PSL institutional infrastructure - next steps

[MattHJensen]

I hope everyone has had a chance to read PSL’s first newsletter and semi-annual report, a draft of the report is currently available for discussion on GitHub. Thanks to @Peter-Metz for taking the lead writing these documents, to everyone who participated to their discussion on GH, and, of course, to everyone who has contributed to producing the achievements detailed within. We have achieved many of the goals that were discussed in late summer.

Now that the Library has a strong foundation, I’d like to propose that a useful next step would be for PSL to apply to join NumFOCUS as a fiscally sponsored project.

NumFOCUS is an open source foundation that sponsors many projects in the open source data science ecosystem. Joining NumFOCUS would provide several benefits to PSL-cataloged projects.

• PSL-cataloged projects could accept tax-deductible charitable contributions. NumFOCUS provides financial infrastructure and administration for a 10% overhead.
• PSL-cataloged projects would gain visibility with a wider audience via NumFOCUS’ communications channels. NumFOCUS email, social media, and conferences would complement PSL's.
• PSL-cataloged project contributors would gain entry to a broader community of open source data science participants.

Joining NumFOCUS is not our only option. Other options to consider include partnering with a different fiscal sponsor, establishing a new non-profit institution, or forgoing a centralized facility for accepting non-profit donations.

My position is that we should apply to join NumFOCUS because the above benefits are worth the additional bureaucratic overhead and institutional risk associated with partnering so closely with another institution. I also believe that NumFOCUS would guide us to put in place useful institutional structures of our own, such as a formal leadership council. Feedback on NumFOCUS from the maintainers of other NumFOCUS projects has been uniformly positive.

In order to apply to join NumFOCUS, we would need to establish a leadership council that has the “authority to manage the technical and artistic direction of the Project and the program activities of the Project.” Council decision making must be majoritarian.

In practice, the leadership council would need to agree to fiscal sponsorship and potentially sign associated documents, as well as set and enforce policies surrounding fundraising. Beyond this, I think the council could take on or not take on duties and oversight of PSLmodels/PSL as it sees fit.

My proposal is for the projects cataloged by PSL and the PSLmodels/PSL project to each have one seat on the leadership council, with the person in the seat appointed unanimously by the project maintainers. If a project does not wish to appoint anyone to the leadership council, it need not.

This approach would mean that, if we are successful growing the number of projects cataloged by PSL, the membership of the leadership council will also grow large and new members will be added suddenly and unexpectedly when they catalog their projects with the library. If the council wishes to change course at any time, it could through a majority vote.

I hope we can discuss this proposal and any alternatives in this thread over the next 7-10 days. If you are happy with proceeding and do not have any follow up questions or comments, please feel free to leave a +1 or thumbs up.

[martinholmer]

@MattHJensen, I like your proposal very much. You have thought this through carefully and it all makes sense to me. I hope the NumFOCUS folks are interested in what we are doing.

[codykallen]

This proposal seems good overall to me. However, I have a specific question regarding multiple projects. @MattHJensen said:

My proposal is for the projects cataloged by PSL and the PSLmodels/PSL project to each have one seat on the leadership council, with the person in the seat appointed unanimously by the project maintainers... This approach would mean that, if we are successful growing the number of projects cataloged by PSL, the membership of the leadership council will also grow large and new members will be added suddenly and unexpectedly when they catalog their projects with the library.

If someone has multiple projects cataloged with PSL and is the primary/only maintainer, would this structure give that person multiple seats/votes on the Council? I think it would be best to limit it to 1 vote per person.

[towashington]

I totally support the motion to apply to join NumFOCUS.

I also think @codykallen made a good suggestion. Under this modified rule, if Projects A is represented by someone who's also a maintainer of Project B, and if maintainers of B think they might have different voices, they would choose someone else to represent them.

If someone has multiple projects cataloged with PSL and is the primary/only maintainer, would this structure give that person multiple seats/votes on the Council? I think it would be best to limit it to 1 vote per person.

[jdebacker]

@MattHJensen I'm in agreement with the suggestion to apply to join NumFOCUS - it'd be a great home for PSLmodels.

