TaxBrain: Open source economic forecasting
- By Kathleen Hickey
- Sep 14, 2016
As economic policy becomes more complex, it grows less transparent.
To bring some insight into the data and forecasts, the American Enterprise Institute’s Open Source Policy Center (OSPC) has developed a new approach to policy analysis.
The TaxBrain web application lets users simulate and study the effect of tax policy reforms using open source economic models. Developed and launched in April by OSPC, TaxBrain aims “to make economic policy analysis more transparent, accessible and scientific,” AEI officials said.
Today most tax models are proprietary, using licensed software, which “put[s] power in the hands of a few,” according to AEI. These forecasts are used by many government analytic groups working in Congress, the White House, the Treasury Department and independent think tanks to arrive at tax policy decisions affecting millions of Americans.
“I can tell you there are a whole lot of assumptions that are being made [in the models] that aren’t being disclosed,” OSPC Managing Director Matt Jensen told Datanami. “Many of these assumptions are being made by government modelers, but they don’t disclose what the assumptions that they make are, so an external [party] can’t say whether the assumptions line up with the empirical literature.”
Forecasts from these models are used to determine as analysis on how much money a bill might cost or how legislation affects the rich and poor. “Every important expected outcome of a policy is being estimated by these models, and those estimates absolutely affect policy decisions,” Jensen said.
An open source approach gives policy makers, journalists and the general public the information they need to better understand policy, Jensen said. “We are also able to leverage the knowledge and interest of experts and the general public to improve our models and make government better,” he added.
TaxBrain uses the open science Anaconda software from Continuum Analytics, and the models are written in Python. Transitioning economic models from proprietary software like SAS to open source software won’t be too onerous for users, Travis Oliphant, CEO and co-founder of Continuum Analytics, told Datanami.
“The nice thing about Python is it’s very flexible, and you can create interfaces that are closer to what you’re used to,” Oliphant said. TaxBrain users “just see a web application where they can explore and make adjustments and then fire off a model that runs across a cluster of machines that all use Anaconda.”
“The TaxBrain initiative is only the beginning of a much larger movement to use open source approaches in policy and government,” Oliphant said in a Continuum Analytics statement.
Kathleen Hickey is a freelance writer for GCN.