What levels of staffing and supervisory personnel are needed? How can automation be integrated into census field activities to reduce the combined cost of people and systems? How much of the product line can shift from print to electronic media? Answer. We are undertaking a number of changes to bring down the cost of data collection for Census 2000. We are implementing four fundamental strategies that will help make Census 2000 less costly: (1) entering into partnerships with state, local, and tribal governments, the U.S. Postal Service, other Federal agencies, and business and community organizations; (2) making the census process simple; (3) using technology intelligently; and (4) making greater use of statistical methods. We plan to find "best in class” partners to help the Census Bureau plan and implement each census operation. For example, we already are working closely with the U.S. Postal Service to develop an address list. This partnership will reduce the amount of field work required in past censuses to compile the address list. More powerful computers have allowed the Census Bureau to design census questionnaires that are easier for respondents to read and fill out. For Census 2000, our top priority is to deliver a questionnaire to each address. But unlike previous censuses, if you do not receive a form, you will be able to obtain one easily at your local post office, shopping mall, community or civic center, or by using our well publicized toll-free telephone number. We may even let people respond using the Internet. These improvements will help improve response rates, and reduce the number of housing units that must be contacted by temporary field staff. Using temporary field staff to find missing respondents costs as much as eighteen times the cost for those who respond by mail. For Census 2000, we plan to scan forms directly into computers that use software that reads handwriting to reduce the amount of labor intensive keying that has been required in the past. We also are developing sophisticated software that reduces our needs for clerical matching. The use of sampling and estimation are integral to the Census 2000 process. After reaching a target of 90 percent response in each county, we will draw a one-in-ten sample of households that still have not returned a completed questionnaire and use it as a basis for completing the job. We estimate that this process will require approximately 560 local census offices, each of which will be smaller than its 1990 census counterpart. These offices will be primarily concerned with hiring the large temporary work force and will be closer to the available candidates. This will reduce the amount of time and travel involved in monitoring this our field work. The local census offices will open in October/November 1999 and will close in August 2000 after completing their data collection activities. This schedule represents a significant reduction in the amount of time that local census offices will be open in comparison with the 1990 census. Our plan for Census 2000 requires a lower total staffing level than the 1990 census; 285,000 temporary workers in 2000 compared with 300,000 in 1990. Our plan for using sampling and estimation to complete the last ten percent of the data collection task reduces our peak field staffing requirements. We also have streamlined our field management structure by reducing the number of supervisory positions in each local census office. The requirement for temporary field staff to visit households that do not return a completed questionnaire is extremely labor intensive. We are using automation to help reduce the cost of field operations by reducing the number of cases that require this type of contact. By using automation, we have been able to make it easier for respondents to complete a census form, thereby reducing the number of households that require contact by temporary field staff and reducing the cost of our field activities. If we are able to resolve the concerns regarding access to and use of administrative records, we will use our software to match data drawn from administrative records to include people who otherwise do not respond, thereby reducing the need for field enumerators. We hope to account for about 5 percent of the households that did not return a completed questionnaire in this fashion. We also hope to contact as many people as possible by telephone, thereby reducing the number of cases requiring a personal visit. The quality check operation will use laptop computers to reduce the number of visits required to reconcile enumeration problems. Expanding the use of laptop computers to other census activities would increase the cost of field operations significantly, and is not considered a cost effective approach for a large and short-term activity such as the decennial census. For Census 2000, the Census Bureau will reduce its printed product line and increase electronic data dissemination. As a result, census data will be more readily available to the public than ever before. Our current plans include providing printed products for redistricting, congressional districts, and governmental units. We are developing a prototype automated system for users to determine their requirements for the final electronic dissemination system for the Census 2000. Question. The out year estimates in last year's Senate and House budget plans do not provide increases above current levels for the census. Additionally and more importantly, the 1997 Budget Resolution assumes an increase of only $2.7 billion over the next six years for the census, while you are assuming $3.9 billion is needed to complete the 2000 decennial. I don't see that we have any choice but to find cheaper ways to do the census-do you agree? Answer. The Census Bureau cannot conduct Census 2000 at a $2.7 billion funding level and achieve the goals of both cost savings and accuracy. We have already reengineered the entire census process to save almost one billion dollars over repeating the 