Minutes of the June 10, 2005 CPAFS Meeting

COPAFS Chair Sarah Zapolsky started the meeting, and we went to Ed Spar for his Executive Director’s report.

Ed reported that COPAFS is moving from its Duke Street location (shared with the American Statistical Association) to another Alexandria location on North Washington Street. He will distribute the new address and contact information soon.

Ed then briefly addressed the 2006 budget numbers – commenting that the early House mark-up calls for full funding of both decennial census and American Community Survey (ACS) activities. The BEA numbers are not as favorable, and reflect a cut in the Quarterly Services Survey. The House hopes to have its bills sent to the Senate by the end of June, but Ed noted that things are moving slowly, and nothing is completely passed yet. In fact, there is concern that floor amendments could seek to shift some census funding to other programs.

Next Ed drew our attention to some Congressional Research Service projections of how House apportionment could look by 2010 – based on Census Bureau state population projections. As expected, the distribution reflects continued movement from the North to the South. Then Ed drew our attention to Patty Becker’s SEMCC newsletter, which describes a change in the non-response follow up in the ACS. Because the original 1 in 3 sample of non-respondents would have yielded fewer interviews in high non-response areas, the alternative adopted for full implementation samples at higher rates in high non-response areas. The variable rate approach is cost-neutral, and will improve data quality in high non-response areas.

Ed wrapped up by announcing an upcoming Census Bureau/COPAFS seminar on immigration – to be held February 13-14 in Alexandria. The seminar will be technical (related to measurement rather than policy), featuring four technical papers plus discussants. Attendance is by invitation. COPAFS also is co-sponsoring a July 14 conference on the nation’s future statistical needs. There is a discount for COPAFS members.

How the Congressional Budget Office Uses Federal Statistics for Economic and Budget Projections.

Robert Arnold. Congressional Budget Office.

Arnold works in the Economic Analysis Division at CBO, a support agency for Congress, whose mission is to provide objective budget information to assist Congress in preparing and analyzing the federal budget, as well as in assessing the impact of federal mandates. The work is nonpartisan, and does not deal in budget policy.

The Economic Analysis Division produces a lot of economic numbers, including two-year economic forecasts. These are “consensus forecasts,” designed to reflect a non-controversial, or “vanilla” view of the economy’s future. Again, the objective is to keep CBO apart from budget debates—something Arnold said they do with mixed success.

CBO is both a data producer and data user. The forecasts involve many equations, but also a lot of judgment based on numerous economic measures from a variety of statistical agencies. Arnold described many of the economic measures they work with (including household net worth, consumption, imports and exports), but noted that Gross Domestic Product adjusted for inflation is the most important measure of economic growth.

He also explained that inflation is important to budget projections because taxes are levied on income not adjusted for inflation. Inflation also relates to labor costs, and in turn to standards of living. A chart of long term productivity trends illustrated the substantial productivity gains of the 1960s and 1970s, the slowed through the early 1990s, and the recent return to rapid increase.

Arnold described productivity measures as important to budget forecasts, but poorly understood. He cautioned that one must beware of sudden changes in productivity measures. For example, a recent increase in Total Factor Productivity may trace to a recent and uncharacteristic divergence in the employment numbers in household and establishment surveys. These surveys usually track together, but have diverged in recent years. The lower employment numbers from the establishment surveys are those used in CBO’s measure of productivity, and contribute to the recent increase. The question is whether the increase reflects the “jobless recovery” or a lack of employment coverage in the establishment surveys.

Arnold described how income shares are of interest to few economists, but critical to CBO because different types of income are taxed at different rates. Inaccuracies in income share data can lead to sizable errors in revenue forecasts. As Arnold noted, CBO is highly dependent on the quality of the input data it receives from other agencies.

Asked if there are specific data series that he would most like to see improved, Arnold noted that, being nonpartisan, he has to answer questions like this very carefully. In his careful response, he acknowledged that the statistical agencies have a very difficult job, and remarked that perhaps he could have ended his presentation with a call for more BEA funding.

Highlights from “Older Americans 2004: Key Indicators of Well-Being.”

Kristen Robinson. National Center for Health Statistics.

Robinson described “Older Americans 2004: Key Indicators of Well-Being” as a chart book produced by the Federal Interagency Forum on Aging-Related Statistics, that was founded in 1986 to encourage collaboration among federal agencies that produce or use statistical data on the older population. To accomplish this mission, the Aging Forum provides a venue for issue discussion, facilitates the development of new databases, and promotes the dissemination of aging-related data.

Following the lead of a “Kids Forum” publication on “America’s Children,” the Aging Forum produced its first chart book in 2000—presenting charts and data on numerous key indicators for the older population. The 2004 book adds some new indicators, but Robinson commented that they don’t want to add too many, as the intent is to publish “key” indicators, rather than a data book.

The 2004 book organizes 37 indicators into five major sections covering Population, Economics, Health Status, Health Risks and Behaviors, and Health Care. The report is designed to be user-friendly, and the Forum seeks indicators that are easy to understand, based on reliable national data, measured periodically, have an objective connection to the well-being of older Americans, and are relevant to large segments of the population.

