(Daniel E. Ho, Yale Law Journal, December 2012, Volume 122)
(Daniel E. Ho, Yale Law Journal, December 2012, Volume 122)
The Code For America Accelerator program is a fantastic opportunity: $25,000, mentoring by savvy, connected folks, and networking among that rarefied class pursuing the golden apple of bringing to public goods the behemoth transformations we’ve witnessed in the past decade in the private sector, transformations made possible by the big changes in the underlying technological costs and bounds.
(I should not forget to mention the chance to meet Aneesh Chopra, which truly is the wonk-geek equivalent of finding out Arrested Development was coming back.)
Now, I’m a big fan of the idea of sharing as a business model. There are just too many possibilities for transformative change to be hoarding them for a reason as silly as cupidity. These are things I’d love to build, but even if I execute on one of them, who has the bandwidth to do all of them? (Not I.)
But someone can and should build these things.
1. Hack Eligibility and Enrollment
If you’ve ever worked closely with social safety net programs, you understand that determining eligibility and facilitating enrollment are among the most difficult implementation details out there.
On the back end, you face complex eligibility criteria and decision trees, many of which are changing year-to-year (particularly in this budget-slashing environment). Sometimes, you’re also faced with relatively strict audit requirements that, if unmet, will jeopardize the program, but, if implemented haphazardly, can deny people access to your services for minor technical reasons.
On the front end, you have users with highly variable capacity to deal with enrollment. Barriers like language, literacy, and access are huge, and often concentrated among the core population a program’s intending to serve. (How does someone working 70 hours a week with kids find time to go to a social services office with limited hours? Or a library with 30-minute limits on computer use, even for people who can barely navigate within a browser?)
Right now, the best out there are clunky web forms, and the worst are paper forms where you, more or less, require a social worker conversant in the specific requirements of the program to fill it out.
This could change. The specs - while daunting - are fundamentally things we know how to pull off:
I’d start with a small program with specific requirements, so you really hit the user acceptance threshold. But abstract the code enough to build a platform that accommodates the broader E&E requirements of the gamut of programs.
(Also, anyone involved in health reform knows what a massive issue this is. Google some RFPs.)
2. A Mint for Municipal Budgeting
Cities, counties, and other municipalities often have anachronistic budget systems. Very often we’re really talking an Excel spreadsheet on some Windows 98 box (okay, fine; that may be hyperbole, but not far off the mark). It’s hard to implement effective budgeting methods like performance-based systems. But it’s even harder when you’re hobbled by your tools.
And municipal budgets matter! It’s a truism that local governments affect a given citizen’s life the most. Now, for a problem that is structurally the same across tens of thousands of customers (cities, counties, etc.) across the country, there doesn’t seem to be very many cutting-edge tools out there. And even the good ones don’t take advantage of the powerful capabilities that SaaS design (trivial update process) and automated BI/analytics that have almost become commodified in the private sector.
I’d start this by pairing up with some real municipal budget wonks (comptrollers, consulting firms) and figuring out just where their ERP systems fall short. Read Comprehensive Annual Financial Reports (CAFRs), city council budget reports and meeting minutes, and bring a techie’s eye to the use cases.
Then I’d go to a small municipality that’s willing to work closely with you and take the chance on something new: build, get feedback, refine, iterate. Pretty soon you’ve got something most of us budget wonks in localities would die for. And in this budget environment, if you can make a value proposition with net savings, that’s a powerful pitch, even with risk aversion out there.
Heck, once you built this, you could even build a public interface that would make transparency (and, hopefully, more informed public dialogue) dead-easy.
3. Networks to facilitate Jefferson’s democratic laboratories
One of the primary motivating ideas for federalism was that different states (and cities, and counties) could experiment with varied approaches to structurally-similar problems and, collectively, could learn and improve policies through trial-and-error.
There are ample groups that try to facilitate this knowledge dissemination (NCSL, NACO, NLC, etc.), but the infrastructure for sharing and learning is still very basic: policy wonks at these orgs and think tanks write issue briefs, people speak at conferences, yada yada.
