Plan B?

If can’t get fixed in time, how about hiring enough call-center capacity to get the job done?

I’m so old …

[voice from the audience: How old are you?]

… why, I’m sooooooo old that I remember, back before the Good Lord gave us websites, we had things called “call centers” where banks of telephone operators took phone calls and consummated transactions. That was a big improvement over making people come in to physical offices.

So if – once again – a contractor f’d up an IT project and we can’t get working in time, what would it cost to hire enough call-center capacity to get the damned thing done between now and the end of the year?


1. Recall that Obamacare was supposed to be a state-run program, with the federal exchange just a backup for states too dumb or stubborn to do it themselves.  As far as I can tell, the state-level programs are mostly working fine. This ought to teach voters a lesson about voting for teahadi politicians, but probably won’t.

2. I’m more than willing to believe that the WH and HHS blew this one by not paying close enough attention. But go ahead: you try to run a big software project through federal procurement under a hard deadline, and come back and tell me about it.

Contracting-out can be a fine idea. It can also be, and often is, a recipe for disaster, especially when the underlying bureaucracy has been so hollowed out that it can’t do the appropriate monitoring and supervision.  Paradoxically, an agency needs to be stronger and better-managed to run a contracting-out system than it does to do things in house. Of course heads should roll over the rollout, but it’s harder to make heads roll when the guilty parties aren’t on your headcount.


Author: Mark Kleiman

Professor of Public Policy at the NYU Marron Institute for Urban Management and editor of the Journal of Drug Policy Analysis. Teaches about the methods of policy analysis about drug abuse control and crime control policy, working out the implications of two principles: that swift and certain sanctions don't have to be severe to be effective, and that well-designed threats usually don't have to be carried out. Books: Drugs and Drug Policy: What Everyone Needs to Know (with Jonathan Caulkins and Angela Hawken) When Brute Force Fails: How to Have Less Crime and Less Punishment (Princeton, 2009; named one of the "books of the year" by The Economist Against Excess: Drug Policy for Results (Basic, 1993) Marijuana: Costs of Abuse, Costs of Control (Greenwood, 1989) UCLA Homepage Curriculum Vitae Contact:

73 thoughts on “Plan B?”

  1. From what I understand, a big part of the problems with are actually in the way the government IT systems communicate with the insurers’ IT systems, and a call center operator is essentially using the same line of communication to the insurers’ systems. Can anyone who’s tried to sign up by phone confirm this?

    1. I understand that the coordination with a gazillion different data structures, one or more for each insurer in each state, was and remains a big part of the problem. Still, though, this doesn’t seem to cover it very well:

      1) A lot of people haven’t been able to get nearly so far as the comparison shopping stage of the process; given that, why should all the insurers’ different infrastructures even matter?
      2) In the long term, complete interoperability may be desirable – but in the short term, it seems to me this could have been much simpler, especially if instead of agreeing to interface with the idiosyncratic system of each insurer the exchanges had simply mandated that they would provide a data packet to each insurer containing all the necessary information in a preset format, and the insurers would figure out how to handle it.
      Similarly, could all the different details of available plans could have been entered into the exchanges’ systems manually and crudely, rather than requiring the ability to consult the insurers’ computers every time? Could the insurers have been required to provide that sort of data to the exchanges in a mandated standardized format?

      There may be sound legal or regulatory reasons the exchanges couldn’t simply pick data formats and require the insurers to figure out how to conform with them – but it sure seems easier to me than creating a single system able to interact with dozens of differently formatted systems.

  2. I haven’t been following the various analyses of this giant f**kup, and don’t really have much basis for assigning blame to private contractors or govt agencies, etc.

    That said, I expressed the opinion many months ago that a smooth implementation of ACA ought to be one of the administration’s absolute top priorities. I am dismayed – well, more than dismayed, angered – that it doesn’t seem to have been. This is just incompetence, in a crucially important situation. Inexcusable.

    1. Sure, because anyone can just order up the smooth implementation of a huge national cross-sector IT project from Amazon and have it delivered by FedEX. It’s possible someone let this slip. But neither of us has the information that would justify a verdict “just incompetence.”

      1. Anyone, Mark? No.

        Anyone with the resources available to the federal government? I think, yes. And there is a long way between “smooth implementation” and what is going on with ACA.

      2. Amazon does have a scalable web service for hire; IIRC, some guy named Barack Obama used this for a highly successful interactive marketing campaign back in2012. The security part would be the big question.

        1. I don’t think security is necessarily the issue, but (the first couple of days aside) I doubt the problem is resources. It seems like the story we’re hearing was actul buggy software and failures in coordinating dozens of insurance provider database infrastructures – not a problem that coul dbe solved by simply throwing more servers at it.

          1. Warren,

            I wasn’t talking about servers, but rather money in general – that could, I suppose have been used to get better management of the overall project.

        2. I don’t think security is an issue since Amazon apparently just got a huge data storage and management contract with the CIA.

