The budget proposes total expenditures of $1.5 billion ($406 million General Fund, $518 million county funds, and $590 million federal funds) for the AFDC-FC program in 1998-99. This is an increase of 9.4 percent (7 percent General Fund) from estimated expenditures in the current year, primarily due to caseload growth.

Budget Underestimates Proportion Of Cases Eligible for Federal Funds

Under the AFDC-FC Program, the state receives matching federal funds (51 percent of total expenditures) for those cases meeting federal eligibility criteria. Non-federal costs are shared 40 percent General Fund and 60 percent county funds.

The budget estimates that 83 percent of children in FFH placements and 77 percent of children in GH placements will be federally eligible in both the current and budget years. Our analysis indicates that this understates the number of children who will be eligible for federal funds for two reasons:

• Proportion of Federally Eligible Cases Currently Higher Than Budget Estimate. Figures 28 and 29 show that, at last count (October 1997), the proportion of federally eligible cases was higher than the proportion assumed in the budget for both FFH and GH cases.
• Proportion of Federally Eligible Cases Is Increasing. Figures 28 and 29 also indicate that the proportion of federally eligible FFH and GH cases has been growing over the last 18 months. In fact, the proportion of federally eligible cases has been growing over the last several years. According to the department, this trend primarily reflects improvements in county eligibility determination procedures.

Figures 28 and 29 also illustrate our approach to estimating federal eligibility in the current and budget years. First, we assumed that the rate of increase in the proportion of federally eligible cases observed in the latest six months of data would continue for the remainder of the current year. Then we assumed that the federally eligible proportion would remain fixed in the budget year at the level projected to be achieved at the end of the current year. Our approach assumes continued improvements in county eligibility determination procedures in the near term, but acknowledges that there is a limit to future growth in the proportion of cases that are federally eligible.

Based on this approach, we estimate that 80 percent of FFH cases and 84 percent of GH cases will be federally eligible in the current year, and that 81 percent of FFH cases and 85 percent of GH cases will be federally eligible in the budget year. Consequently, we estimate that federal expenditures for the Foster Care program are understated by $10.6 million in the current year and $19.7 million in the budget year, and combined state and county expenditures are therefore overstated by the same amount. Accordingly, we recommend that the budget be adjusted to reflect our estimates, resulting in General Fund savings of $4.3 million in 1997-98 and $7.9 million in 1998-99.

In California, the child support enforcement program is administered by county district attorneys under the supervision of the Department of Social Services (DSS). The federal government picks up two-thirds of county administrative expenditures, and makes incentive payments to states designed to encourage them to collect child support. California passes the federal incentive payments to the counties along with additional state incentive payments. These payments are used to support the county costs of the program.

Child Support Incentive Payment Systems

The Current State Incentive System

Incentive Payment = (Incentive Rate) x (Collections)

The same "flat" incentive rate of 13.6 percent is applied to the total collections for each county. In other words, all collections--TANF and non-TANF--are weighted equally in calculating the incentive payments to a county, and each county's incentive rate is independent of its score on any performance measure. The current system does reward performance in that the incentive payment increases directly with the amount of collections. We note that the budget proposes to continue the "flat rate" methodology in 1998-99.

In response to the provisions of the federal welfare reform act--the Personal Responsibility and Work Opportunity Reconciliation Act of 1996--the Secretary of Health and Human Services recently proposed a new federal performance-based incentive system for child support enforcement. A bill currently in the House--H.R. 3130--contains the basic structure of the Secretary's proposal.

Mirroring the Proposed Federal System

The New Collections Base. Under the federal proposal, the new "collections base" would be the sum of collections on behalf of families who have never received TANF, plus twice the sum of collections on behalf of current and former TANF recipients. This change would give more weight to TANF and former-TANF collections than the current state system, in which all collections are weighted equally.

The New Performance Measures. The new incentive rate would be a function of five performance measures:

• Paternity Establishment. A county may use either (1) the ratio of the number of children in the child support program caseload for whom paternity was established in the year to the total number of children in the program caseload who were born out of wedlock or (2) the ratio of the total countywide number of children born out of wedlock for whom paternity was established in the year to the total number of children born out of wedlock in the prior year.



