Research and Writing
This page collects my working papers, dissertation research, and open source work. For shorter writing on applied economics, causal inference, pricing strategy, industrial organization, and marketplace economics, visit the blog.
Working Papers
Profit-Aware Pricing in Two-Sided Marketplaces: Demand Elasticity, Cost Uncertainty, and Customer Lifetime Value in Brazilian E-Commerce
Dinesh R Poddaturi | 2026
This paper develops an empirical pricing framework for two-sided e-commerce marketplaces, combining demand estimation, profit optimization under cost uncertainty, customer lifetime value modeling, and sequential purchase pattern analysis. Applied to 110,840 transactions across 71 product categories from the Olist Brazilian marketplace, the analysis demonstrates that revenue maximization and profit maximization are fundamentally misaligned in elastic markets. A 40% price cut generates a 160% revenue increase but a 162% profit loss. A key methodological finding is that cost uncertainty dominates elasticity uncertainty. A 5-point COGS error forfeits 70 to 75% of potential profit gains, while a 20% elasticity error preserves profitability across all scenarios. The paper includes a complete implementation framework incorporating experimentation design, causal evaluation protocols, phased rollout sequencing, and risk mitigation procedures.
Keywords: Dynamic Pricing, Marketplace Economics, Causal Inference, Revenue Management, Customer Lifetime Value, Price Elasticity, Two-Sided Markets, Profit Optimization, Demand Estimation.
Estimating the Price Effects of the Blocked JetBlue-Spirit Merger: A BLP Demand Estimation and Merger Simulation
Dinesh R Poddaturi | 2026
In January 2024, a federal court blocked JetBlue's proposed acquisition of Spirit Airlines, agreeing with the Department of Justice that the merger would eliminate Spirit as an independent ultra-low-cost competitor. This paper builds an independent structural analysis of the same underlying economic question, what the merger would have done to prices and consumers. Because no public dataset links individual travelers to their choices, the analysis uses a constructed market-level panel of 41 routes over 12 quarters (492 route-quarter markets). Three routes named in the case record are anchored to real JetBlue and Spirit fares from the BTS Consumer Airfare Report and to real jet fuel costs, and the remaining 38 routes are constructed, with route composition calibrated to the government's own breakdown. The paper estimates a BLP random-coefficients demand model by two-step efficient GMM and simulates Bertrand-Nash pricing under common ownership. On the three anchor routes, the simulation implies average post-merger price increases of about 52 percent for Spirit and 18 percent for JetBlue, and a modeled annualized consumer welfare loss of about $64.1 million. Because the data-generating process is known, the estimated parameters can be evaluated against their true values. The spread of price sensitivity across consumers is recovered well, but heterogeneity in consumer taste for ultra-low-cost carriers is not identified. The paper diagnoses the gap with a grid search and shows that it leaks into other estimates, including a mean price-sensitivity estimate about 12 percent above its true value (about 4 percent when the true taste spread is supplied) and diversion ratios that understate diversion to the rival ultra-low-cost carrier by roughly 5 to 7 percentage points. Results are scoped to the three verified routes and to upward pricing pressure; they are not extrapolated to the government's full set of routes.
Keywords: Merger Simulation, BLP Demand Estimation, Antitrust, Unilateral Effects, Industrial Organization, Random-Coefficients Logit, Counterfactual Analysis, Consumer Welfare, Airline Competition.
Open Source
Switchback Experiment Framework for Two-Sided Marketplaces
Dinesh R Poddaturi | 2026 | GitHub Repository
A Python simulation of switchback experiment design for two-sided marketplace pricing interventions. Implements the switchback methodology for settings where network effects prevent user-level randomization, including treatment assignment, outcome simulation, clustered standard error estimation, and power analysis. Designed to demonstrate how causal inference methods can be implemented for real marketplace pricing and policy evaluation problems.
Doctoral Dissertation
My doctoral dissertation at Iowa State University developed dynamic structural economic models of the U.S. beef cattle industry, with applications to disease outbreak scenarios and agricultural policy analysis. The work focuses on dynamic optimization, structural estimation, and demand systems, methods that transfer directly to pricing, industrial organization, and empirical market analysis. Co-authored with Chad E. Hart and Lee L. Schulz.
A Dynamic Assessment of the Economic Impacts of a Foot-and-Mouth Disease Outbreak on the U.S. Beef Cattle Industry
Dinesh R Poddaturi, Chad E Hart, Lee L Schulz | Presented at Western Agricultural Economics Association, 2022
Develops a dynamic framework integrating rancher optimization, cattle biology, stock replacement, age distribution, and market processes, calibrated to the U.S. beef cattle industry. Estimates the economic impacts of a hypothetical Foot-and-Mouth Disease outbreak by integrating trade, consumption, and supply shocks into the dynamic framework. The framework is generalizable to a variety of production and policy shifts beyond disease scenarios.
A Dynamic Model of U.S. Beef Cattle
Dinesh R Poddaturi, Chad E Hart, Lee L Schulz | Presented at Agricultural and Applied Economics Association, 2020
Develops a bottom-up dynamic economic model of the U.S. beef cattle industry incorporating farmer optimization behavior, cattle biology, age and gender composition, and micro-foundations. Calibrated under naive and rational price expectations to capture observed industry dynamics and project future prices and quantities.
Implementing a National Animal Identification and Traceability Program: Economic Assessment using a Dynamic Model of U.S. Beef Cattle
Dinesh R Poddaturi, Chad E Hart, Lee L Schulz | Presented at Agricultural and Applied Economics Association, 2021
Quantifies the short-run and long-run economic impacts of a mandatory national animal identification and traceability system on U.S. beef cattle producers, including producer surplus changes under different cost-sharing programs and adoption rates.
Additional Research
Autoregressive Tempered Fractionally Integrated Moving Average Model (2018)
Dinesh R Poddaturi | MS Statistics Creative Component
Analysis of the ARTFIMA time series model and its applications to real-world data, comparing model fit and predictive performance against ARFIMA and ARMA specifications.
Development of a CRUSH Margin Calculator for Beef and Pork Markets (2017)
Dinesh R Poddaturi, Garland Dahlke, Russ Euken, Lee L Schulz
Web-based application utilizing CME futures settlement price data to generate expected margins for beef and pork producers. Used daily by beef and pork producers for price risk management.
