The Tyranny of the Average

March 6th, 2013 at 10:14 pm

As someone who works with numbers a lot, I find averages to be tremendously helpful.  I’d be lost without them.*

But there’s one area where averages have been misleadingly used of late, an area where the stakes are extremely high—this is far from academic.  I’m not exaggerating when I tell you that if we don’t straighten out this statistical mistake, a lot of people will be a lot worse off.

I’m talking about the historical average of revenues and outlays as shares of GDP, as shown in the CBO figure below.  The average for revenues is about 18%; for outlays, it’s about 21%.  And both lines are expected to climb above these historical averages in the future.

That forecast has led many policymakers on both sides of the aisle to call for a reset: bring both revenues and outlays back down to their long-term averages.  While giving testimony to the Senate Budget Committee the other day, I heard this from many of the Republicans (and was thus relieved to get a great question about why it’s a bad idea from Sen. Angus King).

The problem with the historical average as a guide toward the future is obvious: there have been fundamental changes in our society and economy that make it an unreliable benchmark going forward.

Demographics, for example.  While the share of persons over 65 was relatively constant over the past 40 years at around 10 to 12 percent, it’s expected to grow by more than half over the next 25 years (this point and many others that follow were made here by my CBPP colleague Paul Van de Water).

And even though we’ve recently seen some deceleration in health care spending, the interaction between the demographics that are baked in the cake and faster-than-average growing health costs implies that we’ll need to choose one of the following options (or some combination): a) revenue and spending shares above the historical averages, or b) significantly diminished support for Medicare, Medicaid, and Social Security.

Option c—continuing today’s level of support for retirement security at historical averages for revenues and outlays—does not exist absent unsustainable deficits.  The only fiscally sound way to make due with revenues of 18% of GDP going forward would be to make deep cuts to social insurance programs, including higher eligibility ages and significantly lower benefits for economically vulnerable retirees.

Van de Water adds these points:

Federal responsibilities have grown since 2000, with developments at home and abroad pushing spending above the average for earlier decades. These responsibilities include homeland security in the aftermath of the September 11, 2001 attacks; aid to veterans of the Iraq and Afghanistan wars, many of whom — especially those disabled in the wars — will need health care and income support for decades to come; the Medicare prescription drug benefit, which Congress added in 2003; and health reform, which extends health coverage to tens of millions of Americans who would otherwise be uninsured…

Spending for interest on federal debt also will be substantially higher in coming decades than it was during the past 40 years. Today — largely due to the wars in Iraq and Afghanistan, the large Bush-era tax cuts [80% of which were recently made permanent—JB], and the severe recession — debt held by the public is almost twice as large (as a percentage of GDP) as in 2001, with a commensurate increase in interest costs once interest rates return to typical levels.

Economist Larry Summers makes similar points, adding that our deteriorating public goods will also demand more investment and that we ought not count on productivity gains to save the day:

In some areas technology could greatly reduce government costs but it is important to recognize that by far the largest parts of the federal budget involve cash or in-kind transfers. These parts are far less susceptible to productivity-enhancing technologies than areas that involve the production of goods or services.

Allow me to wax political for a moment before closing this out.  There are conservatives who for years have been trying to shrink government particularly by going after social insurance.  Most recently, they’ve used deficit arguments as a rationale for cuts, and they’ve been extremely successful in dominating the debate.

By advocating that we lock in revenues at or near their historical average of 18%, they’re trying to lock in option c above: either cut the entitlements way back, disinvest in public goods, struggle to service our debt, or run unsustainable deficits.

In this regard, we should make no mistake: the call for allegiance to historical averages is a call for less government and diminished commitments to current and future generations.

Instead, we need a frank national discussion of the questions embedded in these numbers.  Are we willing to devote more of our GDP to these programs going forward?  And as long as we’re talking straight up about all this, if the answer to that last question is “yes,” we can’t do so by simply taxing the top 1 or 2 percent.

Either way, the place to start the discussion is by recognizing that as much as we love averages, they don’t always point in the right direction.


* Of course, as inequality has gone up so much, the median has become increasingly important as well.  In fact, the divergence between the median and the average is a symptom of rising inequality.


Print Friendly, PDF & Email

11 comments in reply to "The Tyranny of the Average"

  1. Smith says:

    Re: “Are we willing to devote more of our GDP to these programs going forward? And as long as we’re talking straight up about all this, if the answer to that last question is “yes,” we can’t do so by simply taxing the top 1 or 2 percent.”

