Government Statistics: The Ridiculous, The More Ridiculous, And The Completely Preposterous

If President-elect Trump really wants to "drain the swamp" in Washington, one thing that seriously needs to be addressed is the state of the data and statistics on the economy that come out of places like the Commerce Department and the Labor Department.  Serious people every day try to figure out how the American economy and people are doing, and what kinds of policies might make things better or worse.  And all they have to rely on are data and statistics put out by the government, nearly all of which are not suited to the purpose at hand, and most of which are intentionally deceptive in ways that are mostly obvious but overlooked by nearly everyone who uses them, who are mostly in the media and academia.  The deceptiveness is always oriented toward the same goal, namely to support advocacy for increase in the size of the government and the budgets of government programs, and, even more disgracefully, in the context of elections, to support the election of Democrats over Republicans.  

Other than myself, the one guy I have found in American journalism who hammers repeatedly on this issue is John Crudele of the New York Post.  Today Crudele is on the issue once again, in an article headlined "Here's how Trump can easily get the economy humming again."  First point:

The Trump administration should audit all economic data coming out of the Commerce Department, Labor Department and, especially, the Census Bureau.

Amen to that!  Among the more important issues that Crudele has identified in government economic statistics over the years, we have, for example:

  • Just this past September 16, Crudele was immediately on it when Census suddenly reported, in the run-up to the recent election, impossibly large gains in household income and declines in the poverty rate.  (See also my post on the same subject here.)  How could household income have suddenly gone up 5% and the poverty rate declined 3% in a period when GDP was only up 2.4% -- all reported in the final stretch of the campaign, just in time to make it look like President Obama's economic policies were suddenly "working" and to give a last-minute boost to Hillary?  Crudele identified certain methodological changes in how Census was treating answers to certain questions on its surveys as the source of the sudden upticks.
  • Looking back to 2012, Crudele reported another version of about the exact same thing.  In October 2012, in its last reports before the 2012 election, the government stated that the unemployment rate had had a relatively sharp one-month drop from 8.1% in August 2012 to 7.8% in September.  According to a Crudele story from November 2013, those numbers had been "faked" by the fabrication of data: "The Census employee caught faking the results is Julius Buckmon, according to confidential Census documents obtained by The Post. Buckmon told me in an interview this past weekend that he was told to make up information by higher-ups at Census."  
  • Then there are Crudele's many reports on so-called "seasonal adjustments" to economic data, which are essentially made up out of whole cloth and can be manipulated to make the numbers look better when the government wants them to look better and worse when the government wants them to look worse.  From a Crudele article on August 10, 2016:  "The US economy lost 1.03 million jobs in July [2016].  That’s a fact.  But John, didn’t the Labor Department announce last Friday [August 5] that July had a gain of 255,000 new jobs? . . .  [T]he discrepancy . . . was caused by seasonal adjustments. . . .  Were the seasonal adjustments applied fairly and consistently in July?  There’s evidence that they were not."  Plenty more details about the gamesmanship at the link.

These things run from the ridiculous to the even more ridiculous.  You are trying to use the government numbers to get a handle on whether economic conditions are getting a little better or a little worse, and now you find out that the bureaucrats can and do regularly manipulate the numbers in amounts larger than any real underlying changes in order to support favored narratives and help preferred political candidates in upcoming elections.  

Yet even Crudele can sometimes miss the forest for the trees.  So now let's move from the merely ridiculous to the completely preposterous.  I'm referring, of course, to the government's two fundamental counting conventions that underlie all of its most important statistics.  These two conventions are well known, and are there for all to see, but have somehow faded into the background such that nobody really notices them any more.  And yet these two conventions are what fundamentally undermine the usefulness of the statistics and render them deceptive in most of the uses to which they are put.  The two conventions at issue are:  (1) the convention that, in measuring the size of the economy, all government spending on goods and services is counted toward GDP at 100 cents on the dollar, as if each dollar of government spending is as valuable toward creating wealth as a dollar of private spending; and (2) the convention that, in measuring what is called "poverty" and "income inequality," nearly all government spending to benefit low-income people is counted at zero cents on the dollar toward increasing income or reducing measured poverty, as if government spending in massive amounts is completely worthless for the supposed beneficiaries.

Let's start with the poverty convention.  I can't even think how anyone can try to justify what the government does.  When the measure of poverty was created back in the 60s, there barely existed any of the vast array of means-tested support programs on which governments now spend about $1 trillion per year.  As "anti-poverty" spending and programs were created and added by the dozens over the years, the bureaucracy came up with one excuse after another why the spending should not be added to the income or resources of the beneficiaries in measuring poverty.  Medicaid?  It's counted at zero, although to buy your own private insurance you have to earn the money as income first.  Perhaps that one could be argued either way (I'm being generous).  How about a slot in a public housing project?  That's also counted at zero.  In my home county of Manhattan, many slots in public housing are worth $100,000 per year and up if measured by the price others are paying to live in the building next door.  Food stamps?  Other nutrition assistance?  Also counted at zero, although exactly how these things differ in any meaningful way from cash income is difficult to articulate.  EITC?  That one is paid in actual cash.  Many uninformed writers who are not attuned to the obscene level of government duplicity in measuring poverty laughably argue that the EITC is the best program for "reducing poverty."  I'll bet you can't even guess the excuse the bureaucrats offer to justify not counting the EITC against the measure of poverty.  To find out, go to this link.  Out of the approximately $1 trillion in annual means-tested and anti-poverty spending by governments at all levels, essentially the only thing that counts against the measure of poverty is TANF.  That's less than $20 billion per year, or less than 2% of total anti-poverty spending.  TANF alone is never enough to raise a single person out of poverty.  

And then there's the GDP convention.  What it means is that the most crazily wasteful government spending always looks like it adds to economic growth.  The Pentagon buys a $1000 toilet seat?  It adds $1000 to GDP!  Bridges to nowhere?  They add their full cost of construction to measured GDP!  A $4 billion structure as the entrance to a subway station?  It adds $4 billion to GDP!  No wonder politicians are always proposing more "infrastructure spending"!  And even worse, this convention always provides the rationale to oppose cutting any and all government spending, even the most completely wasteful.  GDP will go down!  Not real GDP, of course, but GDP as measured by the duplicitous government bureaucrats.

It doesn't take much deep thinking to figure out why government functionaries like these conventions.  If you want to advocate for more money for anti-poverty programs, you need a high rate of measured poverty to play on the heartstrings of the gullible public.  If you want more government spending and government growth generally, it is enormously useful that increases in government spending count as "economic growth" in all circumstances, and that spending cuts count as "economic shrinkage" in all circumstances, without anyone ever even looking or applying any critical thinking to see if the government is wasting the money.  With several trillion dollars of fake counting now occurring every year under these conventions, this game has gone well past the ridiculous and deep into the preposterous.  Drain the swamp!