FINDINGS AND RECOMMENDATIONS

I. FORECASTING METHODS

The INS did not document and preserve an adequate record of forecasting actions which, in retrospect, could have provided a better understanding of procedures used and their effectiveness. There was no evidence that current or previous forecasting methods had been validated. The INS budget analysts were not trained to develop statistical forecasts. Revenue forecasts should be made using accrual accounting, but some fee account projections were made and analyzed on a cash basis, and others on a modified accrual basis. As a result, responsible staff did not have all the accounting and program information to better support resource allocations, budget reprogrammings and justifications. Data used to develop budgets may be inaccurate, but we cannot determine the reliability of the data until INS has a credible accounting system.

Fee revenue collections comprise approximately one-third of the INS' total budgetary authority. The Department of Justice (DOJ) guidance issued in the User Fee Programs (Supplement to the DOJ Budget Formulation and Execution Calls) prepared by the Controller, Justice Management Division (JMD) in April 1993, states that:

"Fee account budgets shall be prepared annually . . . . Forecasts shall be predicated on anticipated workloads, fee based revenues, and accrued expenditures in accordance with OMB Circular No. A-11, Preparation and Submission of Budget Estimates . . . revenues shall be presented on an accrual basis."

Under an accrual method, revenues are recognized when earned and expenses are recognized when a liability is incurred, without regard to the actual receipt or payment of cash. This provides a better matching of revenue with expenses and enhances the ability to set fees appropriately to recover the full costs of providing a public service.

To calculate the total budgetary resources available for INS use, its Budget Division combines the amount of funds appropriated from Congress with unobligated fund balances from the prior year, and an estimate of revenue for the upcoming year. A variety of methods are used to prepare fee revenue estimates. The specific approaches taken to estimate each fee account's revenue are discussed briefly below:

User Fee. The INS projected User Fee collections based on fees remitted by domestic and foreign carriers in prior periods. Carrier collections were entered into an electronic spreadsheet and the total collected was adjusted upward or downward by an amount estimated to reflect trends in international travel patterns.

Examinations Fee. The INS used a statistical model to project the number of applications and petitions which would be submitted during the budget year. This regression model, which calculates numerical links between multiple variables, was developed using approximately 3 years of historical data. The projected numbers of applications were then multiplied by the amount of fee charged per application, and then reduced by applying an anticipated non-collection rate for some of the application types. When appropriate, estimates were then adjusted to reflect the current trends and global influences. Adjustments were suggested and applied by a working group consisting of a number of internal INS experts.

Legalization Fee. This program was expected to be phased out and collections had been dwindling. To minimize the amount of resources expended in developing estimates, the last available year's collections were reduced by a percentage which was determined after discussions with legal counsel.

Land Border Fee. The INS estimated revenue of $1 million per site per year, without refining estimates by location, activity, traffic patterns, or adjustments to projected site openings. Because projected sites were delayed in opening, revenue was much lower than anticipated each year since program inception. Beginning in FY 1994, INS adjusted revenue estimates due to the revised site opening schedule.

Additional detail on the methodology and accuracy of each of these four methods will be discussed in the section immediately after this overall forecast discussion.

The INS' Budget Division was organized to identify resources needed, develop resource allocation plans and monitor incomes and outflows. General duties were stipulated in the INS Budget Division's mission statement and restructuring plan for each of the different branches. For three Budget Division branches involved in Fees Forecasting, the stated missions were:

The Analysis and Formulation Branch managed the process for the development and presentation of the budget and worked with programs, operations, and policy and planning to develop meaningful measurement criteria.

The Execution Branch was tasked with tracking current year resources and monitoring budgetary information for INS automated systems.

The Fee Policy and Rate Setting Branch was charged with reviewing and developing rate schedules, to ensure revenues are sufficient to cover all costs, and ensuring that procedures are in place for monitoring costs of fee activities.

For an additional discussion on INS' Budget Division see Appendix II.

In the 1994 DOJ Report by the Attorney General on Management Controls, Fee Account Management was designated a High Risk Area/Material Nonconformance. This report stated:

"INS has not put in place the staff, systems, and procedures to ensure that the fees collected are accounted for, properly controlled, and deposited promptly."

