Discover more from Overcoming Bias
How Deregulate Govt?
Government at all levels has limped into the digital age, offering online services that can feel even more cumbersome than the paperwork that preceded them … Government is hamstrung by a rigid, industrial-era culture, in which elites dictate policy from on high, disconnected from and too often disdainful of the details of implementation. Lofty goals morph unrecognizably as they cascade through a complex hierarchy. … Jennifer Pahlka shows why we must … consider what it would mean to truly recode American government. (More)
When I talked to some people about software rot a few weeks ago, many of them pointed me to Jennifer Pahlka’s new book Recoding America. Her book is filled with enough dramatic examples of government failures to turn many readers into rapid libertarians. For example:
By the early 2000s many of the satellites GPS relied on were nearing the ends of their [life expectancy]. The need to replace them offered a chance to update their software and improve the accuracy and available of GPS navigation signals. … awarded … Raytheon a $1.5 billion contract to develop the Next Generation GPS Operational Control System … revised once more, to $6.2 billion … Eventually … the Air Force had to launch new satellites whether the new software was ready or not, so it went ahead without some of the software updates. … GPS users aren’t getting the improved resolution and other benefits that these hardware upgrades were supposed to bring. … US government has spend many bilious of dollars for satellites that in many regards do this same things the old ones did.
But Pahlka doesn’t blame government; she instead blames our habit of having government pay private firms to write software according to govt-set specs. She recommends having govt hire good tech people like her as civil servants, and giving them great freedom to just do things as their way. For example, she celebrates civil servants Yadira and Natalia who produced good outcomes via defying apparent laws:
The policy team gave [Yadira] the same answer they had given Natalia when she was arguing against Facebook for Doctors, “Sorry, Congress says we have to.” Like Natalia, Yadira was told over and over that the proposal was simply not compliant with the law. … When Natalie averted the Facebook for Doctors disaster, she had to pit her interpretation of the law against that of agency lawyers and policymakers. … “We’re not following the letter of the law,” [Yadira] admits. “But we’re producing the results that we know Congress intended.” (p. 215)
Now Pahlka does admit that our current firms-write-software-to-govt-spec habits did have a seemingly reasonable origin:
There was no reason government couldn’t acquire the equipment and expertise to do automatic data processing on its own. But the perception - often justified - was that, without the dynamics of marketplace competition in play, government would become slow, expensive, and bad at what it did. It had happened before. In 1921, for instance, concerned about high prices for steel amid war-driven demand, the US government embarked on an effort to make its own. It did eventually build a plant, but the steel produced was half as good as what was commercially available and cost three times as much. … Commodities like steel are best left to the market. It followed that the same would be true of [computing]. Both A-76 and the Brooks Act sought to make sure that … computing … would be things government bought, not made (p.104)
Even so, Pahlka wants us to, in essence, “deregulate” government:
Government is a vast interconnected, complex, adaptive system made up of countless subsystems that connect in ways that are rarely fully seen or understood. … Our elected leaders keep thinking in terms of money, regulations, and oversight … but they need to be directed at the problem underlying our delivery failures: the lack of deskilled technologists within government who are empowered to make the necessary decisions. … What we need is not more mega projects but incremental, stage-based funding. … Venture capitalists fund startups to start small and learn quickly, then invest more when the startups get traction and show value. … [This] pairs perfectly with small cross-functional teams that practice agile development and focus on understanding and meeting user needs. … Removing mandates can have more impact that adding them. We desperately need to simplify and rationalize the policy that has accrued over many years and bogs down our systems. … This applies both to government services themselves … and in the procurement and hiring procedures involved in building them.… Hold public servants accountable to outcomes over process. … We must put user-focused public services in many roles … and we must connect them across disciplines. (p.261-8)
None of the dozen or so other reviews of Recoding America that I read mention this issue, but Pahlka actually says very little about the crucial issue of how exactly to hold whole government agencies “accountable” via outcomes. Yes, we may be over-regulating government (as well as the private sector), but it seems far from adequate to me to just say to regulate “less”. A few examples of small projects that worked out well when civil servants were given (or took) more freedom seems far from sufficient here.
Nations have at times maybe sorta made outcome-accountability work by allowing corrupt officials to make billions from their positions, positions which they lose if outcomes get bad. Sorta like how venture capital puts a fire under startup leaders. But if we just pay agency heads tiny salaries, and then give them great discretion over how to spend billions, including who to hire and fire when, then at what threshold of bad “outcomes” do we fire those agency heads? Do we really think that the politicians who oversee agencies will give remotely as much attention to this as venture capitalists? What do we do if agency heads claim to be expert and to be doing their best, but they get bad outcomes while directing billions of dollars to their associates who treat them very well after they are fired? Are we okay with that?
Pahlka doesn’t seem to believe politicians today are at all up to this task:
We know why this is difficult … Politicians don’t get reelected for simplifying and rationalizing the gargantuan maze of laws, politics, and regulations that govern service delivery. … At the same time, politicians have many many incentives to pass new laws and policies that (mostly inadvertently) wind up making delivery even harder and to publicly admonish bureaucrats for failures, which makes the bureaucracy even more risk averse. … Politicians rarely get voted out of office for failing to clear the criminal record of former felons or making it hard for the needy to access food benefits. … I worry. There are the outlines of a better path, but very few are choosing it. … Power in our government is extremely diffuse. Each federal agency makes its own independent decisions about technology development and pretty much all the other factors that go into getting implementation right. So do the states. … For the country to choose the path of digital competence … Hundreds of thousands of people in appointed and career positions across a wide variety of functions … would all need to decide to reject the status quo at approximately the same time. [But] wholesale change rarely happens so quickly. (268-73)
Thus, as far as I can tell, the only solution Pahlka sees is a huge unlikely radical revolution in how citizens hold government accountable, toward “outcome” accountability. Yet she isn’t willing to even roughly outline how that might work. She just says to hire good people (she can suggest who) and free them to do what they see as best. And then hope.
Added 2Aug: Pahika responds:
I hope to be doing some work on this over the coming year. I don't really know the answer yet. But I suspect it involves shifting from big bang to incremental funding and learning to ask different questions.