STUMP » Articles » Lies, Damned Lies, and Government Numbers » 19 April 2014, 15:44

No, not statistics. Not all numbers are statistics (which are summaries of aggregate data). Some are accounting. Some are projections.

While I will be touching specifically on Obamacare and public pensions with respect to questionable numbers in this post, it goes way beyond both of these issues.

So, when I heard about the Census survey tampering recently, I was mildly annoyed. The particular survey they’re messing with, the Current Population Survey, is something I use for investigating consumer patterns in the the U.S. for my day job. Indeed, I’m trying to put to bed a 100+-page research report which heavily uses Census data in general, and I also use some of the CPS info (though not the health insurance status item).

But I realize the folly of relying on an unaudited data source, so I also pull survey results from places like Pew, Nielsen, EBRI, and many others.

I do not try to craft my analysis from a single source, and, again, I realize that almost all the data I pull is never audited for quality, and more to the point, there are almost no repercussions for the data providers if they fudge some of their numbers, other than losing credibility as a data source. There will be no lawsuits for messed-with numbers. Maybe somebody could lose a job, but only if the fudging they did is against a preferred narrative, generally.

This was Megan McArdle’s reaction

I’m speechless. Shocked. Stunned. Horrified. Befuddled. Aghast, appalled, thunderstruck, perplexed, baffled, bewildered and dumbfounded. It’s not that I am opposed to the changes: Everyone understands that the census reports probably overstate the true number of the uninsured, because the number they report is supposed to be “people who lacked insurance for the entire previous year,” but people tend to answer with their insurance status right now.

But why, dear God, oh, why, would you change it in the one year in the entire history of the republic that it is most important for policy makers, researchers and voters to be able to compare the number of uninsured to those in prior years? The answers would seem to range from “total incompetence on the part of every level of this administration” to something worse.

No, I am not assuming incompetence in this particular instance.

Back to McArdle:

Even if the administration genuinely believes this is defensible, why would they give anyone reason to believe that it is cooking the books? Because those charges are being made, and they’re a lot harder to dismiss than the complaints about birth certificates or dark intimations that the administration has simply made up its enrollment figures out of whole cloth.

I just don’t get it.

I mean, I can certainly think of explanations, but I can’t quite bring myself to believe the worst of them. Which leaves me with the only slightly-less-utterly-appalling conclusion: At some point, very early on in the process, folks noticed that asking the new questions would make it difficult to compare Obamacare’s implementation year to prior years, and decided that assessing the effects of the transition wasn’t nearly as important as making urgent changes to … questions we’ve been asking basically the same way for a decade and a half.

I’m just sitting here, staring dumbly at her.

I mean, maybe she’s being somewhat ironic in what she wrote, inasmuch she realizes that they really don’t want Obamacare measured on its own merits.

And she is fooling herself if she thinks that voters give a shit about the data.

I know she’s a numbers person, as am I, but I have some understanding of what sways voters, and it’s not numbers. And no, I’m not saying voters are stupid. I’m saying the decision-making process for voting rarely turns on numbers, and that goes for highly educated people, for idealogues, for liberal and conservatives and for middle-wayers. Generally people are swayed by values and feelings, not numbers.

I could bitch about should vs. is, but that’s not even my point.

So let’s think as to why they want to mess with the numbers. Yes, it may influence pundits, and maybe some politicians. But most regular people are going to be thinking about what has impacted them, and the people they know, directly. They will know if friends & family have gotten kicked off coverage, or have seen huge increases in premiums and deductibles, or if their doctors or hospitals are outside their insurance network. They will know if friends and family have been pushed onto Medicaid, because they’re “required” to have coverage and they’re relatively low income. They will know if they’re having a hell of a time getting seen by a doctor. They will know if their hours at work are being cut in order to comply with Obamacare. They will know if their grandma was kicked off the retiree health plan because Obamacare has made it too expensive for the government employers to cover widows.

I guess they’re trying to prevent Democrat donors and politicians from panicking.

Good luck with that.

But more to the point, there are numbers I trust a lot more that I use all the time in my day job.

The statutory accounting.

Insurers have to file all sorts of statements for regulatory requirements. Health insurers have to do it, just like all the other insurers.

One thing they have to file is how many policies they’ve got. There’s another metric called member-months as well.

They’ve got to file information quarterly with regulators. This stuff is audited, and they can get in really deep shit for false filings.

This information is not secret. I use it in my day job. I can’t use my job’s data sources for this blog (we have to pay a good chunk of change to get the data in usable form), but there are sites like LifeHealthPro who cover the industry…and I’m willing to bet as the info comes out, we will see exactly what has happened with the numbers.

While the survey changes make it harder to measure Obamacare by one yardstick, do not imagine that there will be no other publicly-available numbers that can be checked.