• Support Order Establishment. The percentage of cases in which there is a support order.
• Current Support Collections. The percentage of total current support owed that is collected.
• Arrearage Collections. The percentage of cases with arrearages in which past-due support is collected.
• Cost-Effectiveness. Total collections divided by total administrative expenditures.

How the Performance Measure Scores Would Determine the New Incentive Rate. The incentive rate earned on each performance measure would be determined by a specified schedule that depends both on the county's level of performance and its rate of improvement. Figure 30 depicts, for illustration purposes, the incentive rate schedule for the support order establishment measure. For example, if at least 80 percent of a county's cases have support orders, then the county would earn the maximum incentive rate for the support order establishment criterion. If fewer than 50 percent of the county's cases have orders, then the county would generally earn no incentive rate for that criterion. However, if the county had improved its support order establishment performance by at least 5 percentage points over the previous fiscal year, then it would earn half of the maximum incentive rate for that criterion.

Figure 30
Incentive Rate Schedule for Support Order Establishment Proposed Federal Incentive System
Performance Level	Percent of Maximum Incentive
80% and above	100%
70% to 79%	80% to 98% (increases by 2% increments)
50% to 69%	60% to 79% (increases by 1% increments)
49% and below	50% if performance level increased by 5% over previous year, otherwise 0%

The maximum incentive rates for paternity establishment, support order establishment, and current support collections would be one-third higher than the maximum rate for arrearage collections and cost-effectiveness. The incentive rates earned for each of the five performance measures would be added to produce a total incentive rate. The total incentive rate would then be applied to the collections base to determine the incentive payment.

Evaluation of the Proposed Federal System

Besides cost-effectiveness, we found no statistically significant relationship between the proposed performance variables and collections. We also found no relationship between collections and demographic variables (for example, unemployment and per capita income) which, according to some program administrators, might have had an effect on the ability to collect child support.

While we recognize that the proposed performance measures represent important components of the enforcement process, we also note that they are only part of a network of elements in that process. The issue is whether collections will be enhanced by giving program administrators fiscal incentives to place greater weight on particular components of the process than they would in the absence of these incentives or, alternatively, whether the administrators should be left to make their resource allocation decisions without bias toward particular program elements. Our findings suggest that the latter course may be wiser.

Proposed System Would Not Resolve "Case-Closing" Problem. The proposed federal system does not resolve a problem with existing federal regulations, which allow counties to close "old" cases (those in which collections have not been made in three years). A county's performance score on the support order establishment, current support collections, and arrearage collections measures, and the first of the two paternity establishment measures would increase if the county closed difficult cases. Because counties have closed cases at different rates (some close all "old" cases, others keep these cases open), comparisons of performance among counties based on the proposed performance measures may be distorted.

For purposes of measuring county performance, we believe that these cases should not be closed, in order to derive an accurate picture of the program. As long as the federal government permits case closure on this basis, however, it might make sense to follow this practice only for federal reporting purposes in order to compete with the other states for federal incentive funds (until the federal administration addresses the problem).

Proposed System Would Create an Indirect Incentive to Recruit "Never-on-TANF" Cases. As indicated earlier, the proposed federal system would reduce the weight assigned to never-on-TANF collections in the collections base. Nevertheless, by making incentive rates a function of performance on never-on-TANF cases, it would create an additional indirect incentive to "recruit" such cases into the county program. (As noted above, non-TANF parents have the option of using either the county district attorney or a private attorney.)

Some counties have begun to recruit these cases--which tend to have relatively large orders that are easier to enforce--by setting up programs that immediately refer all court orders directly to the county program. Through such programs, counties could increase their performance scores and earn a higher incentive rate on all collections. Consequently, a county's fiscal reward for pursuing never-on-TANF cases would include both the additional incentive payments earned on the increase in the collections base and the additional payments earned from the increase in the incentive rate applied to all collections.

The potential problems are twofold. First, recruiting never-on-TANF cases may divert county resources from the enforcement of TANF cases, where the custodial parent does not have the option of using a private attorney. Second, this diversion of resources may result in lower TANF collections, which partially offset the government costs of TANF grant expenditures. In light of this, the Legislature may want to consider an incentive system that contains greater rewards for TANF collections.

