The U.S. Census Bureau’s surveys and data underlie many of our national statistics, and their results are used to guide federal policymaking and business decision making. But the agency is once again under attack in Washington, with budget cuts threatening some of the bureau’s core functions and an ideological challenge brewing from some of the most hardcore, antigovernment congressional Republicans.
Census at a glance
Most people think about the census only in the context of the decennial count of all Americans, mandated by the Constitution and used to determine apportionment of representatives to Congress. But the Census Bureau’s duties go beyond a once-every-ten-years headcount. Some of this work is summarized below.
Census Bureau surveys
The economic census is conducted every five years and surveys businesses in all industries. Nearly 4 million American businesses were sent surveys by the Census Bureau in late 2012. Questions asked of businesses are the result of consultation between the Census Bureau, the Census Scientific Advisory Committee, the U.S. Bureau of Economic Analysis, and the U.S. Bureau of Labor Statistics. Beginning in 2010 the Census Bureau also reached out to about 1,000 trade associations for input.
American Community Survey
The American Community Survey is an annual sampling survey that has been conducted since 2005. The American Community Survey succeeded the “long-form” census, a detailed survey received by a portion of the population for the past 150 years as part of the decennial census process. Questions on the American Community Survey concern population characteristics, as well as housing and economic indicators. All of the questions asked on the survey have a specific federal purpose.
The Census Bureau also conducts other demographic and economic surveys, often in partnership with other federal agencies. Such surveys include the Current Population Survey, which is used to determine measures such as the unemployment rate; the National Employer Survey, which collects data on employee training, among other statistics; and the American Housing Survey, which provides information on housing inventory and characteristics.
The automatic across-the-board cuts known as sequestration, combined with additional cuts enacted in the continuing resolution funding the government through this September, will reduce the Census Bureau’s requested budget for fiscal year 2013 by nearly 13 percent, to $845 million from a requested $970 million. At the same time, the American Community Survey, or ACS—the annual sampling survey that replaced the “long-form” census after the 2000 census—is in the crosshairs of the House Republicans who last year voted to fundamentally alter and defund the survey.
Both of these efforts are certainly pound foolish, but they can’t even be said to be penny wise. Census Bureau statistics are essential to the effective functioning of our government and our economy; the bureau is responsible for accurately collecting, analyzing, and reporting much of the data underpinning crucial government operations.
At a time when we need more data-driven decision making in government, it shows an utter lack of foresight on the part of policymakers to reduce financial and political support for the agency.
Cutting the Census Bureau’s budget and eviscerating the American Community Survey will lead to less-informed government and a more expensive decennial count in 2020, as well as endanger data sources essential to government, researchers, and businesses alike. Let’s consider each of these issues in turn.
Reckless budget cuts to the Census Bureau will lead to less-informed government
The Census Bureau faces an estimated $46 million budget cut under sequestration, the automatic across-the-board spending cuts that took effect at the beginning of March. In a letter last month, acting Commerce Secretary Rebecca Blank wrote that sequestration would force the agency to delay the economic census, the once-every-five-years survey that forms the basis for a wide range of economic indicators from gross domestic product, or GDP, to unemployment. Blank described the move as “putting at risk our ability to take accurate stock of current economic conditions and well-being and potentially impacting policymaking and economic decisions in the private sector.”
It’s unclear exactly how the Census Bureau will actually deal with sequestration now that the cuts are beginning to take effect. Blank’s letter also mentioned reducing contract work, letting open positions in the census workforce go unfilled, and potentially delaying or canceling tests of new methodologies for conducting the decennial census—tests that are necessary to implement cost-saving measures in 2020.
Delaying the release of data from the economic census is a problematic proposition at a time when Congress is potentially planning to address large-scale economic issues such as comprehensive tax reform. The most recent economic census data date to 2007, before the Great Recession. “We have one of the fastest-moving economies in the world,” said Phil Sparks, co-founder of the Census Project, an advocacy group for local governments, businesses, and other census advocates in an interview with The Huffington Post. “To have data that is this dated and dog-eared, we’re just tying one hand behind our backs economically.”
