by Gene Giannotta
The Obamacare rollout has been a study in how not to introduce a major web service to the world. From reports of the debacle, there seemed to be a lack of proper coordination as well as adequate testing. For a system so complex, that was obviously a recipe for disaster.
But while much of the rhetorical whirlwind out of Washington has centered around the question of who’s to blame for the mess, the “who” is largely besides the point. The question we should be asking is why. Just like the other October failure of government - the shutdown - the reasons for HealthCare.gov not working are not a simple matter of finding a person or two who screwed up and firing them, as cathartic as that might be for politicians and others looking for symbolism rather than solutions.
Any entity as large as the federal government, with as complicated a system of intertwining bureaucracies and competing interests both within that network as well as outside, in the realm of public opinion and congressional oversight, is probably more likely to end up with some wires crossed when a project as massive and high-profile as HealthCare.gov goes live.
After all, this isn’t the only botched IT project the federal government has presided over. The Post’s Walter Pincus, a longtime observer of the national security complex, pointed to a recent example under the Defense Department’s purview:
In this case, it was the Air Force. In 2005 it began a program designed to integrate into one system about 240 outdated computer networks at 600 locations that didn’t communicate with each other. It was to manage things such as equipment inventories, contracting, financial administration and personnel assignments.
The Air Force first estimated that the Expeditionary Combat Support System would cost $5.2 billion. On Nov. 14, 2012, it said it was canceling ECSS after spending up to $1.03 billion. The system “has not yielded any significant military capability,” according to an Air Force statement e-mailed to reporters. The Air Force estimated it would need $1.1 billion more to complete one-quarter of the originally designed program. Even so, it would not be ready until 2020.
So imagine if the Obama Administration had delayed the opening of the site until it was “ready.” In that alternate reality, we might be seeing hearings full of politicians wondering where the heck it is…in 2016.
Big bureaucratic entities, whether the federal government or private corporations, tend to face many of these kinds of issues. And in the case of the ACA, implementation requires federal, state, and private organizations all working together, with the added scrutiny of hyper-partisan politicians in Congress waiting to pounce on any slip-up as a sign of its fated implosion.
So in many ways, when commentators and critics point out HealthCare.gov’s flaws, they miss the real phenomenon we should be focusing on.
Just like the shutdown, the Obamacare rollout lays bare the simple fact that unlike a lean startup or small businesses that can afford to wait to release their product when it’s “ready,” government is dictated by political forces that can be far less patient or forgiving. To be sure, politics abounds in private industry as well. The business world is as rife with tensions between competing interests and personalities as the world of governments. But the ACA combines those two atmospheres together to create a uniquely problematic mess.
Consider these points, from Ezra Klein, when he addressed Republican demand for someone to point the finger at:
How about Senate Republicans who tried to intimidate Sebelius out of using existing HHS funds to implement Obamacare? “Would you describe the authority under which you believe you have the ability to conduct such transfers?” Sen. Orrin Hatch demanded at one hearing. It’s difficult to imagine the size of the disaster if Sebelius hadn’t moved those funds.
How about congressional Republicans who refuse to permit the packages of technical fixes and tweaks that laws of this size routinely require?
How about Republican governors who told the Obama administration they absolutely had to be left to build their own health-care exchanges — you’ll remember that the House Democrats’ health-care plan included a single, national exchange — and then refused to build, leaving the construction of 34 insurance marketplaces up to HHS?
Now, this isn’t meant to pin the blame on the GOP, but rather to point out that Obamacare’s implementation - including the construction and reliability of the exchange web sites - hinges on the same sort of politicking that has made it impossible to craft long-term budgets and nearly brought the United States to the brink of default on multiple occasions over the past couple years.
But that also isn’t meant as an indictment of government-run projects, either. After all, it didn’t work out too well when the DoD outsourced the “systems integrator” role on several of its own projects.
TechCrunch’s Gregory Ferenstein, meanwhile, has plenty of criticisms, one which gets to the heart of my point about politics:
The company hired to build HealthCare.gov’s failing database, CGI Global, is an established government contractor (established enough to have actually lobbied Congress on the Affordable Care Act). Even though Canada had previously fired the firm for a botched $46.2 million medical registry system in 2011, CGI Global was still contracted to the build the technical keystone of the U.S. healthcare law.