      3. It’s a matter of priorities. The NSA reportedly has a mammoth system that smoothly interoperates with every computer system connected to the internet. Apparently it’s more important to the federal government to be able to spy on everyone than it is to help facilitate health care insurance for it’s citizens. Who knew?

        1. “More important to the Federal Government?”

          Perhaps you meant “more important to the hostage-takers in the legislative branch who control the purse-strings.”

          And like you said … who knew?

  3. 1. “You might want to avoid signing up on Day One.”
    2. “Yes, some workers are being required to use exchanges — but not these exchanges.”
    3. “Expect to be confused.”
    4. “Don’t bother asking our staff for recommendations.”
    5. “Blue states do it better.”
    6. “Abuse our honor system at your peril.”
    7. “You’ll still pay for this, even if you don’t use it.”
    8. “We’re a magnet for hackers and con artists.”
    9. “You might not be able to keep your doctor.”
    10. “Competition is for the greater good — except when there aren’t any competitors.”
    10 things Obamacare won’t tell you: The health exchanges, central to the law, are also its biggest mystery

    1. A) The days when the Wall Street Journal was a credible news organization have long past, and in the case of anything connected to its editorial-and-opinion arms (such as this sort of newsertainment / news analysis is) never existed.
      B) I specially like the companion story “Individual Policy Pricing Soars Under Obamacare”, which (a) runs counter to a number of other stories; and (b) relies on the utter fiction that “individual policies” existed before Obamacare. In fact, anybody who’s looked into it found that Rescission and Pre-existing Conditions, not to mention Coverage Caps, meant that before Obamacare “individual policies” were almost guaranteed to be useless to those subscribers and would-be subscribers in greatest need of them.
      C) As to the clickbait listicle you’re so interested by:
      1) Really? A massive software project that you shouldn’t sign up to on day one? This is unprecedented! Less snarkily, it is genuinely bad. But hardly shocking, nor damning.
      2) If you read the text, you’ll find this has bugger-all to do with the ACA.
      3) Again, read the text. It is 350 words of utter meaninglessness. I was going to summarize its key points – and found there genuinely weren’t any.
      4) In some states, hostile politicians have made it difficult or impossible for people to help you navigate the system. This isn’t so much a defect within the ACA as it is Republicans being monsters.
      5) See (4). Also, Duh.
      6) Really? Making false statements in order to obtain money might have consequences? Especially when you do it in writing on a government form? There goes my lucrative sideline selling vacation properties in Florida. Also, Duh.
      7) Imagine that. A social-welfare program that takes money from the better-off and gives it to those in dire need. An amazing discovery, and so well hidden that it was hotly and openly debated when the law was being written.
      8) This has nothing to do with the ACA; it’s evil people being evil – this time, not Republican officeholders. Probably not, anyway.
      9) Change insurers or change plans, you might lose your current doctor. This is news to anyone who’s been asleep since about 1960.
      10) This is the only one of the ten points that might remotely deserve to be taken seriously. It is worth noting however that the only specific example cited is in Mississippi. As in, a state with maximal resistance to making the ACA workable, from the state government – a state that has refused the federal dollars that would have subsidized a lot of young people shopping for insurance. Also note that if us lefties had our way there would be a Public Option to take care of this point.

      F’gawds sake, if you want to be informed seek better sources than an overhyped multipage clickbait listicle from the newsertainment arm of Murdoch’s empire. There are big problems with the implementation of the ACA; Ezra Klein was practically committing ritual suicide a few days ago. I’m sure you can find serious, devastating, deeply sourced and highly insightful critiques out there someplace. But that list is purest horseshit, a waste of your time and anyone else’s who happens across it, with scarcely a meaningful point across ten pages heavily laden with advertisements, nor any significant effort in reporting by the author.

      1. My response is shorter: My family of four now has health insurance, thanks to a $600 subsidy that enables us to purchase quality HMO care that would have been unthinkable before, especially due to my multiple pre-existing conditions. The CA website had some issues at first, but I was able to get signed up the second week.

        The 21st century right-wing fever dream is noxious and pluming.

      2. you have the W S Journal totally wrong
        as N Chomsky noted many years ago, the NEWS side is excellent, cause the biz CEOs who read the journal need reality to make decisions; it is the op ed that is fantasy
        the news is still pretty good – often better then the N Y Times (I don’t get wapo or mc clatchy papers)

  4. Srah Kliff of Wonkblog is following this closely, with actual trial logons. And surprise surprise, is getting better already.
    The state sites, often created by the same contractor, seem to work fine. The SNAFU is a nine-day wonder.

    1. The SNAFU is a nine-day wonder.

      Then I’d like to take this moment and thank the Cuban-Canadian Senator from Texas for directing the public’s wrath onto himself during that nine-day window…
      Shyster Ted chose the perfect time to try to bankrupt everyone’s retirement accounts and trash the finances of our country.