On a monthly basis, INS' Accounting Operations and Reports Analysis Branch prepared a Status of Funds Report that summarized for each fee account the bank deposits made, obligations incurred, and the unobligated balance. This report was used to compare actual receipts to estimated collections for the User Fees, and less regularly for the Examinations Fees. In our judgment, the actual number of inspections performed, as reported in INS' Performance Analysis System could also be used regularly to compare the amount of services rendered with the amount of fees collected.

Training in statistical techniques and forecasts was not provided to INS budget analysts. Automated computer depictions and alternative scenario comparisons were not readily available or encouraged. As part of this audit, we explored various options to automate and enhance forecast methodology, such as using different databases, electronic spreadsheet packages, and specialized statistical forecasting software. There has been rapid development of computer tools and software along these lines recently.

We invited INS officials to attend a demonstration of the capabilities of a computer statistical package that we reviewed. We believe that this program could provide a relatively inexpensive automated system to enable INS to create, maintain, and evaluate an historical database. Some INS officials that attended the software demonstration agreed with the need to move forward with the purchase of this or a similar tool. INS could reassign or detail in-house experts from other offices, such as the Statistical Analysis Branch , to develop necessary specialized tools and approaches for its Budget Division. Using such a specialized computer forecasting package has the advantage of demonstrating anticipated effects or impacts under different scenarios, from a legislative change, to an international crisis, or a simple fee adjustment. Some disadvantages of specialized tools would be the training expense and a greater reliance on those with this special training. As an alternative to developing in-house expertise for the User Fee and Land Border Fee from among the available staff, the INS might consider using an outside source to develop revenue projections.

Overall Effectiveness

Annual fee account receipts have been somewhat unpredictable in total. The total annual fee collections ranged from $390 million in FY 1991 to $532 million in FY 1994. However, INS estimated annual User Fee revenue within approximately 10 percent of actual receipts from FY 1992 to FY 1994. Occasionally, when there was a significant disparity between forecast and actual collection activity, the Budget Division was asked by INS management or Congress to describe the estimating technique used. There was no evidence that staff compared alternative forecasting techniques or evaluated the quality of information used or rejected, other than these occasional briefings. The only regular documentation of forecast effectiveness was in INS' budget submissions and reprogramming memoranda.

As mentioned previously, Fee Account Management was designated a High Risk Area/Material Nonconformance in the 1994 DOJ Report by the Attorney General on Management Controls. In this report, INS determined that an appropriate corrective action would be to develop and implement revenue forecasting models and validate those revenue forecasting methodologies for each fee account. Our review disclosed no documented evidence of validation efforts for forecasting methodologies used by the Budget Division.

We believe that INS should document its revenue forecasting procedures and periodically validate their effectiveness through a comparison of results with estimates. Reasons for differences should be determined and documented as part of this validation effort. Staff should perform post-analyses to fine-tune each forecast method, document what actual data were used, and why other sources of information might have been rejected.

INS should continually explore other sources of information which might become available for use and analysis. During interviews with Customs staff, we learned of an automated system at U.S. ports of entry developed by Customs to track inspection activity. Discussions with INS staff indicated that they were aware of the Customs database, however, INS considered its own information to be sufficient.

Fee account activity results from complex international issues requiring specialized knowledge and sophisticated modeling techniques to predict. INS has recently created a working group to develop Examinations Fee forecasts. Working groups combine the knowledge and judgment of budget, program, and statistical personnel among others, to explore circumstances, spark discussions, and highlight events that otherwise might not be considered. The use of the working group is an excellent idea that should be expanded to the other fee accounts. The result of this collaborative technique is the ability to react quickly and effectively. Such an approach would be nearly impossible to translate to a computer simulation or model. Therefore, we suggest that working groups or teams be established for each fee account and used to refine estimates, regardless of what initial method is used.