Summary of Findings on Federal Proposal. In summary, we find:

• The performance measures do not, with the exception of cost-effectiveness, demonstrate a statistically significant relationship with the principal variable that reflects the program's objectives--collections.
• The performance measures do not resolve the "case-closing" problem, which gives a distorted picture of program performance and makes county comparisons difficult.
• Although the proposed system reduces the relative weight applied to never-on-TANF collections in the collections base, it creates an additional indirect--and potentially strong--incentive to recruit such cases into the program. This could result in the diversion of resources away from the enforcement of TANF cases.

An Alternative Incentive System

The Incentive Rate. Under our proposed alternative, the incentive rate would be a function of two performance measures:

• Cost-Effectiveness. Total TANF collections divided by total administrative expenditures on TANF cases.
• Administrative Effort. Total administrative expenditures on TANF cases divided by average TANF/California Work Opportunity and Responsibility to Kids (CalWORKs) Family Group caseload.

Each county's performance score would be the product of cost-effectiveness and administrative effort:

Performance Score = (Cost-Effectiveness) x (Administrative Effort)

A county's performance score would determine the incentive rate that would be applied to its collections base when determining the amount of the incentive payment.

The Collections Base. The collections base would be identical to the base in the proposed federal system: the sum of collections on behalf of families who have never received TANF, plus twice the sum of collections on behalf of current and former TANF recipients. Note that although each county's incentive rate would be a function only of variables related to TANF cases, incentive payments would continue to be made on non-TANF collections as well because they would contribute to the collections base.

Evaluation of the Legislative Analyst's Office Alternative Incentive System

Avoids "Case-Closing" Problem. Because counties have taken different approaches toward the practice of closing child support cases, using this caseload to calculate administrative effort could create a distorted picture of performance. Until a uniform measure of the program caseload is required and reflected in the program reporting systems, we propose measuring performance using the county average TANF/CalWORKs (Family Group) caseload as the case measure. By using a measure of caseload that is outside of the control of the county child support program, the "case-closing" problem would be avoided. This measure, moreover, should have a high correlation to the child support TANF caseload because all CalWORKs (formerly Aid to Families with Dependent Children [AFDC]) cases must be referred to the county for child support enforcement.

Eliminates Indirect Incentive to Recruit Never-on-TANF Cases. Because the incentive rate is determined only by enforcement activities on behalf of families that are currently receiving TANF, the LAO alternative would remove the indirect incentive to recruit never-on-TANF cases. Counties would continue to earn incentive payments on the never-on-TANF collections, but these collections would not affect performance scores and incentive rates.

Net Fiscal Effect. We note that if proposals being considered at the federal level are adopted, our alternative may not maximize the federal incentive payments earned by the state in the short run. This is because our proposal bases incentive payments on variables directly related to collections rather than on the federal performance measures. In other words, there may be a tradeoff between maximizing federal incentive payments and maximizing program effectiveness, at least in the short run. In the long run, however, we believe that our alternative is more likely to increase collections, which would lead to higher federal incentive payments. Moreover, because its performance measures are based on TANF collections, the LAO alternative would give counties a greater incentive to increase these collections and, therefore, would encourage recoupment of the public costs of TANF/CalWORKs grants. Finally, by eliminating the indirect incentive to recruit never-on-TANF cases into the program, the LAO alternative would reduce the risk that program resources will be diverted from TANF cases and toward cases that could be enforced through the private sector.

Impact of the Proposals on County Incentive Payments

Figure 31 (see page 140) compares county performance in 1995-96 under (1) a flat incentive rate (similar to the methodology in current state law and the budget proposal for 1998-99), (2) the proposed federal incentive system, and (3) the LAO alternative. The figure reveals significant differences in incentive payments under the proposed federal system and the LAO alternative for some counties.

Figure 32 (see page 141) illustrates the reasons for these differences by focusing on the performance of Orange and Fresno Counties. (Orange County's performance on TANF collections has improved substantially since 1995-96, so the figure is used for illustration purposes and does not reflect the county's current performance.) Orange County scored higher on the federal performance measures, and would have earned a higher incentive rate than Fresno County under the proposed federal system. However, Fresno would have earned a higher incentive rate under the LAO alternative because it had better cost-effectiveness and higher administrative effort.

