Two Views on the HealthCare Exchanges
CBS News has a story on the ObamaCare website, with the bold from me:
Healthcare.gov launched more than a week ago, and while millions of Americans have signed into the site, not many have been able to actually sign up for insurance because of glitches with the website.
…
No one knows how many people have managed to enroll because the administration refuses to release those numbers, but the website’s launch has been rocky.Media outlets have struggled to find anyone who’s actually been successful. The Washington Post even illustrated that sought-after person as a unicorn, and USA Today called the launch an “inexcusable mess” and a “nightmare.”
White House officials initially blamed the problem on an unexpectedly high volume as they had more than 8 million hits in the first week, but after it went offline over the weekend for repairs, officials now acknowledge other problems.
…
However, computer experts say the website has major flaws.“It wasn’t designed well, it wasn’t implemented well, and it looks like nobody tested it,” said Luke Chung, an online database programmer.
Chung supports the new health care law but said it was not the demand that is crashing the site. He thinks the entire website needs a complete overhaul.
“It’s not even close. It’s not even ready for beta testing for my book. I would be ashamed and embarrassed if my organization delivered something like that,” he said.
So if nothing else, this is bad news for those who championed greater government involvement in health care, and said the people worried about it were a bunch of paranoid nuts who really, deep down, just wanted poor people to die, right? Nope, here’s Krugman from October 1 in a post titled “Good Glitches”:
So, very early reports are that Obamacare exchanges are, as expected, having some technical glitches on the first day — maybe even a bit worse than expected, because it appears that volume has been much bigger than predicted.
Here’s what you need to know: this is good, not bad, news for the program….
The big fear has been that a combination of ignorance and misinformation would keep people away, that they wouldn’t sign up either because they didn’t know that insurance was now available, or because Republicans had convinced them that the program was the spawn of the devil, or something. Lots of people logging on and signing up on the very first day — a day when the Kamikaze Kongress is dominating the headlines — is an early indication that it’s going to be fine, that plenty of people will sign up for the first year of health reform.
Yes, there may be some negative news stories about the glitches. But Obamacare is not up for a revote. As Jonathan Bernstein says, the only thing that matters is whether it works. And today’s heavy volume is yet another sign — along with abating health costs and below-expected premiums — that it will.
You see, in a liquidity trap, if people can’t sign up for a program, that’s a good thing for the program.
Less snarkily: This is just Klassic Krugman: With confidence he is downplaying the “glitches” when he obviously has no idea what he’s talking about. It’s not that Krugman is a pretty good amateur computer programmer, who saw the source code and simply has a difference of opinion from Luke Chung. No, Krugman just took the White House’s word for it that the glitches were due to high demand, then ran to his blog to defend the program. And, in fact, he said something that is probably flat-out false: “Lots of people logging on and signing up on the very first day…” We don’t know if it’s false, of course, because the most transparent Administration in history isn’t releasing any details about President Obama’s signature accomplishment.
So again, when progressives wonder, “Man, why do libertarians hate Krugman so much? It’s just because he’s so right about inflation, I guess” no, that’s not it.
I didn’t want to comment on this post, because you’ve said it all. But you had me scrambling for the “like” or “share” or “+1” buttons, and I’ve only just noticed that I can’t seem to find them anywhere. I’ll share it the old-fashioned way, but you gotta add some sharing options to your blog.
I’m going to have to disagree here. When I want to share a link, I just copy it and paste it into the site I am sharing it with. Those buttons do some click tracking and other shady stuff beyond one-click sharing, while making the site load slower, with more javascript code running on your machine.
Well I guess you mean you don’t want to make it too easy for the NSA to make a complete psychological profile of you.
Fair points, all around.
As a professional computer programmer, I have to say this – the days of demand crashing websites are practically over. This isn’t the launch of some big new online video game, where each person ties up the resources of a server constantly while they play. They log in, submit a few database requests, and log off. No, Krugman’s just talking out his ass. Again.