In our judgment, the Budget Division should develop a consolidated monthly status report for all fee accounts to highlight the status of any new or proposed fees, legislative changes, or a shifting of the Agency's priorities. This monthly status report should be used to periodically brief INS senior management, prepare budget documents, and assist working groups in their forecast assessments. The use of such a briefing document should naturally foster an analysis of trends and reasons for changes, thus enhancing INS' ability to forecast for the future. Significant variances between forecasts and actual collections, including probable causes for such variances, should be discussed and attempts made to improve forecasting techniques.

RECOMMENDATIONS

We recommend that the INS Commissioner:

1. Use an accrual method to prepare all fee revenue forecasts, then document and validate each method, at least annually.

Immigration and Naturalization Service Comments

INS agrees. Currently, INS accrues revenues on a monthly basis and estimates for Immigration User Fees are made on both an accrual and cash basis. Accruals have been based on collections; however, future accruals are to be based upon inspections workload statistics. It should be noted that in order for INS to move to an accrual basis for this account, INS financial systems must be capable of establishing revenue and expense accruals in order for the status of the account to be accurately displayed. Until that time, the account must continue to be managed and controlled for budgetary purposes on both an accrual and cash basis. The INS will formally document each estimating method and will validate each method by evaluation and comparison with actual results.

Status of Recommendation

Resolved, based on the INS determination that User Fee revenue forecasts should be prepared on an accrual basis using inspection workload statistics. To close this recommendation, please provide us with a description of the methodology to be used to determine reasonable accruals for each fee account, along with copies of any implementing memoranda. In addition, please provide a copy of the first annual documentation and validation for each fee account.

2. Explore other statistical analysis methods, such as commercial computer software packages that were developed specifically for forecasting workload activity to project revenue.

Immigration and Naturalization Service Comments

INS agrees. INS has recognized that fee forecasting is an area where there can be continuous improvement. Over the past 3 years, INS has made strides in improving fee forecasting methodology and will continue this effort. It should be noted that regardless of the statistical software package used, it is not always possible to assume constancy of data over time since immigration history is laden with unpredictable political and policy decisions that impact data and estimates. However, INS will consider commercial statistical software packages which are currently available, and any new software as it becomes available, in the effort to enhance revenue forecasting.

Status of Recommendation

Resolved, based on INS' intention to consider commercial statistical software packages in its effort to enhance revenue forecasting. To close this recommendation, please provide us with the results of your consideration of software packages.

3. Expand the use of working groups or teams to each fee account and collect and distribute working group minutes.

Immigration and Naturalization Service Comments

INS agrees. The establishment of the working group for the Examinations Fee Account's revenue forecasting has proven to be a useful tool. Recognizing this, a working group for the User Fee was established and the first meeting was held on January 25, 1996. We will continue to hold these meetings regularly. In addition, a working group will be established for the Land Border Fee Account. Formal working group meeting minutes will document each session. The meeting minutes will be distributed to working group members and appropriate management officials on a timely basis.

Status of Recommendation

Closed, based on INS' establishment of a working group for User Fees.

4. Develop a quarterly report on fee account activity that highlights trends, comparisons, evaluations, and available fund balances.

Immigration and Naturalization Service Comments

INS agrees. A report similar to the one recommended highlighting these issues for fee revenues has been prepared and used in every INS financial review at the end of each quarter. These reports will be prepared and issued on a regular basis, and they will be expanded to include additional information including data on revenue trends.

Status of Recommendation

Resolved, based on the agreement that INS will augment existing quarterly reports to include information on trend comparisons, evaluations, and available fund balances. To close this recommendation, please provide us with a copy of a regular quarterly report issued that includes the expanded information.

USER FEE FORECAST

Because INS used an analysis of actual cash collections and deposits, User Fee activity was evaluated on a cash basis, not by using an accrual method. Carriers are required to submit their fee collections to INS on a quarterly basis. Bank deposits of these collections usually lag well behind actual passenger travel dates; in some cases, fees were not deposited until 3 to 6 months after passengers paid their fees. INS used carrier payment trends to estimate revenue earned. As a result, INS was not complying with DOJ budget requirements to use accrual methods in estimating budget revenue. This could also significantly diminish INS' ability to timely detect changes in international travel patterns, should such a change occur, and thereby reduce their ability to accurately forecast fee revenue.

