Open Government Austin

When you assume ...

False assumptions lead to false estimates.

I wrote earlier about how the city's assumption that the Open Government Online amendment required them to put constituent email online in real time caused them to overstate their cost estimate. (The charter amendment itself actually only requires that they archive it for a longer period of time - an expense, but not on the scale the city describes.)

To assume that the city's incoming email would have to go online in real time would cost a lot of money, but the amendment doesn't require that. When you make that assumption, though, there's a lot of extra, expensive stuff you get to build into your cost estimate. So how much would we save by not doing it? Here's Peter Collins (from the closed-caption transcript of the city council meeting on the city's website) on what costs he assigned in his $36 million cost estimate to that phantom expense (sorry about the caps, that's how it is on their system):

I WOULD HAVE TO HAVE A CONTENT MANAGEMENT TEAM THAT WOULD SOMEHOW MANAGE ALL THE E-MAIL. AND, FOR EXAMPLE, JUST FOR COUNCIL, THE MAYOR'S OFFICE, COUNCILMEMBERS, MAYOR PRO TEM AND CITY MANAGEMENT DOWN TO THE ASSISTANT DIRECTOR LEVEL, JUST TO THROW A LITTLE VOLUME OUT, JUST EXTERNAL E-MAIL AND INCOMING E-MAIL IS ABOUT 100,000 A MONTH. THAT DOESN'T COUNT INTERNAL E-MAIL GOING BACK AND FORTH MAYBE TO THE HEALTH DEPARTMENT OR MAYBE TO HUMAN RESOURCES, SO THAT NUMBER COULD GET A LOT HIGHER.

AS FAR AS YOUR TELEPHONE LOGS, IT'S PROBABLY ABOUT 26 CALLS A MONTH ALONE THAT WOULD HAVE TO BE LOGGED AND NOTED. BUT AS FAR AS WITH THE REALTIME, THE OTHER APPLICATIONS, WE ACTUALLY HAVE, OUR COUNTER SYSTEM IS REALTIME. WHEN WE DO CERTAIN TASKS IN THAT, IT IS AVAILABLE FOR OTHER FOLKS TO SEE. BUT AGAIN, IT ALL DEPENDS ON OUR NETWORK, THE BANDWIDTH, HOW WELL OUR FIRE WALLS ARE WORKING IN SOME CASES, THE ROUTERS.

IT'S A COMPLETE PACKAGE TO PROVIDE REALTIME. THE QUOTE HERE OR THE ESTIMATE -- AND IT IS AN ESTIMATE. WITHIN THAT ESTIMATE, WHAT DRIVES COST TREMENDOUSLY IS REALTIME. YOU PAY A PRICE FOR REALTIME. IT'S A PREMIUM. BUT ALSO IN THIS THAT THERE IS -- WE UNDERSTAND THAT THERE IS NO WAY TODAY THAT I COULD PROVIDE AN APPLICATION THAT CAN READ AN E-MAIL AND MAKE THAT JUDGMENT CALL THAT I'M NOT GOING TO VIOLATE ANYBODY'S PRIVACY. SO THAT'S THERE'S WHY THERE'S ALSO A STAFF FIGURE IN THE ESTIMATE TO FIGURE OUT WHO IS GOING TO DO THAT WORK TO START FILTERING CONTENT BEFORE OR SOMEHOW IS IN REALTIME GET TO THE WEB TO WHERE SOMEBODY CAN ACCESS IT. SO IT'S A VERY CHALLENGING PART TO DO. AND THE BIGGEST MISSING FACTOR FOR US IS THERE'S NO PROGRAM -- NO ARTIFICIAL INTELLIGENCE IS GOING TO COMPREHEND THAT.

So if you assume that sort of elaborate system is required, then yeah that can get really expensive. But if you assume all that's not necessary, if you assume all that's required is that email is archived (there's a whole section specifically spelling out requirements for electronic communications), how much would that reduce the $36 million estimate? My guess is quite a lot, but you can't tell how much from the city's cost figures (pdf).

So that's how spurious arguments by lawyers become spurious budget estimates by IT experts - the lawyers interpret the language in an inaccurate, politicized way, then the IT guy costs the estimates based on the assumptions he's given. At the end of the day, though, the number developed is more of a political figure than a business-like estimate. It's not that implenting the amendment wouldn't require any front end investment, but the city's cost estimate overstates the amount by a lot.

The city also misreads the amendment to require immediate transition to online disclosure of all public information. This leads to millions in short-term consulting costs for immediate implementation. A less sweeping, aggressive interpretation, though, might note that only a few areas of city business must be posted online within one year of passage; the remainder must be done as “expeditiously as possible” and "to the greatest extent practical." This “practical” basis means implementing public disclosure of city information on the internet on a timeline that makes sense, in a practical, cost-effient manner, not based on impractical, over-the-top extremist interpretations.

Don't forget that this proposal is a charter amendment. After it's approved by voters, the city council gets to pass ordinances that interpret how narrowly or broadly these passages will apply. At that time we can assume the most vocal critics on city council then will have their proper chance to decide which records should or sometimes shouldn't go online - e.g., when they think it wouldn't be possible or practical. Some on the Austin city council feel this amendment means proponents don't trust them, but in this and many cases, decisions about what records to put online will, at the end of the day, be entirely in council's hands.

posted by Gritsforbreakfast @ Saturday, March 11, 2006