Academic Research
Quick navigation: AI · Labor · Industrial Organization · Tech
X API Pricing and the Maintenance Slowdown of Third-Party Python Tooling
Abstract: Using public PyPI and GitHub histories for a curated set of Twitter-dependent Python tools and comparison API libraries, this project finds a sharp post-2023 maintenance slowdown in the exposed set. The strongest signal appears in monthly commit activity, with release counts moving in the same direction. The result is observational but offers a clean public-data benchmark for platform-shock effects on complementor innovation.
Keywordsplatform governanceAPI pricingdeveloper toolsGitHubPyPIRepo: tech-econ-twitter-api-pricing-third-party-tool-innovation. Executive summary: Executive Summary.
ChatGPT, Stack Overflow, and GitHub Issue Burden in Open-Source Repositories
Abstract: Using public GitHub issue threads and Stack Overflow Q&A activity, this project measures whether ChatGPT-era assistance changed unresolved issue burden in open-source repositories. The empirical design links repository-month issue outcomes to language-specific Stack Overflow activity and AI-exposure proxies, with robustness checks across project size and maintainer intensity. Results suggest faster initial triage but mixed resolution depth, implying AI can compress support latency without fully reducing long-tail maintenance burden.
KeywordsChatGPTStack OverflowGitHub issuesopen sourcemaintenance burdenPyPI's Mandatory 2FA and Release Velocity in High-Impact Python Packages
Abstract: Using public PyPI and GitHub histories for high-impact and comparison packages, this project tests whether PyPI's critical-project 2FA requirement slowed release activity. The baseline and robustness estimates are small and statistically weak, with no visible release-velocity collapse in the exposed set. The takeaway is intentionally cautious: stronger account security does not appear to have obviously choked package maintenance in this public sample.
KeywordsPyPI2FAsoftware supply chainpackage maintenanceplatform policyRepo: tech-econ-pypi-mandatory-2fa-maintainer-participation. Executive summary: Executive Summary.
AI Assistance and the Distribution of Software Output: Evidence from a Calibrated Team-Week Panel
Abstract: This paper examines whether team-level AI assistance adoption is associated with changes in junior developers’ share of observable output using an 18-month real-data-first team-week panel. A v3 identification sweep (strict/balanced/broad proxies across TWFE, event-study, stacked DiD, and reweighted DiD) supports a balanced-specification baseline as the most defensible design, while maintaining conservative observational interpretation rather than causal overclaiming.
KeywordsAI assistancesoftware teamsjunior developersreal-data-firstidentification sweepRepo: tech-econ-ai-assisted-coding-junior-output-concentration. Executive summary: Working Paper Draft.
AI Capex Shocks and Semiconductor Pricing Dynamics (Synthetic Weekly Build)
Abstract: Using public market data from 2020–2025, this project shows that hyperscaler stock movements (MSFT, GOOGL, META, AMZN) are strongly associated with semiconductor returns, and that the effect is materially amplified for chipmakers with higher AI exposure. The result is a practical exposure-weighted framework for portfolio risk and positioning decisions.
KeywordsAI capexsemiconductorspanel datareproducible pipelineempirical workflowRepo: tech-econ-ai-capex-semiconductor-pricing. Executive summary: Real-Data Executive Summary.
U.S. AI Job Exposure and Wage Growth
Abstract: This paper asks whether occupations that appear more exposed to artificial intelligence experienced meaningfully different wage-growth patterns after the post-2022 AI acceleration. Using a public occupation-level panel, the analysis finds that the baseline effect is small and statistically imprecise, pointing to a cautious conclusion: broad AI exposure measures are useful for screening labor-market risk, but they do not yet establish a large wage break across major occupation groups. The practical lesson is to resist both hype and panic and to treat the current evidence as an early reduced-form signal rather than a decisive causal verdict.
KeywordsAI exposurewage growthTWFEevent studylabor economicsRepo: tech-econ-us-ai-jobs-exposure-wage-growth. Executive summary: Executive Summary.
Algorithmic Pricing, Market Conduct, and Antitrust Risk in U.S. Major Cities
Abstract: Does algorithmic pricing lead to tacit collusion and higher rents? Using a multicity quasi-experimental design and unsupervised machine learning on Airbnb data, this project finds that Smart Pricing does not raise average market levels. Instead, it acts as a productivity multiplier for sophisticated hosts, drastically increasing forward-looking price volatility and complex price discrimination without inflating baseline rents or creating neighborhood price umbrellas.
Keywordsalgorithmic pricingAirbnbfuzzy RDDinstrumental variablesantitrustRepo: Collusion-and-Algorithmic-Pricing. Executive summary: Algorithmic Pricing, Market Conduct, and Antitrust Risk in U.S. Major Cities.
Algorithmic Pricing and Demand Elasticity in App-Based Grocery Delivery
Abstract: This project studies how algorithmically adjusted delivery fees are associated with conversion, retention, and ordering-frequency outcomes in app-based grocery markets. The empirical workflow implements immutable fee-version extraction, fixed-effects and difference-in-differences estimation, event-study lead-lag diagnostics, and heterogeneity analyses across customer segments. The framework is intended as an auditable empirical template for marketplace pricing-policy evaluation.
Keywordsgrocery deliveryprice elasticitydifference-in-differencesevent studymarketplace designRepo: tech-econ-grocery-pricing-elasticity. Executive summary: Algorithmic Pricing and Demand Elasticity in App-Based Grocery Delivery.
Broadband Expansion × Labor Outcomes
Abstract: This study examines whether regional broadband subscription intensity is associated with labor-market outcomes, with emphasis on remote-work participation and income dynamics. Using county-year panel data, the analysis estimates baseline fixed-effects models and sensitivity checks, and then evaluates design credibility through diagnostic and reproducibility modules.
Keywordsbroadband adoptionremote workincomefixed effectsregional economicsRepo: tech-econ-weekly-lab-run. Executive summary: Broadband Expansion × Labor Outcomes.
Can Community Funding Replace Intellectual Property?
Abstract: This theoretical project asks whether recurring community-funding mechanisms can substitute for selected incentive functions of formal intellectual-property regimes. A repeated-game framework is used to characterize equilibrium conditions under alternative assumptions on commitment, enforcement, and free-riding.
Keywordsintellectual propertycommunity fundingrepeated gamesinnovation policy