The Immigration User Fee Account was established in 1986 by the Omnibus Appropriation Bill, P.L. 99-591. Individuals traveling into the United States on commercial aircraft or commercial vessels are charged for inspection services performed at air and sea ports of entry or for preinspection services performed at certain places outside of the United States. Sea passengers arriving in the United States from Canada, Mexico and the Caribbean are exempt from paying the User Fee. Prior to December 1994, the fee was $5 per individual, currently the fee is $6 per individual.

The INS does not directly collect fees from passengers. Air and sea carriers, travel agents, tour wholesalers, or other entities who issue tickets collect the fees from nonexempt passengers, and remit fees to INS through specified banks on a quarterly basis. Because carriers do not always itemize submissions to INS, but may simply provide totals, User Fee revenue may include an unidentified portion of interest that properly should be sent to the Treasury.

The INS estimates of future User Fee collections were calculated by entering prior period deposit information into a spreadsheet to determine a trend line. Trends were adjusted either upward or downward as appropriate for expected variations in international traffic. In addition, the Status of Funds Report prepared by INS' accounting section was also used to evaluate cash deposits from carriers. Beginning in FY 1996, a monthly collection report from carriers was used to supplement available information on User Fee receipts. Carriers who receive more than $50,000 in User Fees in any month, other than one of the four payment months, are required to report the amount received to INS. This monthly reporting requirement was intended to provide more timely information on actual User Fee collections so that INS management can respond more efficiently to unforeseen occurrences. The FY 1995 and FY 1996 receipt estimates were determined using the FY 1994 collections as the base and adding an expected international passenger traffic increase of 6.6 percent per year. This expected increase was determined through an average of industry expectations reported in several commercial and governmental publications. The User Fee forecasting approach has remained substantially unchanged for the last several years.

Forecasting Effectiveness

Undisplayed Graphic
The 2 year revenue forecast has been within ten percent of actual receipts for the last 3 years. The 2 year revenue forecast for FY 1991 was under forecast by $65 million, due primarily to an unanticipated change in the classification of passengers to be charged fees. The 1 year forecast was within six percent for the last 4 years. The following chart depicts actual User Fee revenue collections compared to INS' 1 year and 2 year forecasts.

To compare User Fee collections to other sources of information, we contacted several other federal agencies that are involved in international travel. This included the U.S. Customs Service and the Department of Transportation (DOT). The DOT receives current information directly from the carriers on the number of international flights into the United States and the number of passengers on each flight. DOT information is available on a monthly basis and in a variety of formats. This information includes a database with individual fields encoded in standard format. We requested and received the DOT database of international flights from calendar years 1991 to 1994.

We reviewed the monthly information from the DOT International Form 41 on the arrival of passengers from foreign ports into the United States, from calendar years 1991 to 1994. The Directory of Transportation Data Sources listed this database as the "T-100 International Markets (Data Bank 28IM) Database." The chart below displays the number of passengers arriving at U.S. ports of entry on a monthly basis from FY 1992 to December 1994, the end of our test period. (For this chart, FY 1992 includes June 1991 through May 1992, FY 1993 includes June 1992 through May 1993, etc.) The resulting plot demonstrates that the traffic data has a discernible seasonal variation and a small, but evident, increase each year during our review period.

Undisplayed Graphic
Although a small percentage of international passengers arriving in the United States are exempt from paying the User Fee, we assumed the quarterly and annual effect of these exemptions would be immaterial and did not attempt to adjust for this small amount. We multiplied $5 times the number of arriving passengers per month [ Individuals traveling into the United States were charged $5 prior to December 1994. User Fee was increased to $6 per person after December 1994.] to get an approximation of accrued revenue for each month. We combined these monthly accrual figures into quarterly accruals and assumed that they would not be deposited until 3 months later. We called these quarterly estimates "Anticipated Revenue." The next chart compares quarterly anticipated revenue to actual quarterly User Fee collections.

