I examine whether corporate managers learn from institutional investors through direct interactions at investor conferences. While institutional investors often possess industry expertise, there is limited large-scale evidence on how they can provide useful information to corporate managers.

I found that managers seek more direct interactions with institutional investors at conferences when they have a greater need for information about their firms’ product markets and supply chains. This relation is stronger when managers expect investors to be knowledgeable about these topics. I show that the information learned from conference interactions is reflected in managers’ subsequent decisions in that they issue forecasts more frequently and accurately. I also found that their personal stock trades during the seven days following conference participation earn higher abnormal returns than non-participating insider trades in the same firm-month. Overall, my paper shows that managers acquire decision-relevant information from their interactions with their firms’ institutional investors.

This research program examines how organisations constrain individuals at the intersection of three domains: leadership, ethics, and job design.

First, I study how leaders experience constraints on their behaviour, and how they respond to constraints that might limit their effectiveness. I examine examples such as constraints imposed by a leader’s own misconduct, and further constraints on how a leader decides to discipline misconduct from their subordinates. I also examine sociocultural constraints inherent in leader roles, such as gender stereotypes that restrict access to leader role attainment.

Second, I explore how individuals respond to organizational attempts to control unethical behaviour. Organisations use monitoring, rules, sanctions, among other policies to deter misconduct. I examine how these regulations influence individual choice, such as whether individuals choose behaviours that make them worse off, and how individuals adapt to these regulations over time.

Third, I investigate how both traditional and more modern job characteristics constrain individuals’ work experiences in unexpected ways. For example, I examine how the lack of autonomy at work may have negative spillover effects on employees’ personal wellbeing. In the context of COVID-19, I also examine how new aspects of the work environment, such as the presence of cleaning protocols, can constrain individuals behaviour and performance at work.

The role of the financial sector in society is a double-edged sword. On one hand, it can channel resources to productive uses that boost economic development, create jobs, alleviate poverty, and reduce inequality. On the other hand, it is subject to booms and busts in credit, and these can lead to a misallocation of resources, a drop in living standards, and economic stagnation.

Despite several decades of empirical research on these topics, surprisingly little is known about when and why financial sector activity is associated with good and economic outcomes and when it is not. I propose to make progress on these questions by developing new datasets for research in finance and macroeconomics that measure outstanding credit, value added, output, and input-output linkages for many sectors in the economy and a broad sample of around 120 countries.

These data will help me answer key questions about macro-financial linkages and their policy implications, such as:
1. Which sectors account for major booms and busts in credit? Which activities are associated with debt-fueled booms that foreshadow economic crashes and financial crises?
2. What is the role of financial liberalisation in the surprising decline of manufacturing sectors around the globe, sometimes called “premature deindustrialisation”?
3. Can industrial policy and credit policy successfully target key sectors in the economy, and are such sectoral subsidies successful in engineering economic growth?
4. Do lending subsidies and government-directed credit affect firms’ employment and growth outcomes? Do they subsidise more or less productive firms?

Telecommuting policies have diffused widely across organisations over the past few decades. Specifically, as of 2016, a third of all workers in the U.S. had the option to work from home at least part of the day and 23% of employees worked some or most (10-99%) of their usual hours at home. The current COVID-19 global pandemic has instigated a massive experiment in telecommuting around the world. While scholars and practitioners have long debated the potential benefits and costs of implementing telecommuting policies, we still have limited knowledge of when and how telecommuting impacts employee performance.

One reason for this lack of knowledge is that telecommuting arrangements for each employee are seldom recorded in a form that can be analysed. In addition, prior literature on telecommuting relies mainly on case studies, interviews, and surveys to evaluate the effect of telecommuting in organisations and thus is often unable to provide causal evidence on the impact that telecommuting has on employee performance.

