Are we spending too much on seniors and too little on kids? Many will recognize this as a classic either-or fallacy (what about tax breaks for the wealthy…?) But with Ron Brownstein, Ezra Klein and Charlie Cook all repeating the Urban Institute statistic that federal spending on seniors is nearly seven times that on children, the idea that seniors are crowding out children’s programs is catching on in Washington. Meanwhile, Urban’s estimate that state and local governments spend nine times more on kids than on seniors hasn’t gotten the same attention. Overall, it appears that government spending on seniors is roughly double (or less) that on children, though this measure includes Social Security, which is almost entirely funded through worker contributions.

Few progressives would dispute the need to spend more on children, especially low-income children. This should be the take-away from the Urban Institute’s “Kids’ Share” reports, the latest of which was published last July but is still frequently cited. Unfortunately, its findings are used mainly to justify cuts to Social Security and Medicare rather than expanding worthy kids’ programs.

Reporters love the generational warfare angle, which Urban isn’t above exploiting. The report is rife with selective comparisons to federal spending on elderly and disabled adults, as if there were only three groups of recipients of government spending and the size of the government pie was, by necessity, fixed or shrinking.

Much of the information needed to dismiss the crowding out idea is in the report itself, though you have to do a little digging. For starters, even if federal spending on seniors were seven times that on children, state and local government spending on children is nearly nine times that on seniors, according to Urban’s estimates. This is because most education spending is at the state and local level, as is much spending on children’s health, whereas Social Security and Medicare are federal programs. All told, the report estimates that per capita government spending on seniors in 2008 was a little over double that on children.

Even this two-to-one ratio is misleading, though. It excludes many tax expenditures, even though most economists and budget experts make no distinction between direct spending by government agencies and targeted subsidies administered through the tax system. This is no small matter: tax expenditures are on par with discretionary spending in terms of importance (slightly larger before the recession, slightly smaller since) and three of the five biggest programs and subsidies benefiting children—the Earned Income Tax Credit, the Child Tax Credit and the dependent exemption—take the form of tax expenditures.

In its measure of federal outlays, Urban counts refundable tax credits like the EITC but not provisions that only reduce taxes like the dependent exemption. Urban does provide an alternative measure that includes most federal tax expenditures on children, which increases the measure of federal spending on children in 2011 from $376 billion to $445 billion. This should still not be viewed as an exhaustive measure, however: the authors acknowledge ignoring tax subsidies for employer-provided health insurance for children as well as dependent allowances in some unemployment insurance programs, which would increase the amount to $465 billion. There are likely to be other omissions.

The more expansive measure that includes most tax expenditures is used to track the children’s share of the domestic federal budget since 1960, excluding spending on defense and international affairs. Since defense spending has shrunk considerably since the Cold War, this helps obscure the fact that federal spending on children, including tax expenditures, grew from less than 2 percent of GDP in 1960 to more than 3 percent in 2011, even as children declined as a share of the population from 38 percent to 27 percent. In other words, the “kids’ share” has more than doubled when adjusted for population share. Though there’s nothing wrong with this, it may explain why Urban points the finger at seniors in an effort to build a media-friendly case for expanding children’s programs.

While the share of the population under 20 is projected to keep shrinking, the share 65 and older is projected to rise with the aging of the Baby Boom generation. Fortunately, we prepared for this by building up the Social Security trust fund, a fact that goes unmentioned in the report. Drawing down the trust fund shouldn’t crowd out spending on other programs unless we place artificial limits on total government spending or exist in a state of autarky, which is what economists call a closed society that does not lend to or borrow from other countries.

Also, this assumes spending on seniors should be the same as spending on children. In traditional societies, those who are too young or too old to work are supported by family members. In modern societies, most costs associated with children are still borne by parents. However, most workers expect to live off earned benefits and savings in old age rather than being supported by their grown children.

A government pension increases government spending on seniors even if fully funded through worker contributions. Most people would consider Social Security retirement benefits a return on workers’ contributions rather than a drain on public resources, except to the extent general revenues are used to subsidize the program. In 2011-12, general revenues were used to replace part of the Social Security payroll tax that had been temporarily cut as part of an economic stimulus measure, though this does not count as spending on seniors using the Kids’ Share methodology.

This raises complicated issues of timing and intent. The payroll tax cut was designed to help workers, not seniors, and could as easily have taken the form of a government check to workers, in which case no one would wonder whether it subsidized Social Security. The point isn’t to argue that the payroll tax cut was misclassified, but rather to illustrate the arbitrariness of such measures.

A more pertinent and disturbing example is the president’s budget, which simultaneously proposes cutting Social Security benefits while expanding savings in taxpayer-subsidized retirement accounts. Though individual retirement accounts are a much less efficient way to fund retirement, this form of back-door Social Security privatization would artificially reduce spending on seniors by Kids’ Share standards, both because individual accounts are privately administered and because the tax subsidies accrue mainly to working-age adults, though the goal of these subsidies is supposed to be promoting retirement security.

When it comes to federal spending on seniors, the only valid concern is that Medicare costs are high and rising and, if not contained, will put pressure on other government programs. However, this is not due to lavish benefits, since seniors’ out-of-pocket costs, even with Medicare, remain much higher than that of younger households, but rather the fact that health tends to decline with age. Nor is it a problem limited to Medicare, which is more efficient than private health insurance programs. Rather, the problem is our quasi-private health system, which costs roughly double that of other countries without providing better health outcomes.

The upshot is that Urban does not provide enough information to accurately compare per capita public spending on seniors and children. The best measure in the Kids’ Share report is closer to two-to-one than the widely quoted seven-to-one, though this measure ignores certain tax expenditures. In any case, the senior share includes Social Security benefits which are paid for through worker contributions and which constitute the biggest source of income for most seniors. Since most children’s living expenses are not channeled through a government program, and since seniors also have more costly health problems, we should expect seniors’ share of social insurance programs to be higher, just as we should expect the kids’ share of education expenses to be higher.