      1. Barack Obama is lucky in his enemies. Bill Clinton faced Newt Gingrich, a multiple sleazebag but as clever as himself. Barack Obama was given Sarah Palin, Mittens, and Ted Cruz.

  5. I haven’t seen anybody criticize this launch who actually compared it to something real; I have seen a large number of comparisons to massive private-sector screw-ups.

    1. Isn’t this always the case. Government is tasked with offering something of quality to *all* citizens, usually with lower budgets and heightened scrutiny, and gets compared to highly selective sectors with limited scope, able to charge market rates.

  6. I don’t see how you can do large one-shot projects in-house. Something like this has to be contracted out.

    Unfortunately, governmental agencies are at a disadvantage to the private sector in contracting out. They are obligated to provide due process, which means that as a practical matter, they can’t exclude a bid simply because they have had bad past experience with the contractor. (Debarment is a legal process that the prospective debaree fights like a crazed weasel.) Government bidding has to satisfy recondite procurement laws, which gives a real advantage to the few firms that specialize in knowing these laws. Many of these laws are necessary, to keep contracts away from brothers-in-law. Many are not. Finally, since the government is also typically a monopsonist (especially in defense and road contracting, but not so much here), revolving-door corruption is especially likely, as it is in the military.

    I’m surprised that government contracting works as well as it does, especially given the hollowing out of the federal bureaucracy that has been proceeding on a more-or-less bipartisan basis since Nixon’s days.

  7. Federal (IT) projects have inherent constraints even when done well.

    1. Except under unusual circumstances, it has to go with the lowest bidder.

    2. Government compensation limits its access to (computer) talent.

    3. The Comptroller General opinion on severable versus non-severable contracts require that a unitary projects like a piece of custom software be paid for from a single year’s appropriation. Not being able to spread the cost across multiple Fiscal Years indirectly limits the funds available.

    1. In re: your point three–Big multi-year projects are frequently (usually?) done via IDIQ contracts, which are funded a year at a time via delivery orders for segments of work.

  8. As an independent policy analysis consultant, I always reminded my clients that the work they were contracting out to me would generate work for THEM. They didn’t believe me, of course.

  9. I compare the health exchange with another similar sized but much less complex project I was intimately involved with. In 2006 the FFIEC (Federal Financial Institutions Examination Council, a bank regulatory agency) mandated that online banking sites were required to implement two-factor authentication. I worked for the largest commercial provider of software supporting this feature and saw the way this rolled out at consumer banks across the country. Remember, this was changing the way consumers logged into their banking sites for over a third of the consumer banking accounts in the country. The software had to be integrated with the banks’ online systems by each institution’s development staff. The thing that is strikingly similar to the health exchange rollout is that this led to a huge registration spike at every bank the system was rolled out to, and especially in the first wave of deployments, everything that could go wrong, did. Even if the bank’s site software worked as it was meant to, there was a large spike in help center calls, which were staffed expecting, say, 1/2% of customers to call with issues and instead saw 5%. Notice that lots of people were fine, but even the small percentage that weren’t caused great consternation due to the much higher load. The process of instituting the software was not at fault, it was that any new software has some issues when deployed and when the software is a gate to naive users receiving a highly necessary service even an issue affecting a small number of users generates a stream of exception handling much greater than that to be expected in the steady state growth of the system. I think it is silly, if inevitable, that chicken little is running around saying that the sky is falling at the health exchange website.

    1. Wait … two-factor authentication is supported for consumer accounts? Chase, Citibank, and others don’t appear to have any mention of this on their websites.

      1. It is carried out via soft tokens and recognition of your browser/computer combination. You may remember a time when you had to choose a picture to identify the banking site to you (not the two-factor part but an aspect of the offering), and choose challenge questions.

        1. I get a challenge question every time I log onto my checking account from a different computer.

  10. I don’t think the sky is falling, and doubt that come Jan 1 a single person will have been deprived of a single day of coverage on account of this.

    That doesn’t mean that this important public project has been managed correctly. Maybe the HHS IG should look into it, in February or March after things have settled down.

  11. Of course, it would help if the Federal government were allowed to pay competitive salaries for first-rate talent and skill.

    1. Mark, I would like to see an expanded discussion on this point. Up, above, DCH said, “Except under unusual circumstances, it has to go with the lowest bidder.” You suggested something similar in your post here.
      My question is, is it true? I find it hard to believe. Is the practice of awarding bids almost exclusively to the lowest bidder unique to government contracts?

      We have a heavy equipment construction business and we are (almost) never the lowest bidder, and, yet, our order book remains full. Granted, we’re not talking about the government, but, why would the criteria be different? I understand that the government has to be stricter in abiding by existing rules & regulations, but, that’s just bookkeeping.

      Are things such as quality, reliability, process, experience, familiarity, and/or reputation simply not considered? Since I have little experience with the government contract bidding process I have no basis to argue the point authoritatively. But, it seems that what you and DHC are saying is that when it comes to awarding government contracts, price is virtually the only consideration.