The economic census isn’t the only Census Bureau survey that directly impacts government and business decision making. Data collected in the much-maligned American Community Survey help direct more than $400 billion in federal money to states and localities each year. Highway planning and construction, special education, Section 8 housing assistance, and transit grants are among some of the program areas where ACS data guide state-by-state spending.
“We already suffer too much from what might be referred to as ‘policymaking by anecdote,’ where lawmakers seek to pass legislation before sufficiently examining the severity—or sometimes even the existence—of a perceived problem,” American Enterprise Institute scholar Andrew Biggs said in 2012 testimony before the House Oversight and Government Reform Committee supporting the American Community Survey. “Reducing the quantity and quality of data available to policymakers, analysts and researchers threatens to exacerbate this problem.”
Cutting the Census Bureau budget today means a more expensive count in 2020
While the census gets the spotlight during the year of the decennial count, the Census Bureau is far from idle during the rest of the decade. In addition to running surveys such as the economic census, the Current Population Survey, and the American Community Survey, the early and middle parts of the decade give the agency time to analyze on the recently completed decennial survey and begin planning for the next population count.
This planning period is critical if the Census Bureau is to counteract the rising cost of counting each household. In 2010 it cost $96 per household to conduct the decennial survey, up from just $39 in 1990. Absent actions to control the cost of the count, the 2020 census could cost as much as $25 billion overall, according to the Government Accountability Office. That’s a 92 percent increase from the Census 2010 price tag of $13 billion.
The Census Bureau has pledged to keep the cost of counting each household from rising in the 2020 decennial census. In order to do that, the agency will need to design and pilot programs and statistical methods that could save money on the decennial count, and they will need to launch those pilots in the next few years in order to be ready for 2020. Budget cuts today could doom the potential for offering an Internet response option for the decennial count. This option allows households to reply to the census online, cutting printing, labor, and mailing costs considerably. In addition, some Census Bureau tests have shown that the Internet option also boosts the initial response rate of households, which reduces the need to hire door-to-door canvassers to follow up with nonresponding households. Budget cuts could also prevent the Census Bureau from finding cost-effective ways to use administrative records—data collected by other agencies through things such as tax records and benefit claims—while preserving individual privacy and survey accuracy during the 2020 count.
The Census Bureau simply will not be able to deploy potentially cost-saving methods in the 2020 decennial census if they do not have adequate time and resources to field test those methods well in advance. As a result, the population count could wind up costing billions more than it needs to cost.
Without being able to test these new approaches, the Census Bureau in the later part of the decade will have only two choices: ask Congress for more money to conduct an accurate count, or diminish the quality of the survey. The Census Bureau doesn’t want either of these outcomes, and both possibilities could easily be avoided if Congress adopts a more long-term view of the census budget.
The deficit hawks in Congress should take note. Not all budget cuts are created equal; accurate and timely data on the economy, the demographic composition of the population, and socioeconomic measures are necessary for efficient government and data-driven decision making. These are compelling reasons why policymakers should protect the Census Bureau’s budget.
The debate over the American Community Survey is more about messaging than governing, and will have real-world consequences
For 150 years a long-form census was distributed to a sample of the population as part of the decennial census process. The survey asked a range of detailed questions about everything from housing characteristics to how respondents commuted to work in order to guide federal policymaking. After the 2000 census the Census Bureau transitioned the long-form census from a decennial count to a continuous sample survey, which is conducted on an ongoing basis to provide more current socioeconomic data. This is the American Community Survey, an annual exercise that provides policymakers, researchers, and businesses with granular, real-time information about our constantly changing nation.
