraisals are at or near the contract value, while only 7.6% are below the contract. We develop an RC model and test several predictions using a novel data set of over 3 million appraisals. Our results confirm that appraisers produce more favorable appraisals in reciprocity for past business from the lenders, and their response is stronger for relational appraisers. Lenders penalize appraisers for unfavorable appraisals by assigning less business or terminating the relationship. The penalty decreases as uncertainty over the appraiser type diminishes. These findings demonstrate that the incentives of a financial intermediary in the RCs setting can be misaligned due to its negative externality on third parties.

"Banking Competition and Shrouded Attributes: Evidence from the US Mortgage Market." (with Sumit Agarwal and Vincent Yao)

Abstract: We document that banking deregulation leads banks to offer lower initial rates on adjustable-rate mortgages to attract borrowers, but banks also shroud these contracts by increasing back-loaded resetting rates. More shrouding can be explained by higher proportion of naïve borrowers following the deregulation and banks shroud more where there are more naïve borrowers. These results support the theory that sophisticated firms can exploit consumer biases by designing exploitative contracts. Although competition reduces firm revenues and benefits consumers initially, shrouding helps banks offset vast majority of the initial losses.

"Consumption and Portfolio Rebalancing Response of Households to Monetary Policy." (with Sumit Agarwal, Yeow Hwee Chua, and Pulak Ghosh)

Abstract: Using micro-level administrative data from India, we show the consumption and portfolio rebalancing response of households to changes in interest rate. Exploiting variation in the timing of expiry of term deposits, we find that when interest rate falls, households increase consumption by 6 percent and increase their fraction of wealth into risky assets by 36 percent after the expiry of term deposits. While both existing and new investors “reach for income” by investing in high dividend stocks, new investors also “reach for yield” by investing in high beta and more volatile stocks. The effects on consumption and risky investment are smaller for term depositors with automatic renewal. Households with more liquid wealth have a smaller consumption effect but a larger portfolio rebalancing effect on risky investment. These results highlight that term deposits contract rigidity and household wealth affects the monetary policy pass-through.

“Intentional Bequest Motives and the Choice of Annuity.” (with Haoming Liu and Shenghao Zhu)

Abstract: This paper identifies the intentional bequest motive by exploiting the choices between two compulsory partial annuity plans in Singapore, the default Standard Plan with a higher monthly payout but a lower bequest, and the Basic Plan with a lower monthly payout but a higher bequest. Because actively choosing the Basic plan increases  committed bequests but not precautionary savings, individuals choose the Basic plan only if they want to leave bequests intentionally. Using our own household survey data, we find that about 20 percent of people choose the Basic Plan, which provides direct evidence for the existence of the intentional bequest motive. Better educated individuals, and individuals with children are more likely to intentionally leave bequests. Consistent with the altruism motive hypothesis, individuals with poorly educated or financially insecure children are more likely to leave bequests. Contradicting to the reciprocity hypothesis, the frequency of children's visit has no impact on the probability of leaving bequests. To differentiate between motivations for leaving bequests, we further ask individuals to allocate real high stake lotteries to their children and study the correlation between the allocation and children's behavior. We find that most people allocate lotteries equally across their children, in line with the motive of the joy of giving and against the reciprocity motive. Our structural estimation shows that households have strong bequest motive and bequests are luxury goods. 

"Can Partial Commitment Increase Pension Contribution? A Field Experiment in Sri Lanka."

