SECTION 8 ADMINISTRATIVE FEES: A REPORT TO CONGRESS June 1994 FOREWORD Since 1975, the Department of Housing and Urban Development (HUD) has used a complex methodology to determine reimbursements for public housing agencies (PHAs) that administer the Section 8 rental assistance program. The current fee structure, which is tied to fair market rents (FMRs), has little bearing on the actual cost of administering the rental assistance program, and results in a highly inequitable method of reimbursement for many PHAs. Pursuant to the HUD Demonstration Act of 1993, the Department has prepared this study, Section 8 Administrative Fees: A Report to Congress, to evaluate the fees that PHAs receive for administering the certificate and voucher programs. Data collected by HUD and others over the past decade show distinct differences among various types of PHAs in their ability to cover administrative costs. Although average balance figures may conceal wide extremes in surpluses and deficits, large urban PHAs historically have been over-reimbursed for administering Section 8 rental assistance programs, while the smallest PHAs have been under-reimbursed. The dramatic decline in surpluses for all PHAs over the past decade suggests that these reimbursements have not kept pace with the costs of administering Section 8 rental assistance. HUD's 1995 budget proposes to simplify the current system for reimbursing PHAs for administrative expenses. Because fair market rent levels have no apparent relationship to the cost of administering the certificate and voucher programs, the Department's proposal would decouple FMRs and administrative fees. HUD also would apply a less generous payment to the largest PHAs and establish floor and ceiling reimbursement levels that would limit fees for PHAs in high FMR areas and increase fees for PHAs in low FMR areas. The data contained in this report can help form the basis for developing an administrative fee system that more accurately reflects the costs of administering the Section 8 voucher and certificate programs. HUD looks forward to working with Congress during the Fiscal 1995 authorization and appropriations process to develop such a system. Michael A. Stegman Assistant Secretary for Policy Development and Research EXECUTIVE SUMMARY Overview Since 1975, the Department of Housing and Urban Development (HUD) has provided payments to public housing agencies (PHAs) to administer the Section 8 rental assistance programs. Historically, some agencies, particularly large PHAs, have been over-reimbursed for this function while other agencies, predominantly smaller agencies and rural PHAs, have received reimbursements that were consistently less than the costs they incurred in administering voucher and certificate assistance. The fact that some PHAs reap substantial profits while others incur heavy losses suggests an important need to reevaluate the current system of fee reimbursement. Pursuant to the HUD Demonstration Act of 1993, the Department has prepared a report that examines the extent to which the fees that PHAs receive for administering rental assistance cover the costs they incur. However, the report makes no effort to determine what the costs of administering the certificate and voucher programs should be, nor does it indicate the extent to which PHAs use administrative fees to cross- subsidize other allowable housing activities. As requested by Congress, this study also examines the potential impact of Federal mandates on the cost of administering the rental certificate and voucher programs. Background HUD's Section 8 certificate and voucher programs provide affordable, private market housing for 1.3 million households, most of whom are very low-income families. In 1993, almost 2,600 PHAs administered these rental assistance programs. Over 54 percent of these agencies are very small, operating programs of fewer than 200 units. In contrast, the 235 large PHAs -- those with 1000 or more units -- are responsible for more than 660,000 units. The Department provides administrative fees to cover activities that PHAs undertake in administering the rental voucher and certificate programs, including intake functions (e.g., issuance of vouchers and housing search activities), maintenance of assistance, and termination of payments. In 1993, the average ongoing reimbursement equaled roughly $44 per unit month of assistance; the total national cost was an estimated $659 million. The Department calculates administrative fees as a percentage of HUD-determined Fair Market Rents (FMRs). Each PHA receives a reimbursement based on a blended rate to reflect levels set before and after 1989. In 1993, the national average was approximately 7.6 percent. Although FMRs reflect fluctuations in local rental market conditions, they have little bearing on the actual cost of administering the Section 8 rental assistance program. This link between administrative fees and FMRs results in a highly inefficient and inequitable method of reimbursement for many PHAs. Interpreting the Data The most accurate information on costs comes from a very detailed study of administrative activities of large, urban PHAs conducted by Abt Associates in 1986. Given its "time and motion" data reflecting the actual levels of effort by PHA staff, the Abt study is a unique and extremely valuable resource on administrative costs. Abt reported that large urban PHAs received considerably more in administrative fees than they spent in administering the Section 8 certificate program. While the study estimated average total expenses at $333 per recipient year, total administrative reimbursements averaged $431 for the PHAs sampled. This estimated surplus of $98 per recipient year amounted to 23 percent of total administrative fees. Because small and rural PHAs comprise 90 percent of all PHAs and serve over one-half of the Section 8 rental certificate and voucher recipients, the study of large urban PHAs in 1986 does not represent the full range of PHAs. Moreover, these results may no longer be valid even for large, urban PHAs. The HUD analysis builds upon the Abt findings by collecting data from actual PHA operating statements during the 1980s and early 1990s. These operating statements greatly enhance the information obtained from the Abt study because they reflect program experience in all types of PHAs and show trends over time. When linked with basic program data, this information also enables the Department to evaluate the adequacy of these fees by geographic area, demographics, and program size. Based on the analysis of operating statement data during the late 1980s and early 1990s, HUD found the following: * Between 1984 and 1987, the average PHA showed a sharp decline in its surplus from 20.3 percent to 10.1 percent. This decline occurred in almost all types of PHAs; only in PHAs with high FMRs was the decline insignificant. The general decline of this surplus coincided with a reduction in the ongoing administrative fee from 8.5 percent in 1984 to 7.65 percent in 1985. * In 1992-1993, the average PHA had a fee surplus of 9.7 percent. Large urban PHAs had surpluses of 16.2 percent, while small urban and small rural PHAs had surpluses of 8.0 and 7.2 percent, respectively. In contrast, the smallest PHAs -- those with 200 or fewer units -- had an average loss of 1.4 percent. State and regional agencies also fared poorly, with an average loss of 0.8 percent. * Current operating statement data show wide variations in fee surpluses. While 31 percent of all units in PHA programs showed a surplus of 15 percent or more, about 40 percent of all units were in programs that incurred a loss. Roughly 14 percent of units are in programs with losses greater than 15 percent. Even among large urban PHAs, 31 percent of units were in programs that lost money; one-third of these units were in programs that incurred losses exceeding 15 percent. * Most PHAs have been able to build up operating reserves that can be used to support their Section 8 program or for other housing-related purposes permitted under State law. The average PHA has accumulated reserves equal to 52 percent of annual fee income. Reserves range from 65 percent for large urban PHAs to 28 percent for PHAs with programs of under 200 units and 21 percent for state and regional PHAs. Federal Mandates The most significant new Federal mandate has been the Family Self-Sufficiency program (FSS), which requires PHAs to coordinate the provision of training and services for a portion of their residents to help low-income families achieve economic self- sufficiency. By the end of FY 1993, 1,000 PHAs had mandates to implement such programs as adjuncts to their rental housing assistance programs. The costs that PHAs will