Minutes of the March 9, 2007 COPAFS Quarterly Meeting

COPAFS chair Ralph Rector started the meeting by introducing Ed Spar for the Executive Director’s report.

Ed Spar. Executive Director’s Report.

Noting the very full program, Spar talked only briefly about the budget numbers—leaving it to the presenters to cover the details. He also introduced Steven Dillingham, the new director at the Bureau of Transportation Statistics. Spar then recalled the presentation on plans for 2000 census geography that was not happily received at the December meeting. Word at that was that we could expect a Federal Register notice on the topic early this year, but there has been no notice yet. Spar’s understanding is that it is currently somewhere between Census and OMB, but when asked, OMB’s Brian Harris-Kojetin said it had not reached his desk. COPAFS members and affiliates will be alerted when the notice is issued. Spar then noted that the FCSM methods conference is scheduled for November 5-7, and that a conference on incentives (in survey work) is in the works for early next year. The topic involves important issues involving response rates and statistical bias, and has not been the subject of a conference since 1992. Spar reported his understanding that the Secretary of Commerce has stated that the naming of a new Census Bureau director is imminent, and without elaborating, outgoing director Louis Kincannon said that was consistent with what he has heard.

Dates for this year’s remaining meetings are June 8, September 14 and December 14.

Today’s meeting focused on the budget issues faced by statistical agencies. In the morning session, we heard from a panel of presenters from agencies involved with economic data, and the afternoon panel focused on demographic statistics.

Panel One: Economic Data

Philip Rones. Bureau of Labor Statistics.

Rones commented that BLS fared better with the final 2007 budget numbers than they had feared, but that challenges remain. For example, they have long needed to make improvements to the Consumer Price Index, where data collection still relies on 1990 census data. As Rones put it, the sample is very old, and they are not collecting information where they need to. But with the budget held at FY2006 levels (with some minor relief), they cannot make improvements of this type. Asked if at this point, it might make sense to just wait for the 2010 counts, Rones, argued that those counts are still years away, and that updating from 1990 data is an improvement that should have priority. Of the $15.2 million budget shortfall, about $7 to 8 million is for normal operations, so BLS has had to cut costs related to travel, training and equipment. There are proposals to cut some products from some surveys, but Rones said they are “loath to do that,” and know that users would object strenuously. But he reiterated that the situation could have been worse, as the relief they did get has prevented the need for furloughs and a hiring freeze. And they are even moving forward with some development projects, such as experimental data on employee hours and earnings.

Steven Landefeld. Bureau of Economic Analysis.

Landefeld observed that the BEA budget is much smaller than that of BLS, but noted that like many other agencies, they had been flat funded from 2006. With costs increasing, he explained that flat funding is really a cut, and that, like BLS, BEA is having to cut travel and similar expenses – although in their case, not training. Still there is pressure to improve on statistical products of current interest, such as offshoring, but much of this work has fallen by the wayside. However, Landefeld noted that these are one-year cuts, and BEA has a long list of projects that they hope to get back on track in 2008. Landefeld made the case for the importance of R&D work at BEA, and noted that the 2008 budget proposal includes $2.1 million for R&D on measures of “intangibles,” data for specific industries (such as aerospace), and for work on the international and regional accounts. Looking even further ahead, he noted that budget priorities will seek to address improvements to measures of medical expenditures and educational output, as well as quarterly GDP by industry, and further work on offshoring.

Question for Louis Kincannon

At this point, Ed Spar noted that, while scheduled on the “demographic” panel, the Census Bureau wears two hats, so he asked Census director Kincannon for a quick update on the Economic Census. Kincannon noted that they did get the 2007 funds needed for the Economic Census, but said the more important news was an $8 million initiative to improve data on the services sector. The objective is to increase the coverage of this sector over that achieved by the Economic Census and other sources.

Howard Gruenspecht. Energy Information Administration.

Gruenspecht noted that EIA is a statistical agency, but in contrast to many other statistical agencies, they also do a lot of analysis (he described their work as 80 percent statistical and 20 percent analytical). He presented a chart showing the relationship between natural gas storage data and gas prices to illustrate the real world importance of the data they provide—that the data really matter. Gruenspecht commented that they could always use more money, but that the EIA budget situation is not bad, as they not only received an increase over continuing resolution flat funding, but an increase over the original amount requested. As Gruenspecht put it, when there are concerns about energy and energy prices, there is a lot of emphasis on Department of Energy data. EIA is looking for a $14.4 million increase for 2008, which they hope to devote to rectifying some quality problems with petroleum data, implementing mandated biodiesel and ethanol surveys, and achieving a number of other priorities related to cyber security, product improvements, and the tracking of gasoline blending activity. But while budgets have not been a major concern, Gruenspecht noted that “personnel has been a big deal for us.” They have to do a lot of hiring to compensate for heavy attrition, and have a lot of older workers who “could walk out the door tomorrow.”