Undisplayed Graphic Undisplayed Graphic

The two plots vary during most quarters, but the size of the variations appear somewhat similar. To determine the relationship of the total amounts, we then plotted both sets of data using a cumulative approach.

The adjacent chart demonstrates that towards the end of our review period the total anticipated revenue equals the total actual deposits.

The totals in the DOT T-100 database of international travelers would appear to compare favorably with the totals in the carrier collections. Using the DOT T-100 data to predict revenue on a cumulative basis could possibly give INS better timeliness and accuracy on a annual basis, but this type of analysis might be even more valuable on a quarterly basis. Analysts could project the amount of revenue that will be submitted by the carriers one quarter in advance, and this process could provide a better estimate of accrued revenue, particularly for the end of the year financial close-outs.

Although INS has had the most success in forecasting User Fee revenue, there are other methods and sources of information that could be used to improve it even further. We suggest that information, such as the DOT database, could be used to predict revenue more accurately based on our comparison over a 4 year period. Advantages of the DOT database are timeliness, standardized format of data, and the flexibility of use for analysis. It is relatively easy to sort and analyze these numbers by global travel trends, seasonal and annual variations, or U.S. ports of entry. This information would also give INS some assurance that the amount of receipts collected is reasonable based on the total number of arriving passengers. It would also provide a better indication of accrued revenue for the end of the year close-out and the financial statements.

Periodic carrier audits are performed by the INS to determine if collection submissions are reasonably accurate and timely. Customs also performs periodic carrier audits. A Memorandum of Understanding between Customs and INS to share the responsibility for auditing carriers, share relevant information, and to compare carrier fee submissions between the two agencies, was suggested and drafted. We believe that INS should proceed with this agreement and continue to explore other sources of information for comparisons and reassurance.

RECOMMENDATION

We recommend that the INS Commissioner:

5. Use the DOT T-100 monthly data to analyze passenger traffic and accrued revenue on a cumulative basis. The data could be used by the working group to assist in predicting the revenue and as auxiliary evidence to validate User Fee forecast method.

Immigration and Naturalization Service Comments

INS agrees that this data could prove to be a useful tool, but it is not the only tool which should be analyzed and evaluated. The T-100 data does not provide information related to fee exemptions. INS' own Performance Analysis System for inspection data will be used in conjunction with the DOT T-100 data. INS will begin to use the T-100 data cumulatively in analyzing User Fee revenues.

Status of Recommendation

Resolved, based on the agreement that INS will utilize the DOT T-100 data cumulatively in analyzing User Fee revenues. To close this recommendation, please provide us with examples of ways INS is using the T-100 data in its analysis of User Fee revenues.

EXAMINATIONS FEE FORECAST

Examinations Fee forecasts were developed using a computer model that compared and averaged the submission activity of 45 fee-related applications and petitions. A working group analyzed and adjusted the results of the computer forecast, when appropriate. Although working group meeting minutes were usually prepared, they were not always done timely or in enough detail for members to recall, after a period of time, why forecast numbers were changed.

The Examinations Fee or Immigration Examinations Fee Account was established by the DOJ Appropriations Act, 1989, P.L. 100-459. All deposits to the Examinations Fee are available until expended to reimburse the appropriated account for expenses incurred in providing immigration, adjudication, and naturalization services. Applications for immigration, nationality and citizenship benefits, and naturalization are received and adjudicated at service centers, field offices, numerous sub-offices, and some land border ports of entry. Applications were submitted by mail or in person at field offices or service centers. Receipts were collected by INS employees and contract employees and deposited to banks, usually on a daily basis.

First attempts to develop a forecast model for the Examinations Fee included the creation of a spreadsheet to track the amount of application fees collected. The number of applications submitted and other workload information was extracted monthly from the Performance Analysis System. [ An audit of the Performance Analysis System was beyond the scope of our audit, therefore we did not verify the accuracy of the information contained in this system. However, nothing came to our attention which indicated the data were inaccurate. ] A simple graphic plot and analysis was used to estimate workload trends and associated revenue. The Performance Analysis System was also used to evaluate the backlog of applications to assist in managing resources. After an initial evaluation of the spreadsheet forecast, INS confirmed that a more sophisticated approach was necessary.