The proposed federal system would have rewarded a county like Orange that scores highly on the federal performance measures. How-

Figure 31
Comparison of Three Incentive Payment Systems Hypothetical Example Based on 1995-96 Data Twenty Largest Countiesa
(In Millions)
County	Flat Rate (Current Law)	Federal Proposal	LAO Alternative
Los Angeles	$24.2	$8.7	$16.9
Orange	8.2	10.1	8.3
Alameda	7.4	9.8	7.7
San Bernardino	6.8	5.6	5.6
Santa Clara	6.8	8.1	7.3
Fresno	6.7	7.3	8.0
San Diego	6.4	8.3	5.5
Sacramento	6.4	5.7	5.5
Riverside	6.0	5.3	5.0
Kern	4.3	5.7	4.5
Ventura	4.3	4.8	6.1
Contra Costa	3.8	4.3	3.8
San Joaquin	3.5	4.0	3.4
Stanislaus	3.3	4.7	3.9
Tulare	3.0	3.5	3.5
San Francisco	3.0	3.9	3.3
San Mateo	2.4	2.6	2.6
Sonoma	2.3	2.9	3.0
Santa Barbara	2.0	2.6	2.4
Monterey	2.0	2.6	2.3
Total (58 counties)	$136.3	$136.3	$136.3
aAs measured by total child support collections.

ever, an examination of program data indicates that Orange was an average performer in 1995-96 on measures related to collections, such as the AFDC (or CalWORKs) recoupment rate. On the other hand, the LAO alternative would have rewarded a county like Fresno that manages to combine relatively high levels of administrative effort and cost-effectiveness, which reflects the ability to collect a substantial amount of child support in an efficient manner. To phrase it another way, counties that have only a moderate level of collections but do well on the cost-effectiveness measure by holding costs down to a very low level will not fare well on the administrative effort measure; and counties that have high levels of administrative effort but whose collections do not keep pace with this effort will not do well on the cost-effectiveness measure.

Figure 32
Federal Incentive System and the LAO Alternative Effect on Incentive Rates Hypothetical Example Based on 1995-96 Data
	Orange	Fresno
AFDC recoupment ratea	13.8%	17.3%
Administrative effort (AFDC)	$360/case	$451/case
Cost-effectiveness ratiob (AFDC)	2.61	2.74
Proposed federal incentive rate	8.4%	7.2%
LAO alternative incentive rate	6.9%	7.9%
aChild support collections on AFDC cases divided by AFDC grant expenditures.
bCollections per $1 of administrative expenditures.

Finally, we conclude that in addition to being superior to the proposed federal system, our proposal is preferable to an incentive system based on a single flat rate applied to collections--as provided by existing law for the current and budget years only. This is because under our proposal the counties will have a specific incentive to increase their administrative effort, and to apply this effort so as to increase collections in a cost-effective manner.

Legislative Analyst's Suggested "Performance Enhancement Process"

• Identify those counties that are (1) performing poorly and (2) showing a relatively low level of improvement in performance over the prior year. We define this group as counties that (1) rank in the bottom quartile on the LAO performance score and (2) are below the median in improvement in that score over the prior year.
• Among the counties that are not performing well and not showing adequate improvement in performance there are two strategies. First, for counties that are making a low level of administrative effort, the department and the county would be required to establish a three-year expenditure plan designed to reach targeted levels of administrative effort. Second, for counties with relatively high levels of administrative effort but a relatively low level of collections, the Department of Social Services would be required to conduct a program review and provide technical assistance. For counties in the latter category (which are characterized as having low cost-effectiveness ratios), the program review would attempt to discover the causes for the county's low level of collections, examining management practices such as the use of automation and the allocation of resources. We note that the intent is not to sanction these counties but to assist them in improving their performance.

Conclusion

Accordingly, we recommend that legislation be enacted establishing a child support incentive payment system in which county incentive payments are a function of county administrative effort and cost-effectiveness. In addition, we recommend that legislation be enacted establishing a performance enhancement process for counties that rank low in performance.

