MOH provided the Budget Analyst"s Office with two Microsoft Access Databases that contain information on the Downpayment Assistance Loan Program (DALP). The first database contains records for FY 1998-99 and FY 1999-00. The second database contains records for FY 2000-01 and FY 2001-02. The Budget Analyst sampled 44 of a total 215 DALP recipient records, in order to determine the accuracy of the data contained within the databases. MOH uses the information in the databases to help manage the DALP and to produce various reports. In order to complete our test, we checked the accuracy in each data field within the databases against the corresponding 44 DALP paper files at the MOH offices.

The primary source document used for comparison from the manual files was the DALP loan application. The DALP loan application has been altered over the course of the Program to add additional data elements, such as social security number, census tract, supervisorial district and borrower occupation. Of these data elements, we were able to locate other sources of verifiable forms within the files for the social security number and borrower occupation fields.

The Budget Analyst found errors in the DALP database. However, the Budget Analyst did not find inaccuracies in the manual, master loan files that are used to process DALP loans. None of the errors cited below resulted in an ineligible household receiving a loan or a loan greater than what the recipient was eligible for.

Of the 44 sample records, 20 records, or 45.4% of the total, contained at least one error. In addition, three records, or 6.8% of these records, contained two errors. The errors, which are of varying degrees of significance, are as follows:

· There were four errors each in the "Household Income" and "Interest Rate" fields. (The correct interest rates were used in the loan document computations. It is only in the DALP database that the errors occur).

· There were three errors each in the "Male/Female," and the "% Median" [meaning percent of the median San Francisco income] fields.

· There were two errors in the "Social Security Number" field.

· There was one error each in the "Address," "% Down," "First Loan," "Down Payment," "DALP Loan," "Purchase Price," and "Ethnicity," fields.

As previously stated, the errors contained in the DALP database are of varying degrees of significance. Some involved simple transpositions, such as recording "Household Income" as $34,446 instead of $34,466 and recording "First Loan" as $167,960 instead of $167,930. Others, such as the examples that follow, are of greater significance:

· Recording "Household Income" as $14,635 instead of $25,704, and as $37,563 instead of $55,956.

· Three instances of recording "% Median" as 1, instead of 59, 71, and 91, respectively.

The errors in "% Median" are particularly significant, because they cause the reported income characteristics of the Program"s beneficiaries to be misrepresented. For example, for the three instances of incorrect recording cited above, the value of 1 in the "% Median" field instead of the correct values of 59, 71, and 91, would incorrectly lower the overall value of the reported average percentage of median income for all DALP recipients.

As previously stated, we sampled 44 of 215 DALP records. The sampling method consisted of selecting every fifth record of the population. Although we cannot statistically extrapolate an error rate to the population of DALP records, such precision is not necessary in this case: the sample size is large enough to demonstrate that the population of DALP records contains errors, and that those errors can result in reports that are inaccurate.

The DALP database is used for recording information on all of the significant characteristics of a DALP loan, including the recipient"s income, the purchase price and location of the residence, and the gender and ethnicity of the recipient. The information contained in the DALP database is used by MOH for analyzing Program outcomes and also for compiling reports, such as the Annual Report to the Board of Supervisors. Therefore, it is incumbent upon MOH to insure that the DALP database contains accurate data.

Accordingly, the Budget Analyst recommends that MOH perform a 100 percent review and correction of the data contained in its database, and that MOH develop procedures for entering new data that include safeguards against faulty recording.

ANSWERS TO SPECIFIC QUESTIONS

As stated in the Introduction to this report, Motion MO2-48 included eight specific questions concerning the effectiveness and efficiency of the Affordable Housing and Home Ownership Bond Program. In this report section, we recite the questions and provide answers.

Question #1: How many new net units, as opposed to rehabilitated units, of housing have been or will be created?

Answer: Based on our review of the loan or grant agreements for each individual development, a review of building permit applications, and site visits, we have tabulated the number of new net units in the category indicated as shown in Table III on the following page.

As shown in Table III, a total of 861 newly constructed units have either been completed, are currently in construction, or are in the planning stage with loans or grants already extended. A total $25,405,207in loans or grants has already been made to the 10 listed development projects that are completed. An additional $20,045,933 in new commitments or pending commitments⁵ has been made to five of the ten projects shown in Table III for a total of $45,451,140 ($25,405,207 plus $20,045,933).

In addition, MOH has made commitments or pending commitments in the amount of $12,877,814 to three additional new projects shown in Table IV, which are planned to provide a total of 481 units. Thus, the amount that has been loaned, granted, committed, or is pending commitment for new developments is $58,328,954 ($45,451,140 plus $12,877,814). Therefore, our estimate for the number of newly constructed units to be supplied with Proposition A funding support is 1,342 units (861 units plus 481 units).

When the sum of $58,238,954 for newly constructed units is added to funds loaned, granted, committed, or pending commitment to units and beds acquired and rehabilitated, and to newly constructed beds in group housing, the approximately $85 million allocated for new and rehabilitated rental housing developments is exhausted.