incur in implementing the FSS programs reflect the coordination and management of services, which includes identifying service providers and enlisting their support, taking tenant applications and determining eligibility to participate in FSS, and assessing the needs of eligible tenants. Wages of FSS service coordinators comprise the principal costs of the program. Estimates of the potential annual cost of implementing the Section 8 FSS programs range from $31.7 million to $79.2 million, depending on the number of families that can be assisted by a service coordinator. The law authorizing the Family Self- Sufficiency program permits HUD to adjust the administrative fee to reflect the cost of implementing FSS. Approximately $8.4 million is available for the Section 8 FSS program in FY 1994, and the Department has proposed $17.3 million for 1995. Other program mandates have the potential to increase PHA administrative costs. The portability feature of the Section 8 program allows families to move outside a PHA's jurisdiction while continuing to receive assistance. However, the Department estimates that less than 5 percent of all families receiving assistance at any one time exercise the portability feature. Moreover, PHAs may receive additional reimbursement for costs associated with the portability provision in some circumstances. In addition, several special set-asides of rental certificate and voucher funds require PHAs to operate under several different sets of rules, which may impose an additional administrative burden. Recommendation for Change The information in this study underscores the need to reformulate the manner in which PHAs are reimbursed for their administrative costs. Data collected by the Department and others over a period of nearly a decade reflect clear differences in the ability to generate operating surpluses among different types of PHAs. As part of its FY 1995 budget proposal, HUD is proposing to simplify the current multi-tiered system for reimbursing PHAs. The Department's plan would increase the level of reimbursement for smaller PHAs operating in non-metropolitan and rural areas that traditionally have been underfunded, while realizing savings from programs operated by large urban PHAs with high FMRs that have generally been overfunded. Further, the proposal would decouple FMRs and administrative fees, thereby protecting PHAs from downward adjustments in FMRs. I. INTRODUCTION Statutory Mandate This report complies with Section 11(b) of the HUD Demonstration Act of 1993, which required the Department to study the costs incurred by Public Housing Agencies (PHAs) in administering the Section 8 rental certificate and voucher programs. Specifically, the Act required HUD to: 1) consider variations in costs attributable to geographic area, the tenant population, and the number of units administered by the PHA; and 2) analyze the costs associated with Federal mandates such as the Family Self Sufficiency (FSS) program. The Act also required HUD to submit its findings in conjunction with the Department's 1994 legislative recommendations. Section I of this report provides an overview of the administrative fee system, including a description of the fee amounts and the method of reimbursement. This section also documents the data and methods used in the report. Section II presents the findings of the study on administrative costs based on recently completed research and a collection of actual PHA operating statements. Section III evaluates the impact of unfunded Federal mandates on PHA administrative costs. PHA Administrative Fee System Program size. The Section 8 rental certificate and voucher programs provide rental assistance to approximately 1.3 million low-income families living in housing owned and managed by private landlords. PHAs administer these programs under contract with HUD. Currently, almost 2,600 agencies administer certificate and voucher programs. In 1993, the number of families receiving Section 8 rental assistance exceeded the number residing in public housing. Program characteristics. Most PHAs serve a single city or county. Many are also local housing authorities formed many years ago to develop and operate public housing programs. However, some local PHAs were formed specifically to administer Section 8 assistance. In addition, 137 state and regional PHAs with multi-county service areas provide assistance to approximately 220,000 families. Despite the misconception that certificate and vouchers work only in certain types of markets, the Section 8 program assists low-income households in every state and metropolitan area, and in most non-metropolitan counties, including many remote rural areas. Program cost. In FY 1993, the average administrative reimbursement to PHAs under the rental certificate and voucher programs was $44 per unit month of assistance, or about 10 percent of the average monthly cost of housing assistance subsidies of $400 per family. The aggregate national cost of the administrative fee system was approximately $659 million in FY 1993. Method of fee reimbursement. The methodology for reimbursing PHAs has evolved since the start of the Section 8 rental certificate program in FY 1975. At that time, there was little historical information on which to base the reimbursement system because earlier programs did not assist families who sought their own housing in the private market. The original system, which was designed largely to cover the labor costs of administering the rental assistance program, assumed that these expenses varied in relation to the costs of renting decent quality housing. The original fee reimbursement consisted of a $275 per unit start-up or "preliminary" reimbursement, and an ongoing administrative fee equal to 8.5 percent of the two- bedroom Fair Market Rent (FMR) that HUD established for the local area. At first, most PHAs found that the system provided ample reimbursements for the costs they incurred. An early study of PHA administrative fees from 1978-1980 found that the reimbursements were generous enough for most to generate surpluses and build up operating reserves. In fact, growing anecdotal evidence by the mid-1980s suggested that some PHAs were able to build up quite large reserves. As a part of a government-wide initiative to reduce administrative expenses for the certificate program, the Federal Government reduced PHA administrative fees from 8.5 percent to 7.65 percent of the FMR in 1985. With the introduction of vouchers in 1985, the Federal Government set the ongoing fee at 6.5 percent of the FMR under the assumption that PHA administrative responsibilities under the voucher program would be less costly than administering the certificate program. However, voucher demonstration research published in 1990 indicated that administrative costs were about the same under the two programs. In 1987, Congress enacted a law requiring that PHAs be reimbursed at 8.2 percent of the FMR, subject to availability of annual appropriations. However, each year's appropriation has provided for the 8.2 percent fee only for allocations made in or after 1989, and not for the assistance allocated from 1975 through 1988. As a result, each PHA now receives reimbursement based on a blended rate, using 8.2 percent of the FMR for allocations of certificates or vouchers made starting in 1989, 7.65 percent for allocations of certificates made prior to 1989, and 6.5 percent for allocations of vouchers made prior to 1989. The national average blended rate for all PHAs was approximately 7.6 percent in 1993. Linking administrative fees to FMRs. Calculating the administrative fee as a percentage of the FMR is a growing problem for both PHAs and HUD. Under this system, PHAs bear the burden of tracking the numbers of families receiving assistance under allocations made before and after 1989 for both certificates and vouchers. Also, FMRs are subject to fluctuations in local rental market conditions that often have no bearing on the cost of administering the program. One significant source of fluctuation in FMRs is the periodic revision of FMRs when more accurate American Housing Survey or decennial census data become available. The 1993 revision resulted in final FY 1994 reductions for one-half of all metropolitan areas and three-fourths of all non-metropolitan areas. Some areas experienced reductions exceeding 10 percent and even 20 percent of the 1993 FMRs. To avoid sudden decreases in PHA administrative fees, Congress stipulated in 1993 that the FY 1994 fee would be based on the larger of 1) the 1993 FMR, or 2) the 1994 FMR, but not to exceed the 1993 FMR by more than 3.5 percent. Data Sources This report provides information from three primary studies -- a 1986 study