All the right tools are out there, so handling several million small transactions per day is perfectly achievable… but each new generation of programmers needs to re-learn the same old lessons just done slightly differently. Each new generation of managers wants to hire a new batch of programmers, probably all carefully screened for political correctness given the sensitivity of the project.
Don’t underestimate how badly bottlenecked a design can become once those middle-ware scripts start to stack up.
Since no one is giving any figures, we don’t know how many visits we are talking about here… but I can assure you there will be a long queue of equipment vendors offering to fix the thing.
“All the right tools are out there, so handling several million small transactions per day is perfectly achievable… but each new generation of programmers needs to re-learn the same old lessons just done slightly differently.”
The private sector doesn’t generally have a problem starting up a marketplace site.
This explains part of the problem. 195 seconds to load the page (3 minutes) is pretty bad. Looks like firstly they loaded it up with way too much CSS and JS crap (and yeah I do see that in the private sector all the time) and also for some reason one particular file is a ridculous bottleneck.
http://weaselzippers.us/wp-content/uploads/JW54ij22.png
It looks like they are causing a js file to be loaded as if it was html… why? Of course that’s going to cause a bottleneck! Especially when it’s nearly half a MB.
Yeah, that struck me as odd too, maybe a javascript function is loading some additional script in the background? Then perhaps internally evaluating the secondary script, as some sort of module loader?
I’ve seen so much junk code in my life, this doesn’t seem particularly shocking. They will no doubt make some rapid improvements by removing the worst of the rubbish then run around like heros when they get the page load time under one minute.
Bloatware cannot be contained, it is like the blob movie, it squeezes through small cracks then expands out into the next room, and the next. I strongly suspect there’s and equivalent effect with bloating of government departments, and corporates too for that matter.
Let’s also keep in mind, we cannot just *assume* that 100% of the traffic is from people looking to sign up. I’ve heard three separate conservative talk radio programs run segments of “Host attempts to login to Obamacare website and fails” for entertainment value. Obviously some people just want to see how the site works, and what their options are. There have also been many stories about “sticker shock” from people (mostly young and low-income people) who are surprised to find their options were *far* more expensive than what they were offered previously. Presumably, some of these people aren’t going to sign up either.
In other words, high traffic =/ high demand. The fact that a lot of people are viewing the Obamacare website doesn’t necessarily mean they all want to sign up for it, and furthermore, some of the people who WANT to sign up for it probably won’t, due to the high price.
So, no surprise here, the government is incompetent. They can’t get a website to work and they can’t properly anticipate demand. I’m sure they’ll get the hang of it and everyone will love the low premiums, low out-of-pocket co-pays and deductibles, and really love “free” health-care…oh, it’s just such a rosy picture..or is it my glasses? 🙂
Oh, and by the way, if the system can’t tell them second-by-second the number of successful sign-ups, there’s another sign of incompetence.
Jon Stewart recently had Kathleen Sebelius on who refused to give him a straight answer about why the individual mandate, like the employer mandate, can’t be delayed for a year. The stupid thing is that Stewart gave her a hard time about government incompetence and then said that the only answer is a complete government take over of the healthcare system, i.e. single payer. He just really doesn’t see the contradiction in the words coming out of his own mouth.
You would think that with all of their mathematical economists that they could have pinpointed a range of the number of people that would have tried to sign up.
Jon Stewart is a bucket of sarcasm masquerading as intelligence. Not impressive.
“Jon Stewart is a bucket of sarcasm masquerading as intelligence.”
Thank you!
Even the CBS story is BS since they conflate “sign into” with “creating accounts”, and just you wait, “creating accounts” will eventually be conflated with purchasing actual policies. Millions (or, more likely, a hundred thousand tried 20 times based on my experience) attempted to access the high level web address of healthcare.gov.
My own experience- it took me a week to create an account after 10 or more failures (and this doesn’t even count the 20 or so times I couldn’t even get the site to take me to the “apply button”, and since I created that account, I have been unable to login to it or reset password or userid. I was told to create a second account about 30 minutes ago using a different userid and password (apparently all the accounts created before are useless), but the site failed to create the account both times I tried- a repeat problem I had at the end of last week.