But organizing this information in a structured way, so it’s trivially easy for a policy staffer to understand the whole landscape of experiences around a specific policy issue. There is a mildly awesome version of this for policymakers implementing health reform set up by the National Academy for State Health Policy called StateReforum.
Social network-style sites that organize and structure this kind of experiential policy knowledge to streamline the process of getting actionable info on your issue could be hugely impactful. So I say: carpe eventus.*
(* Google says this means “seize experience”; YMMV)
4. Hack the stakeholder process
A major issue, in both the development of legislation and the process of implementing through regulation and programmatic management, is how to effectively utilize the knowledge of affected groups like businesses, advocacy organizations, and other stakeholders who often know a specific issue at a much more granular level of detail than the people tasked with governance.
Improving the process of soliciting comment, navigating and processing input, and responding could be valuable across an array of situations. I imagine a web app that’s essentially a shared document you can mark up, a la Word - but with much more robust capabilities designed around the use cases of policy formation.
For public comment, even just being able to quickly hone in on all the comments made on one provision - rather than in lengthy letters (generally not mapped section-to-provision) - would be great. You could even get quick analysis of which sparked the most comment or controversy, and automatically review different groups’ input on that specific issue side-by-side, structuring and streamlining the comment review.
For more “sensitive” policy formation (legislative), one could limit to only privately-invited entities to review and edit draft text. You could even let them respond to one another’s comments and not have to be the middleman.
And on the “private” side, there’s a larger potential consumer base among interest groups, law and consulting firms, and others who want to be able to effectively manage internal discussion and mark-up of such documents, but keep that input private. This could subsidize what may be a free/low-margin product offered to policymakers without big budgets.
The Federal Register already provides their publications (including proposed rules, comments/responses, and final regulations) in XML, and increasingly such information comes in this or other structured formats.
What do these ideas have in common? Two primary things:
I’m still toying with whether to apply to the Accelerator with another geek friend. But quitting day jobs is tough, despite my love for geeking out on Python. So maybe, just maybe, I’ll use one of these ideas in its infancy.
But the truth is I’d be just as happy for someone else to build any or all of these, because we in the wonk community would kill for them.
And in a world of scarce and diminishing resources, imagine what increasing the underlying productivity of governance by 10% would mean for the people served by government services? Think on that.
(If you’re interested in this kind of stuff and have thoughts you don’t want to leave in a comment, feel free to shoot me an e-mail.)
[Edit: 4:32 ET - minor grammar fix]
Imagine for a moment a world where all of the Repair shops and Automobile Mechanics in the country formed an association. As part of that union, they all agreed that no one would work on any vehicle unless the car owner signed an arbitration agreement. Same goes for the hiring of mechanics — they had to sign as well.
Now imagine when you had a dispute over a repair, you went to the Repair Garage & Automobile Mechanics Arbitration Association. How do you think that would turn out? That is the FINRA arbitration system, only instead of disputes over $600 repairs, its $100,000 of losses — in some cases millions of dollars. What are the odds you will get a fair hearing in this private, opaque, non transparent, literally Wall Street owned system. Its a national embarrassment, a legal sham.
Welcome to the Kangaroo Court of Wall Street.
In the first part of the paper, D&S analyze the optimal tax rate on top earners. And they argue that this should be the rate that maximizes the revenue collected from these top earners — full stop. Why? Because if you’re trying to maximize any sort of aggregate welfare measure, it’s clear that a marginal dollar of income makes very little difference to the welfare of the wealthy, as compared with the difference it makes to the welfare of the poor and middle class. So to a first approximation policy should soak the rich for the maximum amount — not out of envy or a desire to punish, but simply to raise as much money as possible for other purposes.
Now, this doesn’t imply a 100% tax rate, because there are going to be behavioral responses – high earners will generate at least somewhat less taxable income in the face of a high tax rate, either by actually working less or by pushing their earnings underground. Using parameters based on the literature, D&S suggest that the optimal tax rate on the highest earners is in the vicinity of 70%.