Filed under: Uncategorized
One of the standard lines about transparent liar and MIT Professor Jonathan Gruber is that he just thought normal uncontroversial economic thoughts that all economists do, like that the favorable tax-treatment for paying people in Health Plan rather than in Money is bad and ought to go away.
I happen to agree with that, by the way. I have never liked the tax advantage for paying people in Health Plan. Now, I’m not sure how or why it can somehow be Objective Economics to dislike that policy (economics can surely highlight that that policy distorts but whether that distortion is good or bad (it’s bad IMHO) is still a matter of opinion not of ‘economics’ right?). But if that’s what I’m supposed to think then I’m happy to add my voice to the chorus of All Economists – and Gruber, purportedly – in disliking a tax-advantage for paying people partially in Health Plan.
But now let’s look at another policy Gruber favored and favors – his ‘baby’, in fact – the employer mandate. What does the employer mandate say?
It says that if you are a business and employ 50 or more people, you must either (a) package Health Plan into their compensation, or (b) pay a ‘penalty’.
Um. Hold up. Wait a second. This is the same damn thing as a tax-advantage for paying people in Health Plan, which Gruber supposedly hates like all good economists!
Don’t believe me?
A tax advantage for paying people in Health Plan means that you can pay someone $X take-home salary, or give them $Y take-home salary plus a Health Plan worth $Z, and the cost to you of doing the latter is lower. Or, equivalently, for the same cost-per-employee you could give them $Y cash + $Z Health Plan instead of $X cash, and $Y+$Z > $X.
A ‘penalty’ for not paying people in Health Plan means that you can pay someone $Y in take-home salary plus a Health Plan worth $Z, or you could not give them a Health Plan and perhaps all else equal give them $X>$Y take-home salary instead, but there would be a penalty for doing the latter, so the net cost to you is higher. Or, equivalently, for the same cost-per-employee if you give them all cash $X instead of $Y cash + $Z Health Plan, then $X < $Y + $Z.
Folks. These are the same things, economically. In either instance, an employer is incentivized to pay people partly in Health Plan. Gruber hates that, we are repeatedly told, except he also insisted on it.
He insisted on it because his (alleged) model came out with better econometrics if he switched boolean_EMPLOYER_MANDATE = TRUE on line 3700. We get that. He built a giant model which allowed him to switch on and off policies, including some that (like the employer mandate) are patently unconstitutional and indeed fucking fascist. (For all I know, his model also has a boolean_EVERYTHING_IS_APPROPRIATED_BY_THE_STATE and a boolean_SLAVERY…do we know it doesn’t?)
The problem is that his ‘baby’ fascist policy is economically equivalent to tax-favoring Health Plans in every way, and as an economist he must realize that. Indeed we are repeatedly told that he cares only about economic effects in all this so surely he’s not dumb enough to not see the equivalence between the employer mandate and the 50+ year tax policy he wanted to get rid of. So how to reconcile these things?
He must have figured the employer mandate would eventually wither on the vine and become a non-factor. How? Because as the ‘Cadillac tax’ drifted to affecting more and more people, more and more employers would reach the crossover point beyond which just paying the penalty is actually cheaper than paying in Health Plan + Cadillac tax. This must have been what he wanted to happen in the long run, even if the mandate was necessary in the short run to make the take-up and the model scoring look politically palatable.
This is the only coherent way I can see to interpret the design of his policy. It’s designed to push everyone onto ‘exchange'[sic]-sold Obamacare plans eventually. Just as most of my righty wacko compatriots conspiratorially theorized from the get-go. I should have believed them.
Filed under: Uncategorized
This is something people now should be asking about Obamacare & Gruber but won’t.
I have been writing about and trying to call attention to the ‘risk corridors’ for a long while now, probably longer than anyone else. I pointed out that they are complex derivatives similar to synthetic CDO tranches. I questioned whether they were being ‘priced’ with any accuracy. I noticed when HHS announced they had no plan for the contingency that (as would seemingly be more likely than not, a priori, just from first economic principles) the government would have to make risk-corridor payouts. I worried that they are the real elephant in the room when it comes to the true cost of Obamacare.