      Like I said, this concept is foreign to me, and, is an irrefutable recipe for disaster.
      Can anybody here expand on this subject, anybody with first hand experience?

      1. I was once in a business that relied almost entirely on government contracts, though with states or municipalities, and not the federal government. In those cases the RFP’s generally awarded points based on a number of factors, of which price was just one, albeit heavily weighted. Matters such as references, experience, and financial stability were also taken into account.

        In one case we lost a bid because the agency awarded a competitor just enough “experience” points to overcome our price advantage. When we complained they told us that they had indeed rigged the process for the winner. Apparently the expiring contract, awarded to a newcomer mostly on price, had been a disaster, and the new award was going to a firm they had dealt with successfully in the past. I believed them, and found it hard to quarrel with the decision. I might well have done the same.

        I don’t know whether federal contracts awards are similarly flexible.

        1. I prepared o lot of proposals for Federal software projects. They always required long,detailed, technical proposals documenting technical approach, corporate experience, personnel experience, and management plan. For each of those sections the RFP provided great detail to guide the competitors in preparing their proposals in conforming to the scoring criteria, which were also provided in the RFP.

          After the initial submissions were graded, very weak proposals were eliminated and the remaining competitors went through a second round, called Technical Clarification. The govt would send questions relating to every section and subsection where the original proposal lost points. The questions were detailed and pointed, allowing the responses to recapture many of the scoring points left on the table in round one.

          At the end of the scoring criteria section was an allocation of scoring between Technical and Cost. What I always saw in the RFP was this sentence: “To the extent that technical scores are nearly equal, cost will have more weight in the final determination.”

          The process of clarification was (is, I assume) known as “leveling.” Contracting officers, of course, denied it, but we competitors all knew:

          If your company truly doesn’t belong, you’ll lose out in the first round. But if you’re moderately competent, then the two step technical process is generally a test of your Prpoosal Team’s prowess in sweeping up the evaluation points to get you to the cost competition.

          (Apologies in advance for any typos. Typing that long reply on an iPhone is no fun.)

      2. A lot of what I do is represent federal contractors. There are plenty of contracts where it really is just price — lowest price technically acceptable — but for anything particularly sophisticated, you’d really expect to see best value.

        1. Thank you all for your thoughtful responses. I think I’m getting the flavor of how these contracts are awarded. Unfortunately, I can’t say I’m encouraged in believing that, in the end, the “best man” wins.

          While reading the comments I was reminded of an incident that occurred some years ago while I was a negotiator for one of the major tire companies. A certain city had put out a bid request for their tire needs in the coming year. These bids were somewhat tiresome to prepare in that the individual quantities were quite small, but, the total quantity was too large to dismiss.

          The contract called for a twelve month price guarantee, however, my company had a price increase scheduled approximately six months into the term of the contract. So, my bid was a two-tiered offer, six months at the current price, and, the rest at the higher price.

          When the day arrived for the contract opening and they opened my bid, they refused to accept it. The simple math required for them to arrive at the net prices was too cumbersome for them to deal with. The total value of the contract was not considered. So, I asked them to skip my bid, go on with the other bidders, and, that I would give them my offer before they were finished. I called my headquarters for advice, and, my comptroller said, “what the heck, there’s no time to calculate a new bid…..go with the higher prices.”

          And, what do you know; I won the bid, regardless. But, rather than being thankful for the success, I was pretty angry. This was at a time when all municipalities in the country were going through very difficult times. And, as a tax payer, I was mad that a group of non-elected bureaucrats made a decision that would cost their citizens a lot more money simply because “that’s not the way we do things here.” The clerks, not the citizens, decided that an extra cost of $42,000 was a fair price to pay for not having to do a minor calculation.

    2. I’m not sure if it’s that or more of an indictment of the current state of software engineering in general. In theory, even taking into account the need to interface with third-party systems and the high loads of such a service, this isn’t exactly rocket science from a pure programming perspective. It’s actually pretty mundane stuff for the most part. What matters is that Fred Brooks’ “The Mythical Man Month”, published almost four decades ago, a book with essays about the management of large software projects, is still considered current material in software engineering courses.

      There are, as far as I know, at least two different subsystems to One is the frontend, the other the backend. The frontend is actually working just fine. It’s basically a static site created with Jekyll that serves files via Akamai. While the content delivery could have been optimized some more, Akamai has been serving huge static sites to large crowds for years. This isn’t the part that’s creating the problems. This was contracted out to a firm called Development Seed. They may have been working in a garage, but the result looks just fine.

