Testimony: CPI-U is the best measure of inflation for low-income households
Amanda Dunker
Testimony by Amanda Dunker, CSS Health Initiatives Senior Policy Associate
Submitted to Dr. Nancy Potok, Chief Statistician of the United States
RE: Directive No. 14
Dear Dr. Potok:
Thank you for the Office of Management and Budget’s (OMB) request for comments on appropriate measures of inflation for the Official Poverty Measure. The Community Service Society of New York (CSSNY) has been an unwavering voice for low- and moderate-income New Yorkers for over 175 years. Throughout our history, we have focused on preserving upward mobility and ensuring a basic quality of life for all New Yorkers.
OMB asks for comments on using the chained Consumer Price Index (the chained CPI) or the Personal Consumption Expenditure Price Index (PCEPI) to account for inflation in the Official Poverty Measure instead of the basic Consumer Price Index (CPI-U). CSS believes that the chained CPI and the PCEPI are inappropriate measures of inflation for the Official Poverty Measure because they do not reflect the consumption patterns of low-income Americans as well as the CPI-U. Adopting either alternative inflation measure would result in cuts to many programs that set a floor on the standard of living in the United States, including Social Security, Medicaid, the Supplemental Nutrition Assistance Program, and even access to school lunches for children. These cuts would occur because the chained CPI and the PCEPI would underestimate the impact inflation has on low-income households. The threshold for receiving assistance would become lower and lower in real dollars over time and thus many people experiencing poverty would be cut off from assistance.
Further, it is unclear that OMB has the authority to change how inflation is calculated for the Official Poverty Measure. CSSNY believes that a legal change of such magnitude in terms of effecting the standard of living for tens of millions of Americans should be debated in Congress and changed statutorily, not through administrative rule-making.
CSS strongly recommends that the OMB working group look more broadly at methodological problems with the Official Poverty Measure that cause it to grossly understate the level of income needed to maintain a basic standard of living in the United States. Otherwise, we ask that the working group “first do no harm” by maintaining the use of the CPI-U to adjust the Official Poverty Measure.
I. Background
The Purpose of Adjusting the Official Poverty Measure for Inflation
Measures of inflation are used to maintain the real value of dollars despite changes in prices. For example, when used to adjust tax brackets from year-to-year, measures of inflation ensure that the government does not raise taxes by omission without Congressional approval. In 1919, $30,000 could buy several houses. If we hadn’t adjusted our tax brackets for inflation over the years, people earning $30,000 a year would be taxed as though they were among the wealthiest people in the country. Similarly, measures of inflation ensure that we do not reduce the impact of programs like Social Security over time. Congress agreed that Social Security should provide a certain level of purchasing power to solve the problem of poverty among the elderly. Adjusting Social Security for inflation ensures that the program doesn’t shrink away to nothing as prices rise.
The Official Poverty Measure is used to understand the thresholds at which Americans cannot meet basic needs. It is calculated by the Census Bureau. It is necessary to adjust it every year for inflation to ensure that it captures the real value of income, that is whether or not a certain level of income provides enough purchasing power to obtain a basic standard of living.
The inflation adjustment to the Official Poverty Measure is especially important because the base measure understates the amount of income needed to keep families healthy and safe. In 2018, the poverty threshold for a family of four was only $25,465. Americans report an inability to fulfill basic needs at incomes much higher than this. For example, only 12.3 percent of people in the United States lived at or below the official poverty rate in 2017. However, 40 percent of adults in the United States were struggling to meet at least one basic need, such as food, health care, housing, or utilities. OMB must be careful to avoid making changes that would enlarge the gap between the Official Poverty Measure and the economic realities of maintaining a household in the United States.
Differences Between Measures of Inflation
The CPI-U is derived by examining the prices of a standard “market basket” of goods. Researchers obtain the prices for hundreds of items in every part of the country every year through surveys. Many statutes refer specifically to using the CPI-U to adjust the poverty level for inflation. (For example, 42 U.S. Code § 1395w–114 (C)(ii) referring to 42 U.S. Code § 9902 (2).)