[MaxGhenis]

Applying to join NumFOCUS SGTM. I agree with the sentiment above that the leadership council voting system may deserve some fleshing out.

[donboyd5]

On its face it looks very attractive, but there are still questions I would want to know answers to, if I were PSL, before committing 100%.

To me, the attractions are:

1. Association with, and ability to learn from, and possibly collaborate with, a set of mostly-huge already-sponsored projects. This could help with fundraising, too - auras would rub off on PSL. I didn't realize NumFOCUS even existed, but it turns out they do and have a lot of household- and near-household names involved. They all seem like heavyweights. I was impressed to see that NumFOCUS includes Nobel economist Thomas Sargent's QuantEcon. Their corporate sponsors are huge, too.
2. Having a fiscal administrator that can handle all of the nightmares of receiving money (this is not tongue-in-cheek) and administering this money.
3. The benefit, when seeking money, of having that fiscal administrator charge only 10% (a university research fund might charge 40% or more). This leaves more money for actual project work, as opposed to administration, and also makes potential donors more-willing donors.

My questions and potential concerns are:

1. All of the fiscally sponsored projects seem one step closer to technical computing and mathematical tools, and further removed from policy tools, than PSL. Doesn't mean that's a problem - this could be a latent area of interest for them, but it also could suggest that NumFOCUS's priorities lie elsewhere. I'd want to understand more about NumFOCUS's priorities going forward.
2. All of the fiscally sponsored projects are huge. My guess, not knowing any of the numbers, is that they are vastly larger than PSL. There must be some reason for this. Suggests that we may not be their focus for fiscal sponsorship. Or is it possible that they provide only a partial list on their website, leaving smaller ones off? I don't think so, but perhaps I got it wrong.
3. Even the non-sponsored but affiliated projects seem huge, but maybe we're closer in size to them. Again, not necessarily a problem at all - certainly it would be great to be in the company of either the sponsored projects or the affiliated projects - but it would be good to understand.
4. I wanted to understand how much money runs through them - how much their sponsored projects benefit from fiscal sponsorship - but I haven't been able to make a lot of sense of their financial information. Their annual report suggests that about $2 million runs through them in support of projects, if I read it correctly. (See numbered pp. 34-36 -- about the most non-intuitive presentation I have ever seen, including a log scale for financial report information, which no one ever uses, with no mention of it other than the x-scale. No totals, either, just details.) (Their 2016 budget suggests it has grown substantially from 2015 - see bottom of page.) Two million dollars can't possibly be anywhere near enough for their sponsored projects, so (unless I am reading the report incorrectly) those projects must be raising a lot of their money elsewhere. I'd want to understand why.
5. Or, is it possible these sponsored projects raise funds that don't show in the annual report in the way we'd expect, but that the projects still benefit from fiscal administration by NumFOCUS and still benefit from the 10% expense charge?
6. Given that PSL may want to raise money in ways that fall outside the NumFOCUS fiscal sponsorship grantor-grantee structure, it would be important to understand how this could affect flexibility. For example, could PSL raise money, via contract, to develop a product that becomes open source? Could several organizations in a state pay PSL to develop income tax data and an income tax model that are specific to the state, with public open-source code and data (perhaps coupled with a report to the payors that is non-public)? Could such a project benefit from NumFOCUS fiscal administration and the 10% charge? If not, what other legal structures would PSL need in place so that it could accept money outside of the NumFOCUS framework, in addition to accepting traditional donations within that framework?
7. I see that there are two kinds of fiscal sponsorship, the comprehensive model and the grantor-grantee model. There probably are pros and cons to the two approaches; the former seems like more work to set up and to persuade NumFOCUS of but easier to live with, while the latter seems to be the opposite, possibly requiring additional legal entities for PSL.

None of these are red flags (at this point, anyway), although numbers 1 and 6 on the questions list might be potential yellow flags. It would seem to be worth doing detective work on a few of these questions before committing 100%. The best way to do that detective work, in my opinion, would be to talk directly to NumFOCUS, and to talk to 1 or 2 of their sponsored (and possibly affiliated) projects.