      The backend is an entirely different story. While you could argue that problems may lie with the contractor, a firm called CGI Federal [1], the more likely explanation is that the backend combined the most bloated technologies known to humanity and then had to create deliverables based on a strict deadline. The sad part is that CGI Federal may not have had much choice in their choice of technologies: while server software for sites like (huge loads, security requirements) isn’t quite a monoculture, it’s not as though software developers have a great many choices aside from the few dominant players in that particular market. Especially with a strict deadline and limited opportunities to adapt other software. I could go on a rant here, but that would likely be too technical to be of interest too most readers here.

      We do have a similar situation that’s currently flying under the radar (no pun intended) with respect to the F-35 software, which is also an awful mess. That’s also something that I could rant about at length, but also something that’s probably too technical for most people here.

      [1] Yes, technically was built based on an IDIQ contract started in 2007. No, this is really not George W. Bush’s fault, but a problem with how government contracting works.

        1. COBOL code? Where do you find people that know this stuff? “Specialist in obsolete software” sounds a good trade to be in.

          1. You’d be surprised. IBM supports COBOL actively with compilers and libraries on a number of platforms, and a very large number of enterprise applications is written in COBOL. COBOL has been extended with support for XML, Unicode, etc. so that programmers don’t have to stop writing COBOL applications.

            And, honestly, COBOL is not any worse than Excel sheets with hundreds of thousands of lines of VBA code attached. If I had the choice to kill either COBOL or Excel+VBA, I’d get rid of the latter (though it’d be a close call).

            But, as Goethe said: “Die ich rief, die Geister, werd ich nun nicht los.”

          2. Easy.

            Just send out help wanted notices to all Medicare and Social Security recipients. There’s millions of us who (used to) know this stuff. I myself wrote some of the most entangled code ever seen.

          3. “COBOL has been extended with support for XML, Unicode, etc.”

            Fuel injection, power steering, automatic transmission, seat belts, disk brakes, etc., etc. on a Model T. 🙂

  12. The exchanges software did not go out for bid. The contract was awarded to CGI, a contractor that already had a Government Wide Acquisition Contract, sort of a pre-vetting.

    Big software projects are almost always late. In private industry, the problem is handled by delaying rollout. A delay was impossible here, so they had to release software that was not ready and hadn’t been tested. If you put out software that hasn’t even been beta tested, it is almost certainly going to be a disaster.

    I don’t see the point of using a call center instead of a web site. What software would the call center use? How would that software be any better than what is being presented to consumers? Do you expect call center operators to do database operations that are normally handled in software? No, no. The problem is not the top level of code, but lower levels, and any call center operators would have to use the same software because they’d be doing the same thing. And that code is not presently working for consumers, nor, we hear, does it work when call center operators try to use it.

    1. The link shows that it was a competition, and that CGI beat out 3 other companies for the job.

    2. ” A delay was impossible here”


      Democrats just brought the nation to a financial crisis in order to prevent a delay; A delay would have been as easy as passing one of those continuing resolutions.

      Numerous features of the program with black letter implementation dates, (Such as the employer mandate) have been delayed by executive fiat. Not change to the law, just a decision that the President didn’t have to follow it.

      Maybe you didn’t want to delay the rollout, but delaying it was eminently possible easy even. Don’t pretend “politically undesirable” = “Impossible”, nothing prevented a delay except the pride of the Democratic leadership, which couldn’t admit a program Republicans said wasn’t ready for prime time really wasn’t.

      1. “Democrats just brought the nation to a financial crisis in order to prevent a delay; A delay would have been as easy as passing one of those continuing resolutions.”

        Please STFU, Brett; the adults are speaking. A delay would have been as temporary as a provisional dictatorship.

        1. Yeah, like the provisional dictatorship that decided the employer mandate wouldn’t kick in on the date the law dictated it would, and could have done the same with the roll out date, had doing so not been politically distasteful.

          Impossible my ass. You just didn’t want to delay it.

          1. We’ll see more of this conspiracy theory. However, I wouldn’t exclude a fine thread of causation running the other way. That is, knowing that a debt limit crisis was virtually certain in the first fortnight of October, and a shutdown quite likely, the White House decided to take the risk of rolling out the websites on the day originally planned rather than going for a one-month delay, with no guarantee of a perfect launch. Compare Eisenhower sticking to June 6 1944: the weather was only just good enough for a landing, the Germans, lacking weather buoys in the Atlantic, thought it ruled one out. Several key German generals too the opportunity of a break.

        2. @barry

          forget the passage you quoted, that’s not the part of his comment that turns the world upside down, here’s the money quote–
          “. . . nothing prevented a delay except the pride of the Democratic leadership, which couldn’t admit a program Republicans said wasn’t ready for prime time really wasn’t.” mr. bellmore’s insinuation that the republican party wanted to delay implementation of the affordable care act so as to give all the parts a chance to be ready isn’t the most shameless crap he’s tried to peddle here but it’s in the top 1%.

          1. “mr. bellmore’s insinuation that the republican party wanted to delay implementation of the affordable care act”

            If they didn’t, why did they write it into a continuing resolution, and send it to the Senate?