Underlying this concern was the (unanswered) question of who exactly set up the risk corridors and how they structured them and whether they knew what the hell they were doing. As I wrote,
Hey, don’t worry, I’m sure that the unnamed legislative aides who wrote ACA have the actuarial and financial knowledge to price Synthetic CDO Tranches appropriately. It’s not like the government is getting ripped off or anything. They probably have law degrees so hey.
This is weird and complicated! Try to picture who, within the government, set this up and decided upon these numbers (“108 percent” or 8 percent corridor width; where did “75 percent” come from; etc.) Then ask yourself whether they knew what in the hell they were doing when setting up these, effectively, multibillion-dollar, unbounded-payoff derivatives contracts. Then, try to sleep at night.
Earlier in the year, much hay was made of a CBO report that purported to project that risk corridors would make, rather than lose, money for the government over the coming years, to the tune of some $8 billion. The blogger from ACA Death Spiral thought this estimate made no sense and implied unrealistic profits for the profitable insurers. I then replicated this conclusion on a hacky spreadsheet, as is my wont.
It was even less tenable given that the administration had asked Congress for $5.5 billion to pay the ‘risk corridors’ in year-1; why did something destined to make $8 billion over 3 years need us to pay $5.5 billion in year 1?, no single journalist asked anyone in any position of power. But I did:
…Obama is telling us that 80%+ of Obamacare insurers will lose money next year, and that insurers will lose an average of 25%. BUT WAIT THERE’S MORE. Since (obviously) we all believe the CBO number, this means insurers are due for a massive turnaround in the following 2 years: they’ll make so much that we will net $13.5 billion, more than making up for this year-1 $5.5 billion loss … and become a business that makes 12-13% in profits on every dollar of premiums they collect.
This was, to say the least, an odd scenario to be believing in. It’s difficult to believe. Yet it is the scenario being painted by (a) the CBO score and (b) Obama’s year-1 funding request. So the question arises – which no journalist asked anyone – what was the CBO score based on?
Was it, perhaps, based on Jonathan Gruber’s modeling wizardry? His “GMSIM” or similar? And if so, is it correct, or is it proving as inaccurate as “GMSIM” has been in (e.g.) predicting the number of signups?
If I found it hard to believe at the time that ‘unnamed legislative aides’ and people with ‘law degrees’ would have had the wherewithal to model synthetic-CDO tranches, I feel some justification in that. But if CBO was advised in their ‘risk corridor’ modelig – as they were in the overall ACA modeling – by, oh say, an MIT ECONOMICS PROFESSOR who SPECIALIZES IN ‘HEALTHCARE ECONOMICS’ (and/or his former students who work for CBO, of course) that would make the whole story far more believable. In fact, at this point from what we know, Gruber’s modeling was so crucial to the entire process that it seems unlikely the ‘risk corridors’ would have been modeled by any other technique.
Which means we can safely infer that the prediction of “$8 billion profit over 3 years” came from Gruber or a Gruberesque model.
And, as ACA Death Spiral (and I) have already pointed out, that prediction looks retarded to anyone schooled in a modicum of basic financial mathematics. Which, of course, ‘healthcare economists’ might not be, who knows.
In any event, some independent or external corroboration of the (Gruber’s?) (ridiculous and unbelievable) CBO score of the ‘risk corridors’ might help give us comfort that this program isn’t destined to blow a hole in the budget. Particularly if it’s indeed the case that the rosy scenario they currently paint effectively came from an ideologue who ‘would rather have the law’. It would sure be swell if we had journalists who would ask these questions, to someone.
Not ‘healthcare journalists’ like Sarah Kliff, or Voxsplainers like Ezra Klein, mind you. Real journalists, I mean.