But let us think of other numbers that have been used by governments for supposed decision-making, which indicates some flummery. Megan McArdle again links to the particular story of interest:

Detroit doesn’t have any money. Rather, it doesn’t have enough money: It cannot both pay its debts and run the city. That’s why the city filed for bankruptcy some months back, over the outraged protests of pensioners who stand to lose a big chunk of their monthly checks.

Today, those pensioners are breathing a little easier. The city seems to have struck a deal that will leave their pensions in better shape than previously feared. Instead of cutting those monthly checks by as much as a third, the city will trim them by less than 5 percent.

How did they make such a big change? In part by assuming higher investment returns:

“The city initially factored in rates of 6.25 percent and 6.5 percent for the two funds, but eventually agreed that the funds could be presumed to do better — 6.75 percent — because of an improved outlook based on the funds’ 2013 performance as compared with 2012.”

The city also promises lower or no cost-of-living adjustments, and less in the way of retiree health benefits. And the proceeds from a proposed deal, not yet agreed upon, to save the Detroit Institute of Arts’ collection by soliciting a mix of state and private funds to pay off the pensioners.

Naturally, many other creditors are unhappy. And taxpayers should probably be wary, too. Raising the assumed rate of return on your pension funds is every government’s favorite way to lower its required pension contribution, but if the assumptions don’t pan out, the city will still be on the hook for the pensions — or back in bankruptcy again. It’s particularly risky if you just had a nice big run-up in the stock market, which seems to promise untold future riches. That, of course, is usually when the market is preparing for a correction.

To put it shortly, it’s not their problem. That is, it’s not the problem of the decision-makers here. Other municipalities have gone through bankruptcy, like Vallejo, California, promising no cuts to pensions.

And then finding out it’s still a problem.

Here is John Bury on some similarly skewed analysis due to government numbers on public pensions:

Alicia H. Munnell, director of the Center for Retirement Research at Boston College, in a WSJ blog believes that benefit cuts for New Jersey retirees should not be considered in part because “New Jersey benefits for current employees are now significantly below the national average and employees pay most of the costs.”

This statement strikes one as odd since the latest valuation reports show employee contributions at $1.93 billion while government contributions are at $2.98 billion but Ms. Munnell is obviously referring to the annual accrual of benefits known as the Normal Cost which, according to the July 1, 2013 actuarial reports, does show employee contributions covering 65% of that Normal Cost for the five largest plans ($1,874,252,148 out of $2,897,259,254 in this spreadsheet). But there is a reason for that.

Public plan actuarial reports are not designed to value benefits honestly but rather to use whatever gimmicks won’t get you kicked out of the AAA (and some that should) to get the contribution amounts the people who hired you want (ideally $0).

…..
New Jersey employees are not picking up most of the cost of their pension benefits. They are picking up most of the costs of benefits valued using dodgy assumptions and none of the costs of the massive ($157 billion?) unfunded liability that has resulted from this perversion of actuarial principles.

This is partly the origin of the huge disputes in the actuarial community over public pension valuations.

I got entangled in the public pension debate when I realized the gulf between valuations of promises made for stuff like annuities and life insurance and the valuations of promises for public pensions.

Heck, I even had some issues with private pension valuations, but interestingly, because of all the checks and balances there. The possible groups overseeing misstatement of pension valuations: IRS (companies get some tax relief from sponsoring pensions, so if the company misstates the pension value, that’s problematic), SEC (for publicly-traded companies), PBGC (the entity that “insures” private pensions – they take over private pensions if they go belly-up). To a certain extent the bond market is also involved.

But what about public pensions? How much can we rely on their numbers, even if we accept the current actuarial approaches?

Oh look, some can’t even afford to be audited

It’s hard to believe that there are no audited financial statements for the North Carolina state pension, the seventh largest public fund in the nation with $87 billion in assets. I’m unaware of any other public pension that completely lacks financial statements audited by either an independent accounting firm or the State Auditor, or both.

(As a result of a forensic investigation of the pension I’m undertaking on behalf of the state’s largest employee association, the State Employees Association of North Carolina, I stumbled onto this unlikely state of affairs).

Certain limited pension information provided by State Treasurer Janet Cowell for the state’s Comprehensive Annual Financial Report (CAFR)—included in the massive 300-page CAFR—is supposedly audited by the State Auditor’s office. Good luck finding it. However, no audit of TSERS investments has ever been performed and there are no publicly-available separate financial statements for the pension.

This represents a major material weakness in the CAFR which is relied upon by ratings agencies, municipal bond holders and the federal government in providing assistance to states.

Yeah.

So, let’s just say I’m not swayed by the “fact” sheet put out by the White House. That’s so cute, thinking we think the numbers you spew are at all credible.

Yes, we know your media lapdogs will faithfully bark to the master’s tune.

But those who have real money decisions relying on accurate data, we’ll seek other sources, thanks very much.