conducted by Abt Associates, Inc., of large urban PHAs, an HUD analysis of PHA operating statement data from the late 1980s, and a HUD analysis of PHA operating statement data from the early 1990s. Appendix B provides a more detailed discussion of the data used in this report. II. COSTS OF ADMINISTERING SECTION 8 RENTAL ASSISTANCE Historically, some agencies, particularly large PHAs, have been over-reimbursed for administering Section 8 rental assistance programs. In contrast, many smaller and rural PHAs have received reimbursements that are consistently less than the costs they incur in administering these programs. The fact that some PHAs reap substantial profits while others incur heavy losses suggests an important need to reevaluate the current system of fee reimbursement. Pursuant to the HUD Demonstration Act of 1993, the Department has examined the extent to which the fees that PHAs receive for administering rental assistance cover the costs they incur. However, this report does not suggest how much it should cost to administer a rental certificate or voucher program. Similarly, the analysis does not indicate whether PHAs, either individually or collectively, are over- or under-staffed. A. Administrative Costs Study of Large Urban PHAs The 1986 study Administrative Costs in the Housing Voucher and Certificate Programs provides the most thorough and accurate information available on administrative costs associated with the rental certificate and voucher programs. Conducted by Abt Associates, Inc. as part of the Freestanding Housing Voucher Demonstration, the study documented components of administrative costs and collected data on the actual PHA staff time spent on administrative functions. From a sample of 18 large urban PHAs, Abt collected reliable data on initial certificate/voucher issuance and housing search activities for 16 sites, and collected data on ongoing administrative costs for 13 sites. Using this information, Abt made national projections for large urban PHAs. Administrative Costs is useful because it reflects expenses actually incurred in running program activities. These expenses included not only the immediate usage and salary costs of program staff, but also all of the supervisory, overhead, fringe benefit, and non-labor costs associated with program administration. Abt collected information on these costs from the Demonstration PHAs' operating records and accounts, records of administrative activities collected for Demonstration households, and special records of PHA staff time collected for this study. Table 2.1 presents a typology of PHA administrative activities used in this study -- intake, maintenance, and termination. Intake included all activities associated with initial participation by a family, including receiving the application, determining eligibility, briefing the family, and conducting the housing search. Maintenance activities included the ongoing PHA activities to support continued assistance for the family, whether leasing in place or moving (with continuing assistance) to another unit. Direct casework labor comprises roughly 30 percent of the costs PHAs incur in administering the rental assistance program. Administrative Costs quantified the average cost of performing each type of intake and maintenance activity, the time it took to perform the activity, and the implied hourly wage for the staff performing each task (Table 2.2). Other costs included staff time not directly attributable to intake and maintenance tasks, costs of PHA management and support staff, fringe benefits, overhead, and other non-labor costs (Table 2.3). As illustrated in Table 2.4, large urban PHAs received considerably more in administrative fees than they expended for administering the Section 8 certificate program. Assuming that PHAs expanded their programs by 12 percent each year based on incremental allocations of assistance from HUD, costs for administering the certificate program averaged $333 per recipient year. This consisted of $261 for administering ongoing slots (including turnover) and $72 for administering new intake slots. In contrast, total reimbursements averaged $431 per recipient year, which consisted of $400 from ongoing fees plus $31 from new intake slots. The difference of $98 reflects a surplus of 23 percent. By lowering the assumed new intake rate to 6 percent, the estimated administrative surplus reaches 28 percent of fees (Table 2.5). Expenses and Fees by Major Activity. According to the Abt study, PHAs generally were over-compensated for their ongoing expenses but under-compensated for their intake expenses. Ongoing expenses averaged roughly 65 percent of their fees for these activities, while expenses for new intake slots averaged about 232 percent of the reimbursement. Assuming a 6 percent rate of growth of new intake slots, the $139 surplus of ongoing activities per recipient year far outweighed the deficit of $21 for new intake slots. In estimating expenses for ongoing activities, the Abt study examined the impact on administrative costs when people terminate (or "turn over") their certificates. Reimbursement for ongoing activities covers both turnover costs and normal maintenance costs for ongoing slots. Abt found that administrative fees for ongoing activities enable large urban PHAs to meet their combined expenses for both the turnover and maintenance of ongoing slots. For the average large, urban PHA, the Abt study suggested that the administrative fee structure in 1986 was more than adequate for PHAs to meet ongoing and total administrative expenses associated with the rental assistance programs. However, these averages can conceal important differences among individual PHAs. Assuming that new intake slots grew at 12 percent per year, the performance of the 13 PHAs varied substantially. Nine of the thirteen PHAs with both intake and ongoing expense data generated a total surplus of at least 27 percent of their fees, and 6 of these PHAs enjoyed a surplus greater than 40 percent. On the other hand, three of the thirteen PHAs showed a deficit in their total expenses relative to total fees. With a growth rate computed as the higher of their 1985 and 1986 rates of certificate growth, 9 of the 13 PHAs showed at least a 31 percent surplus, while 2 PHAs showed a deficit. Impact of Tenant Population and Other Factors. There are no factors that easily predict PHA surpluses. Table 2.6 provides median characteristics of PHAs grouped by surplus size. These results, as well as results through more rigorous statistical testing, suggest that the magnitude of surpluses for large urban PHAs in the sample are unrelated or only marginally related to such characteristics as the two bedroom FMR of the area, the proportion of the zero and one bedroom recipient units (a proxy for elderly households), the turnover rate of existing recipients, the ratio of public housing to Section 8 households served, or the ratio of the county's FMR to local government wages. B. Evidence from PHA Operating Statement Data Because small and rural PHAs comprise 90 percent of all PHAs and serve over half of Section 8 rental certificate and voucher recipients, the study of large urban PHAs in 1986 does not represent the full range of PHAs. Moreover, these results may no longer be valid even for large, urban PHAs given the date of this information. The HUD analysis builds upon the Abt findings by collecting data from actual PHA operating statements during the 1980s and early 1990s. These operating statements greatly enhance the information obtained from the Abt study because they reflect program experience in all types of PHAs and also because they show trends over time. When linked with basic program data, this information also enables the Department to evaluate the adequacy of these fees by geographic area, demographics, and program size. Appendix B provides a more thorough discussion of the advantages and disadvantages of each data source. 