When your dealing with a large project such as the health care exchange. Its easy for executives to improperly plan for capacity issues. This could have been avoided with simple metrics and using a confidence interval to determine the appropriate service level. My Firm accounts for these possibilities 100% of the time.
Peter
@
Operationconsultinggroup.com
When they socialize medicine outright, I hope they turn to you, Peter.
“It’s just because he’s so right about inflation”
Hey, as a Sumnerized true-believer in the EMH, I also scoff at inflationista doomsaying! But I also don’t hate Krugman like a lot of other people do, I just wish he was more like the 90s version of himself. Kind of like Alice in Chains.
I also don’t identify as a libertarian, but I’m not a proggle either, so close enough.
At least we agree about Alice in Chains.
Much as you “diss” it, this is a perspicuous example of why Bayesian inference needs to be taught, especially to intellectual ideologues. Krugman doesn’t think it’s enough to say that “oh, the program will still be fine, even despite these hiccups.”
No, to him, the hiccups are evidence that the system is working! Apparently, if numerous people had signed up, with no problems reported, that would be evidence of Obamacare’s failure?
No, it’s heads Krugman wins, tails you lose. Krugman doesn’t play by the rules of us mere mortals.
To be fair to Krugman, he is saying that if no one showed up it would be evidence of failure. You can always find technical fixes for a computer system, but genuine widespread lack of interest would be unfixable.
Mind you, we don’t actually know how many people did show up before the web site fell in a heap, so maybe it was only a few thousand.
Which still points to their failure. They have been screaming ever since debate on this law began that that millions of people have been wanting insurance, but just cannot afford it. They then fail to design a site to handle all of the hits? Pathetic. No wonder they work for government. They’re failures.
To be fair to Krugman, he is saying that if no one showed up it would be evidence of failure.
So he doesn’t violate Bayescraft on an orthogonal issue. He’s still saying that the problems reported are *evidence* of success, which would imply that it would be evidence of failure if no problems were reported.
The correct approach here is to say that the failures were not sufficient evidence to shake our optimism in the program, *not* that they somehow vindicate it!
Silas when have I dissed Bayesian inference? I’m not denying that I ever did, I’m just not sure what you’re talking about. The time Bryan Caplan used it regarding the bailout, I just thought he used it in a goofy way.
Well, here’s what happened then:
Bryan_Caplan: Okay guys, be sure to put all your cards on the table. Tell us *now* whether you think high employment in a year would be evidence *for* or *against* the bailout being helpful.
You: What a crock! No matter what unemployment is like in a year, I still would think the bailout was a bad idea.
Me: So, Bryan_Caplan’s post doesn’t apply to you; he’s only trying to prevent people from using both eventualities as evidence for their position. If you don’t plan to use the unemployment rate as evidence (i.e. you regard it as uninformative regarding the helpfulness of the bailout), you’re clear.
You: No, I totally get it, and he’s wrong.
Me: ???
That wasn’t how it happened, Silas. Bryan was saying he thought if unemployment went way up after the bailout, then he’d agree it was more likely that it helped than he would have thought beforehand. *That* was what I was saying was goofy.
But it’s not. You can raise your credence in a belief without considering it thereafter plausible. That is, Byran could go from assigning 0.01% probability to 1% probability.
I think you were equating “raising credence” with “raising credence to the level of plausibility”.
You claim you would still put low credence in the bailout’s helpfulness after an unemployment surge. That’s fine: it just means you think changes in unemployment are uninformative, or that they’re informative, but not in a way that would make you believe (have high credence in) the bailout’s helpfulness.
Bryan_Caplan was just trying to say that if you want to use future unemployment stats as evidence *for* your beliefs later, then you need to commit *now* to how you plan to treat the sign change.
You claimed to “get” Bayesian inference but at the same time not understand the above.
Silas I don’t want to sound like a jerk but I got through NYU’s doctoral econ program, right? I passed tests involving formal game theory models relying on Bayesian models of beliefs.