            Why should you care why they wanted to delay it? The point is simply, you COULD have delayed it, if you’d wanted. Let’s have none of this nonsense about how the roll out date couldn’t be delayed. That’s just flat out wrong. I’d go so far as to characterize it as a lie. Of course it could have been delayed.

          2. mr. bellmore, if you want to quote me then either quote me completely or use ellipsis to show that the quote has in some way been edited. of course if you had quoted the rest of my statement you would have had to repeat my complaint about your comment which was your implication of benevolent or benign motives for the republican attempt to delay the affordable care act. of course it matters why they wanted to delay it because they wanted to make sure that no one got any of the benefits of the act so the republicans could continue to insist that there would be no benefits of the act well into the next election cycle. you know this perfectly well and your attempts to seem ignorant and disinterested put you at the very edge of mendacity. i repeat, the edge of dishonesty. lie to yourself if you must but don’t piss on my feet and tell me it’s raining.

      2. Ugh. Brett, we won. Against the forces of stupidity and backwardness, we won. And we will again. The morons on your side tried one of the dumbest and least successful ploys possible, and ended up with LESS than if they’d acted like reasonable adults and just negotiated. Get over it.

  13. Here’s “brainwrap”‘s brave one-blogger site that tracks reported signups for all states, giving links to sources. The total at the end of week 2 is 467,419: well ahead of the Administration target of 500,000 by end October. The total includes Medicaid signups as well as private insurance, and the definitions vary, possibly a lot. It’s almost all due to the states running their own exchanges, and mainly just six: CA, KY, MA, NY, OR and WA.

    1. “Includes” Medicaid signups” may e the understatement of the century. The data here is highly murky, but there is reason to believe that the vast majority of the numbers listed – especially for California and Oregon – are Medicaid enrollees not insurance enrollees under the ACA exchanges. Oregon in particular is reporting very high numbers of units but at present has a state insurance exchange website that is beset by technical problems every bit as serious as those of the federal system.

      While expanding Medicaid may well be a fine and laudable public policy goal on its own terms, it is NOT evidence of the success of the core insurance exchange model at the heart of the ACA. That more people sign up for Medicaid when the threshold for enrolling in Medicaid is lowered should surprise precisely no one, especially when some of the states that are expanding their Medicaid program are making a big marketing push, via the food stamp program, to encourage sign-ups.

      But the ACA itself will fail if the insurance exchanges don’t attract a large volume of enrollees, which at present they are not doing – even in those states which are running their own state-level exchanges. And specifically, the ACA will fall is the insurance exchanges don’t attract a large volume of relatively young and healthy enrollees. There is zero data being published to shed light on the demographics of enrollees to date (though I find it hard to believe that if the data thus far were favorable to the ACA cause that HHS wouldn’t be aggressively publicizing it). But if you think that a healthy 25 year old is just as likely to spend hours navigating a buggy, frustrating website to enroll in coverage as is a 50 year old with serious pre-existing conditions then you have an interesting mental model of the world to say the least.

      1. Indeed, a major point of the individual mandate was to force the young and healthy to buy vastly over-priced insurance, so as to subsidize those older, sicker people who were getting the heavily subsidized insurance. The point was to screw over the young and healthy. If the young and healthy don’t sign up, the whole financial model for the ACA doesn’t work.

        That this was obvious from the beginning was one of the reasons for supposing that it wasn’t intended to work.

        1. i noticed a typo in mr. bellmore’s comment of 10/20 at 7:00 am. allow me to fix it

          indeed, a major point of the individual mandate was to force the young and healthy into getting health insurance which might seem over-priced before the application of any possible subsidies so as to convince insurance companies that they could survive no longer being able to exclude pre-existing conditions and impose lifetime limits on coverage. the point was to keep the insurance industry involved in the business of health insurance rather than end up with a government run, single-payer system. if the young and currently healthy don’t sign up then a single=payer sytem becomes a much more obvious necessity.

          that this was obvious from the beginning was one of the reasons for supposing that this was a plan devised by conservatives and republicans. whether that means that it wasn’t intended to work depends on the motives of the right wing think tanks that came up with it in the frist place.

          1. “… was to force …”

            Government is all about force. Although, in this case, no one is being forced…yet.

        2. a major point of the individual mandate was to force the young and healthy to buy vastly over-priced insurance, so as to subsidize those older, sicker people who were getting the heavily subsidized insurance.

          It makes no sense to talk about insurance being overpriced without mentioning the time period covered by the policy. There are two reasonable ways to look at this:

          1. Participants pay more than an actuarially fair premium early in life, and this is balanced by paying less later on. There is nothing wrong with this so long as the later lower premium is available when the time comes. It is useful to think of Obamacare as, in some respects, lifetime insurance. There is no reason why annual premiums ought to match annual expected benefits. Indeed, given the sharply increasing nature of health care costs over a lifetime this sort of partila levelling of premiums is a good idea.