1. Certificate Results in the Late 1980s Operating Surplus. As illustrated in Table 2.7, the average PHA showed a sharp decline in surplus from FY 1984 to FY 1987. Nationwide, the average surplus fell from 20.3 percent in FY 1984 to 10.1 percent in FY 1987. Almost all PHAs experienced a decline of similar magnitude. For example, large, urban PHAs saw their surplus decline from 22.8 percent in FY 1984 to 13.1 percent in FY 1987. PHAs with high FMRs in FY 1987 saw the smallest decline from 21.2 percent in FY 1984 to 18.4 percent in FY 1987, which may result from the above-average growth of their FMRs from FY 1984 to FY 1987. The general decline of the percentage fee surplus occurred at the same time as the reduction of the ongoing fee from 8.5 percent to 7.65 percent of the two-bedroom FMR. Consider a PHA with $300 of ongoing fees per recipient year in FY 1984 and $240 of ongoing expenses. Suppose that all of its HUD fees and expenses came from its ongoing (as opposed to preliminary expense) account. In FY 1984, this hypothetical PHA would have generated a surplus of $60, or 20 percent. If the fee were reduced to $270 per recipient year by 1987, the net surplus would fall from $60 to $30 by 1987, a reduction of 50 percent. The most plausible inference is that the average PHA did not find slack in its operations to maintain the FY 1984 surplus level with its lower fees. Geographic Area, Tenant Population, and Other Factors. As illustrated in Table 2.7, the average surplus for large, urban PHAs exceeded the national average by several percentage points for FY 1984, FY 1986, and FY 1987. In contrast, the FY 1987 average rural PHA surplus of 2.9 percent was considerably lower than the national average of 10.1 percent. Because PHAs with an operating surplus of five percent or less might be holding expenses down to avoid a deficit, the budgetary constraint on rural PHAs in FY 1986 and FY 1987 might be even greater than that shown in the operating statement data. There are plausible reasons for the relatively greater pressure of expenses on the fees of rural PHAs-- the greater geographical dispersal of their units and households and the greater proportion of affordable units likely to fail quality standards. At the aggregate level, Census data and American Housing Survey data show a higher likelihood of these problems in rural areas. High-FMR PHAs averaged a surplus of five to eight percentage points above the overall average in FY 1986 and FY 1987, while low-FMR PHAs averaged a surplus four to six points below the overall average in FY 1984, FY 1986, and FY 1987. For FY 1986 and FY 1987 combined, the surplus for low-FMR agencies was identical to the surplus of 4.2 percent for PHAs in rural areas, which reflects the considerable overlap between these groups. Another reason for the relatively tight budgets in low-FMR areas is that their FMRs grew only 2.7 percent per year during this period, while FMRs in high-FMR areas soared by 10.1 percent per year. Finally, high-FMR agencies fared better than their low-FMR counterparts because wages that drive expenses vary less than the FMRs that determine reimbursement fees. Using the state and local government wage index as a proxy for PHA wages, average monthly wage for low-FMR agencies was 66 percent of that for high-FMR PHAs, while the FMR for low-FMR agencies was only 51 percent of the FMR for high-FMR agencies. Table 2.7 illustrates that PHAs with low FMRs relative to wages have a below-average operating surplus, whereas PHAs whose FMR is relatively high compared to their wages generate an above- average operating surplus. In comparison to the FY 1987 nationwide average surplus of 10.1 percent, PHAs with a low FMR to wage ratio had an operating surplus of 7.2 percent, while those with a high FMR to wage ratio had an operating surplus of 14.4 percent. PHAs that administered Section 8 rental assistance but operated no public housing units show an operating surplus several points above the overall mean, whereas PHAs with extensive public housing units showed an operating surplus slightly below the national average. PHAs with a high ratio of public housing units were somewhat more likely to show higher expenses and a lower operating surplus than PHAs without public housing units, which may indicate that some PHAs were cross- subsidizing non-certificate housing activities. 2. Rental Voucher Results for the Latest Two PHA Fiscal Years To update these findings from the late 1980s, HUD collected and verified operating statements for rental certificate and voucher programs from a sample of 535 PHAs in 1993. HUD asked field offices to submit this information for the latest two years, which in most cases was for FY 1992 and FY 1993. Unfortunately, a problem resulting from methods that PHAs use to prorate costs between the various parts of local certificate programs has made that program data unusable for purposes of this analysis. Reliance on administrative cost data from the rental voucher program alone would not made sense in the late 1980s, because the program was relatively new and still maturing. However, because the voucher program is well established and is virtually indistinguishable in terms of administrative functions and cost from the certificate program, the voucher information adequately reflects activity in both Section 8 rental assistance programs. Appendix B provides a more detailed discussion of the operating statement data for 1993. As illustrated in Table 2.8, PHAs received $40.72 per unit month to administer voucher assistance and reported $36.77 per unit in expenses. The percentage operating surplus for all PHAs was $4.05, or 9.7 percent. Consistent with earlier research, large urban PHAs fared much better than other PHAs, with an average operating surplus of 16.2 percent. Small urban and small rural PHAs generated average surpluses of 8.0 and 7.2 percent, respectively. Surprisingly, state and regional PHAs seem to have the greatest difficulty operating under the current fee system. During the latest two fiscal years, these agencies incurred a slight deficit of 0.8 percent on average. Table 2.8 confirms that the fee surplus varies significantly by size of PHA. During this two-year period, PHAs with less than 200 units actually lost money on average. PHAs with 200-999 units experienced a 9.4 percent fee surplus, which is roughly equal to the average for all PHAs. PHAs with more than 1,000 units generate larger than average surpluses of 10.7 percent. The large state agency PHAs, which often break even or generate only small surpluses, pull down the average for the 1000+ group relative to the results already reported for large urban PHAs. The results for the latest two fiscal years are very similar to those found for a smaller sample of PHAs operating certificate programs in 1987. Table 2.9 compares the fee surplus in 1987 to that calculated for a subset of the most recent sample -- the 288 PHAs that have valid data for both time periods and both programs. For these PHAs, the average surplus was 10.8 percent in 1987 and 8.6 percent during the latest two years. Large urban PHAs experienced about a 14 percent surplus in both time periods. Small rural PHAs also showed no significant change. Surpluses for state/regional PHAs declined from 5.6 percent in 1987 to 1.2 percent, and surpluses for PHAs with less than 200 units decline from 6.7 percent in 1987 to 1.9 percent for the latest two fiscal years. PHAs that spend less than the amount they receive under fee reimbursements add any net income to their operating reserves. Most PHAs have reserves: only 9.5 percent of units are in PHA operating statements reporting a negative average reserve value. Five percent of units in large, urban PHAs are in programs with negative reserves, compared with 7 percent for small urban PHAs and 12 percent for small rural PHAs. Once again, State and regional PHAs seem to do less well, with 20 percent of units in programs with negative reserves. As illustrated in Table 2.10, the average PHA has accumulated average reserves equal to 52 percent of annual fee income. This ranges from 65 percent for large urban PHAs to 28 percent for PHAs with under 200 units and 21 percent for state and regional PHAs. Once funds are placed in the operating reserves, PHAs may withdraw the money for any housing-related purpose allowed under State law. A large difference between the maximum and average reserve amounts would indicate that significant amounts have been withdrawn by the PHAs. As shown in Table 2.10, the lower relative reserves for State and regional agencies do not result from significant withdrawals of funds. Their maximum reserves were only about 25 percent above their average reserves; this is similar to the result for all PHAs. Like the Abt study, the HUD analysis also found that, even among large urban PHAs, some agencies have expenses in excess of fees. As illustrated in Table 2.11, roughly 70 percent of units administered by large urban PHAs are in programs with a surplus, including 30 percent in programs with a fee surplus of 25 percent or more. On the other hand, 31 percent of units in large urban PHA programs have lost money during this period, including 4 percent in programs with losses exceeding 25 percent. Some PHAs with an excess of expenses over fees might be "living off their reserves" -- intentionally using large reserves to provide a higher level of services. For example, some PHAs may use their large reserves to hire additional staff. While this may occur in some