The Trump administration is considering a unilateral change to how the government measures poverty. If adopted, the change would reduce by millions the number of people eligible for Medicaid, parts of Medicare, SNAP, and other benefits over the next 10 years.

One of the questions this raises is whether a progressive president could make the same change in the opposite direction. In other words, could a President Warren or Sanders make millions of more people eligible for Medicaid and other benefits without Congressional approval by increasing the poverty line?

The short answer is probably “yes,” as long as the change increases the accuracy and relevance of the statistical poverty line. (See the background section below for a discussion of some of the legal and technical questions involved in Executive Branch decisions to change the poverty line in ways that impact program eligibility in a non-trivial manner.)

At the same time, it would be a departure from the norm that the Executive Branch doesn’t unilaterally restrict or expand access to public programs by reducing or increasing the official poverty line, a norm that the Obama administration abided by when the Supplemental Poverty Measure was established as a distinct, more modern measure of poverty in 2010.

If the Trump administration breaks this norm, as it is currently considering, then the next progressive president should have no qualms about increasing the official poverty line in a way that more accurately reflects social consensus about the income needed to live above poverty and has the side benefits of moving us closer to public health insurance for all and a more inclusive welfare state.

In a 2016 survey conducted by the conservative American Enterprise Institute (AEI), Americans were asked, “[What is the] highest annual income a family of four can have and still be considered poor by the federal government.” The average response was $32,293 — an amount 33 percent higher than the poverty line used for federal programs that same year. Increasing the official poverty line by 33 percent may seem like a lot, but there is little question that it would produce a poverty line that is more accurate and relevant than the antiquated one.  

Technical and Legal Background

The federal government produces two closely related, but distinct poverty measures: 1) one for statistical purposes established by the White House’s Office of Management and Budget (OMB) in Statistical Policy Directive 14, and used by Census to produce annual poverty estimates; and, 2) one for programmatic purposes established by Congress in 42 U.S.C. § 9902(2) and cross-referenced in various parts of the federal code that authorize specific benefit programs.

According to 42 U.S.C. § 9902(2):

The term “poverty line” means the official poverty line defined by the Office of Management and Budget based on the most recent data available from the Bureau of the Census. The Secretary [of HHS] shall revise annually … the poverty line, which shall be used as a criterion of eligibility in the community services block grant program established under this chapter [and the many other federal programs that cross-reference this definition]. The required revision shall be accomplished by multiplying the official poverty line by the percentage change in the Consumer Price Index for All Urban Consumers [CPI-U] during the annual or other interval immediately preceding the time at which the revision is made.

In short, the “poverty line” for programmatic purposes is the “official poverty line defined by the Office of Management and Budget” (the statistical poverty line) and “revise[d] annually” by HHS using the CPI-U.

Both the statistical and programmatic poverty lines are updated in the first month of each calendar year by increasing the statistical poverty line for the calendar year prior to the last calendar year by the change in the CPI-U. So, in January 2019, the statistical poverty line for 2017 was increased to reflect the change in the CPI-U between calendar years 2017 and 2018 to produce both the latest statistical poverty threshold and programmatic poverty threshold.

The most notable difference between the statistical and programmatic poverty lines involves labeling. The statistical poverty line released in January of each year is labeled using the prior calendar year; the programmatic measure is labeled using the current calendar year. In other words, the 2018 statistical poverty line is roughly equal to the 2019 programmatic poverty line. (“Roughly” because the programmatic poverty line has some other minor differences that make it easier to use for programmatic purposes).

Although federal law clearly specifies the use of the CPI-U for annual updates of the programmatic poverty line, the Trump administration’s position appears to be that their power to “define” the statistical poverty line means that they can use the Chained-CPI-U or other methods, instead of the CPI-U, to update the statistical poverty line each year.

At the same time, there is little question that HHS would still be required to use the CPI-U to produce the programmatic poverty line each year (unless Congress rewrote the federal law specifying its use). But since the programmatic poverty line is produced by applying the CPI-U to the statistical poverty line, any change to the statistical poverty line would ripple through to the programmatic poverty line.

The Chained CPI-U increases more slowly than the CPI-U, so if the Trump administration required its use for annual updates of the statistical poverty line, the ripple effect would mean significant reductions over time in the number of working-class people who are eligible for Medicaid, Medicare Part D, SNAP, WIC, and other social programs.  

Setting aside the precise differences between the CPI-U and the Chained-CPI-U — for more on that, see Dean Baker here and here — does the Trump administration (or a future progressive administration) have the legal authority to unilaterally change the poverty measure in a way that cuts (or expands) Medicaid and other programs for millions of working-class people over time?

In theory, the power to “define” the statistical poverty line and thereby the programmatic poverty line (which Congress seems to have acknowledged) includes the power to rebenchmark the measure by increasing the threshold to reflect contemporary social consensus about what it means to be poor. If this is the case, a progressive president could rebenchmark the statistical poverty line to increase it substantially, at least as long as the increase makes the measure more accurate and reliable, and then use this higher threshold to produce the programmatic poverty line. Since the current poverty lines use a 1963 benchmark or baseline threshold and have only been updated for inflation since then, such an increase could be justified as reasonable and more accurate on a number of grounds.

Does the power to define the poverty line also include the power to ratchet the programmatic poverty line back over time by switching to the Chained CPI-U? This is less clear cut. First, using the Chained CPI-U would clearly make the base-1963 statistical poverty measure even less accurate over time. Second, since Congress explicitly specified the use of the CPI-U for producing the programmatic poverty measure each year, it seems convoluted to add the Chained CPI-U to the mix. Congress knows how to rewrite laws to require the consistent use of the Chained CPI-U. It did exactly that for the tax code in its 2017 tax bill, but it has never specified the use of the Chained CPI-U for means-tested programs or Social Security.

However the legal issues pan out, the Trump administration clearly shouldn’t change the poverty line in a way that cuts Medicaid and other benefits over time for millions of working-class people. But if it does, the next progressive president shouldn’t shy away from increasing the poverty line and extending benefits to millions more working-class and middle-class people.