This is a periodic newsletter of the interesting things we’ve seen and what we are thinking about in open source policy analysis.
OSPC-incubated projects featured in Green New Deal analysis. In a recent AEI Economics Working Paper, AEI’s Aparna Mathur explores policy proposals to accomplish some of the stated goals of the Green New Deal. Mathur first advocates for a carbon tax as a means for achieving climate goals. She then explores possible solutions for raising revenues, reducing economic inequality, and increasing work incentives. Using Tax-Calculator,* Mathur estimates revenue and distributional impacts of increasing the top marginal income tax rate, expanding the earned income tax credit and the child tax credit, and introducing a universal basic income. Mathur then uses the Paid Family Leave Cost Model* to estimate the cost implications of various paid family and medical leave scenarios, a policy mentioned in the Green New Deal. Link
Test your doodling ability with artificial intelligence. To play “Quick, Draw!” you have 20 seconds to draw a doodle on your computer, and an open source neural network developed by Google tries to guess what you are drawing. The model is trained with a database of over 50 million doodles from users like you. See if the algorithm can recognize your doodles or contribute to the dataset with your own drawings. Link
An unsurprising drop in charitable giving. A year ago, AEI’s Alex Brill and Derrick Choe used Tax-Calculator* to predict a decline in annual household charitable giving of $16.3 – $17.2 billion as a result of the Tax Cut and Jobs Act’s increased standard deduction. In a recent blog post, Brill and Choe compare their 2018 prediction to a new estimate from Giving USA, which reports a decline of $15.5 billion in household charitable giving over the past year. Link and link
The case against differential privacy. In a Twitter thread, Steven Ruggles (University of Minnesota) makes the case against differential privacy (DP), a new privacy formulation for the 2020 Census that we have discussed in this newsletter several times. Ruggles uses specific examples of DP-protected data to argue that DP may make tabular data unusable for most applications of small-area data – such as drawing school boundaries or assessing racial segregation – and is not appropriate or feasible for producing microdata. Link
Open source software in self-driving cars. Tier IV, a Japanese startup that develops open source software for self-driving cars, recently raised over $100 million to facilitate commercialization of its technology. The software is already used by hundreds of companies, auto manufacturers, and government agencies. Link
Policy Simulation Library (PSL) June newsletter. The monthly PSL* newsletter is complete with user highlights and model updates. Link
* These projects are attendees or graduates of OSPC’s incubator program.
Edited by Matt Jensen and Peter Metz