Budgeting Issues

Budget Overestimates Child Support Incentive Payments

The Governor's budget assumes that child support collections on behalf of non-CalWORKs families will increase by 51 percent in 1997-98 and 29 percent in 1998-99. Based on these estimates of collections, the budget projects the amount of incentive payments which will be made to the counties in the current and budget years.

We surveyed the child support enforcement programs in 13 of the largest counties in the state and found that non-CalWORKs collections in the first six months of 1997-98 have increased by 18 percent over the same period in 1996-97. We anticipate that non-CalWORKs collections will continue to increase, due to the effect of certain policy changes adopted in 1997. However, given the actual experience to date, it seems unlikely that these collections will grow enough to register a 51 percent increase between the current and past years.

While it is difficult to predict the impact of the 1997 policy changes, we believe it would be reasonable to assume an increase in the growth rate to 29 percent for the second half of 1997-98 and all of the budget year--the same rate of growth assumed in the budget for 1998-99. Based on these assumptions, we have projected non-CalWORKs child support collections for the current and budget years. Figure 33 (see page 144) compares our estimates with the Governor's budget, as well as showing actual collections in 1994-95, 1995-96, and 1996-97.

Based on our assumptions, we estimate that the General Fund expenditures for non-CalWORKs child support incentive payments are overstated in the budget by $20.3 million in the current year and $26.3 million in 1998-99. Accordingly, we recommend that the budget be reduced to reflect these estimates.

Budget Overestimates Effect of Arrearage Distribution Changes

Child support collections made for custodial parents receiving CalWORKs grants are used to offset the public costs of the grants (except for the first $50 per month which goes to the parent). In addition, the custodial parent currently must permanentlyassign to the state the rights to "pre-assistance" arrearages--that is, collections on arrearages that accrued prior to a family going on aid. Permanently assigned pre-assistance arrearages are used to offset the family's CalWORKs grants, even if the arrearages are collected after the family goes off aid.

Chapter 270, Statutes of 1997 (AB 1542, Ducheny, Ashburn, Thompson, and Maddy), in response to provisions of the federal welfare reform act, requires changes in the distribution of pre-assistance arrearages beginning October 1, 1998. Under Chapter 270, a family that goes on aid after October 1, 1998 must temporarily assign the rights to pre-assistance arrearages to the state--that is, until the family goes off aid.

The budget incorrectly assumes that all pre-assistance arrearages will be paid directly to the family (instead of assigned to the state), beginning October 1, 1998. Under Chapter 270, however, pre-assistance arrearages collected while a family is still on aid will continue to be assigned to the state. In addition, pre-assistance arrearages collected for families who went on aid prior to October 1, 1998 remain permanently assigned to the state, even if the arrearages are collected after the family goes off aid.

Consequently, the budget overestimates the amount of pre-assistance arrearage collections that will be paid directly to families, and underestimates the amount that will offset the cost of CalWORKs grants. We estimate that expenditures for the arrearage distribution changes are overstated in the budget by $61.5 million ($26.4 million General Fund). Accordingly, we recommend that the budget be reduced to reflect these estimates, for a General Fund savings of $26.4 million

California Food Assistance Program

The budget proposes an appropriation of $24.3 million from the General Fund for the cost of coupon purchases and program administration. This is a decrease of $16.8 million from estimated expenditures in 1997-98, mostly attributable to noncitizens attaining citizenship. We note that the President, in his budget for federal fiscal year 1999 (October 1998 through September 1999), proposes to restore federal food stamp benefits to certain noncitizens, including the groups covered in California's state-only program.

In December 1997, there were 322,756 aged, 21,360 blind, and 668,513 disabled SSI/SSP recipients.

Caseload Growth Is Overestimated

The Governor's budget assumes that the SSI/SSP caseload will increase by 1.4 percent in 1997-98 and 3.1 percent in 1998-99. The projected growth rate of 3.1 percent for the budget year includes 2.1 percent for basic caseload growth, and 1 percent for the net impact of other factors such as the 1997 federal law change that made legal noncitizens residing in the U.S. prior to August 22, 1996, but not yet receiving SSI/SSP, eligible for SSI/SSP if they become disabled.