UPDATE 11/18: By the way I clearly hadn’t been keeping up with the risk corridors, as something the administration did in the spring caused CBO to revise its estimate (unsurprisingly) downward, from +$8 billion profit to $0. Or in other words, the administration did something vis-a-vis insurance companies that caused the US Gov PnL to be -$8 billion and (by implication) insurers’ PnL to be +$8 billion, versus the prior estimate. I won’t pretend to understand how/why but you should go to acadeathspiral.org and check out the most recent post for more. And of course, this noise just makes the question of whether CBO is using Gruber’s model more ripe than ever.
Filed under: Uncategorized
I have some questions about “GMSIM”.
When MIT Professor Jonathan Gruber sucked up contract after contract by this and that state, and eventually the feds, to analyze Health Care Things, we’re told he performed the work using code he (humbly) called “GMSIM”, where in a surprise twist the “G” does not stand for ‘Generalized’ or ‘Governmental’ or ‘Generous’ oh no it turns out it stands for “Gruber”.
Here are some of my questions.
What is this “GMSIM”? What’s in it? Can we see the code? What language(s) is it written in? Who has access to it? How robust is it? Who can modify it? Is there version control? Do they do unit testing? Do they backtest it against known history? Does it estimate its own error bars? Is there a theoretical underpinning – with theorems – to its ‘synthetic firm’ approach? If so, what is the expected accuracy of this approach?
I mean, how do we know its output is worth a lick of salt and bears any relationship whatsoever to reality? Which, doesn’t that matter? If not why not?
Here’s a big reason why that’s important. I gather all these contracts and grants didn’t involve Jonathan Gruber & staff (& graduate students) handing over a piece of software to the client for them to run. No, they just involved Gruber & students running “GMSIM” and then telling the client(s) what it said. Telling them via some ‘report’. In other words, the deliverable was the output, not the software. That’s what this guy made millions of taxpayer dollars handing over to government bureaucrats and contract-manager types.
But in any given case, how do we know that ‘report’ and the numbers in it weren’t just made-up? How do we know they were any more reliable than the Bernie Madoff fund’s performance report?
Hell, how do we even know there is such a thing as “GMSIM”?
Is this an unfair question? Maybe it was a priori. But now, I don’t think so. Jonathan Gruber purports to be disturbed by the sausage-factory that is the legislative process, but nevertheless, he has made it clear that if there’s a choice between behaving ethically and being party to that sausage-factory, he would ‘rather have the law’. What if he decided that the Noble Purpose of Having The Law necessitated making things up? Do we know that he didn’t? Is his character so obviously beyond question that we cannot ask this?
Not in my estimation.
“Oh, Sonic, he’s a professor, all this work has gone through peer review.” Sorry. From what I can see Gruber’s career at this point is (primarily) based on running “GMSIM” over and over again and, after collecting checks for it, writing papers describing what “GMSIM” said and the implications of “GMSIM” saying that thing. Peer review would have reviewed those papers for their, like, good-paper-ness. It wouldn’t be a regular or normal feature of the peer review process (I’d never done it, or experienced such a thing) for the reviewers to even look at the actual code. The role of such reviewers is more like saying: Well, if he says he has this “GMSIM”, and that’s what it says, this is perfectly acceptable paper describing that.
So who has looked at “GMSIM”? Ironically, it wouldn’t be the peer reviewers or his academia peers who elevated him to such high status, put him on editorial boards, named him to this and that society or award, no.
Would it be the various contract officers who oversaw his contracts? Naw. Generally these will be middle-aged bureaucrat ladies who have worked in government their entire careers. Maybe they got periodic status updates from him and then checked off that he delivered this & that in their Microsoft Project file. At most.
Would it be the sponsors who actually asked for the work at the various agencies? Well maybe, after all they would be the ones signing off on his work and the deliverables. The problem is that if he made up numbers they wanted to hear and handed them to them, that might have suited their purposes just fine. We can’t infer from acceptable fulfillment of any given contract that the client even saw the code let alone checked the code because – again – the code was not the deliverable, some numbers were. And all involved at every stage of this process – Gruber, and the client(s) – would have only cared about whether the numbers met their purpose. Which evidently was not any sort of veracity or modeling-fidelity to the real world, but rather, getting some law passed (and showing that they had done ‘analysis’ of its impact).