1990 census process. The plan we have developed with our stakeholders and advisors will provide the most accurate data for redistricting purposes, be open, be viewed as fair, and be as cost-effective as possible. We could not stand behind the accuracy of a $2.7 billion Census 2000. We could implement a census that would cost less than $3.9 billion, with nearly the same level of accuracy, but we believe that the American public-whose cooperation is crucial to a successful census-would be uncomfortable with the use of methods that, while less expensive, utilize more statistical estimation. For example, we could spend $3.2 billion to conduct a census using a 70 percent truncation level instead of a 90 percent level. Or, we could begin sampling for nonresponse after a certain date ("direct sampling") rather than at a predetermined truncation level. Again, however, it is not clear that the public would be comfortable with cost saving methods that introduce such high level of statistical estimation. We are anxious to continue working with you to weigh budget constraints against the goals of the census and a high level of public confidence. Question. Constitutionally, we are mandated to conduct an “actual enumeration” every 10 years, which has been your guiding principle at the Bureau. As long as we've been a nation, some way or another a census has been taken every 10 years. But I see that you have all these other projects and programs besides the decennial census that you want money for. You have the Continuous Measurement program, the 1995 Census, Economic programs, the North American Industry Classification System (NAICS)- these aren't mentioned in our great Constitution. Why have you taken on all these other projects? Are they draining limited resources you need for the decennial? Answer. Since the 18th Century, the Congress has directed that information be gathered, not only for Constitutionally mandated reapportionment, but also to monitor the progress of, and promote the general welfare of the Nation. The Census Bureau's budget request reflects the full measure of the responsibilities that we have been asked to fulfill. The Congress has authorized, encouraged, and even required the Census Bureau by law (as with the economic censuses) to collect data beyond the enumeration that the Constitution requires for apportionment. In fact, in the very first decennial census in 1790, Congress included questions that were not required for apportionment. The Census Bureau's non-decennial activities include collecting data that describe the full-range of our national economic activity as well as providing population estimates and current demographic data between censuses. The Census Bureau has become the primary data collector for the Federal Government. This is cost effective not only for on-going data collection efforts, but also for the decennial census. The decennial census is able to make use of the strong infrastructure and common systems of the Census Bureau which would be extraordinarily expensive to duplicate from scratch every 10 years. It is very efficient for the Federal Government to have consolidated these data collection activities in the Census Bureau. Question. Continuous Measurement-you say the information you collect with this program is essential for distributing Federal funds to States through block grants and for assisting them in measuring the performance of these federally funded programs to assure that Federal and State dollars are well spent. But-historically, the only measurement the Bureau provided were based on decennial censuses. Can you explain to me how this Continuous Measurement is helping you? Since you previously relied on the decennial for this information, couldn't we save some money by eliminating this program? Answer. The Congress, other government officials, businesses, and citizens tell us they make decisions with census data every year, not once every 10 years. The Continuous Measurement program means better decisions based on timely and accurate information. The Continuous Measurement program will produce social, economic, and housing data comparable to those collected in the decennial census "long form" every year, rather than once every 10 years. Planners at all levels of government and business will have current data available to make decisions and efficiently allocate funds. The Continuous Measurement program will provide annual data for States, Congressional Districts, and cities or counties of 250,000 population or more. This program will provide new estimates each year based on accumulations over several years for all areas with smaller populations such as towns, townships, census tracts, or block groups. It is essential to collect simultaneously Continuous Measurement and the decennial long-form data in 2000 to measure the differences between estimates from the two programs. Sustaining the Continuous Measurement program will provide data users with much more timely, accurate information and make it possible to eliminate the decennial long-form questionnaire in 2010, an oft-cited goal of the Congress. Question. In a March 7, 1996 memo you sent up here to the Committee, you talk about implementing the new NAICS (North American Industry Classification System) through money provided in your Salaries and Expenses account. How does this effect your [SIC] operations? Answer. For fiscal year 1996, the Bureau of the Census requested $3.9 million to cover the costs of implementing NAICS. Planning calls for full implementation of NAICS to occur over the 1995-2000 cycle of the 1997 Economic Censuses. Since NAICS represents the first complete revision of the industry classification system in over 50 years, the Bureau must assure that the approximately 15 million establishments contained in its data files are appropriately classified according to the new system. The fiscal year 1996 Continuing Resolution, however, contained no funds for the implementation of NAICS. Faced with the programmatic requirement to press