Our review indicates that during the first six months of 1997-98, the caseload actually declinedby about 5,200 cases, or 0.5 percent, and was 1.7 percent below the administration's current estimate for that period. During this period of caseload decline, however, there was an increase in the number of applicants seeking disability evaluations for purposes of qualifying for SSI/SSP. Despite the recent caseload decline, we believe that in the long run growth in the aged portion of the caseload will mirror statewide population growth for aged individuals, and that growth in the disabled portion of the caseload will ultimately reflect the addition of disabled noncitizens who were not yet receiving aid as of August 1996.

Accordingly, we have adjusted the department's forecast to reflect the latest actual caseload data but have followed their projected future trend. After making these adjustments, we project the caseload will decline by 1 percent in 1997-98 and increase by 2.5 percent in 1998-99. Based on these projections, we estimate that the General Fund expenditures for SSI/SSP grants are overstated by $49 million in the current year and $64 million in the budget year. We recommend that the budget be reduced to reflect these estimates.

Budget Proposes to Eliminate State Cost-of-Living Adjustment

Background. Chapter 606, Statutes of 1997 (AB 67, Escutia)--the 1997-98 budget trailer bill for social services--extended the suspension of the state cost-of-living adjustment (COLA) through December 31, 1998, but did not extend the 4.9 percent statewide grant reduction. Accordingly, the 4.9 percent statewide grant reduction ended on October 31, 1997, and the state COLA is scheduled to resume on January 1, 1999.

Calculating the State COLA. The SSP grant adjustments for the state COLA depend on both (1) the California Necessities Index (CNI), which is applied to the combined SSI/SSP grant, and (2) the U.S. Consumer Price Index (CPI), which is used to determine the amount of the federal COLA and is applied only to the SSI component of the grant. (Specifically, the state COLA sets the SSP portion of the grant equal to the difference between the combined SSI/SSP grant as increased by the CNI, less the amount of the federally funded SSI portion.)

Budget Impact of Governor's Proposal. The Governor's budget estimates that the CPI will be 2.6 percent, and that the CNI will be 3.2 percent. Based on these assumptions, restoring the state COLA on January 1, 1999 would result in a six-month General Fund cost of $51.7 million in 1998-99. Based on our review of more recent data, we estimate that the CNI will be 2.84 percent. Using our lower estimate of the CNI, we estimate that restoring the state COLA would result in a six-month General Fund cost of approximately $39 million. We note that the actual cost will depend on both the final CNI and CPI figures.

Impact on Recipients. Figure 34 shows SSI/SSP grants on January 1, 1999 for individuals and couples under both current law and the Governor's proposal. Although the budget proposes permanent elimination of the state COLA, the budget includes the "pass through" of the federalCOLA to recipients, resulting in grant increases of $13 per individual and $19 per couple. After this increase, grants under the Governor's proposal would be 0.8 percent less (for individuals) and 1.2 percent less (for couples) than current law. As a point of reference, we note that the federal poverty guideline in 1997 is $658 per month for an individual and $884 per month for a couple. Thus, under both the Governor's proposal and current law, the grant for an individual would be just above the poverty guideline--specifically, 1 percent above the poverty guideline under the Governor's proposal and 2 percent above the guideline under current law. Grants for couples would be 33 percent above the poverty guideline under the Governor's proposal and 34 percent above the guideline under current law.

Figure 34
SSI/SSP Maximum Monthly Grants Current Lawa and Governor's Proposal
January 1998 and January 1999
		January 1999
				Change from Current Law
Recipient Category	January 1998	Current Law b	Governor's Proposal c	Amount	Percent
Individuals	$650	$669	$663	-$6	-0.8%
Couples	1,156	1,189	1,175	-14	-1.2
aThe grant levels shown in this figure do not reflect the effect of the 4.9 percent grant reduction in low-cost counties (pursuant to current law) because this grant reduction requires relief from the federal maintenance-of-effort requirements. Such relief has not been enacted and the budget assumes no relief in 1998-99.
bIncludes federal SSI COLA of $13 per individual and $19 per couple and application of the state COLA (about $6 for individuals and $14 for couples).
cIncludes federal COLA of $13 per individual and $19 per couple.