There is various documentation on the web (including on his web page) describing the purported model. Block diagrams of its input/output, wall-of-text verbal descriptions of its data sources and demand-curve modeling assumptions, seemingly one such document per client/state (indeed this documentation is probably a big part of the deliverable for which he got paid; make a Word DOC, get paid $300k, nice gig eh?). But it doesn’t describe the actual code in any detail, not even in pseudocode, that one could use to replicate the method. So as far as I know these documents could be made-up fiction.
His current and former graduate students could probably testify to the existence of this “GMSIM”, since it sounds like they may have been the ones doing most of the real grunt work. Adapting “GMSIM” to Vermont, adapting “GMSIM” to Kentucky, slaving away overnight, getting paid grad-student-research-level peanuts for it (but in fairness ending up with a degree & letter of recommendation out of it), etc etc. But have we heard from any of them? Seemingly not. (Which is weird; Gruber’s been at MIT for 20+ years, he should have an army of former students ready to call Sarah Kliff & testify to Gruber’s greatness and wonderfulness and ethical purity etc etc.)
In my view the best circumstantial evidence we can point to that “GMSIM” exists is that it predicted the results of CBO scoring ok. That is, for wargaming purposes, the feds would ask Gruber ‘if we made this change how would CBO score it’, he’d give them an estimate, and by all accounts it’d end up close to the actual answer CBO would end up with. This is strong evidence that Gruber was therefore doing something over and above Making Stuff Up to get numbers to come out. However, we’re now finding out that Gruber seems to have basically hand-held/coached the CBO into making and basing their scoring off of his model, and that he had former students at CBO – so as far as we know both models could be equally fake/vacuous, and just ‘accurately’ replicating each other’s nonsense.
Where is the independent, third-party check on any of this? Why did we hitch our entire nation to the vagaries of one particular central-planner’s model that (I submit) no one else checked with anything resembling a critical eye? Well in the case of lefties who wanted a National Health Care Thing and didn’t give a rat’s ass about details, in a way it’s not that mysterious. (They wanted a National Health Care Thing and didn’t give a rat’s ass about details.)
But is this how “#science” is supposed to work? You send one guy to make a ‘model’, a ‘model’ no one else has bothered to check or even has an interest in checking, and then he tells you what the model (purportedly) says, and you just go with that? Is that the vaunted ‘data-driven’, ‘technocratic’ ‘science’ future to which we are all required to submit our lives now else be damned as a know-nothing cretin who hates science? If so, count me out, and sign me up for hating “#science”. And anyone who does favor such a process is not being scientific at all.
And isn’t there something unseemly about an academic (a) openly campaigning for a complex law which (b) is so complex it will inevitably prod if not in fact legally require state after state, and then the feds, to commission expensive external ‘analysis’ as to its impact from (c) academically-credentialed consultants like him but not only that (d) from him in particular because (apparently) he’s the only one with ‘the model’?
Talk about a sweet setup for the selfless Professor who just cares about his countrymen so much he takes millions of their dollars campaigning for a law that (as a side effect) will require them paying him millions of their dollars. Can I get in on this incestuous government-academic circle-jerk gravy train?
In any event, these are just some of the doubts and questions raised by the revelations that a supposedly neutral economist ‘would rather have the law’. They may be impolite to ask but they are not unfair to ask given that this particular academic has seen fit to impose the vision of the world (allegedly) simulated by his (alleged) ‘model’ upon 300 million of his countrymen, and has been paid handsomely for providing its (alleged) output to our representatives in government.
In light of the above, it would be appropriate for Prof. Jonathan Gruber to release the code behind “GMSIM” to the public – whose output they have paid him for handsomely – for them to look through and understand where that output came from. And of course to dispel any lingering doubts that it exists.
Filed under: Uncategorized
Arnold Kling is, as usual, quite sensible on Gruber. Concludes thusly:
5. I think that the extent to which the attacks on Gruber have become personal is something that every economist, regardless of ideology, will come to regret. I am all for criticizing the ideas and the world view that underlie Obamacare. However, a world in which every economist who steps into the policy arena is subjected to opposition research and “gotcha” attacks is not going to be pretty.