In April 1994, INS' Statistical Analysis Branch developed a statistical forecast (regression) model to project the number of applications and petitions that would be submitted on an annual basis. FY 1995 was the first full budget year this model was used to project collections for the Examinations Fee. This projection was based on activity in 45 fee-related applications and petitions, five of which currently charge no fee, with two additional lines reserved for future use. When developing the computer forecast model, analysts noted that in some months activity varied dramatically. To lessen the impact of a single month's variation, the Statistical Analysis Branch eliminated extreme variations and one-time events in the pattern by replacing some of these numbers with averages. From this modified database, the model produced an annual activity projection for each of the 45 line items. Estimated annual Examinations Fee revenue was calculated by multiplying fees charged per application, by an anticipated collection rate. These collection rates range from 40 percent to 100 percent and were also based on historical behavior; some application fees are not collected due to exemptions, waivers, returned checks, or other circumstances. Annual revenue estimates are then divided into monthly revenue estimates.

The working group periodically compared actual numbers of applications submitted to INS, as reported in INS' Performance Analysis System, to forecast activity. In FY 1993 and FY 1994, INS received 4.5 million and 4.2 million applications respectively for various immigration services and benefits. Receipts for the Examinations Fee totaled $267.6 million in FY 1994, a 2 percent decrease from the $274 million for FY 1993. This decrease was attributed to decreasing application activity.

Forecast Effectiveness

The best and the worst Examinations Fee forecasts noted during our review period occurred in FY 1991. The following chart depicts the actual Examinations Fee revenue compared to INS' 1 and 2 year forecasts.

Undisplayed Graphic


The 2 year revenue forecast was underestimated by approximately $102 million, due to a substantial change in rates that was unanticipated for the 2 year forecast, but the 1 year forecast was underestimated by only $8 million. INS under forecast revenue from FY 1991 to FY 1993. Then, the 1 year forecast for FY 1993 was over forecast by about 18 percent, and the FY 1994 estimates exceeded actual revenue by an average of about 24 percent. Because Examinations Fee revenue forecasts remained imprecise, INS began using a working group approach in April 1994. The statistical model and the working group process was used to adjust the 1 year Examinations Fee forecast for FY 1995. However, since the FY 1995 actual figures were not available when this report was prepared, it is too soon to comment on the effectiveness of the statistical model and the working group.

The Examinations Fee working group was organized to review, analyze, and refine each line of the computerized forecast output. The members of the working group were selected because of their expertise in current legislative affairs, statistics, demographics, and other key disciplines. The combined knowledge and judgment of budget, program, and statistical personnel among others, was pooled to explore circumstances, spark discussions, and highlight events that otherwise might not have been considered. The result of this collaboration is the ability to react quickly and effectively in a variety of circumstances. We believe the working group process would be nearly impossible to translate to a computer model. The use of the working group was an excellent idea that should be expanded to the other fee accounts. Therefore, we suggest that working groups or teams be established for the other fee accounts and used to refine estimates, regardless of the initial method used. However, working group minutes must be documented timely and in enough detail for working group members to recall why adjustments were made.

Staff should validate this methodology and regularly perform post-analyses to refine the forecast method. Excessive reworking and averaging of data used in producing forecasts should be avoided since this could diminish objectivity and possibly impact reliability. Although Examinations Fee receipts were compared to anticipated receipts on occasion, in our judgment, these comparisons and explanations should be prepared and distributed on a regular basis, perhaps monthly.

Fee account activity and forecast results should be recorded and used to further refine the estimation process. INS staff should be encouraged to continually explore alternative approaches. We suggest that INS management investigate other statistical analysis methods, such as, commercial software packages that are developed specifically for forecasting revenue. One such commercial package that was demonstrated for this audit and presented to INS representatives is available through Government approved sources.