Some economists have argued that the CPI-U overstates inflation because it is not updated frequently enough to account for people substituting away from higher-priced goods for lower-priced goods. Fresh vegetables versus canned vegetables is a common example – if the price of fresh vegetables increases, the CPI-U treats the change as though families continue to purchase the same amount of fresh vegetables at the higher price. The chained CPI adjusts the market basket more frequently and so can better capture the reality that higher prices for fresh vegetables might lead families to buy more canned goods at cheaper prices. This results in a lower estimate of inflation for the economy as a whole.
The PCEPI also incorporates substitution more readily and thus is another inflation measure that grows more slowly than the CPI-U. A major difference between the PCEPI, the chained CPI, and the CPI-U is that the PCEPI incorporates survey data from retailers to weight the different categories of goods Americans are purchasing. Another difference is that the PCEPI looks at broader economic units than individual households. For example, it includes health care spending by third parties such as employers on behalf of individuals.
II. The Chained CPI and the PCEPI Are Inappropriate Measures of Inflation for Low-Income Households
The chained CPI and the PCEPI both attempt to overcome the possibility that the CPI-U overstates inflation. However, the theoretical reasons for that overstatement do not apply to low-income households. The main argument against the CPI-U is that it does not capture the ways that consumers change their patterns of consumption when prices increase. For the country as a whole, the chained CPI and the PCEPI may more accurately reflect that behavior. However, when attempting to judge the purchasing power of people living at the poverty threshold, it is important to remember that they likely have the least ability to adjust their consumption patterns.
First, as described above, even at twice the poverty level large numbers of people in America are unable to make basic purchases like paying their utility bills. Households that are skipping basic necessities are unlikely to spend a large percentage of their budget on surplus items that they can easily cut back on.
Second, low-income people in the United States may have fewer options for comparison shopping. There are many instances where living in poverty forces people to spend more on basic necessities than wealthier Americans spend. Issues like lack of transportation make it more difficult to shop at stores with the most competitive prices or the greatest variety of options. For example, many low-income Americans lack access to fresh foods. It is irrelevant for people living in food deserts whether the price of fresh produce is rising or falling, because they are less able to buy fresh produce anyway. Thus, the substitution that many households make in that situation does not apply. It would require a thoughtful and substantive effort to review the market basket used to develop the CPI-U and the chained CPI and assess how elastic the basket really is for low-income households.
The PCEPI shares this problem with the chain CPI but also has other characteristics that make it a better match for wealthy households than for low-income households. The PCEPI’s addition of business expenditures, such as health insurance and pension costs, add components that are unlikely to affect consumption patterns of low-income people. It also adds room for error because for many goods, it begins with the costs of production and then then imputes the price for the final purchaser.
III. OMB May Not Have the Authority to Change How the Official Poverty Measure is Adjusted for Inflation
As discussed above, the statutes guiding the amount of funding for some programs refer specifically to the CPI-U. It is unclear how many programs do this but it is a factor OMB must review carefully before making any changes. Over time, using either the chained CPI or the PCEPI would result in a lower and lower real value for many programs. It will mean that people clearly experiencing deprivation will be cut-off from basic necessities like food and shelter. The purpose of our safety-net is to guarantee basic living standards. The technical details of those programs should not change in a way that makes them substantively different programs in the future. CSS recommends that OMB produce the following for the public and members of Congress before attempting to change how it calculates inflation for the Official Poverty Measure:
- A complete list of the programs that would be affected.
- Estimates of how many people would be prevented from enrolling in safety-net programs like SNAP, Medicaid, Children’s Health Insurance, and Medicare Part D over time, and what the alternatives for those individuals would be for accessing food and health care.
Thank you for your attention to these comments. Please contact Amanda Dunker at adunker@cssny.org or 212-614-5312 with questions.