This is all part of what the vaunted ‘peer review’ system is supposed to help weed out and protect the guild against, I thought.
Jonathan Gruber and his (retarded, I am now convinced) ‘research’ should have been criticized, and ridiculed, and made fun of, by all his peers (assuming they were/are capable?), unless/until he made it more rigorous and made sure it was sufficiently cleansed of ideological bias.
That clearly did not happen.
What did happen? He evidently rocketed up to become a renowned ‘healthcare economist’ [sic], chairing this and that, getting awards for whatever the hell, being called before Congress and paid phat consulting checks for his (supposed) ‘expert’ advice and testimony on all things ‘healthcare’. Pols used him as a credentialed ‘Expert’ to shove the policy they wanted down peoples’ throats, and he – because he’s an ideologue – was all too happy to oblige. “I’d rather have this law than not”, said the disinterested credentialed Smart researcher, who made a comic-book about his awesome objective quantitative research (LOL).
It all traces back to the fact that the academy did not police itself properly. They gave, or through their inaction/laziness allowed to get, this guy and his (retarded) output the mantle of ‘expertise’. If the bad PR spilling off of him now redounds back onto other academicians in his peer group I understand that will be unfair in some cases but they have only themselves to blame.
It’s just flabbergasting to me that guys like this are not highly ridiculed to death by other professors until they are too embarrassed to show their face. Is the academy not capable? Or is it that they all just agree with him, are equally ideological, and equally Machiavellian? In light of what we otherwise know about professors – their verbal dexterity, their intelligence, their petty infighting – couldn’t one be forgiven for inferring the latter?
So no, I don’t feel the slightest bit sorry for other ‘healthcare economists’ [sic] who might get a bad name. The whole field – like the field of climate science – being smeared over this incident would not necessarily be unfair.
Academics, if you want to be taken seriously, if you want to preserve your aspirational social role as high Smart priests who inform the rest of us peons how things should objectively be, don’t let clearly retarded and highly-motivated ideologues hold themselves out as ‘experts’ in your name, in your department, in your field. Police yourselves. Is that bad advice on my part? I think it’s fine advice.
And why did this guy become such an ‘expert’ in the first place? Where did he even come from? What’s so special about Jonathan Gruber, transparent liar? Why was he, among all possible people, the one trotted in front of us all this time? Kling says this:
3. Gruber is paid the big bucks because he has a quantitative model of how insurance health reforms will play out. Relative to most academic economists and policy makers, my level of trust in such models is rather low.
Here’s the way I read ‘paid the big bucks because he has a quantitative model of how insurance health reforms will play out': Gruber’s the only guy who built a spreadsheet of some sort. All the other ‘health economists’ are even more retarded so they didn’t even have that. Hence, one guy with a spreadsheet (which he himself acknowledges fails as a complete model, evidently) able to hand (D) politicians (and before that Governor Romney) some stupid, fake chart or numbers, ends up dominating the discussion and being trotted out as an Expert Mascot in order to push a dishonest policy.
Seriously, I bet the ‘model’ is some kind of gunky spreadsheet. Or something little better, anyway. A mishmash of data being shoved together to reach predetermined conclusions, with tons of ad hoc assumptions. That’s what it sounds like to me. I’d bet on it. And also, that’s what I (now) think of the field of ‘healthcare economics’ [sic]. That it is so retarded that one guy building some stupid-ass spreadsheet was able to dominate it, garner high praise, suck up grants and consulting contracts, and (through little objective identifiable merit) become the ‘architect’ of a gigantic bill that affects and redistributes wealth amongst 300+ million people against their will.
Oh, but I probably just hate #science. That’s the only reason one wouldn’t want to unthinkingly accede to the prescriptions of ‘experts’. ‘Experts’ who, occasionally, let slip the truth that truth itself is malleable whenever and wherever they (if not the public which would be actually subject to it) would ‘rather have this law than not’.