LEGALIZATION FEE FORECAST

Based on the requirement to maintain an appeals unit through calendar year 1996, INS estimated that Legalization Fee collections would only be sufficient to meet minimum funding needs. The Legalization Fee Program was expected to be phased out and fee collections have been dwindling. Several class action suits relating to the Legalization Program are pending final resolution in the courts. Depending on the outcome, the resolutions could generate additional processing, but since the outcome is unknown, no additional resources are being requested. As a result, the Budget Division sought to minimize the amount of resources expended in monitoring this program. There was no documentation of methods used to produce a forecast.

The Immigration Legalization Fee Account was established in November 1986 by the Immigration Reform and Control Act of 1986, P.L. 99-603. The Act authorized the INS to charge fees in connection with the filing of applications for the adjustment of status and to deposit these fees into a separate account. The Act was intended to grant amnesty to certain aliens who entered the United States illegally or lapsed into illegal status prior to January 1, 1982, and who have resided in the United States since then and who are not excludable. These persons could apply for temporary resident status. The law allows those granted temporary resident status to apply for permanent resident status after 18 months upon showing general admissibility and citizenship skills. Applicants submit applications by mail to an INS service center. Receipts were collected and totaled by INS employees and sent to banks on a regular basis.

During 1990, the Legalization Regional Processing Facilities were merged with the INS service centers. INS management completed the majority of applications for permanent resident status during 1991. The 1992 DOJ Appropriations Act provided up to $3 million per year of the unobligated balances remaining in the Legalization account to the Office of Special Counsel for Immigration Related Unfair Employment Practices. All of the 107 legalization offices were closed by the end of 1993, except for the Los Angeles and Manhattan offices. INS estimated that additional filing fee receipts from court mandated actions would cover any additional costs incurred. During FY 1993, approximately 500 to 1,000 of the weekly applications for class membership were accepted as prima facie eligible. The results of a class action suit could cause the reopening of the Legalization Program. The outcome was unpredictable at the time this report was prepared and it is not possible to estimate the number of applications which may be submitted depending on the Court's decision. Resources in the Adjudications and Naturalization Program field offices will be used to process and adjudicate applications of class members to the litigation and all remaining applications for adjustment under the Legalization Program. As of the beginning of FY 1995, there were over 7,000 appeals pending, which consisted of appeals to the denial of applications for temporary residence,termination of temporary residence, and denial of status adjustment to permanent residence.

To estimate future revenue, recent application collection trends are decreased by about 50 to 75 percent, based on expectations from INS legal counsel after assessing recent events, estimates from the adjudication programmer, and the status of legal actions.

Forecast Effectiveness

Since the program extended beyond expectations, actual receipts consistently exceeded forecast revenue during most of our audit period. The following chart compares actual revenue to INS' revenue forecasts from FY 1991 to FY 1994.

Undisplayed Graphic


Because this program is phasing out, we agree with INS' decision not to expend additional resources in developing more sophisticated forecast methods. However, if the class action suit or other movements result in the reopening of the Legalization Program, we encourage INS to adopt techniques and recommendations that were discussed in other sections of this report. We particularly advocate the use of working groups or teams to evaluate relevant factors.

The INS officials also stated that there was some movement toward merging what remains of the Legalization Fee Account into the Examinations Fee Account. We agree that would be appropriate at this time.

LAND BORDER FEE

An appropriate forecast method for the Land Border Fee had not been developed. Revenue estimates were not adjusted to reflect variances in population size, density, or port usage. In addition, available information on decal sales and inspection workload were not always used in preparing revenue forecasts. Additional U.S. land border sites along the northern border, and at least one on the southern border, will be opening. As the Land Border Fee Account grows, the lack of an appropriate forecast method and inadequate fund monitoring practices, could become critical.

The Land Border Inspection Fee Account was established as a pilot program [ This pilot program will expire September 30, 1996, unless extended.] by the DOJ Appropriation Act of 1991, P.L. 101-515. This Act authorized the Attorney General to establish, by regulation, pilot projects under which fees may be charged at one or more land border ports of entry to study the feasibility of charging fees to provide inspection services at land border ports of entry. Collections from this program were restricted to improving or developing traffic facilitation and inspections techniques. Program managers use funds for operating expenses at the test sites and excess funds are used to develop new border control techniques. The fees collected under this Act were deposited into a separate account.