cases, PHAs that incur losses do not have larger than average reserves to accommodate a shortfall in fees. As illustrated in Table 2.12, 82 percent of units in PHAs that incurred losses in the most recent two years have an average reserve value equal to six months or less of annual fee income. In addition, the 18 percent average reserve as a percentage of fee income is lower for these PHAs than the 52 percent average reserve for all PHAs. Neither surpluses nor deficits provide conclusive evidence that the system of fee reimbursements is flawed. There is no standard to determine how much a PHA should spend to administer the Section 8 rental assistance program. While losses at some PHAs may reflect management problems or short-term adjustments to circumstances beyond the agencies' control, there is no guarantee that PHAs with surpluses have provided an adequate level of service. Nevertheless, the weight of evidence from PHA operating statements suggests a need to address some relative inequities in the system that have resulted from the linkage between ongoing fees and FMRs. C. The Administration's Proposal The Administration's 1994 legislative package would substantially revise the method used for administrative fee reimbursement. This proposal would redistribute fee income from large PHAs and those in unusually high FMR areas to smaller PHAs and those operating in unusually low FMR areas. As previously shown in table 2.8, PHAs operating in the highest 15 percent of FMRs have a fee surplus of 15.7 percent, versus 8.2 percent for those operating in the lowest 15 percent of FMRs. This is similar to results found by quintile of FMRs for the certificate program in 1987 (Table 2.7). Table 2.8 also calculates the percentage operating surplus for PHAs that would gain or lose over 5 percent on a per-unit- month basis under the Administration's proposal. PHAs that would gain the most have an average fee surplus of only 2.5 percent for the latest two fiscal years, well below the national average of 9.7 percent. PHAs that would lose the most had an average fee surplus of 12.4 percent, which is higher than the national average, and more than sufficient to maintain adequate operating reserves. The Administration's proposal would allow for unusual expenses of State and regional agencies to be reimbursed over and above the normal per-unit-month fee. The proposal would retain current statutory provisions that allow for additional fees for small PHAs, delivery of assistance within large geographic areas, and extraordinary costs. HUD would approve additional fees only in unusual circumstances, where the PHA documents and justifies the need. Appendix A provides a more comprehensive discussion of the Administration's proposal. III. FEDERAL MANDATES In the context of PHA administrative fees, unfunded Federal mandates represent any major new function that increases the cost of program delivery without modifying the level of administrative fee reimbursement to reflect the new activity or function. The most significant new Federal mandate has been the Family Self Sufficiency (FSS) program, which requires PHAs to coordinate the provision of training and services for low-income families, to help these families achieve economic self sufficiency. By the end of FY 1993, approximately 1,000 PHAs were required to implement FSS programs as an adjunct to their rental certificate and voucher programs. This section provides background information on FSS and discusses the potential effect on PHA administrative costs. This section also describes the potential impact of portability and program set-asides. Family Self Sufficiency Program Background. Established by Section 554 of the National Affordable Housing Act in 1990, FSS uses HUD's housing programs as leverage to create individualized packages of services to help public housing residents and recipients of Section 8 rental certificate and voucher assistance either to reduce their dependency or to become self-sufficient. Although HUD originally defined self-sufficiency as being free of housing assistance and welfare (AFDC, SSI subject to income eligibility tests, Medicaid, food stamps, and general assistance), the Housing and Community Development Act of 1992 revised this program to focus on welfare alone. Not only does HUD evaluate local FSS programs by the number of families who actually achieve self-sufficiency, but also by the number of FSS families who, as a result of participation in the program: * have family members who obtain their first job, or who obtain higher paying jobs; * no longer need benefits received under one or more welfare programs; * obtain a high school diploma or higher education degree; or * accomplish similar goals that will assist the family in obtaining economic independence. Participation Requirements for PHAs and Families. FSS began as a voluntary program for PHAs and Indian Housing Authorities (IHAs). HUD operated the program as a competition using a set- aside of public/Indian housing development funds and Section 8 vouchers and certificates called "incentive units." In FY 1993, the program became mandatory for PHAs receiving additional public housing development funds or Section 8 vouchers and certificates. The 1992 Act made the FSS program voluntary for IHAs. PHAs may be exempted from the program under four conditions: * lack of accessible supportive services funding, including lack of the availability of programs under other Federal programs; * lack of funding for reasonable administrative costs; * lack of cooperation by other units of State or local government; or * lack of interest in participating in the program on the part of eligible families. The Secretary may approve requests for other types of exemptions as well. Participation in the program is entirely voluntary for families. The 1992 Act made it very clear that no family can be adversely affected because of a decision not to participate. Minimum program size. The minimum program size, which is established separately for public housing and Section 8 assistance, is based on new allocations of incremental units. Specifically, the minimum size is equal to the cumulative total of all new public housing development units and Section 8 certificates and vouchers received under the FY 1991 and 1992 FSS competitions, plus the number of incremental public housing units, certificates, and vouchers received in 1993 and in subsequent years. Under the program, PHAs must replace families who complete the FSS program or drop out. There is no tie between any specific dwelling units, the families occupying them, and the program requirements. The number of new incremental units allocated simply determines the number of slots that PHAs must add to their FSS program. Potential cost of administering FSS. The costs incurred by PHAs in carrying out FSS programs reflect the coordination, rather than the provision, of services. This coordination includes: * identifying service providers and enlisting their support; * taking tenant applications and determining eligibility to participate in FSS; * determining needs for education, job training and related services; * assisting families in attaining these types of training and services; * establishing an FSS escrow account for the family; and * implementing the tracking and reporting required by the Department. The FSS service coordinator function is essentially a case management function added to the casework activities described in Table 2.1 through 2.3 of this report. While the administrative costs of these activities are primarily wage costs, there are also fringe costs, and to a lesser extent, overhead and non-labor costs (e.g., office rent and supplies) associated with these activities. The general magnitude of the FSS commitment is reflected in a few simple assumptions about program delivery: * The average wage cost for staff involved in FSS service coordination is $29,750 per year, which reflects an average hourly wage of $11 and an additional 30 percent for overhead, fringe, and non-labor costs. Table 2.2 provides implied average hourly wages found for large urban PHAs in 1986; * The number of families that one FSS service coordinator can assist increases with the size of the PHA; and * The FSS program size reflects the minimum program requirements, based on FY 1992 bonus allocations and FY 1993 incremental allocations. Assuming a ratio of one FSS coordinator per 20 families, the cost of administering FSS is approximately $79.2 million. On the other hand, if one FSS service coordinator could manage the FSS workload for as many as 50 persons, the program cost would be $31.7 million. Table 3.1 provides alternative estimates of cost based on assumptions about the capacity of service