          2. Another idea is that younger participants are buying not just insurnace but the option to keep it, at acceptable cost, regardless of changes in their health. Options are valuable. Options on volatile underlying assets are especially valuable. Here the buyer is getting a call option on a very volatile asset – future health care – at a stike price represented by the insurance premium.

      2. Stephen: If you read past the headline, the disparity in Oregon is that signing up for Medicaid can be completed online, while signing on for insurance is being processed manually as the state realised that the online process was still unreliable. Your assertion that the exchanges are not attracting a large number of enrollees is wishful thinking. Apart from a few journalists, why would 460,000 people subject themselves to a tedious and buggy online registration unless they are very seriously interested? Do you have a demographic breakdown of the early signups to support your contention that they don’t include the healthy young? I thought not.

        Brett: the momentarily healthy subsidizing the momentarily sick is called insurance. For reasons identified by Kenneth Arrow in 1963 and many times explained to you, unregulated private health insurance does not work. That’s why the sane, Burkean, Oakeshottian, Bismarckian, Swiss, conservative approach to health care financing is regulated insurance. On the subsidies between generations: isn’t it a conservative value that young people should look after the old, and the childless help support children?

        1. “On the subsidies between generations: isn’t it a conservative value that young people should look after the old, and the childless help support children?”

          So look after the old that actually need it. The old in general are wealthier than they have ever been. I’m fairly certain that Bill Gates wouldn’t miss receiving his Social Security check.

          Along will all the concern for Walmart employees on food stamps, don’t forget that they’re paying regressive payroll taxes on what Walmart does pay them.

          1. Eliminating SS payments to Bill Gates would be a mosquito-fart in a hurricane, fiscally. Kicking Warren Buffett off Medicare, same thing. Talking about a 0.01% level of “waste” is silly.

            Now, if you want to argue that the top 20%, say, of retirees can do without their “entitlements”, that’s different. We would then have to look up some statistics. Specifically, statistics on the finances of retirees in the 80th percentile of wealth or income. Maybe those people would be just fine without their SS or Medicare, or maybe they’d become a financial burden on their children; I don’t know. What I do know is that their children would rather pay higher taxes to fund “entitlements” than take the old folks in to live with them.


        2. With all due respect – you’re mixing up your numbers. The 500,000 goal for the first month was the goal for the exchanges only, not the goal for the exchanges + Medicaid. If the vast majority of the 460,000 units to date have been enrollees in the expanded Medicaid program (as appears to be the case) then the exchanges are indeed falling well short of their goal. And this is a serious problem, especially if there is any skew in the risk pool of the people signing up for the exchanges.

          I do not, as you say, have a demographic breakdown of early enrollees in the exchanges. HHS does, and is not releasing the data. If the data were favorable to the ACA narrative (i.e. if the early enrollee group were balanced appropriately across high risk and low risk populations) then HHS would have every incentive in a highly politicized environment to release the information. Even if you don’t read anything into that fact, you’re still left with the question of whether high risk populations who have been seeking coverage unsuccessfully for a long time, or low risk populations that have to be pressured into attaining coverage via a mandate, are more likely to sign up in an environment in which signing up is extremely frustrating and time consuming. You’re right – I don’t know. But I’d be willing to wager on the split any day of the week.

          You suggest that the only reason why early Medicaid enrollments are robust while early exchange enrollments are minimal in Oregon is that the former is an easy process while the latter is a difficult process. The implications presumably being that there is likely “pent up demand” for the exchanges that will flood through once the (presumably minor) logistical problems are fixed. two problems with this thesis:

          1) The sign up process for the exchanges will never be as straightforward as the sign up process for Medicaid. Even if the exchange websites were working perfectly it is still going to be more difficult to execute a 3-party transaction (enrollee, government, insurer) than it is to execute a 2-party transaction (enrollee, government), especially given the fact that the exchange model provides some degree of consumer choice that must be navigated. Not to mention the fact that the front end problems with the exchange websites are only a relatively minor contributor to the overall problem, with back end connectivity to multiple insurance company systems (which is of course not necessary for the expanded Medicaid program).

          2) Signing up for Medicaid is, for the enrollee, basically a free lunch. They get coverage at no cost. Signing up for the exchanges requires that the enrollee pay money. Having to pay for things tends to depress demand for them, last I checked.

          1. How do you know the 500,000 was for private insurance only, excluding Medicaid? It’s not a public target or promise, just the forecast in an internal memo of 5 September by Sebelius leaked to the AP, which hasn’t SFIK published the full text. The AP’s article compares a sample of state targets with outturns by October 6; it doesn’t separate out Medicaid.
            BTW, is Medicaid signup on the website exclusive to Oregon?