Ron Bosecker. National Agricultural Statistical Service.

Bosecker describved 2007 as one of the most interesting for his agency, but the second most difficult to deal with in budget terms, since the flat funding associated with the continuing resolution comes in the second most critical year in the funding cycle for the Census of Agriculture. He described himself as a lot less excitable now than a few months ago, before they got the funding needed for the Census. Still, like other agencies, they are cutting back on travel, training and hiring, which is difficult to do during the ramp up to the Census. And there are some program casualties, including the January Labor Survey and a data series on chemical use. The 2008 budget proposal includes a $17.7 million increase to cover Census of Agriculture expenses, but Bosecker said they would still be “treading water” on annual programs, and face significant challenges in completing the Census. In explaining some of these challenges, he reminded us that the definition of a farm involves the sale of $1,000 or more of agricultural products. There followed some discussion of this definition (which dates back to 1972) and why it has not been changed. It seems to be a sensitive issue at the Department of Agriculture.

Thomas Petska. Statistics of Income, IRS.

Petska reminded us that most SOI products are produced at the request of Congress, and involve data on corporations and individuals, as well as non-profits, trusts, and taxes on estates and gifts. The products include edited samples for research purposes, and Petska described the editing process as very complicated. SOI receives data in media ranging from paper to various electronic formats, and significant amounts of data entry work is required. In response to a question, Petska acknowledged that the move to electronic filing has reduced some of the processing burden, but cautioned that the data are not as clean as one might expect, and that problems in electronic data can be difficult to detect. On budgets, the good news is that SOI receives a nominal increase most years, but the bad news is that it is flat in real terms. And SOI, too, has been losing experienced staff to retirement even while having a limited hiring freeze. And as others on the panel had noted, they are challenged to address data quality and data product concerns despite severe funding constraints.

Panel Two: Demographic Data

Before we started the Panel Two presentations, Kathy Wallman from OMB distributed the Analytical Perspectives chapter of the president’s 2008 budget proposal. She noted, however, that the 2008 proposals went forward before Congress acted on the 2007 budget, so some of the 2007 numbers in the analysis are not consistent with the finals. Wallman playfully struggled to bring herself to say that we should “trust Ed’s numbers” (the COPAFS budget sheet) over those in the report.

Jeffrey Sedgwick. Bureau of Justice Statistics.

Sedgwick descdribed the BJS budget as having three big chunks. First is the Criminal Justice Statistics Program, which he described as fortunate to be continuing at 2006 funding levels instead of taking a proposed cut. Second is the National Criminal History Improvement Program, which distributes money to states to improve data on criminal history—for use in applications such as background checks for those purchasing guns. This program also is benefiting from 2006 flat funding as opposed to a proposed cut. The third major budget component consists of Prison Rape Data Collection Activities, which Sedgwick described as an interesting, but little known project, for which data are difficult to collect. The program gets about $15 million per year to collect data on prison rape, and rank facilities from worst to best in dealing with this issue. Echoing the descriptions of other presenters, Sedgwick noted that flat funding over the years amounts to significant budget cuts in real terms (about a one third cut since 1980), and that Congress loves to ask for additional data without providing the additional funding needed. The impacts of this “flat” funding include reduced sample sizes and data precision in major programs such as the National Crime Victimization Survey (which is about two thirds of the BJS budget), delaying some data products, and terminating some programs, such as the Survey of Indigent Defense. In contrast to other agencies, Sedgwick noted that personnel is not a major issue at BJS, as they have already ridden the wave of retirement, and have a relatively young staff. He acknowledged, however, that they do have some problems protecting their skill sets. In the follow up discussion, Sedgwick observed that while Congress is the biggest user of BJS data, they want expressions/evidence of support from other users to justify funding for BJS statistical programs.

Mark Schneider. National Center for Education Statistics.