coordinators. FSS Service Coordinator Funding. While PHAs operating a public housing FSS program may receive additional operating subsidies to cover administrative costs of FSS, agencies operating a Section 8 FSS program must cover the additional administrative costs for this program out of their regular administrative fee reimbursement. Authorization language permits HUD to adjust the administrative fee to reflect the costs of carrying out an FSS program. In FY 1993, HUD made the first awards of FSS service coordinator funds for public housing. The Department provided $26.3 million to PHAs administering FSS programs in conjunction with their public housing programs. However, HUD did not provide funds for FSS programs associated with rental certificates and vouchers in FY 1993. In FY 1994, the Department expects to provide an additional $26.3 million for public housing FSS programs, and $8.4 million to PHAs administering FSS for rental certificates and vouchers. Because only $8.4 million is available for allocation to support Section 8 FSS coordination for FY 1994, the Department has tentatively limited eligibility to smaller PHAs. For FY 1994, the Department will accept applications only from PHAs with 600 or fewer units under management. HUD hopes to expand funding availability to a wider range of PHAs in future years, which is contingent on Congressional approval of the Department's budget proposals. For FY 1995, HUD proposes $38.7 million for public housing FSS and $17.3 million for FSS in certificates and vouchers. Other Federal Mandates The Federal government has made numerous changes in the design of the Section 8 program since the late 1970s, when Section 8 certificates were generally recognized as a relatively simple and straightforward program with reasonable administrative costs and reimbursements. The current rental certificate and voucher programs operate much as certificates did in the 1970s. However, the rules that PHAs must implement have become much more complex. Relative to the late 1970s, the certificate program and voucher program added significant PHA responsibilities in such areas as the rules for maintaining PHA waiting lists (including administration of Federal preferences for admission to assisted housing) and the rules regarding eviction, termination of tenancy, and payment of damage claims. The potential cost impact of these aspects of PHA administration is difficult to quantify and falls outside the scope of this report. However, program simplification is an obvious yet often overlooked means of reducing administrative costs. Portability. The portability feature allows low-income families to move outside of the PHA's jurisdiction while continuing to receive assistance. Since the inception of vouchers in 1985, the Federal government has required limited portability. Statutory changes in 1987 and 1990 extended statewide and metropolitan-wide portability for both rental certificates and vouchers. PHAs must comply with a special notice establishing HUD policy regarding which FMRs, payment standards, and occupancy standards apply. Portability generates higher PHA administrative costs from the "billing" of subsidies and fees from one PHA to another, issuance of monthly checks for payment of billed amounts, as well as the additional telephone calls, correspondence and paperwork associated with this activity. Portability also creates a problem because of the extra staff time needed to understand and explain portability rights and requirements. HUD estimates that less than 5 percent of all families receiving assistance at any one time exercise the portability option. However, when families exercise the portability option extensively, it can add significantly to administrative cost. Similar to FSS, HUD provides PHAs with additional reimbursement for administering the portability provision. In addition to the usual administrative fees, which the initial and receiving PHAs split, the initial PHA may bill HUD for actual expenses of up to $250 per family, when supported by a certification from the receiving PHA. However, this applies only to the extent that PHAs are eligible to claim preliminary expenses for that same year, which limits their ability to claim the additional portability fee. Subprograms and Special Set-asides. Sub-programs and special set-asides potentially add to the administrative costs of PHAs because they require separate applications for assistance, and, in some instances, separate accounting, outreach to special populations, management of separate waiting lists, and more generally, increased staff and PHA supervisory time spent learning the requirements of the program and explaining the features to potential program participants. Shared Housing, Manufactured Housing, and Project-based assistance are all examples of subprograms within the certificate program that have increased the administrative responsibilities of the PHA in recent years. None of these programs account for a significant percentage of units, either nationally or locally. Yet PHAs must be familiar with the program features, train staff, and administer these provision when necessary. The rental voucher program is perhaps the most significant example of a separate program of tenant based assistance that has complicated the responsibilities of the PHA. The existence of two separate programs, with separate budgeting and accounting and different rules for calculation of tenant subsidy and different responsibilities of landlords and tenants, has no doubt caused the PHAs to utilize more staff hours carrying out PHA functions than if all the units were certificates or vouchers. The Administration's 1994 legislative package includes a proposal to merge the rental certificate and voucher programs. APPENDIX A SECTION 8 FEES, SECTION-DESCRIPTION Overview of the Proposal Section 232 of the Department's proposed 1994 reauthorization bill would amend section 8(q) of the 1937 Act to change the way HUD determines fees that are paid to PHAs, including Indian housing authorities, for the costs of administering the Section 8 certificate and voucher programs. This amendment also would limit the preliminary fee to the initial increment of assistance to PHAs that have not previously carried out a certificate or housing voucher program. For these PHAs, the amount of the preliminary fee would increase from $275 to $500. Under the revised system, a PHA would receive a fee for each month for which a dwelling unit is covered by a housing assistance payments (HAP) contract. This initial fee would be 7.65 percent of a HUD-determined base amount for the first 1,000 units, and 7 percent of the base amount for any additional units. Each year, HUD would publish as a notice in the Federal Register the per-unit-month fee amounts that would apply for PHAs operating in each metropolitan area and non-metropolitan county for that Federal fiscal year. The Department would index the per-unit-month fee amounts to wage-inflation or other measurable data that reflect the costs of administering the program. The determination of the "base amount" would build on current practices. As a result of Section 11 of the HUD Demonstration Act of 1993, the base amount used in FY 1994 is equal to the larger of two numbers: (a) the FY 1993 fair market rent (FMR) established by HUD for a 2-bedroom existing rental dwelling unit in the market area of the PHA; and (b) the FY 1994 FMR when higher than the FY 1993 FMR, but not to exceed the FY 1993 FMR by more than 3.5 percent. Under the proposed system, the base amount would be identical to the base amount actually used in FY 1994, but would be subject to a ceiling and floor. The base amount would be used only to determine the fee amount for the initial year. To calculate the ceiling and floor, HUD has examined the distribution of certificates and vouchers across all PHAs in the country. The Department proposes to establish a floor that reflects the 15th percentile of the units administered by PHAs, or $422. The proposed ceiling represents the 85th percentile of rent, or $777. By applying these caps and floors to the base amounts, HUD will increase the level of reimbursement to PHAs serving areas with very low FMRs and avoid over-reimbursement of PHAs providing assistance in very high FMR areas. The proposal would retain authority for HUD to increase the fee if necessary to reflect the higher costs of administering small programs and programs operating over large geographic areas (see Section 8(q)(1) of existing law), and for extraordinary expenses (see Section 8(q)(2)(A)(iii)). In addition, HUD could approve higher fees if necessary to reflect the higher costs of administering the Family Self-Sufficiency program under section 23 of