Below are some past project descriptions by Divisions:
The summer research projects are designed to give the interns an understanding of and spur interest in business research. Research in the business school is highly interdisciplinary and spans many disciplines outside of the traditional areas of accounting, finance, management, marketing, and operations research. These include, but are not limited to, economics, psychology, sociology, political science, computer science, engineering, statistics, and data science.
Projects in Accounting, Economics, and Finance typically require interns with strong backgrounds in econometrics. Some projects may require students to have more specialized skills in data science and computer science, while others may require interest in legal, regulatory, and policy issues. Past accounting projects have studied US hedge funds and their relation to sovereign debt, corporate transparency and firm communication, and investigations in shareholder voting. Economics projects span the areas of development economics, labor economics, industrial organization (I/O), monetary policy, and macroeconomics. Finance projects have investigated the efficiency of capital markets, corporate finance theory, impact of financial shocks on innovation, structure of company boards.
Projects in the Decision, Risk, & Operations (DRO) division are well suited to interns who major in engineering, computer science, statistics, and applied mathematics. DRO professors have looked at how to optimize patient flow through hospitals and ICUs, market matching, optimizing school admissions and public housing assignments.
The Management and Marketing divisions offer opportunities to both behavioral researchers, in sociology and psychology, and quantitative researchers. In management, researchers have investigated the effects of power in hierarchy in organizations, decision making in organizations, effect of gender and diversity in organization, field experiments in hiring, and negotiations. Management projects that require more advanced statistics or applications of machine learning may include gender pay equity in the UK, or exploring the foundations of creativity and success in music. Recent projects in quantitative marketing include studies of the effect of video game lotteries on consumers, developing a method to impose ethical constraints on decisions, and predicting churn from social media (Twitter) data.
Accounting
Time Arbitrage and Investor Time Horizons
Project Overview:
The project is an empirical asset pricing paper on the concept of "Time Arbitrage". We measure the time horizon of active investors and examine if firms with long-term investors perform better than firms with short-term investors. After demonstrating the existence of Time Arbitrage, we attempt to provide a mechanism for why the phenomenon exists.
Responsibilities:
- Data collection for robustness tests (via WRDS, SQL queries, or web scraping)
- Statistical analysis of financial datasets
- Citation management and bibliography organization
- Literature review and synthesis
Required Skills:
- Experience with statistical software (Python, Stata, or R)
- Proficiency in managing and merging large datasets
- Familiarity with Overleaf/LaTeX (for draft preparation) and Microsoft PowerPoint (for slide creation)
- Experience with citation managers such as Zotero or Mendeley is helpful but not required
Decision, Risk, and Operations
Do LLMs Understand Chronology?
Project Overview:
Large language model have been found to be useful in certain forecasting tasks, including predicting stock performance from news sentiment. However, backtesting the performance of LLM forecasts presents a challenge if the backtesting window overlaps with the training window because of the risk of information leakage. Newer LLMs are better than earlier models at following instructions to restrict their use of information to permitted time windows. But to do so they need to understand chronology.
This project will test that understanding through a series of experiments that test increasingly complex notions of chronology. The project requires the ability to query LLMs through APIs. It will also involve formulating and analyzing the results of tests.
Responsibilities:
- Formulate and run API-based queries to test LLMs
- Analyze experiment results and conduct basic statistical analysis
- Evaluate LLM-based financial forecasts
Required Skills:
- Proficiency in Python, especially for running API queries
- Basic statistical and data analysis capabilities
- Understanding of LLM functionality and limitations
Economics
Case Studies in Global Macroeconomic Investing
Project Overview:
International macroeconomics provides an analytical framework that underpins portfolio construction and thematic overlays under both systematic and discretionary trading strategies implemented in speculative short-term investments by hedge funds and the deployment of long-term pools of capital by real money asset allocators and owners in the endowment, family office, and pension-fund spaces. This research project will focus on writing case studies on major global macroeconomic themes for the Core courses Global Economic Environment and Markets & the Economy, as well as my electives in Global Macroeconomic Investing. These case studies are intended to become essential resources for teaching at both CBS and other business schools through their indexing on the CBS and HBS online case catalogues. The professor and student will work in conjunction with Columbia Case Works to develop these studies into published cases for broader impact. We are currently finalizing cases on the US dollar’s reserve currency status, the next steps for the Fed’s interest rate policies, and deglobalization; we are set to begin new drafting efforts on commodities, fiscal policy, and financial crises.
Responsibilities:
- Data collection, organization, analysis, and presentation in tables and charts
- Conduct literature reviews and manage reference databases
- Assist in drafting and outlining case content
- Attend and summarize meetings with research sources
Required Skills:
- Coursework in macroeconomics, international economics, and finance (through at least the third-year undergraduate level)
- Experience with Stata, R, and/or Python
- Strong writing and organizational skills
- Proficiency with Excel for chart creation and data visualizationNon
- CBS free preparatory training: Bloomberg Market Concepts and Bloomberg Financial Markets Training Program
Finance
Labor Market Frictions and Retirement Investment Decisions
Project Overview:
This research examines how labor policies, particularly non-compete agreements (NCAs), influence retirement financial planning, focusing on defined contribution (DC) plan investment allocations. DC plans, now the predominant retirement savings vehicle in North America, transfer investment responsibility to employees. While offering flexibility, many employees struggle with optimal allocation decisions due to behavioral biases and plan design challenges.