To surmount these difficulties in empirically testing the impact of telecommuting on employee performance, I conduct three studies related to telecommuting and employee productivity. In the first study, “The Effect of Telecommuting on Information Acquisition: Evidence from the U.S. Patent Office,” I investigate the association between telecommuting and employees’ information acquisition patterns in an environment where thorough search and acquisition of information is essential. This study presents two contrasting predictions as to the relation between telecommuting and employees’ information acquisition. On the one hand, I predict that telecommuting hampers information acquisition because of reduced communication with colleagues. On the other hand, I predict that telecommuting enhances information acquisition because telecommuting can shift commuting time into work time and enables quieter and uninterrupted working environments. I use the U.S. Patent Office (USPTO) as a setting and seek to conduct empirical tests to address this research question.

In the second study, “Subordinates’ Task Performance and Departure Rates when the Supervisor Works from Home,” I investigate whether office-working subordinates show a lower level of performance when their supervisors work from home, relative to when their supervisors work at the office. This scenario stands in contrast to virtually all the prior literature on telecommuting that focuses on the impact on the performance of telecommuting subordinates working on tasks. In practice, however, many organisations demand employees have several years of work experience on the job for training purposes before they start to work from home, leading to a situation where they need to work at the office but their experienced supervisors work from home. I use the USPTO as an empirical testing ground and conduct empirical tests.

In the third study, “The Effect of an Electronic Monitoring System on Employees’ Productivity in Telecommuting Arrangements,” I examine whether the implementation of an electronic monitoring system affects telecommuters’ productivity. When employees work from home, organisations have traditionally placed a stronger reliance on output-based controls due to the inability to physically observe their employees. Due to recent advances in technology, however, many organisations now use input-based controls. One such input-based control is an electronic monitoring system, which records the number of hours spent working and leaves out any time in which an employee does not perform work activities. I take advantage of a field-research setting in which the organisation adopts an electronic monitoring system for its telecommuting employees.

Occupational positions are hierarchically arranged. Occupying leadership positions reflects one’s social status, which has been shown as an important predictor of health and well-being. To identify determinants of holding leadership positions, or leadership emergence, is a question with both important academic as well as practical significances.

Recently, there have been some remarkable breakthroughs using large samples for whole genome explorations on behaviour, social status and psychological traits, such as intelligence, personality, risk tolerance, educational attainment, and household income. However, the absence of whole genome findings on leadership has generated a seemly knowledge gap for us to form a more comprehensive understanding of relationships among our biological endowment, psychological traits, social status, and well-being.

We plan to use GWAS analyses on participants of European ancestry from the UK biobank. UK Biobank is a large prospective cohort study in the United Kingdom, which follows over 500,000 volunteers with the age from 40 to 69 during their recruitment from 2006 to 2010. For the 502,538 participants, we were able to match O*NET job titles for 274,223 individuals.

The whole genome association approach, the de facto method in studying human traits and diseases in the past decade, will be able to examine all genes of an individual to identify possible genetic associations with leadership, which will help address some long-lasting important debates such as nature vs. nurture causes of leadership emergence, trait antecedents of leadership emergence, gender differences in becoming a leader, as well as the association between leadership and health.

Venting and complaining are common forms of social interaction that have unique interpersonal and intrapersonal functions. Nonetheless, research has made limited headway into understanding venting and complaining among employees within organizations. The present research, therefore, aims to provide three key contributions to existing literature. Firstly, this study potentially challenges the widely held underlying assumption that venting only results in negative consequences.

By viewing the act of complaining through the lens of social exchange theory, I also examine how leader complaining behavior can serve as a social bonding mechanism to enhance the leader-follower relationship, which leads to subsequent prosocial behaviors towards each other. Secondly, I would like to investigate how leaders downward complaining towards their followers can serve as an emotional outlet to promote leader well-being. Lastly, via the norm of reciprocity, I also examine the potential benefits of receiving complaints from one’s leader. Given that the socioemotional benefits generated from complaining, I also test to see if leaders develop a higher level of trust and repay the favour to employees. I will test my research questions through field studies that use the dyadic experience sampling method and experiments with real business simulations.