The Budget Analyst obtained the information on loans or grants for the cited number of new units completed, in construction, or planned for from loan and grant agreements, and from minutes of "Housing Committee" meetings for commitments made. The information on pending commitments was obtained from internal MOH documents.

As noted previously, other sources of funds are also used for some of the projects listed in the following tables. According to MOH, a major objective in allocating Proposition A funds to development projects is to complement funding provided from sources outside of the City, such as Federal funds from the Department of Housing and Urban Development (HUD), so as to maximize the development of affordable housing.

Table III

Net New Units Funded Wholly or Partially with Proposition A Funds

Table IV

Commitments or Pending Commitments for Additional New Housing
Units Wholly or Partially Funded with Proposition A Funds

In addition to the loans, grants, commitments, or pending commitments for additional new housing units wholly or partially funded with Proposition A funds, Table V on the following page shows loans, grants, commitments, or pending commitments for 470 rehabilitated housing units, including base conversion units on Treasure Island and the Presidio, wholly or partially funded with Proposition A funds in the amount of $17,015,741, or an average of $36,204 per unit in Proposition A funding.

Table V

Net Rehabilitated Units Funded Wholly or Partially with Proposition A Funds

Question 2: For those projects in which zero units have been reported constructed or committed to be constructed, but a number of beds have been constructed or have been committed to be constructed, who is being served by these new or rehabilitated facilities?

Answer: "Number of Beds" is the unit of measurement/performance for developments that house individuals not constituting a family unit in the same dwelling unit. Some of the facilities measured on a "Number of Beds" basis could also be measured on a "Number of Units" basis if all residents constituted a family unit. For example, Table VI on the following page lists Swords to Plowshares, which is a Treasure Island base conversion project, as having a 56-bed capacity. Those 56 beds are located in 24 dwelling units. Because the clients served by Swords to Plowshares are unrelated individuals who share a dwelling unit, and who enroll, complete, or otherwise depart the program offered by Swords to Plowshares on an individual basis, the population is accounted for on a "bed" basis. Were the same dwelling units managed by a sponsor who served families living in the dwelling units, the unit of measurement and performance would be a "Dwelling Unit."

Other development projects in the "Number of Beds" category are constructed for group living programs, where residents share common living facilities, such as kitchens and dining areas.

Table VI shows that projects having a total of 264 beds have either been completed (170 beds) or have a commitment of Program funds (94 beds). Currently, there are no bed facilities under construction.

Table VI shows Proposition A funding that has been or is being provided to each development project measured on a "Number of Beds" basis. In addition to the $3,138,019 already provided to Friendship House for its new 80-bed facility, an additional $3,152,000 in Proposition A funds is pending commitment, which would provide a total of $6,290,019 for Friendship House, and which would increase the total allocated for the 264 beds to $8,969,521.

The information on the "Number of Units" development projects was obtained from the development agreements and from MOH"s project files.

Table VI

Facilities Housing Unrelated Individuals Number of Beds Provided

Question 3: How many firefighters, police officers, and teachers have benefited from the home ownership portion of Proposition A?

Answer: Based on the information contained in the "Borrower" field of the DALP databases, a field in which we found no errors in our sample test of the DALP databases, the numbers of loan recipients employed in the occupations of interest are shown in Table VII, below.

Table VII

Downpayment Assistance Loan Program Occupational Categories

Question 4: In the budgets of the completed projects, what were the original total budgets in comparison to the final budgets: and, in addition to Proposition A funds, what other sources of funding were used?

Answer: The original budget and final actual expenditures for projects that have so far been completed with the use of Proposition A funding are shown in Table VIII on the following page. The Hotel Tax Fund was the funding source under "Final Other City Funding" for the John W. King Senior Center ($2,175,980) and Presentation Senior Community ($3,344,766). Housing Opportunities for Persons with Aids (HOPWA) funds and Community Development Block Grant (CDBG) funds were the funding sources under "Final Other City Funding" for the 8^th Street Community Resource Opportunities ($1,811,657).

With the exception of Presentation Senior Community, which has submitted an independently audited "Owner"s Cost Certification" to MOH and to HUD that reflects the "Final Total Budget" cost shown in Table VIII, the figures shown are based on MOH"s record of payment requests, which track all project costs and sources. MOH reports that the Owner"s Cost Certifications for the remaining development projects are expected in July and August of 2002.

For all completed projects shown in Table VIII, Final Proposition A funding equaled the Initial Proposition A Funding with the exception of the 8^th Street Community Resource Opportunities, where Final Proposition A funding of $7,997,539 exceeded the Initial Proposition A funding of $2,834,610 by $5,162,929 or 182.1 percent, despite the fact that the total project actual cost of $18,068,672 was actually less than the total budget of $18,159,296. The reason for this is that the project was funded in stages, related to the development stage and the needs of the project for funding at that time. In the case of the 8^th Street Community Resource Opportunities Project, Proposition A provided initial funding for planning and design of the project, and later provided funding for construction.

Other funding has been provided by the City"s Hotel Tax Fund, by (a) the Department of Housing and Urban Development, and (b) the Veteran"s Administration.

Table VIII

Original and Final Project Funding All Funding Sources for Completed Projects