[hdoupe]

For the podcast people, the communications director for NumFocus was on the Change log podcast this week. They talk some about the different ways projects can be involved with NumFocus, who their donors are, and what types of projects are included.

[MattHJensen]

Thanks to everyone for the discussion. I agree on all of the points raised. Given that there are some outstanding questions, I will set up a call between myself and someone at NumFOCUS to discuss them. Then I'll report back here.

[MattHJensen]

I have a call scheduled with Leah Silen, the NumFOCUS executive director, next Tuesday afternoon (March 12). I will report back afterwards. Please comment here if you think of additional questions.

[MattHJensen]

I just spoke with Leah Silen. She answered several of the outstanding questions above, as well as some additional questions I asked. The answers below are not verbatim, and some of them rely on my outside knowledge beyond Leah's direct responses.

Given that PSL may want to raise money in ways that fall outside the NumFOCUS fiscal sponsorship grantor-grantee structure, it would be important to understand how this could affect flexibility. For example, could PSL raise money, via contract, to develop a product that becomes open source?

The PSL leadership council will need to decide what PSL is ok accommodating, but NumFOCUS is extremely flexible. If there are contractural deliverables, someone from the PSL leadership council must be responsible.

All of the fiscally sponsored projects seem one step closer to technical computing and mathematical tools, and further removed from policy tools, than PSL.

There are several NumFOCUS projects that are are domain-specific tools, such as QuantEcon, Econ-ARK, Cantera, SunPy, and SolarPy. The NumFOCUS board will ultimately decide, but PSL is definitely in the ballpark of NumFOCUS' mission.

Or, is it possible these sponsored projects raise funds that don't show in the annual report in the way we'd expect, but that the projects still benefit from fiscal administration by NumFOCUS and still benefit from the 10% expense charge?

Many projects receive other resources such as grants through universities, staff contributions from private firms (under a host of arrangements), volunteerism, etc. None of these are fiscally administered by NumFOCUS or charged the 10%.

---

New questions from me:

How long does the process take after the application is submitted:?

3-4 weeks.

1. Leah looks over the app and asks for anything that's missing.
2. Leah finds a point person on the NumFOCUS board.
3. The point person on the NumFOCUS board looks through the materials and gets back in touch with the PSL leadership council if anything is missing.
4. The point person presents to the board.
5. The board decides.

Is there an onboarding process for projects that join NumFOCUS as fiscally sponsored projects

Yes.

1. Document sent to PSL introducing key people at NumFOCUS and what they help with
2. Call between key people at NumFOCUS and the PSL leadership council.

Is there anything to keep an eye on in our application.

Make sure we check all the boxes, including having a governance structure set up and a code of conduct.

[MattHJensen]

Given my conversation with Leah, I think a useful next step is for PSL-cataloged projects to appoint their representatives to a PSL leadership council and then that group should have a series of meetings with the following agenda:

• Formalize PSL's governance structure in some simple document.
• Discuss PSL's near and mid-term fundraising objectives.
• Given those objects, decide whether PSL needs a fiscal sponsor.
• Decide whether to apply to NumFOCUS.
• If yes, plan how to complete the application.

If others agree that this makes sense, then it would be helpful for representatives from PSL cataloged projects and PSLmodels/PSL to report back to this issue that they've been unanimously selected by their projects' maintainers to serve on the PSL leadership council.

[jdebacker]

I'll represent the Cost-of-Capital-Calculator on the PSL leadership council. Glad to talk specifies bout governance structures and objectives, but I think it's clear we could benefit from a fiscal sponsor and really like the NumFOCUS model.

[MattHJensen]

I spoke with the other PSLmodels/PSL maintainers (@hdoupe @andersonfrailey, and @Peter-Metz), and I will represent the project on the PSL leadership council.

[martinholmer]

While speaking with @MattHJensen, we decided @martinholmer would represent the Tax-Calculator project on the PSL leadership council.

[rickecon]

@rickecon will represent the OG-USA project on the PSL leadership council.

[hdoupe]

I'll represent ParamTools.