the 1937 Act. Addressing Problems with the Current System The current system of Section 8 administrative fees is unnecessarily complex, unwieldy, and inconsistent with program needs. For pre-1989 allocations a 6.5 percent fee applies for vouchers and a 7.65 percent fee applies for certificates. For both programs, an 8.2 percent fee applies for incremental allocations made after 1988. Research shows that administrative costs for certificates and vouchers are very similar. By making the fee system more uniform, the proposed amendment would simplify the current system and eliminates its most serious flaws. Basing administrative fees on each year's FMRs links administrative budgets to changes in FMRs. Rents are subject to market forces and periodic rebenchmarking that can produce sudden increases or decreases in FMRs and administrative fees without any connection to changes in administrative costs. Erratic and sudden changes in administrative fees can undermine sound program management and can disrupt PHA efforts to provide a high and consistent quality of management and advisory services. Without the special legislative language enacted in the Fall of 1993, the rebenchmarking of FMRs to the 1990 Census would have increased or decreased administrative funding for some PHAs by 25 to 30 percent. The Administration's proposal would solve these problems on a more permanent basis. Under this proposed new PHA fee system, PHAs would no longer face the possibility of sudden decreases in administrative budgets. Small PHAs and PHAs with unusually low FMRs would tend to receive higher fees. Large PHAs and those operating in high-FMR areas that receive excessive fees would experience some decreases. The link between administrative fees and FMRs produces upward pressures on the latter. The primary cost of administering the Section 8 program is paid in wages to PHA employees. These wages are closely tied to local wage costs, but not necessarily to local rental costs. FMR/local wage ratios differ significantly from area to area, with low-FMR areas relatively underfunded and high FMR areas relatively overfunded. Small PHAs and PHAs in non-metro areas tend to have the lowest FMRs and appear to be least-favored by the current system. HUD research indicates that housing costs (and FMRs) are more variable than wages and non-housing costs, and that areas with unusually high or low FMRs receive relatively high or low levels of administrative funding relative to local wage and other non-housing costs. The Department's proposal addresses this inequity by placing a "ceiling" and "floor" on the calculation of the initial fee base. Small programs appear to have difficulties with current administrative fee levels, partly because they tend to operate in low-FMR areas and partly because they are unable to achieve the economies of scale possible in larger PHAs. This is especially true where the small program covers a large geographic area. Under the new system, small PHAs would tend to receive higher fees, and could also apply for additional funds as needed. The Administration's proposal would repeal the current statutory provision in Section 8(q)(2)(A)(ii) regarding costs of assisting families who experience unusual difficulties. The currently used "hard-to house" add-on to fee reimbursements is no longer necessary to ensure that adequate assistance is provided to large families with children. Almost 20 percent of families participating in the Certificate and Voucher programs contain three or more children, and the recent rebenchmarking of FMRs in most local markets has generally increased the FMR applicable for units with three or more bedrooms, and presumably also the availability of such units. The proposal would retain current statutory provisions that allow for additional fees for small PHAs, delivery of assistance within large geographic areas, and extraordinary costs would be retained. However, HUD would approve additional fees only in unusual circumstances, where the PHA documents and justifies the need. The use of a floor in the setting of the initial fee base should help most small PHAs and PHAs serving large geographic areas, minimizing the need for additional fees. The Administration proposes to increase the current preliminary fee limit from $275 to $500 per unit for new allocations. The old limit, which has not been modified since the program's inception in the mid-1970s, is no longer a significant source of revenue because program sizes are now large relative to incremental unit allocations in any one year. Under this proposal, HUD would provide the preliminary fee to PHAs in their initial year of implementing either the certificate or housing voucher program, without documentation by a PHA, which would eliminate unnecessary paperwork. Virtually all PHAs are able to justify the proposed level of preliminary fees in their first year of participation in the program. Implementation Implementation of this proposal will require issuance of a proposed and final rule. HUD anticipates initial implementation by FY 1995. To avoid administrative problems associated with sudden changes in fees and to recognize the previous rebenchmarking of FMRs, this proposal would extend the fee rates applicable in FY 1994 until the Department implements a final rule for this legislative proposal. APPENDIX B DATA DISCUSSION Overview This report provides information from three primary sources -- a 1986 study by Abt Associates, Inc., of large urban PHAs entitled Administrative Costs in the housing Voucher and Certificate Programs, an HUD analysis of PHA operating statement data from the late 1980s, and a HUD analysis of PHA operating statement and other data from the early 1990s. This information is organized at three levels: Limited data for the universe of PHAs. Using information from special extracts of the Section 8 Management Information System (MIS) and other sources, the Department collected data on the total number of certificates and vouchers, the year of allocation, and FMRs for all PHAs. The data reflect a total of 2,540 PHAs with total fund reservations made for 1,191,254 units as of June, 1993. Table B.1 presents information on the PHAs included in this report. This data file includes 1,378 PHA operating programs of 200 or fewer units, which account for about 10 percent of all units. The 235 large PHAs -- those administering programs of 1,000 units or more -- are responsible for 58 percent of all units. Intermediate size PHAs comprise the remaining 32 percent of all units. Neither the data nor the findings of this study reflect information from the New York City Housing Authority, which administers the largest single program in the nation. Given the size of the New York City Housing Authority programs and the unique nature of that housing market, HUD often presents findings for these assisted housing programs separately from findings for the rest of the nation. Because presenting separate findings for New York would have required HUD to publish the annual operating statements for that agency, the Department chose not to include this information in the report. Detailed data for 2,100 PHAs. For 2,100 PHAs administering 96 percent of all units, the Department has collected additional information identifying such factors as geographic location, tenant population, presence of public housing, and local government wage levels. Operating Statement Data. For a representative, stratified sample of 535 PHAs, HUD has collected and validated Section 8 rental certificate and rental voucher operating statements (Form HUD-52681) for the latest two PHA fiscal years, generally for FY 1992 and 1993. The Department has linked these data with: 1) a previously collected, representative sample of FY 1987 operating statements for Section 8 certificates for 350 PHAs; 2) FY 1986 and FY 1984 operating statement data for 250 of these same PHAs; and 3) the data on PHA and community characteristics referenced above. For statistical analyses, HUD weighted sampling strata based on the total number of certificates and vouchers (certificates only for 1987 and prior years) and the geographic status of the PHA jurisdictions. Comparison of Research and Operational Data The data available through PHA operating statements offer some advantages over the types of data available through the Abt study of large, urban PHAs. First, the data are available for many more PHAs than the 13 to 16 PHAs in included in the 1986 study. Second, unlike the Abt study, which required making assumptions about the growth of new intakes to compute total estimated costs and fees, the PHA operating statement data provide direct measures of total expenses and fees and their balance. Third, the operating statements