“I think procurement in the federal government is broken. It favors incumbents and the status quo over the lean start-ups in terms of its archaic procurement rules and regulations,” Vivek Kundra, former U.S. chief information officer, told The Washingtonian.
Startups simply don’t have the knowhow to get around the oddly complicated procurement rules—or the congressional ties to curry favor. As a result, a mediocre contractor charged an astounding $93 million for a botched job.
Yes, but how to “fix” the system? Those large contractors, relying so much on government dollars, won’t just let go of their clout without a fight. And if smaller firms and start-ups could gain access to a more competitive procurement system, the results probably won’t be immune to the same logic of the existing reality. The very nature of competition (and rational human behavior) means that once one reaches the goal, they’re likely to try to hang on, preferably with the least amount of effort (i.e., the least costly route).
Evan Burfield, co-founder of DC startup incubator 1776 and formerly in charge of the company tasked with creating Recovery.gov, was also critical of the procurement system, in an op-ed this week. But while he holds up Recovery.gov as an exemplar of how government can work with startups on smart, effective web site projects, it was not without criticisms in its early days.
Burfield concludes by saying:
The best result that could come from the attention on HealthCare.gov is if Congress seeks to reform how the government procures IT contracts. Americans of all parties want an effective government that costs taxpayers less. Procurement reform would be a win for all of us.
A few weeks ago, I referred to “collective action problems” when discussing the politics of the shutdown and whether it was a “rational” outcome:
Essentially, it means that individuals, by pursuing their own self-interest, will in turn collectively harm the good the group as a whole. For example, lots of individuals thinking “well, what I do won’t make that big an impact, so what the heck” and going ahead with bad investments or irresponsibly getting rid of waste could end up causing terrible effects on the economy or the environment.
What this means is that individual rationality does not necessarily translate to social, or group, rationality. It might be in my best interest (benefits outweigh the cost, and hence “rational”) to litter, or waste a negligible amount of money (to me) on a bad financial bet. But taken together, all of those bad bets and tossed wrappers pile up.
The same logic can apply to procurement. After all, one man’s “cronyism” is another man’s “sticking up for my constituent.” Or fundraiser, as the case may be. The point is that Congress, as a body of 535 individuals with separate incentives, policy ideas, and interests, probably won’t use this moment as an excuse to streamline a process that probably works pretty well for each individual. Contractors know how to divvy up their processes to align what’s good for them with what is good for multiple members of Congress. The classic examples are in Defense contracting, but they exist everywhere.
In other words, we may agree with one part of the problem but the processes involved are so complex that they defy the kinds of easy solutions those diagnoses imply. But we prefer simplicity, so answers that make things sound easy are far more palatable. Like saying the administration should have relied more on startups, a more competitive contracting system, a more centralized approach, or some combination of these.
Or comparing Obama’s government with Obama’s campaign. My Votifi colleague Matt Sarge touched on this earlier this week, saying, “The consensus seems to be that the ACA website is an undertaking unrivaled by anything attempted on the campaign trail.”
The Post’s Brian Fung also tackled this problematic comparison:
Yet political campaigns are geared to do one thing, and that’s to win. Everybody who’s involved in a campaign shares a common interest, and to the extent that their tasks vary, staffers and volunteers nevertheless operate as a team. As the president’s critics often point out, however, governing is different from politicking. All agencies are supposed to work toward the common goal of providing for the public’s welfare. In reality, it’s a messy landscape of cross-cutting political interests and battles over pride and budgets. That makes coordination a lot more difficult. Even when the agencies themselves have agreed to cooperate, their infrastructure might not.
It’s a good sign that we all agree on the need for a more agile, responsive, and coherent process when it comes to government projects generally and contracting in particular. But that doesn’t mean we’ll agree on how to get there. And so maybe we can get those processes more agile, responsive, and coherent, but in order to do so we’ll have to make peace with the fact that they - and the system they exist within - will continue to be annoyingly imperfect and inherently complex.