  14. On an economics blog, one would think that people would be sensitive to survivorship bias.
    In this case, massive corporate screwups vanish; I know of multimillion or even multi decamillion projects that just vanished.
    Or, if the launch is bad, it gets fixed

    second, most of the problems are due to the climate created by republicans; in particular in the 40s and 50s, gov’t had large numbers of engineers who could check projects (mostly highways and stuff like that)
    then these dept’s were outsourced, and oversite was by contractor; the results – a prime example is “the big dig” in boston – a disaster.
    (another amusing rule: highway projects are quoted lowest cost, maintenence is a separate unrelated cost. so , natch, the contractors build crappy roads, and then their buddys come in and fix em 5 years later)

    1. Three weeks and I still get an absolutely blank user profile page when I log into my account on the federal exchange.

  15. The “call center solution” is wildly unrealistic (and not so much for reasons of cost). Specifically:

    1) There might not be enough underutilized call center seat capacity in the entire United States to support this operation. IIRC, the goal for ACA exchange sign-ups is about 7 million people. The sign up process, under the best of circumstances, is a very long process indeed. Enrollees need to provide multiple pieces of information that they may or may not have at their ready disposal. They need to be informed of the range of coverage options they have, and any questions they have need to be answered. Then they need to actually be enrolled in their plan of choice. Note that many (though not all) exchange enrollees will be, by definition, uninsuraed people who have little understanding of how health insurance works. I would be shocked if you could do that in less than 30 minutes (standard issue customer service calls to the cable company can easily average out at 15 minutes – and that’s for something much less complicated than ACA sign-up). At 30 minutes per call this would imply 3.5 million man-hours. If a seat can be evenly utilized 80 hours a week (which it almost certainly can’t, given that real world volume at call centers peaks over the lunch hour and in the early evening), and if there were a full 16 weeks available before the March 1 penalty deadline (which is almost certainly impossible – see #2 below) then that implies the need for over 2,700 seats. And this mostly in Q4 when U.S. contract call centers are at their busiest supporting seasonal peak demand from the retail sector. Granted such call centers would not need to handle all 7 million sign-ups (as the websites will carry some of the burden) but they also will need to field calls from people who ultimately don;t sign up and thus do not cunt against the 7 million.

    2) It takes time to stand up a working call center operation. I once worked for a company that contracted out a call center program involving maybe 300-400 agents (note that the scenario above implies the need for thousands of agents). Planning for the process began in late summer, a vendor was selected in mid November, and the go-live date was mid March. This was considered an exceptional rush job on the part of the vendor. Typically when companies hire out major call center operations the process takes 12-24 months. And that’s for call types that are much simpler than ACA enrollments and for which there is widespread existing precedent. A 2,700 seat operation requires probably 10,000 agents to be hired, trained and put to work (recall that the 2,700 seat math implies 80 hours a week, and that’s before anyone quits, takes sick time, vacation time, etc. And these jobs are rarely 40 hour positions). And given that agents in such an environment would need to learn multiple insurance companies’ systems, the training process would be exceptionally long – likely 4 weeks or more.

    3) Call center agents would still need to access insurance company systems, so all of the “back end” issues that are plaguing the exchange websites would still be present, to one degree or another.

    1. I think you’ve buried the call centre option. The Administration is committed to, they just have to make it work. Of which there is a good chance, if the White House doesn’t screw it up by demanding so many progress reports and scapegoats that the people doing the work have no time or concentration left to actually do it. “Turn off the alarms so I can hear myself think”, as the supervisor says in the China Syndrome film.

  16. I feel compelled to point out something that no doubt others must be saying. The common wisdom on the internets and pundit fueled talk shows: Troubles with the operation of the web site are being trumpeted as the failure of the Affordable Care Act. Software is not a law. The software can be rewritten without changing the intent of the law.
    I just thought that should be mentioned so nobody forgets it. While this is a political problem it is not a legal problem.

    Now I’m going to wade into weeds I know nothing about.
    It seems to me that if the troubles with the web site have to do with interfacing with all of the old software used by various insurers, what is needed is a system requiring participating insurers to plug their information into the government’s web page.
    If I were in charge (God help us) I would be taking a two pronged approach: Keep patching the existing system while accepting it as a lost cause. Get a team building the new system as quietly and quickly as possible.

    Thank you for your kind attention. Now back to our regular comments.

    1. Or that maybe the faulty implementation is a feature of having made private insurance the centerpiece of the ACA and not a bug? In other words, can anyone really be surprised that a patchwork of private insurance policies that really are not intended nor can actually compete effectively with one another to provide health care to the entire population at a lower cost than a public option?

    2. Fred: “..a system requiring participating insurers to plug their information into the government’s web page.” That sounds to me like the approach taken by the HHS medical IT team, which has put a lot of effort into information protocols. They have been blessed by a a lack of externally imposed deadlines, so have not been forced to roll out stuff that wasn’t ready. So far they seem to have avoided the disasters of the top-down megaproject for unified IT in the English NHS.

  17. October 23: via TPM, anecdotal evidence of improvement at from one frustrated user from North Carolina. One swallow and all that, but one swallow is pretty enough.

Comments are closed.