Varkara LLC is a New York consulting firm that builds economic models, statistical analyses, and software for government agencies. Every project delivers a working model, the code behind it, and a written report.
Varkara LLC builds economic models, statistical analyses, and software for government, nonprofit, and private clients.
We build models that show how a proposed change in tax law or program rules will affect households, government revenue, and public budgets. We specialize in microsimulation. A microsimulation model starts from a large, representative sample of real households, applies the policy change to each one, and adds up the results.
We analyze survey and administrative data, evaluate programs, and build forecasts. Every forecast gives a best estimate and the full range of outcomes around it.
We write analysis software in Python, R, C++, and Stan. We build automated processes that clean, check, and load an agency's data. We modernize older code so it runs faster and is easier to maintain, with full tests and documentation.
Economic research and written reports for government, nonprofit, and private clients. Every report comes with the code that produced its results.
Our two principals lead and deliver every project. Together they bring more than two decades of modeling experience.
Managing Partner
Steve Bronder is a statistician and software engineer with experience building statistical models and software to answer practical questions in areas like public health, credit risk, and economic statistics. He is a co-author of Stan, a programming language for statistical modeling that researchers use in fields such as economics, ecology, epidemiology, and market research. Stan has been cited in more than 1,000 published studies of COVID-19. Steve currently works as a Research Scientist at the Flatiron Institute. Previously, he worked as a Principal Data Scientist in Capital One's Financial Institutions Group, building statistical models for structured products. He has provided computational mathematics consulting to Columbia University, including statistical consulting for the Bureau of Labor Statistics. He holds a master's degree in applied statistics from Columbia University and a bachelor's degree in economics and mathematics from Duquesne University.
Managing Partner and Lead Economist
John Ricco is a public finance economist with more than a decade of experience building economic forecasting models to answer practical policy questions in topics like tax, public health, and education. With more than a decade of experience at top think tanks, John is a nationally recognized policy expert who is regularly tapped as a source by major media outlets (e.g., the New York Times and the Wall Street Journal) and as a speaker by regional economic conferences (e.g., the Pacific Northwest Regional Economic Conference, where he gave the 2026 keynote address, and the University of Michigan's RSQE Economic Outlook Conference). Outside his employment, he has provided independent consulting on sports economics and contract research on tax policy. He holds a bachelor's degree in economics from Duquesne University and a master's degree in computer science from the University of Pennsylvania.
Our principals speak regularly at economics and statistics conferences and policy forums.
John Ricco. Keynote address, "The Impacts of Tariffs," 60th Pacific Northwest Regional Economic Conference, Portland, Oregon
Steve Bronder. Co-presenter, "Embedded Laplace Approximation in Stan," StanCon 2026, Uppsala University
Steve Bronder. Co-instructor, "Contributing to Stan: A Developer's Guide to the Core, Interfaces, and First Contributions," tutorial at StanCon 2026, Uppsala University
John Ricco. Panelist, "Trade & Tariffs: Where Are We Now and What Should We Expect?," 73rd Annual RSQE Economic Outlook Conference, University of Michigan
John Ricco. Session chair, "Tax Evasion and Enforcement: Measurement and Estimation," 118th Annual Conference on Taxation, National Tax Association, Boston
John Ricco. Panelist, "Distributional Effects of Tariffs," webinar
John Ricco. Panelist, "The Child Tax Credit," Tax Policy Seminars
Steve Bronder. Speaker, "From Bayesian Inference to LLMs: Modern C++ Optimizations for Reverse-Mode Automatic Differentiation," CppCon 2025
John Ricco. Presenter, "PSL Demo Day 55," Policy Simulation Library
Steve Bronder. Panelist on Stan development, StanCon 2024, University of Oxford
Steve Bronder. Co-presenter, "Scaling Stan's performance with reduce_sum in practice," StanCon 2020 (virtual)
Steve Bronder. Co-presenter, "The State of GPU Computation Support for Stan," StanCon 2020 (virtual)
Steve Bronder. Co-presenter, "The State of GPU Computation Support for Stan," StanCon 2019, University of Cambridge
Steve Bronder. Co-presenter, "GPU optimized math routines in the Stan Math library," StanCon 2018, Aalto University, Helsinki
Steve Bronder. Presenter, the mlrfc R package for forecasting, R/Finance 2017
Send us a message to discuss a solicitation or a project, or email us directly.