📬

Proposal: A Common App for EA Funding

TL;DR

Manifold Markets spent ~$3k in opportunity costs just applying for grants. A common application, shared between different funders in the EA space, could save time for applicants and surface more projects to funders.

Background

  • Sources of grant funding Manifold has applied for:
  • Ballpark estimate: writing these applications up took about 1 fulltime day each
    • Let’s call it 7 days * 8 hours * $50/hour = $2800 spent on writing applications
    • Seems high? Yeah, it seemed high to me too. Some funders (e.g. EA Funds) try to be nice and explicitly say “this should only be a 30min process”
      • But think about incentives: if we apply for a $100k grant, then a 5% increase in success rate is worth $5k. So if writing a good grant takes our probability of winning from 20% to 25% (which feels conservative), it’s easily worth at least one person’s entire day to do it.
      • Also: think back to how long you spent on each of your college applications!
  • Applying for grants sucks (through no fault of the funders)
    • Takes a long time to hear back
    • You get rejected a lot
    • Wastes time that you could be spending on Actually Doing Things

Proposal: Create a Common App for EA funding

Examples of existing Common Apps

  • Common App for undergrads
  • Triplebyte for software engineer jobs
  • Angellist for startup angel investments
  • National Resident Matching Program for med school residents?

V0: A new standard

  • Two or more grantmaking orgs standardize on a set of questions to ask (eg Google Form). Here are the common questions we’ve seen:
    • Two sentence intro
    • One page project explainer
    • Bios of people running the project
    • How much money do you want
    • What you’ll do with the money
  • Opt-in public database where other granters can see proposals (eg Google Spreadsheet)

V1: A new website

  • Funders can set up profiles, listing:
    • What they’re looking for
    • Funding timelines
    • Decision criteria
  • Applicants can submit grant proposals to multiple funders at once
  • Grantmaking orgs can browse through public proposals on an ongoing basis

V2: And beyond

  • The platform could allow funders to coordinate on particular projects
    • Share vetting notes and decisions
    • Split up what portion of funding gap to fill
  • Grant proposals could be opened to the public ala Kickstarter?
  • Quadratic funding could be used to (see Gitcoin Grants)

Problems a Common App could solve

  • Applicants save time applying to different funds
  • Applicants can discover more funding sources
    • Navigating EA funding has been tricky, e.g. we thought Manifold (not being a 501c3) was ineligible for SFF, until Nuno mentioned Rethink Charity. And I’m fairly deep into EA!
  • Funders can see more project applications
  • New funders like Scott can get started without needing to invent infrastructure from whole cloth
  • Funders can see who else is funding and adjust accordingly, reducing over & underfunding

Cost effectiveness estimate

warning, here be dragons

There were 656 applicants to the first round of ACX grants.

  • If 10% of those continue on to apply to EA Funds,
  • and 10% apply to the FTX EA Fellowship,
  • and a common application saves 1 hour per additional application
  • that’s 130 hours saved, or about 3 weeks of full-time work

3 weeks of full-time work is also my rough estimate of how long V0 and V1 would take to build out; so after a single grant round, the Common App would pay for itself. ACX and FTX have both indicated interest in future rounds, and EA Funds presumably will be open for a long time.

(This calculation doesn’t include funder time saved, or the positive externality of more/better proposals being funded through a standardized system. You may also want to discount for a less than 100% chance of success, and the opportunity cost for an engineer-hour vs the average applicant-hour.)

Caveats

  • Funders may have specific criteria they’re looking to fund
    • But just like the college Common App, funders can add supplementary questions, essays, or interviews
  • Increasing the total number of applicants could increase filtering load on funders
    • Solution: funders with too much work add more essays, or delist themselves
  • Big chunk of value of winning a grant is in the prestige of winning
    • But individual funders can still publicize the grants they chose to distribute
  • Another chunk of grant value is in the network of other winners
    • Again, funders can still do this
  • Getting any marketplace/network off the ground is hard (aka cost of convincing funders to get on board)
    • E.g. Would Tyler Cowen want to run Emergent Ventures through this common app?
    • Don’t think this is a huge issue; a common app is already usable if any single funder uses it, and valuable once 2+ funders do
      • Also, the common app could branch out of EA, and work for charities in general (and hopefully draw in more people towards EA causes)
  • Opportunity cost of building this out
    • 2-4 weeks of 1 full-time engineer
  • A new common app might compete for funding mindshare
    • image
  • Maybe this is what EA Funds already is?
  • Unknown unknowns: maybe I don’t understand the funding space
    • Haven’t done much grantmaking myself; would love to chat with grantmaking orgs!
    • Maybe grantmakers already have too many applications to review
    • Maybe most grant applicants only fit for one or two categories, and have an easy time discovering those

Who should implement this?

  • EA Funds, probably
    • ACX Grants already runs through EA Funds infrastructure
  • Lightcone/LessWrong team, maybe
  • Manifold could build this as a side project... if somebody wants to fund us to do it 😉