Macro, micro, and externalities

Alistair Croll
Published in
6 min readOct 14, 2017


Over the last few months, working on FWD50 has given me a chance to dive deeply into issues of government transformation. I’d spent some time working with the Presidential Innovation Fellows and Code for America over the past few years, but hadn’t really understood the challenges and potential of digital technology.

One of the things I’ve realized is that, at their core, most issues in government policy (and, for that matter, business) come down to two things: Tension between the many and the few; and the creation and regulation of externalities.

Macro versus micro

Most Western governments have two fundamental goals: Democracy, which takes a macroscopic view that the will of the majority should be the way the country is governed; and constitution, whose microscopic view protects the rights of the individual.

A democracy without constitution is little more than a mob, treading on the minorities. And a constitution without democracy is a collection of special interests, each clamouring for its own advantage. One of the first things I learned about in civics class was Rawls’ Veil of Ignorance: If you were to design a society, you should do so without knowing who you’ll be in the society once you’re done. You might be a pawper or a prince; a criminal or a saint; a member of the majority of the minority. It’s a good thought experiment.

The correct answer, however, is far more complicated. Nearly every major public policy discussion falls on a spectrum between free-for-all Libertarianism and a totalitarian Nanny State. (Happy to argue about the location and width of these arrows, naturally.)

Policy also makes strange bedfellows, with many issues (recreational drug and alcohol use, abortion, etc.) not appearing where you’d think they’d lie on this spectrum. Jonathan Haidt’s The Righteous Mind talks about how fundamental beliefs in things like purity or tribal affinity govern these policies far more than we think; his TED talk is an excellent introduction to these concepts:


An externality occurs when a cost or benefit is not properly captured by the system it’s in. For example, heavy trucks damage roads far more than cars, but if that wear and tear isn’t properly captured in taxation, the truckers derive a benefit from it because they aren’t paying their true cost.

Capturing true costs is difficult. Imagine that two people pay $300 a month each in car insurance. One of them drives a thousand miles a week; the other uses their car rarely, driving a total of 20 miles a week. The risk of accidents increases with miles driven, but the coverage is billed by hours of car ownership.

When a government doesn’t deliver a workable solution for its services, the private sector can step in to help. Tax preparation software is a good example of this—taxation is a government function that every citizen must use, but if the government’s systems aren’t acceptable, then the private sector can step in. Scheduling medical visits in a public healthcare system seems like a government function, but there are dozens of medical scheduling applications.

Government creates externalities when it lags behind on innovation of the things the society has decided it should deliver; and it reduces externalities when it regulates those externalities—for example, by legislating pollution, or imposing a carbon emissions tax.

Where do we draw the line?

Every society is essentially an argument about where we draw the line between public and private. One approach is to insist that the government delivers all its services; another is to let the free market deliver services where doing so is lucrative, letting government ensure that the rights of the marginalized are still taken care of.

For example, a commercial organization might build a portal to access government data, but the government might provide printed copies for those without computers, or a voice-based system for the blind. But this in turn can lead to marginalization if those government-run services fall behind, disadvantaging those who can’t afford the commercial solution.

Much of what we see as politics is the negotiation of acceptable political discourse, known as the Overton Window. It’s the realm of acceptable political discussion, and it’s tugged in either direction by activists. LGBTQ rights weren’t on the agenda a century ago, but thanks to activism, today it’s controversial for a politician to oppose them. The same is true of slavery, suffrage, and myriad other issues over the centuries.

Digital government changes this dramatically

In putting together content for FWD50 over recent months, I’m struck by just how fundamentally technology forces us to rethink all of this. Consider my earlier example of two car drivers. Today, we have the ability to capture miles driven, and charge drivers based on the risk they accrue. In fact, a number of startups such as Metromile are doing just this.

Just a few of the fundamental shifts with which digital government must contend include:

  • The personalization of a service to an individual.
  • The reduction in costs thanks to digital delivery.
  • The ability to analyze and track every interaction.
  • The notion of representation and direct democracy in a connected population.
  • The organization of society by personal affinity rather than geography.

A good definition of technology is “solving a human problem.” When we create a technology (such as the internal combustion engine) we create—and need to regulate—externalities (through seatbelts, speed limits, taxation per mile, insurance.) The gap between the creation of a technology and its regulation often leads to profit and calamity. Algorithmic trading made Wall Street rich, at the expense of homeowners; leaded gasoline made cars run smoothly, but may have created a spike in crime.

Tech outstrips governance

Unfortunately, while technology has accelerated, our legislative systems haven’t kept up, particularly in the Common Law model where precedent takes time to work its way through the courts. In communications alone, we’re innovating at an exponential rate. We took 10,000 years to come up with writing; 1,000 for the printing press; 100 for radio; and 10 for the smartphone. Technological novelty has increased, but the governmental innovation hasn’t kept pace. As Lawrence Lessig observes, we’re now seeing the impact of commercial journalism on democracy, with regulators scrambling to catch up.

I’ve been teaching a class on data science and critical thinking at HBS. In a discussion of Cathy O’Neill’s How Algorithms Rule Our Working Lives, we talked about some of the issues with automated processing of job applications. An astonishing 72% of US resumés are never looked at by a human, and the algorithms that screen out the bulk of applicants may be horribly biased.

But at the same time, that’s in response to technologies that let users send out hundreds of resumés with a click of a button where once they might have sent only a few, hand-typed, applications by post. Digital applications are the HR equivalent of asymmetric warfare, and companies have responded with automation of their own.

The real challenge for policy makers in a digital era is this: As digital technology creates an asymmetric threat of new externalities and widespread injustices, government needs to respond in kind. Those responses need to be both macroscopic (scaleable, exponential systems that inherently adjust their models and are self-reinforcing) and microscopic (regulation, governance, appeals process, and oversight.)

Photo by Oscar Keys on Unsplash

If Rawls were here, he’d tell us to put the blindfold back on, and to reconsider everything we’ve built as a society—knowing what we now do about ubiquitous connectivity, automation, artificial intelligence, clean energy, and the myriad other innovations in Pandora’s digital box.



Alistair Croll

Writer, speaker, accelerant. Intersection of tech & society. Strata, Startupfest, Bitnorth, FWD50. Lean Analytics, Tilt the Windmill, HBS, Just Evil Enough.