3 ms·

I strongly believe that open sourcing the analyses that are intended to influence policy would be a big step to mitigating the 'money problem' in Washington. If

I strongly believe that open sourcing the analyses that are intended to influence policy would be a big step to mitigating the 'money problem' in Washington. If we could see exactly which assumptions policy researchers rely on in their analyses, we could judge the analyses more easily and more accurately. Disclosure: I work at the open source policy center at aei.

Karmage 10y ago
Open data should be the rule everywhere. The internet was created to share academic and scientific data. I dream of a world where, if you make a scientific claim, that you are bound to release the data you used to arrive at the conclusion.
- matthjensen 10y ago
  I agree. In the case of fast moving policy debates, open analytical code is also important. There are also situations when government agencies like the Congressional Budget Office, the Joint Committee on Taxation, and many many others use private data sources like tax return data. In these cases, I think they should be required to release the analytical source code as well as data dictionaries that allow others to create dummy or synthesic data. They should also release detailed summary statistics when possible-- after taking differential privacy concerns into account.
- nl 10y ago
  This isn't the same thing at all. Open Data is great! But the OP is talking about open-source models, so one can understand the assumptions that it makes. In many cases this is much, much more important than the data itself. For example, economic models usually rely on assumptions about changes to growth rates/interest rate/unemployment rates etc etc. These aren't really data - they may be based on data, and projected forward based on various assumptions. Often these assumptions are reasonable, but occasionally they are completely absurd.
nl 10y ago
This is really good stuff. People should pay more attention to this[1]. It is a complete, open source[2] mode to allow people to experiment with changes to the US Tax System. I'm not from the US so I can't comment sensibly about how accurate or reasonable it is, but it is more than I've seen anywhere outside the FRBNY DSGE macroeconomic model[3] [1] http://www.ospc.org/taxbrain/ http://www.ospc.org/taxbrain/ [2] https://github.com/open-source-economics/Tax-Calculator https://github.com/open-source-economics/Tax-Calculator [3] https://github.com/FRBNY-DSGE https://github.com/FRBNY-DSGE

Karmage 10y ago

Open data should be the rule everywhere. The internet was created to share academic and scientific data. I dream of a world where, if you make a scientific claim, that you are bound to release the data you used to arrive at the conclusion.

matthjensen 10y ago
I agree. In the case of fast moving policy debates, open analytical code is also important. There are also situations when government agencies like the Congressional Budget Office, the Joint Committee on Taxation, and many many others use private data sources like tax return data. In these cases, I think they should be required to release the analytical source code as well as data dictionaries that allow others to create dummy or synthesic data. They should also release detailed summary statistics when possible-- after taking differential privacy concerns into account.
nl 10y ago
This isn't the same thing at all. Open Data is great! But the OP is talking about open-source models, so one can understand the assumptions that it makes. In many cases this is much, much more important than the data itself. For example, economic models usually rely on assumptions about changes to growth rates/interest rate/unemployment rates etc etc. These aren't really data - they may be based on data, and projected forward based on various assumptions. Often these assumptions are reasonable, but occasionally they are completely absurd.

nl 10y ago

This is really good stuff. People should pay more attention to this[1]. It is a complete, open source[2] mode to allow people to experiment with changes to the US Tax System. I'm not from the US so I can't comment sensibly about how accurate or reasonable it is, but it is more than I've seen anywhere outside the FRBNY DSGE macroeconomic model[3] [1] http://www.ospc.org/taxbrain/ http://www.ospc.org/taxbrain/ [2] https://github.com/open-source-economics/Tax-Calculator https://github.com/open-source-economics/Tax-Calculator [3] https://github.com/FRBNY-DSGE https://github.com/FRBNY-DSGE