    I refer to figure 3. on this page showing share of income for the top 1% rising from 10% in 1980 to 20% today. The 10% rise represents 6% of GDP. So just taking half of that back in taxes from the 1 or 2% does in fact solve the problem. What’s the end result? The 1% still has a 15% share of income, probably still too large. Tax rates (income, capital gains, estate, and corporate) are nearer to 1950-1980 (before Reagan began era of large peace time deficits).

    Only if high tax rates induce the 1% and corporations to share profits and raise wages, should other income groups face new taxes, especially top 10%.

  2. save_the_rustbelt says:

    One of my stats profs in grad school had two very popular lectures:

    How to Mislead People with Accurate Statistics
    How to Mislead People with Averages

    Economists are, of course, notorious. Sorry.

  3. Fred Donaldson says:

    The frank discussion that you mention should be based on why we have a middle class median individual income of about $32,000 a year and much less than that when retired, and we have one guy in this country who has a personal wealth equal to three thousand, million pounds of quarters. If you loaded that coin onto trucks, each with a ton of quarters, and lined them up with ten feet between them, that cavalcade of quarters would stretch farther than from Paris to Beijing on regular roads.

  4. Robert Goldschmidt says:

    One of the reasons we need more government is to make up for less wage purchasing power. Since 1970, the cost of key elements of the American Dream, measured in months of median wage, have taken off.

    The cost of a year of in-state college has grown from 0.5 months to 5 months.

    The cost of a year of family health insurance has grown from 1 month to 6 months.

    In the mean time corporate profits as a percentage of GDP are up 5% and wage are down over 6%.

    If we want to reduce government spending, we need to first bring wages and profits back in balance.

  5. Jim Dotzler says:

    Jared… One reason for higher spending on healthcare as a percent of GDP is that Baby Boomers are starting to retire and get costly. This has rendered Washington apoplectic with worry about affording it — demanding cutbacks in spending rather than higher levels of taxation or debt.

    But here’s my question: What happens to all these numbers when Boomers die? Won’t the numbers magically improve as the elder deciles shrink? If so, should we be worried at all about medical-based deficit spending now, now, now?

  6. maggiemahar says:


    I agree that hostoric benchmaarks are not a good guide
    going forward– but not because the population is aging.

    The aging will happen, as you note, over 25 years. We have time to adapt, and we can, as Sweden and Germany (to a lesser extent) have adapted.

    Conservatives have used the “aging of the boomers” to fear-monger.

    Most imoportantly, research has produced a consensus that one out of three Medicare dollars are wasted–on unncessary tests and surgeries, overpriced drugs and devices (particularly devices) and preventable medical errors in our

    The ACA will make structural changes that can squeeze out much of this waste.

    Most importantly, the ACA moves us away from fee-for-service “reimbursements(which rewards providers for “doing more) and toward a reimbursement system that rewards sproviders only if they are able to provide better outcomes for less.

    See Atul Gawande’s piece “Big Medicine” in the New Yorker. There, he describes how his hospital is now required to have “skin in the game.” His Boston hospital has signed contracts with insurers and Medicare that require that they meet quality goals- at a lower price. If they don’t, his hospital loses “millions of dollars.”

    (I would advise ignoring the first half of Gawande’s article on “The Olive Garden’ etc. I love his writing, but food is not his area of expertise.)

    What’s important about Gawande’s piece is that it makes it clear that patients already have “skin in the game” (literally)., We don’t need to increase their risk.

    But providers need to have “skin in the game.”

    Finally, we really can cut Medicare spending without hurting beneficiaries. If you have time, please take a look at this HealthBeat post:

    Jared, no doubt you are aware of everything I’m saying, but I posted this comment hoping to alert some of your readers. (And to urge you to write a little more about how, if we improve the quality of care, we can cut costs without drawing blood.)

    Better care and lower costs go hand in hand.

    • Jared Bernstein says:

      All very important points, Maggie, ones I’ve been meaning to emphasize more, so thanks!

  7. Kevin Rica says:


    If everybody has to keep their historical average shares, by how much would we need to reduce the 1% share of national income?

    BTW — the best comment that I ever heard about averages was from your old boss, Robert Reich, who noted that he and Shaquille O’Neil were an average of 6′ tall.

    • Kevin Rica says:

      Sorry — I should proof read better. I meant to ask:

      If everybody has to keep their historical average shares, by how much would we need to reduce the “top one percent” of income earners’ share of national income?

      I hope that is clearer and more numerate.

  8. purple says:

    I think the more philosophical of conservatives think that increased social benefits spending will crowd out U.S. military power. And they fret over a world without U.S. domination. In a certain way, they may be right.