Suppose Bryan Caplan says, “I think there’s a 95% probability that I’m the coolest economist on earth,” and I say that’s crazy talk. Would you accuse me of not understanding probability theory?
If you want to find the actual post I will be happy to look at it, but until I see evidence I’m sticking with my prior belief that I understood Bayesian rules.
I doubt this exchange has led Landsburg to update his priors.
If you said “95% is an invalid probability” (rather than saying that *that specific* number is questionable), then yes, that would indicate poor understanding of the concepts.
Likewise, if you said it was crazy talk that you have to update in *favor* of the bailout conditional on high unemployment, given that you plan to update *against* the bailout if there is *low* unemployment, then yeah, that would suggest poor understanding of bayescraft.
Especially if, as was the case, you seemed to be confusing an increase in credence with a high level of credence (see previous). Your specific words at the time were something like, “so the feds give lots of money to failures — no way will I believe that helped the economy!” Which in now way contradicted what Bryan_Caplan was saying.
If you want to bother finding it, I’ll re-read what I said at the time.
“but genuine widespread lack of interest would be unfixable.”
Pretty sure they already accounted for this by making it illegal NOT to sign up.
I mean, I guess if this happened, it might be evidence that the fine/penalty/tax is too low and they need to threaten people with firing squads or something, so even that could still be “fixed.”
I was expecting a lot of people to just shrug and pay the fine. That still may happen, but I guess people are at least attempting to examine their options.
Right, but my point is, you said that a genuine lack of interest would be “unfixable” but I don’t think that’s true.
Government always has a “solution” to people not doing what it wants them to do. But the solution might have to change from a small fine, to a large fine, and eventually to jail.
In any case, they’re already on the record that this is absolutely NOT a voluntary program. They don’t care whether you’re interested in it or not, aside from the fact that if you are, that happens to create positive PR for them. But no matter, if you aren’t, they’ll just force you to anyway.
“…when he (Krugman) obviously has no idea what he’s talking about”
That could have been the entire post.
this stands out to me: “White House officials initially blamed the problem on an unexpectedly high volume as they had more than 8 million hits in the first week,”
8 million hits is not a big number. Craigslist, for example, serves 50 billion page views per month. Plus, it just doesn’t sound like a number too big to expect between serious inquiries and curiosity seekers.
http://www.craigslist.org/about/factsheet
-Is that like the Ministry of Peace?
It occurs to me that the MSM performs a function a bit like the QA testers for large healthcare websites – you hope they do some research, find problems/issues and inform those involved trying to fix it before it causes too many problems.
Imagine if your QA testers never see any problem with what they’re supposed to test?
Now that I view the MSM as the QA for government, it all starts to make sense.
btw – follow me on twitter @TheHootyman
This x 1000:
It’s not that Krugman is a pretty good amateur computer programmer, who saw the source code and simply has a difference of opinion from Luke Chung. No, Krugman just took the White House’s word for it that the glitches were due to high demand, then ran to his blog to defend the program.
Don’t miss the big picture. This website talk is a distraction and although I can’t believe it was done on purpose, it serves them well.
1. I already saw Juan Williams on Fox saying it’s all the fault of Congress. They didn’t allocate all the (additional) funds the Administration requested for the implementation phase.
2. Forget about the “lets let them roll it out and when disaster hits the people will come demanding we overturn it” garbage. The standard Republicrats remedy will be applied: The Law is great, Even if it’s not the masses are addicted to it and the only humane solution is to throw more money at it.
3. If the TwiddleDems are lucky they will even get a TwiddleReps Congress and president to to the shoveling of money.
4. This ends up being a single payer system financed by an additional payroll tax or an increased income tax just as it is in so many other countries.
5. Get ready to stand in line for the doctor, set up your cancer operation months (or years) in the future and get only those medicines and treatments the Gods of Washington think you require and deserve. Unless you have money of course (lots and lots of it) in which case the parallel medical systems will be happy to help you.
I’ve lived through this in other countries. If you take the road to France don’t be surprised when the lights of Paris appear on the horizon.