Schneider described NCES as having two major budget lines. The first is statistical – consisting of about $90 million per year for data on schools, students, adult literacy, and other topics. The second is assessment – also consisting of about $90 million per year for performance measures for schools. As Schneider described it, all lines in the NCES budget have friends—including some in Congress—so it is difficult to eliminate anything even when costs increase and budgets are flat. Schneider, too, described the “effective budget cuts” associated with flat funding, and noted that his agency also has an aging staff, with many nearing retirement. Or as he put it, we have young workers and old workers, but few in the middle. And like many agencies, NCES is having to cut back on equipment purchases, as well as travel and training. But Schneider ended on an optimistic note saying the 2008 budget looks more promising, as it may provide funding for a number of surveys and enhancements that he described.

Jennifer Madans. National Center for Health Statistics.

In perhaps the bleakest presentation of the day, Madans explained that the NCHS budget has still not been finalized, as it is part of the CDC budget. Their budget is split into two major components – 1) programs and 2) leadership and management, and Madans noted that they are not sure how much can be devoted to programs. The centralization of functions within CDC is further complicating the NCHS budget picture, and they have had to absorb numerous budget cuts in recent years. Again, we heard that “level funding” (for NCHS over 15 years) is a cut in real terms. Over half of the NCHS budget is in personnel—leaving very little discretionary money for R&D and data improvements. As Madans described it, they have pretty much exhausted the short term fixes, and are now at the point where they will have to start making real cuts in major programs. For example, the vital statistics program is in need of a redesign, but instead is being cut to the point where data collections may cover less than full years, and there is even discussion of the possibility of collecting some vital statistics only every other year. And all NCHS surveys, including the National Health and Nutrition Examination Survey and the National Health Interview Survey, are in trouble, and at risk of reduced sample sizes. In response to a question of possible private funding (such as from biotechnical firms), Madans explained that they are prohibited by law from taking money from for-profit companies. And in responding to an almost rhetorical question on why her agency is being hammered with so many cuts, Madans commented that maybe “health statistics don’t get the respect that economic statistics get in this country.”

Mary Frase. Science Resources Statistics, NSF.

Frase explained that Science Resources Statistics is a small part of NSF, and that their budget situation is better than most of those we had heard about. They are charged with collecting and disseminating data on scientific enterprise – including money and human resources devoted to scientific enterprises. Frase reported that the SRS budget will increase about $3 million over a previous total of about $27 million, and that they are actually challenged to determine how to spend the money. But while program funding is not a problem, all is not easy, as the budget for salaries has been flat, and they are not sure how much they will be able to devote to travel and training. So with no prospects for increasing staff, SRS will be challenged to monitor the growing number of contracts made possible by relatively favorable funding for programs. Frase concluded by describing some of their survey work, and their plans to add a question (field of degree) to the ACS. In response to questions on how they could add a question to the ACS without legislative mandate, Frase said they did pursue legislation, but that OMB now is acting as the gatekeeper additional ACS questions.

Louis Kincannon. U.S. Census Bureau.

Kincannon distributed a document summarizing changes in the Census Bureau budget from 2007 to 2008. The census director explained that most of their budget goes to the “periodic” area, which includes the decennial census, with its large and very predictable budget cycles. He noted that much of the ramp up for the 2010 census will hit in 2008 with dress rehearsal activities, and they also are adding to the ACS budget to beef up nonresponse follow up. Other budget highlights for 2008 include continued ACS data collection for group quarters, the Economic Census, and the Census of Governments. Kincannon noted that there is no 2008 funding for SIPP (other than planning and development), but that in 2009, they expect to collect data retrospective to 2008.

While the favorable resolution to the Census Bureau’s 2007 budget allows them to look ahead, Kincannon described the resolution as a close call – with the period of 2006 flat funding bringing them to within a few weeks of severe actions. He warned that similar flat funding (or delays in approving required funding) would be more difficult to absorb for the next three years, when there is even less flexibility in the census process. The Census Bureau is working hard to educate Congress on this reality. With respect to human capital, Kincannon reported that the Census Bureau also faces the risks associated with a workforce with many nearing retirement. He also expressed concern about data quality control in household surveys as budgets are squeezed, and said they hope to get an exemption from regulations that would require the finger printing of the 600,000 temporary employees they expect to hire for the 2010 census. But in contrast to other agencies, Kincannon said that the Census Bureau has taken a relatively hard line against expectations that they do more work with less funding.

Concerns from COPAFS Constituents

No issues were raised, and the meeting was adjourned.