CaLaw: Lessons from a legal hackathon


Legal informatics is here; even if it isn’t really defined yet.  Whether it’s through Online Dispute Resolution startups or creating more accessible legal databases through open source hackathon, the legal field is finally ready to succumb to a revolution it has been able to stave off for two decades now.  And once the attorneys give, so will every other last redoubt.  To wit: my first legal hackathon.

My first legal hackathon, indeed, hackatahon of any type, was with #CaLaw held at the MapLight organization in Berkeley.  Organized by Ari Hershowitz of Tabulaw, #CaLaw is “a collaborative effort to make California’s laws more accessible to everyone.”  For those that are unfamiliar, a hackathon is generally described as X.  For #CaLaw, that meant developing tools to make the California legal code more accessible.

For instance, as of now, there is no not for profit open source database of the California statutes.  There are for profit entities that provide the information, but of varying quality.  In addition, California law requires the publication of revisions and proposed revisions of legislation every night, work performed by Leginfo or the Legislative Counsel of California.*  But it’s obvious the website has not been updated since its launch in 1994.

Enter #CaLaw.

#CaLaw brought together hackers and attorneys and even a lobbyist (thank god, because he actually could explain how Sacramento really works) to come up with tools for the public.  For instance, two of the participants (Brad & Grant) have been building  Legix grabs data from Leginfo and puts it into more accessible formats, such as XML.  These two are the co-founders of a for profit legislative research and tracking tool.

In any case, as my first hackathon, I spent most of the time observing and learning.

The best lesson, however, was the refutation of what was considered the most obvious: money = influence.  But let me backtrack a minute, our initial though was that we can track influence over proposed legislation, by identifying money spent (donated to solicitations) with final changes to and proposed changes to the law.  For example, a measure to revise an existing law in favor of, say car companies, would come with clear donations from car companies to the sponsoring politicians.


Perhaps the best thing that came out of #CaLaw was the insight that what we expected did not actually exist.  Indeed, when we began, it appeared simple.  At the core, our goal was to identify influence, i.e. how money spent by lobbyists influenced legislation.  Our approach rested on several assumptions: (1) we can track changes to the law; (2) we can identify which politicians proposed the changes and how they voted; and, (3) we can track money spent by lobbyists in favor or against proposed legislation.  From a data standpoint it made sense, thanks to Leginfo we knew what laws were changed, how often revisions occurred, when the previsions were proposed, and by whom.  And, thanks to MapLight (and in part their API) we knew which politicians voted for which legislation and how money flowed.

Simple, right? Voila we can identify influence.  Nope.

There were two problems, (1) that’s not how the system works in practice; and, (2) MapLight’s data was insufficient.  The first problem was the insightful piece, the second problem simply illustrates the difficulty in acquiring that most important of factors in developing useful applications: data that is proper, accurate, relevant, and sufficient.  Consequently, building the tool wasn’t possible.  But the lessons were still powerful.

Problem 1: even where data was sufficient, there was no direct relationship between money and influence.  We began with a simple purpose: find sections of the code with multiple revisions and lots of money given to one side or the other and identify the final measure.  First issue, most proposed legislation did not have much revision in committee –one change if any.  That meant money wasn’t causing a change in the measure, e.g. watering down a measure that increased regulation.  Second issue, not a lot of money was pouring into proposed revisions.  That is to say, a proposal wasn’t followed by a huge grant of money in support or against a proposed revision.

Turns out, Sacramento doesn’t operate necessarily the way we think (or our group presumed), at least not directly.  Groups need not give money directly to a politician in support or against legislation.  In many, if not most cases, special interests wrote the legislation themselves, the politician is just proposing it.  So no money is spent.  And, where such legislation is not proposed by the sponsoring politician, special interests need only express their support or opposition; once indicated, the politician knows that should they go against the interest group, else they might be cut out of more general fundraising opportunities or other means of support.

Problem 2: MapLight tracked data on some legislation, but not others.  Some of this was because MapLight operates on a short budget and has to pick and choose where to shine a light, e.g. on the most high profile legislation.  But MapLight is also limited by the current system.  Reporting requirements of special interest lobbying are slippery to say the least.  For instance, “lobbying” must be reported, but “public education” does not need to be reported  the difference is slight –lobbying is expressing support or opposition, public education is expressing a view point on impact of legislation, same effect, but legally distinct –but allows special interests to hide their effective stance, let alone expenditures.

Lessons from the hackathon –all is not lost.

These setbacks aside, the event was great.  As Ari noted, it wasn’t until he organized the event that he was able to gather all these people together and demonstrate that there was a community of developers, attorneys, policy makers and those that just care.  Baby steps as it were.

But there are also more practical lessons as well.

1. Open formats and democracy.  There must be a push for open formats of all State information, laws and agency data.  Similar to the Federal governments efforts via, all State data should be made public in an open standard format, with few exceptions.  Such an effort would be the legitimate inheritance of the open government movement and indeed, democracy in the 21st century.

2. Searching for influence.  Although our assumptions about money and influence weren’t quite correct, we still learned something, how Sacramento works.  That’s a lesson that should be passed on as fast as possible, because democracy demands an educated society.  Moreover, it sheds some light on how money does influence politics and is able to avoid the light.  At least now we know where not to look.  And, although no tools were built, once the lesson was learned, new applications and tools were discussed and new ideas generated.

3. Open formats and the economy. Our efforts were civic minded, our group wasn’t seeking to make money.  But it was pretty clear how money can be generated if the State would adopt an open information platform.  Following the lines of, developers would be able to leverage the legal codes or data generated by state agencies, to give or sell.

4. New discovery tools. Government regulations must create clear standards for reporting information, all of it.  In so doing, non-special interests groups can focus their, limited, resources on getting to the most valuable information.  For instance, reporting by lobbyists now must be done in a PDF format.  That’s good.  Although makes it difficult to pull data from the reams of digital paper special interests submit, technology is making it easier.  For instance, Google’s GSA tool or the Mini (pictured) can easily sift through and find the relevant information.  But it makes it easier to create wrappers to present that data if the original standard is simple and universal.


*From the Leginfo website:

What Legislative Information is Available?
California Law requires that for each current legislative session, the following information be made available on the Internet:

  • The Legislative Calendar
  • The schedule of legislative committee hearingsA list of matters pending on the floors of both houses of the Legislature
  • A list of the committees of the Legislature and their members
  • The text of each bill introduced, including each amended, enrolled, and chaptered form of each bill
  • The history of each bill introduced and amended
  • The status of each bill introduced and amendedAll bill analyses prepared by legislative committees in connection with each bill
  • Any veto message concerning a bill
  • The California Codes
  • The California Constitution
  • All statutes enacted on or after January 1, 1993
  • Rules of the Legislature are rules the Legislature adopts to govern its own legislative procedures. These rules are available on the Internet, as is information provided by individual Legislators. In addition to legislative information for the current session, information for the prior legislative session is also available.

All of the information is directly related to the legislative process and how a bill becomes law. Once laws are passed by the Legislature they can have a direct impact on your life. In a representative form of government, citizens possessing knowledge about issues and the process of government can affect the policy decisions under consideration by the Legislature.
Prior to selecting any of the legislative information, it is important that you understand the legislative process and how and where the legislative information fits into the process. A review of Appendix A, Overview of the Legislative Process and Appendix B, Glossary of Legislative Terms should be helpful.


Be first to comment