When we wrote last year’s year in review blog post, PolicyEngine had only three months and one country model (the UK) under its belt. Still, we showed how our transparent, software-driven approach to public policy analysis could transform society through our range of features, use cases, and partnerships.
This year, we launched in the US, improved our data and modeling architecture, and contributed real time analysis to unfolding policies. We also invested in new partnerships and set the stage for a groundbreaking 2023. Here are some of our highlights.
Launching new policies and models#
In March, we had our biggest launch to date: PolicyEngine US. We initially showed how current policy and custom reforms affect individual households, then added population impacts in July using the Current Population Survey.
PolicyEngine US model of the End Child Poverty Act, a bill from Representatives Mondaire Jones and Rashida Tlaib
But the US is not just one launch: each state has their own benefit program rules, and most have their own income tax as well. We added state income tax models in Maryland, Massachusetts, Oregon, New York, Pennsylvania, and Washington. In addition to video-filled launch posts for New York and Washington, we dove into programs like Oregon’s Earned Income Tax Credit and Exemption Credit, Maryland’s Standard Deduction, and Supplemental Security Income.
Also in March, we applied gradient descent to reweight the UK microdata, making PolicyEngine the most accurate UK microsimulation model. And last month, we forked OpenFisca into our own microsimulation framework, improving the stability and extensibility of our software.
Showing PolicyEngine in action#
Either through us or other analysts, PolicyEngine provided analysis for dozens of policies in 2022. The UK’s rapid leadership changes created a flurry of proposals, which we number-crunched in stride. After analyzing May’s Cost of Living Support Package, we turned to proposals from then-Prime Minister candidates: Rishi Sunak’s Basic Rate cut and Liz Truss’s marriage allowance rise. When Truss took 10 Downing, we studied her energy bill price cap, stamp duty cut, and cuts to Income Tax, National Insurance, and other taxes.
In the US, we modeled programs in the Inflation Reduction Act like clean vehicle tax credits, and more recently, we recorded how we updated our model’s SNAP rules to reflect the omnibus bill reforms. We also shed light on the giving season by showing how charitable contributions affect taxes.
Chart built with the PolicyEngine Python package showing the impact of the electric vehicle tax credit before the Inflation Reduction Act
We shared PolicyEngine far and wide, from the Green Party of England and Wales’s PolicyFest to the US Congressional Budget Office. We presented PolicyEngine at Georgetown’s Beeck Center, the Basic Income Guarantee Conference, Berkeley Initiative for Transparency in the Social Sciences, Tax Economists Forum, Civic Tech Chat, Move Nevada Forward, the Society of Actuaries, and various UK parliamentary groups.
From left: Scott Santens, PolicyEngine research assistant Aleida Baumgartner, and PolicyEngine co-founder and CTO Nikhil Woodruff, at the Basic Income Guarantee Conference
For researchers, we wrote about how we estimate the effects of UK carbon taxes and how to use PolicyEngine to simulate basic and guaranteed income policies. PolicyEngine powered research projects at the Centre for Policy Studies, , UBI Center, Center for Growth and Opportunity, Maryland Child Alliance, End Poverty Make Trillions, Centre Think Tank, and more.
And for developers, we also wrote about how to use our API, which is used in products from the Fund for Guaranteed Income and . This developer documentation made our hackathon projects with DemocracyLab and EA Global more productive, too.
Left: PolicyEngine team at the DemocracyLab Hackathon in Seattle. Right: PolicyEngine/GiveCalc team at the Effective Altruism Global Hackathon in Berkeley.
Earlier this week, we released our plan to improve our US model’s accuracy with modern data science techniques. We’re excited to start that work in early 2023, and we’ve also been exploring other opportunities to leverage the recent explosion of artificial intelligence tools to make policy analysis more robust, accessible, and even delightful.
Much more is on our horizon. We’ve been hard at work on a ground-up rewrite of our app, which will improve usability and add features and state tax models. We also expect to model all remaining US states and PolicyEngine Canada (at least a prototype) next year. At that point, PolicyEngine will provide real time analysis on a wide range of policy conversations. If you’re a policy analyst looking to amplify your work with best-in-class tools, we’d love to help.
This wouldn’t be possible without our donors (Dylan Hirsch-Shell, End Poverty Make Trillions, Innovation Network for Communities, and ), partners (Center for Growth and Opportunity, UBI Lab Northern Ireland, Maryland Child Alliance, OpenFisca, and Policy Simulation Library), advisors, and dozens of open source contributors. We’re grateful for the support of each of these individuals and groups, most of whom joined us this year. If you’d like to accelerate our development and growth, you can make a donation or get in touch.
Thanks for a wonderful 2022. We can’t wait to share more next year!