The indicators are presented in chart form with explanatory text, and the data are provided (with definitions) in the appendix. The book is not published annually, but the indicators are updated more frequently on the website

Robinson noted that they have distributed about 15,000 of the chart books so far to policymakers, researchers, teachers, and others. Asked about efforts to publicize their findings, she commented that publications about children tend to get more press attention, but that they seem to be reaching their target audience without much press attention.

An Update from Capitol Hill.

John Cuaderes. House Subcommittee on Federalism and the Census.
David McMillen. House Committee on Government Reform.

Cuaderes introduced himself as the new staffer with the new Subcommittee on Federalism and the Census, and commented that the previous subcommittee was focused on technology, and perhaps paid too little attention to the census and data collection. He described how he had been asked to serve, and has spent the past three months trying to craft a good balance for the new subcommittee’s agenda. Cuaderes said we can expect a number of hearings on the census during the September-December time frame, and conveyed the subcommittee’s awareness that they need to reach out to experts outside the Beltway.

McMillen commented that the new subcommittee beginning is a refreshing one, and noted that most of the critical 2010 census decisions will be made in the next two years, as the Census Bureau takes the position that design elements not in the 2006 tests will not be in the 2010 census. Also, many census contracts have to be established in the next few years. McMillen also noted that by 2009, the American Community Survey will be pumping out 21,000 tables of data each year, and that it will be a major challenge to keep track of that much output.

Asked what are the most controversial and pressing census oversight issues, McMillen noted that the two are not the same. Cuaderes cited technological initiatives, noting that IT contracts have not always worked out, and could put the 2010 census at risk. Congress, he argued, needs to make sure that the Census Bureau gets what it is paying for in such contracts. He also noted that they will be looking at how the Census Bureau responds to things that do not go well in the 2006 census tests.

McMillen also described the importance of Census Bureau partnerships with local governments in the address list development effort, and commented that there are troubling signs, as the Census Bureau has not provided specifics on its plans for the 2010 address list. As McMillen put it, “there will be hell to pay” if address list improvements are not achieved, and emergency funding is required to complete the 2010 count.

Discussion followed about how confidentiality restrictions may dampen LUCA (Local Update of Census Addresses) participation. McMillen commented that the reversal of those restrictions would require congressional action that the Census Bureau would oppose, and Cuaderes added that such legislation would not make it out of committee.

Asked about topics for future hearings, Cuaderes said nothing specific has been established, but that we might expect something on rural areas, ACS, MAF/TIGER, the use of handheld computers, and LUCA.

In response to a question about MAF/TIGER balance, McMillen noted that most of the $80 million is going to TIGER, and that not enough is going to the master address file. McMillen sees the issue of whole household misses as a serious problem, but Cuaderes was more optimistic, and argued that the Census Bureau has done a good job of identifying and addressing issues from the 2000 census.

Geographic Area Issues for the 2010 Census.

Michael Ratcliffe. U.S. Census Bureau.

Ratcliffe started by suggesting a revised title for his presentation – “Census 2000 Geographic Concepts and Criteria.” The idea was not to rehash old issues, but to review the 2000 concepts and definitions as a means to identifying issues for the future. And with the ACS eventually providing data every year, new issues will be involved. Census 2010 geography will have to be useful not just for decennial data, but also for ACS data released beyond 2010.

With this in mind, Ratcliffe went through a list of 2000 geographic definitions, and let the dialog develop. There was considerable discussion of the extent to which block groups and tracts are defined by local participants. It varies by area, but that’s nothing new. The more provocative question was whether block groups and tracts (which are stable between censuses) should be updated more frequently to accommodate the frequent release of ACS data. For example, a tract with 1,000 households in 2010 might have 10,000 households by 2016, and could be split into several smaller tracts. Questions were raised about the presentation of 5-year moving averages for a moving set of small areas, and while it may be achievable in technical terms, users could become confused by an abundance of small area data presented in a variety of geographic rosters of varying vintages. There also was comment on the difficulty of transitioning to new rosters (not to mention mapping and geocoding) every year. In response, Ratcliffe explained that discussion at Census has been in terms of a single mid-decade update – not annual updates.

The issue of geographic updates is even more interesting for ZIP Codes, where changes are frequent and at the discretion of local postal officials. The question is whether the Census Bureau should update its ZIP Code Tabulation Areas (ZCTAs) to reflect these changes. Ratcliffe confirmed that the Census Bureau has not updated ZCTAs since those used to present 2000 census data, and observed that ZCTAs are not currently a tabulation area for the ACS. However, he noted that the Census Bureau is still determining the tabulation geographies for the ACS, and that the possibility ACS/ZCTA updates is something for consideration.

The session concluded with discussions related to the experiences of COPAFS attendees with specific geographic areas.

Concerns of COPAFS Constituencies

No concerns were raised, and the meeting was adjourned.