[MattHJensen]

Thanks everyone. I think the meeting ought to be open to anyone who wants to attend, but to keep the scheduling manageable, I'll email @hdoupe, @martinholmer, @rickecon, and @jdebacker to find a time that works for the council members.

[MattHJensen]

The first PSL leadership council meeting will take place on Wednesday, March 20, at 2:30p Eastern (tomorrow). All are welcome. See this comment for the agenda.

Join Webex meeting at:
https://ospc.my.webex.com/join/matt.jensen | 628 792 770

Join by phone:
+1-510-338-9438 USA Toll
Access code: 628 792 770
Global call-in numbers

[MattHJensen]

Thanks to @martinholmer, @jdebacker, @rickecon, @hdoupe, and @MaxGhenis for participating in our first leadership council meeting on Wednesday, in which we drafted bylaws (#107), decided to adopt the NumFOCUS code of conduct (#108), and decided to apply to NumFOCUS for fiscal sponsorship.

If you have the opportunity, please review and provide feedback on a draft application to NumFOCUS here -- I would prefer if you use comment mode rather than directly editing the document. Assuming that there are no major changes, I hope to submit this application early next week.

[MattHJensen]

I would like to send the NumFOCUS application to Leah Silen (NumFOCUS executive director) for feedback by the end of this week. If you have any suggestions on the draft please leave it in the google doc by this Friday at noon.

[jdebacker]

The draft looks good. I just made a few small edits.

[MattHJensen]

While we wait as the NumFOCUS board reviews our application, we might begin thinking about the fundraising capabilities we'd like PSL to offer. To facilitate the discussion, I drafted an overview of an initial set of capabilities that we could choose offer.

PSL Fundraising Capabilities -- Discussion Draft

The first page outlines the three capabilities, and the second page proposes several questions to begin a discussion about these capabilities or others we might need. I'll leave this for a few days to see how active the discussion is over here, and then I'll set up a meeting of the leadership council to decide how to proceed.

[MattHJensen]

We will have a leadership council meeting on Tuesday, June 4th, at 3p eastern to discuss PSL fundraising capabilities. This meeting is open to everyone, regardless of your current status on the leadership council.

Join Webex meeting at:
https://ospc.my.webex.com/join/matt.jensen | 628 792 770

Join by phone:
+1-510-338-9438 USA Toll
Access code: 628 792 770
Global call-in numbers

[MattHJensen]

NumFOCUS on-boarding has just begun. They need a short description of our project, like the examples you see on this page. Ideas?

[rickecon]

First stab:

Policy Simulation Library develops open source policy modeling tools, encourages collaborative contributions, and makes tools accessible to a broad group of users.

And definitely include a read more button with a link to https://www.pslmodels.org/about.html.

[MattHJensen]

Agree on the link to the about page.

What about this? Builds on Rick's.

PSLmodels is a collection of models and other software for public policy simulation. The Policy Simulation Library (PSL) community encourages collaborative contribution and makes the tools it develops accessible to a diverse group of users.

[MattHJensen]

The Policy Simulation Library (PSLmodels) is a collection of models and other software for public-policy decisionmaking. PSLmodels is developed by independent open source projects that meet standards for transparency and accessibility.

[MattHJensen]

The Policy Simulation Library (PSL) is a collection of models and other software for public-policy decisionmaking. PSL is developed by independent projects that meet standards for transparency and accessibility. The PSL community encourages collaborative contribution and makes the tools it develops accessible to a diverse group of users.

[MattHJensen]

Thanks everyone very much for your participation in this issue and discussion.

The current status of our institutional plans are that:

• PSL has partnered with the NumFOCUS Foundation as a NumFOCUS affiliated project.
• Some PSL community members are pursuing funding for various aspects of PSL development through non-profit research institutions such as the University of Chicago, University of South Carolina, and the State University of New York, as well through for-profit research institutions such as the Open Research Group.
• Some PSL community members are currently funded through the general funds of their institutions.
• We will continue to monitor and reassess whether/when the PSL community needs a dedicated non profit.

I am closing this issue for now, but please feel free to continue this discussion anytime over at our new PSL forum.