We conduct several randomized controlled trials in more than 200 villages in Sri Lanka to study whether incentives and partial commitment pension designs generate higher participation and savings in the micro pension. In Experiment I, individuals are randomly assigned to a control group, a free installment group, and a matching group. We find that a free installment for the first month contribution increases the pension participation from 8 percentage points to 34 percentage points, and increases the pension contribution by 4 times.  A 100% matching for the first month contribution also increases the participation and contributions, but the effect is smaller. We show that the results can be explained by that free Installment group attracts more present bias agents, and present bias agents are more likely to participate in the pension when they do not need to pay first month contribution. In Experiment II, we further compare a full commitment pension including only one commitment account with a partial commitment pension contract including a liquid account and a commitment account. Individuals are randomly assigned to four groups: full commitment pension with high withdraw penalty, partial commitment pension with low withdraw penalty, partial commitment pension with high withdraw penalty, and a choice group in which they choose full commitment or partial commitment with high withdraw penalty. We find that, partial commitment pension with high withdraw penalty and the choice group have 9.9 percentage points and 7.7 percentage points more participation compared to full commitment pension, respectively. We show that high commitment contract attracts more sophisticated agents, and higher degree of commitment increase the pension participation more for sophisticated agents. These results are consistent with the theory of optimal illiquidity.

“The Impact of Housing Credit on Personal Bankruptcy.” (with Sumit Agarwal)

Abstract: We use a linked housing transaction dataset and a personal bankruptcy dataset to study the impact of housing credit on personal bankruptcy in Singapore. Using a difference-in-differences (DD) approach, we find that an increase in housing credit increases the likelihood of personal bankruptcy by 0.15-0.22 percentage points for house buyers who have more exposure to the housing credit increase. To investigate the mechanisms, we show that the observed effect is unlikely to be driven by composition effects and selection for irresponsible buyers. The effect is mainly due to the increasing debt burden from an increase in housing credit. Moreover, we find that those who file for bankruptcy are more likely to sell their houses before bankruptcy with low return, suggesting that home owners use house sales to smooth consumption and mitigate negative shocks.

“Information Provision, Patient Sorting, and Healthcare Quality.” (with Nan Yang, Junjian Yi, and Ye Yuan)

Abstract: We examine a mobile outpatient appointment app---a light-weight information technology innovation---launched by Chinese hospitals. The objective is to assess the effect of information provision and a streamlined appointment process on hospital operations and the alignment of healthcare supply and demand. Using a longitudinal dataset on hospital operation and a difference-in-differences model, we document that the app increases completed hospital consultations by 9.5%, by boosting registrations by 4.8% and reducing appointment cancellations by 3.4%. The app improves queuing efficiency in overcrowded hospitals and draws demand for underutilized ones. It results in a better match between patient demand and hospital service by directing patients to the hospital and department more suitable to one's medical conditions and to less busy days. Subsequent app launches shorten average waiting time for booked appointments.

"An Experiment on Reference Points and Expectations."

Abstract: I conducted a controlled laboratory experiment to test to what extent expectations and the status quo determine the reference point. In the experiment, I explicitly manipulated expectations by exogenously varying the fluctuations of lagged beliefs and tested whether expectations affect risk attitudes. I also exogenously varied the time of receiving new information and tested whether and how fast individuals adjust their reference points to new information. Moreover, I derived and tested a new theoretical prediction to distinguish two expectation-based reference dependent models: fixed reference points and stochastic reference points. I find that both lagged expectations and the status quo influence the reference point significantly. While reference points are sticky to lagged expectations, subjects adjust their reference points quickly after receiving new information, which further confirms the role of expectation as the reference point. Moreover, both reduced form and structural estimation support the model of the stochastic reference point reflecting full distribution of expected outcomes rather than that with a fixed reference point.

“Competing Sources of Reference Points: Evidence from a Real-Effort Experiment.” (with Sebastian Berger, and Lorenz Goette)

Abstract: A key open question for theories of reference-dependent preference is what determines the reference point. Previous research has delivered independent evidence regarding expectations and social comparisons. In many settings, however, both expectation-based and social reference points may be relevant. It is unclear which of the two actually matters more when they compete. In a real-effort experiment, we jointly manipulate the rational expectations of subjects regarding their own earnings as well as the earnings of a partner. We find that effort provision is significantly affected by individual expectations and not by social comparisons. If earnings expectations for one are high, subjects increase effort to increase their earnings. The partner's earnings expectations do not affect effort. Our results suggest that individual expectation plays stronger role in determining reference points than social comparison under competing sources of reference points.