NCAs restrict employees from joining competitors after leaving their jobs, reducing labor mobility and limiting wage growth opportunities. These constraints shape employees' long-term financial planning by influencing their portfolio allocation decisions. In particular, our hypothesis is that restricting labor mobility reduces the co-integration between labor income and business cycles, as agents cannot move to better jobs when the economy is booming. The classic life cycle model of portfolio construction then implies that agents will invest in more risky assets such as equity, allowing them to compensate for the lost income growth opportunities in good economic states.
By investigating this relationship, the research aims to uncover how labor market frictions shape retirement planning behaviors and long-term financial outcomes. We utilize a granular data of 401(k) portfolio decisions at the pension plan level and will start with regression analysis of the hypothesized relationship. We then construct a life cycle that incorporates costly job turnover opportunities to quantify the welfare loss of NCAs while taking employee stock allocation decisions into consideration.
Responsibilities:
- Clean and merge large datasets
- Construct variables and perform regression analyses
- Conduct robustness checks on model estimates
- Assist in proofreading an economic model (no modeling required)
Required Skills:
- Proficiency in regression analysis using Stata
- Ability to manage and manipulate large datasets in Python or Stata
- Familiarity with optimization models in microeconomics and financial economics
Management
Facilitating Psychological Safety in Professional Teams of Biologists & Data Scientists with Generative-AI
Project Overview:
Group-based, imaginative play (Sutton-Smith, 2009) has recently been explored as a means for teams to foster stronger team bonds and more effective collaboration. This play enhances team functioning by reducing anxieties surrounding intergroup and interpersonal differences, and fostering creativity, trust, and a sense of psychological safety (Ubaka et al., 2023). Building on previous work done with MBA and executive teams, we evaluate a new tool for facilitating group-based, imaginative play and fostering inclusive leadership behaviors in teams—art generative-AI. Art generative-AI is a promising tool for fostering psychological safety in teams as it allows team members to engage in interpersonal risk-taking and novel relationship building behaviors while having fun together. In addition, preliminary evidence suggests that team members learn to better communicate with one another as they engineer prompts to communicate with the AI tool, and that rapid prototyping with AI helps team members utilize their unique perspectives to overcome challenges. In the next iteration of the project, we will study the effectiveness of a group-based, imaginative play exercise facilitated through art generative-AI for enhancing psychological safety and team performance in professional teams of biologists collaborating with data scientists. We will assess both short-term and longitudinal effects.
Responsibilities:
- Conduct literature reviews
- Assemble and clean datasets
- Conduct descriptive statistical analysis
- Data collection (i.e., running studies)
- Create/edit experimental materials and visualizations
- Data content coding (qualitative)
Required Skills:
- Ability to work in dynamic, fast-paced environments (i.e., customer service skills)
- Strong problem-solving, initiative, and resourcefulness
- Basic understanding of psychology (able to interpret paper abstracts from relevant journals)
Preferred Skills:
- Experience with data processing and regression in R and/or SPSS
Marketing
Project Overview:
The intern will work on mainly two projects involving experimental and observational data.
This first project examines the causal impact of comment sections on social media posts portraying diversity of other inclusive marketing visuals. Conducted in two phases, it involves an online field experiment with over 3 million Facebook and Instagram users. Phase 1 collects and analyzes comments to assess how political ideology influences engagement with posts. Phase 2 evaluates how exposure to different comment narratives affects subsequent user behavior, including engagement (comments, views), intent (clicks), and off-platform actions (newsletter sign-ups, donations, petitions). Our methodology expands on traditional A/B testing by manipulating comment visibility and leveraging election data as a proxy for political ideology.
The second project studies the impact of reviews in platforms like Tripadvisor and Yelp on product design and horizontal differentiation. It will involve the collection and analysis of reviews and menu data from major review platforms.
Responsibilities:
- Assemble and clean large datasets
- Conduct regression analysis
- Conduct text analysis (natural language processing)
- Conduct literature review on field experiments on social media platforms and review platforms
- Draft and polish paragraphs in academic style
Required Skills:
- Knowledge of statistical softwares or programming languages (R, Stata, python, etc.)
- Familiarity with text parsing and analysis (NLP) and/or machine learning techniques.
- Familiarity with statistical tools such as linear regression
- Familiarity with working with large panel and cross-sectional data
- Familiarity with the concepts of experiments in social sciences
- Excellent communication skills
- Previous exposure to digital marketing and experience in managing online marketing campaigns on Facebook are a plus