For this experience sampling field study, I will be using a multi-wave multi-source survey involving leader-follower dyads in various companies. Additionally, to better understand the causal link between complaining and the social bonding process, I will manipulate how individuals complain to others by putting participants into a business simulation using e-confederates.

The Crowdsource project involves a critical investigation into the research process through the exploration of how a large group of analysts (~100) approach a common set of research questions. Using a common data source, which has been used for the publication of more than 100 refereed journal articles, we ask the analysts to address the same four research questions. The analysts are able to approach the research questions using independent decisions on variable definitions and model specifications, for the statistical analysis of the common data source. After compilation of the results from the various analysts, we explore the variance to the estimated results for each of the four research questions. We then also investigate how prior beliefs are connected to the results generated by individual analysts.

By making this form of investigation, we hope to contribute to ongoing debates about what constitutes rigorous and reliable methods for quantitative investigations in the Social Sciences. In recent years, there has been an increasing appreciation for the need to have a serious introspection into our research processes. Crowdsourcing, pre-registration, and reproducibility and generalisability are increasingly becoming part of our common lexicon of accepted research practices. We contribute to this groundswell of introspection and redefinition of processes through our deep investigation into the research process for large-N, secondary data research.

This paper investigates household consumption and market concentration in the US retail sector. We start by documenting four main facts with store- and household-level consumption micro-data from 2004-2016: 1) rising national aggregate concentration, 2) negligible change in regional concentration, 3) rising household concentration and heterogeneity across households, and 4) increased one-stop shopping. We show that households have increasingly concentrated their spending in fewer stores despite a growing number of stores in regional markets. We explore how underlying factors from both the supply and demand side drive these trends.

We use three main empirical strategies to estimate the effect of local retailer availability, product assortments, and pricing on household retail concentration: two event studies using the entry of superstar chains and household migration respectively, and an IV that builds on uniform pricing and similar assortment within retail chains.

Our preliminary results suggest that the entry of supercentres, a greater variety of products, and lower relative store prices can boost household retail concentration. However, only a small proportion of the change in concentration during this period can be explained by movements in these supply-side factors, which implies that demand-side factors may play a potentially important role. We will further construct a structural model which incorporates both supply- and demand- side factors to highlight the implications for market power and welfare.

This research project studies information aggregation problems (defined broadly) in the context of managing marketplaces and online platforms. More specifically, it has three streams.

The first stream studies active learning problems through the lens of sequential experimental design. In an active learning problem, the decision maker (DM) can proactively influence the data flow to make the data aggregation process more efficient. Under a sequential hypothesis testing framework, I strive to develop the general methodologies associated with it.

The second stream aims to study how to apply the active learning methodologies to platform operations, such as preference learning and ranking & selection. A concrete example is to design surveys/questionnaires on an e-commerce platform. I also explore modelling and estimation problems to capture the customer preferences on the platform.

The third stream aims to study dynamic learning and information aggregation problems in the presence of strategic disruptions. Both the theoretical framework and its applications in platform operations are considered, such as fraud clicks and fake orders, among others.

Reinforcement learning (RL) has emerged as one of the most important fields of AI and represents a step towards building autonomous systems with a higher-level understanding of the visual world. For instance, one of the recent advances is the development of AlphaGo – a computer programme that combines advanced search tree with deep neural networks. These neural networks use RL techniques to take a description of the Go board as an input and process it through a number of different network layers containing millions of neuron-like connections. In this research, we seek to explore the use of RL techniques in the field of Operations Management and replicate its successes in this area.

Real-world applications in Operations Management require RL algorithms to act in the presence of model and data uncertainty. During learning process, it is likely that the agent executes sub-optimal actions that may lead to unsafe/poor states of the system. Exploration is particularly brittle in high-dimensional state/action space due to an increased number of low performing actions and noisy observations. We seek to build on Robust Optimisation techniques to build exploration in approximate RL setting. To ensure robustness during learning, we will explore the robust policy and value iteration algorithms.