forward with implementation of NAICS but with no funding for it in fiscal year 1996, the Bureau elected to discontinue certain lower priority economic statistics reports and programs in order to free up funds. The details of those reductions were contained in the March 7 memo cited above. But the reductions did not offset totally the $3.9 million originally requested. NAICS implementation is moving forward at a pace slower than planned or desirable. The Bureau's fiscal year 1997 budget submission includes a request for $5.5 million for NAICS implementation. As a critical step in the process, the fiscal year 1997 request includes $3.5 million for verifying by mail the correct classifications for up to 700,000 establishments, the largest businesses most directly affected by the new classification structure. If the fiscal year 1997 request is approved, with comparable outyear funding through the end of the 1997 Economic Censuses cycle, the Bureau believes that it can implement NAICS, a step that would enormously improve our ability to depict all the components of the U.S. economy. SUBCOMMITTEE RECESS Senator GREGG. The subcommittee will now stand in recess until tomorrow morning at 9:30 when we will receive testimony from the Federal Bureau of Investigation, the Drug Enforcement Administration, and the Department of State. [Whereupon, at 2:25 p.m., Wednesday, May 15, the subcommittee was recessed, to reconvene at 9:30 a.m., Thursday, May 16.] DEPARTMENTS OF COMMERCE, JUSTICE, AND STATE, THE AGENCIES YEAR 1997 AND RELATED FOR FISCAL APPROPRIATIONS THURSDAY, MAY 16, 1996 U.S. SENATE, SUBCOMMITTEE OF THE COMMITTEE ON APPROPRIATIONS, Washington, DC. The subcommittee met at 9:30 a.m., in room S-146, the Capitol, Hon. Judd Gregg (chairman) presiding. Present: Senators Gregg, Lautenberg, and Bumpers. DEPARTMENT OF JUSTICE FEDERAL BUREAU OF INVESTIGATION STATEMENT OF LOUIS J. FREEH, DIRECTOR ACCOMPANIED BY: JAY A. BRIXEY, ASSISTANT DIRECTOR, FINANCE DIVISION, FEDERAL BUREAU OF INVESTIGATION STEPHEN R. COLGATE, ASSISTANT ATTORNEY GENERAL FOR ADMINISTRATION, JUSTICE MANAGEMENT DIVISION, DEPARTMENT OF JUSTICE MICHAEL J. ROPER, DEPUTY ASSISTANT ATTORNEY GENERAL/ CONTROLLER, JUSTICE MANAGEMENT DIVISION, DEPARTMENT OF JUSTICE ADRIAN A. CURTIS, DIRECTOR, BUDGET STAFF, JUSTICE MANAGEMENT DIVISION, DEPARTMENT OF JUSTICE OPENING REMARKS Senator GREGG. Why do we not begin. We are going to be joined by Senator Hollings and probably Senator Byrd, but the Director is kind enough to be here early, so let us begin. The way we have been running these hearings is fairly casual. To the extent you want to make an opening statement, you are welcome to do so. To the extent you want to just make comments, then we can put your statement in the record. That is fine, too. The floor is yours. OPENING STATEMENT Mr. FREEH. Thank you, Mr. Chairman. Thank you for the opportunity again to appear before you here and present our issues. This committee historically, and Senator Lautenberg, in particular, have given the FBI tremendous support and understanding, as well as leadership. I have a written statement I would like to submit for the record. I would like to take a few minutes to give you an overview of our 1997 budget and some of the important issues. We have had, as you know, a very busy year in terms of hiring the people that we need to perform the law enforcement functions of the FBI. Thanks to this committee, we will by the end of the year have hired all of our agents to the peak 1992 level. In fact, we will be slightly over our targeted staffing level of 10,700 very, very briefly at the end of the fiscal year, which will be the first time in many years that we were up to our agent complement. We also expect to have all of our support employees on board by the end of the year. We have prioritized the hiring and training of agents, 1,200 of them within the last 12 months. We have not given a secondary importance to support hires, but we certainly had our focus primarily on the law enforcement investigators. But we are bringing on board our support employees. There are six areas with respect to the enhancements that we have requested in 1997. Perhaps I could just go through those very, very briefly. 1997 BUDGET REQUEST In total, the 1997 budget request to the committee is for $2.8 billion. It represents a $227 million increase over the 1996 request. The No. 1 request is for a $100 million appropriation for the implementation of the Communications Assistance for Law Enforcement Act [CALEA] of 1994 to ensure that law enforcement maintains court-authorized access to electronic surveillance to perform its most complex cases. In many ways, this is not simply an FBI request, nor a Department of Justice request. The electronic surveillance that the carrier compliance request will fund deals not only with law enforcement but with national security. The work that we do in the counterintelligence area as well as the assistance that we give the intelligence community is really dependent on the access to conversations, as well as data, which go over the various electronic networks. The $100 million specifically would be used to enforce the CALEA Act. That is the money that will go to the manufacturers and the carriers to retrofit their networks to ensure that we can maintain our court-authorized access. The majority of criminal wiretaps in the country are actually done by State and local authorities, not by the Federal Government, which is why this is a request for the law enforcement community at large. The majority of all of those requests are drug cases, which are the most difficult, the most complex, and where this particular technique is so valuable. I emphasize again, as I have during the history of this issue, that the act and this funding gives the FBI no new powers. All it does is ensure that the authority that Congress gave to the FBI and the other law enforcement agencies in 1968, which is the authority to listen to criminal conversations pursuant to a court order, is not re |