indicate the PHA's operating reserve, which provides some indication of the accumulated fee surpluses (or deficits) that PHAs have generated over the years. Admittedly, this measure provides a very conservative estimate of the adequacy of fees, because PHAs are permitted to withdraw funds from the reserve for other housing- related purposes permitted under State law. However, there are also some disadvantages to using operating statement data. First, operating statement data are exceedingly difficult to work with in a research mode. Preparers of the forms sometimes do not often follow instructions, for example entering information on the wrong line or improvising additional line items and entries. The forms sometimes contain arithmetic errors, missing values, and illegible entries. Moreover, there is frequent confusion in the reporting of positive and negative values in the operating reserve section of the statement. Coopers and Lybrand experienced all of these problems in their 1981 study of administrative fees. Furthermore, operating statement data demonstrate considerable volatility in PHA expenses, particularly in the ability to generate an operating surplus from year to year. The causes of this volatility are unclear. Several years worth of data need to be examined for any individual PHA, and probably also for groups of PHAs, to understand the relationship between fees and expenses. An additional disadvantage of operating statement data is that the data do not always represent actual costs incurred in carrying out Section 8 activities. Instead, they might represent expenses shared with other PHA activities. HUD instructs PHAs to prorate shared costs accurately, but there are no fixed guidelines on the methods of proration to be used, and PHAs do not document the method they actually use. The importance of proration of expenses is discussed below. Finally, as operational documents designed for non-research purposes, operating statement data lack the measures for success rates and turnover rates used in building up estimated intake and ongoing expenses in the Abt study, and provide no evidence on the true costs of servicing new intake slots. Some intake costs are reported under the category of "preliminary" expenses, but most are shown under "ongoing" administrative expense. Essentially, surpluses and deficits from new intake activities are submerged in the accounts of ongoing activities. Abt's Administrative Costs study and HUD's analysis of operating statement data differ in more than just method -- they sometimes differ markedly in their bottom line. For example, of 12 large urban PHAs for which estimates of FY 1986 expenses and fees can be compared, 5 had dissimilar net balances. In all five cases, Abt estimated much lower expenses and much higher net surpluses than those reported by the PHAs on their operating statement for the same year. This systematic difference does not in any way invalidate the methods and results used in the Administrative Costs study. Seven of the 12 PHAs with a range of low to high surpluses showed similar expenses and net balances in the two sets of data. The 5 PHAs with dissimilar results might have been cross-subsidizing housing programs other than the Section 8 program. All five of these PHAs in 1986 serviced more public housing units than certificate units, and four agencies serviced at least twice as many public housing units as certificate units. In contrast, of the seven PHAs with similar research and operational results, four serviced fewer public housing units than certificate units. To the extent that PHA operating statements mix costs from different programs, they might portray Section 8 expenses and balances less reliably than the research data for comparable PHAs. However, if agencies use Section 8 fees to cross-subsidize other housing programs, the operating statement data provide a useful caution to the finding of substantial over-reimbursement for large urban PHAs. An additional caution, applicable to both the operational- and research-based methods, is that they take as a given the quality of housing services actually provided by PHAs. They make no judgement about the adequacy of these services or of the impact of lower or higher fees on these services. The Problem of Proration In order to update the findings from the late 1980s, the Department began a new data collection effort from another sample of PHAs. HUD instructed field offices to submit the latest two years of PHA operating statements for the rental certificate and voucher programs. The sample included all PHAs for which data had been submitted and verified in 1987, any PHA that had crossed the large urban threshold (1000+ units) since 1987, plus a special 10 percent sample of very small agencies (under 200 units), because that group had been under-represented in the earlier data collection. In all, HUD collected and verified data for 535 PHAs. The Department assembled a total of 1,840 operating statements, including 1,012 forms for the certificate program and 828 for the voucher program. Unfortunately, a problem involving methods used by PHAs to prorate costs between the various parts of local certificate programs has made the certificate program data unusable for purposes of this analysis. Prior to FY 1990, PHAs submitted to HUD each year a consolidated operating statement for each rental certificate program and one for each rental voucher program. In order to comply the 1987 Housing Act provision regarding separate contracts for individual funding increments, the Department made accounting and automated system changes that necessitated separate PHA operating statements. In 1990, HUD began to require PHAs to submit separate operating statements for any contract renewals and an "ongoing" statement for the remainder of the PHA program. For rental vouchers, most PHAs submitted multiple operating statements to HUD for fiscal years 1990 and FY 1991. As a result of changes in HUD accounting and automated systems, PHAs resumed sending HUD consolidated statements in July 1992. For the certificate program, however, HUD accounting and automated system changes were not completed until 1994, and PHAs have continued to submit multiple statements. In May 1994 the Department issued a notice allowing for consolidated statements in the certificate program. Given the proliferation of operating statements resulting from these changes, and the need to submit this report to Congress on a timely basis, HUD's data collection included the consolidated statements for Rental Voucher programs, but only the largest increment of units (generally, the "ongoing" statement) for certificate programs. Originally, the Department assumed that most PHAs would prorate costs among its increments on a per- unit basis, and that the largest increment would accurately reflect PHA-wide costs. This assumption has turned out not to be true. With the funding of renewal increments, the operating reserve for that increment begins at zero, even though the PHA may have hundreds of thousands of dollars of accumulated reserves in the remaining parts its rental certificate or voucher program. When a PHA begins to administer assistance under this new increment, there is a natural tendency to avoid over-spending because this would cause the agency to report a negative operating reserve for that part of its program. Similarly, a PHA with negative reserves on the new increment would not want to go too far "into the red." Under such circumstances, a PHA with adequate reserves in its ongoing increment might prorate any excess of expenses over earned fees to the ongoing increment, drawing from reserves there. A manual comparison of operating statements collected with those submitted by PHAs for renewals strongly suggests that this type of proration has in fact occurred during the time period included in this data collection. HUD believes that rental vouchers, which have had consolidated operating statements since July 1992, are reasonably representative of PHA operating conditions during the recent period. The analysis of operating statements for the latest two fiscal years addresses the adequacy of fees and compares results with the experience of the late 1980s solely on the basis of rental voucher programs administered by PHAs. As a measure of total tenant based activity, results based on vouchers might slightly overstate expenses and understate the fee surplus. For any units allocated prior to 1989, voucher reimbursements are made at a lower percentage of the FMR (6.5 percent) than for certificates (7.65 percent). The blended fee reimbursement (unit-weighted) is estimated at 7.28 percent for vouchers, 7.71 percent for certificates, and 7.62 for the two programs combined.