In 2012 the Republican-controlled House voted first on an amendment to make responding to the American Community Survey voluntary rather than mandatory, which would make the survey less accurate and much more expensive to administer, as more manpower would have to be devoted to following up in person or by phone with nonresponsive households. In the same debate, the House voted on another amendment to defund the survey entirely, claiming that it was unconstitutional. The Senate didn’t take up either proposal, and the American Community Survey was preserved.
But Rep. Ted Poe (R-TX) and Sen. Rand Paul (R-KY) have now introduced new legislation concerning the American Community Survey. Much like the 2012 amendment, the legislation would make ACS response voluntary. But it goes further than that. First, it would require the Census Bureau to explicitly note in the survey instructions that responding to the bulk of the form is not mandatory. Second, it would bar the Census Bureau from asking questions related to religious affiliation or practice. Because the agency is already prohibited by law from asking these questions, an amendment such as this would merely waste congressional time and resources.
Poe claimed in a press release that the survey is “intrusive” and that he hears from “countless Texans” who “feel intimidated” by the survey. In last year’s floor debate over defunding the American Community Survey, Iowa Republican Steve King summed up his support for the amendment to defund the survey, saying, “I think it’s important to have the information, but it’s important that people have freedom and liberty and we do not have an intrusive federal government that would impose a fine on people if they didn’t let the information come out about whether they had a flush toilet.”
Every question asked on the American Community Survey, however, must have a direct federal purpose. Housing questions, for instance, are used to determine fair market rents for housing-assistance programs, while questions about educational attainment and household income are used to help direct federal education dollars to low-income areas. The Census Bureau provides extensive information on its website about how the government uses the data collected. A 2010 analysis found that seven of the questions in the American Community Survey have been asked in every long-form census since 1850.
Businesses also use ACS data in a range of decisions, including where to build new stores. In a Bloomberg Businessweek story about last year’s attempt to defund the survey, the Chamber of Commerce’s chief economist, Martin Regalia, is quoted as saying that the survey “is especially important to some of our bigger members for trying to understand geographic distinctions and other granularity in the economy.” Tom Beers, the executive director of the National Association of Business Economists, said that ACS data prevent businesses from “flying blind.”
Simply put, there is no valid policy motivation behind legislation to make the American Community Survey voluntary. Not only is the Census Bureau already barred from asking questions related to religious affiliation or practice, but making the American Community Survey voluntary would also raise costs and make the results of the survey less accurate. We need only look to our neighbors to the north for evidence that voluntary censuses don’t work.
Canada conducts a short- and long-form census every five years, and after the 2006 count, the conservative Harper government made responding to the long-form survey—but not the short-form survey—voluntary. Canada is the only country to have made part of its census process voluntary, and the results are staggering. In 2006, when the long-form census was last mandatory, Statistics Canada, our neighbor’s Census Bureau equivalent, saw a 94 percent response rate. In 2011, when that same survey was made voluntary, the response rate dropped to just 69.3 percent. What’s more, the survey cost about $30 million more than its predecessor because Statistics Canada increased the sample size in an attempt to maintain data reliability.
The Census Bureau produced a report on the expected costs of a voluntary American Community Survey following last year’s attempts to defund it. They found that maintaining the current level of data reliability if the American Community Survey became voluntary would require a larger sample size and, ultimately, a bigger price tag—about $90 million higher in 2013.
Since the policies proposed in this legislation would be actively detrimental to the American Community Survey—or outright illegal—the astute observer is forced to conclude that the legislation is more about messaging than it is about governing.
The Census Bureau needs champions. When the House voted last year to defund the American Community Survey, business groups and researchers on both sides of the political spectrum spoke out in passionate defense of the survey’s purposes and the Census Bureau’s professionalism. But that broadside was just clearly the beginning of challenges to the agency’s ability to gather and publish data on the well-being, population, and socioeconomic status of the nation—data essential to the effective functioning of our government and our economy.
Kristina Costa is Speechwriter to the Chair of American Progress. She also conducts research and writes on social impact bonds and other topics in government reform.