Filed under: Uncategorized
Unethical behavior in research and scholarship strikes at the heart of the scholarly and educational enterprise. A shared understanding of expectations and responsibilities is, therefore, critical — not only to the quality of the research enterprise but also to the collegial life of this community. Academic misconduct can take many forms, including fabrication or falsification of data…
Oh, no reason.
Filed under: Uncategorized
Here’s something I can easily believe about the ACA’s ‘established by the State’ language: That it would have been caught, and removed, in some sort of more formal reconciliation/editing process. Had the bill had such a thing.
Sure. It probably would’ve. I don’t see why it wouldn’t’ve.
I think this is one of the implied claims of the anti-Halbig ACA knee-jerk-defenders, and as such, I guess I agree with it. It might even be kind of what people mean, or think they are saying anyway, when they say ‘no one who voted for ACA’ (or, somehow more hilariously, ‘no one who reported on ACA’) thought the subsidies applies only to state ‘exchanges’ [sic].
Now, I say ‘kind of’ because that’s still a stupid statement. Of course no one who voted for ACA had read the actual text of it with any attention in the first place (why would they have? it was a National Health Care Thing put forth by (D)s and that’s all Smart People needed to know), so to speak/speculate about what they ‘thought was in it’ is just about as relevant and interesting as for me to state what I ‘think is in’ Middlemarch.
But nevertheless, stipulated: a more lengthy and careful process around the creation of that statute probably would have resulted in the removal of the ‘established by the State’ language, replacing it with language unambiguously granting subsidies to any Obamacare-plan recipients.
1. Let’s pause to note that this claim, even if true, is – contrary to what lefties think – totally and easily compatible with the Halbig story. All the Halbig story says is that the ‘by the State’ language was originally intended to be a ‘carrot’, to incentivize states creating ‘exchanges’ [sic], at some earlier stage of thinking behind the creation of the bill. So to say – if this is what lefties are saying – it ‘would have been removed later’ – is not actually a rebuttal to anything. Fine, maybe it would’ve!
2. But – and this is the important part – it wasn’t. It’s still there. You know, in reality. (You Reality-Based people know what reality is right?) So now here we are (in reality). The text of a statute says a thing and that statute was passed by both houses of Congress and signed into law by the President. That’s the law now.
The stupid-lefty position amounts to this: “Let’s ignore the fact that the law says that thing. Let’s just go with what we Smart People know it would have been replaced with in a hypothetical, counterfactual careful bill-vetting process. Forever. The text of the bill will stay what it is. But the entire machinery of government and bureaucratic apparatus will just act as if that text says a different thing than it actually says. Now, and forever, indefinitely.”
Folks: that is INSANE. That is not how things are supposed to work. If that ‘logic’ stands then we no longer have any claim to be a nation of laws in any meaningful way. After all, at any given time, whatever collection of people who happen to hold power could ‘act as if the law says’ any fricking thing they want to. If we’re now allowed to read bills according to what we ‘know the text would have been replaced with if [whoever] had fixed it’, I have plenty of suggestions. That 39% tax bracket? Oh that just means 3.9%. Those gun-owner registration laws? They really mean gun non-owner registration.
Don’t like it? Tough. We now interpret texts of laws according to what we Think They Would Be Replaced With In A Hypothetical Fixing-Process In Our Minds. That’s just jurisprudence 101. There is no rebuttal and no objective argument against this, once, like, reality has been abandoned.
Insane, and lawless. Oh, sorry, I meant ‘progressive’.
The bottom line is this. If you think ACA ‘would have been fixed’, I’m totally prepared to agree with you. But the answer to that observation, even if correct, is to go out and freaking fix it! Not to just act as if it has already been fixed, leave it on the books unfixed, and call it a day. What the hell is that?
It’s not smart, it’s not even Smart. It’s just dumb. Don’t be dumb.
Filed under: Uncategorized
Via Twitter: Jonathan Gruber confirms what Obamacare critics say.
Screw you too, Jonathan Gruber.