Prior to FY 1996, INS was operating Land Border Fee sites at Blaine, Washington; Point Roberts, Washington; and Detroit, Michigan. At these sites, applicants were charged $25 per vehicle to obtain a decal that permits use of a traffic lane dedicated to commuters and frequent crossers. Vehicles using the commuter lane were defined by INS as low risk frequent crossers. This population represented an average of 23 percent of the total vehicle traffic processed annually at the Blaine, Washington site. Applicants for the decals paid in person or mailed applications to the participating land border site. Receipts were collected, totaled, and sent by INS employees to banks on a regular basis. The INS opened the first dedicated commuter lane on the southern border in the first quarter of FY 1996 at Otay Mesa, California.

The INS also developed a program to facilitate the entry of pre-screened local border residents through selected Northern border ports of entry during periods when the port is closed to general traffic. The Automated Permit Port program was designed to allow entry by means of an automated access system. The program would be utilized at low-volume, limited-hour, low risk locations on the Northern border, using different forms of automated technology (such as voice verification systems, video inspections systems, card access systems, hand geometry, or other forms of biometrics) depending on specific location requirements. INS had planned to open several automated sites in late FY 1995 on the Northern border, and several more sites in FY 1996. Although none were opened in FY 1995, several began opening during FY 1996.

Some program managers in the land border inspection program receive weekly activity reports from the dedicated commuter lane ports that show the number of inspections performed, and the number of decals sold and issued. These weekly activity reports appeared to be timely and accurate. Our comparisons of data contained in these weekly reports and the Performance Analysis System monthly reports indicated that the two sources were consistent. We also compared monthly deposit totals in the Financial Accounting and Control System (FACS) for the Land Border Fee Account to bank deposit records at one dedicated commuter lane site and found significant inconsistences at one site. These differences resulted from delays in entering deposit information into FACS. As a result, management did not always know on a timely basis how much revenue was available in this fund balance. However, INS did not use either the weekly activity report or the Performance Analysis System monthly report on a regular basis to evaluate land border workload activity or forecast revenue.

Forecast Effectiveness

The chart below shows INS estimates over this period. When the pilot program began, INS management anticipated revenues of $1 million per site, per annum. A revenue over forecast of nearly $3.4 million resulted, in FY 1993.

Undisplayed Graphic
Since the land border program began, there have been delays in opening new sites and in FY 1994 estimates were adjusted accordingly.

While the Land Border Fee program is not a major program that generates a substantial amount of fee revenue, the INS intends to significantly expand the number of sites using dedicated commuter lanes and automated ports. With expansion, actual receipts will grow and inexact revenue forecasts will become more serious. Relevant factors should be evaluated when estimating revenue for future ports, such as size of the local population, historical records of the number of weekly border crossings, and the average time for each vehicle to pass through the inspection point. The documentation and analysis of this type of information should lead to more realistic and accurate revenue forecasts.

We found that land border receipt activity was not collected and evaluated regularly by INS. INS staff in both program operations and budget indicated that other offices were responsible for monitoring the collection of fee receipts. Such ambiguities could become significant as the size of the land border account grows. Responsibilities between offices and staff members need to be more clearly defined so that information is available without duplication or omission. The use of a working group or team appears particularly appropriate for refining Land Border Fee revenue estimates.

RECOMMENDATION

We recommend that the INS Commissioner:

6. Direct program managers, budget analysts, and accountants to share information and develop taskings for monitoring activity and identifying key evaluation factors.

Immigration and Naturalization Service Comments

INS agrees. A memorandum is being prepared for the Deputy Commissioner's signature to formalize these working groups, including one for the Land Border Fee Account. These issues are to be included in the scope of the working group's activities.

Status of Recommendation

Resolved, based on INS' issuance of the memorandum discussed above formalizing working groups for each of the Fee Accounts. To close this recommendation, please provide a copy of the final objectives and requirements developed for each working group.

#####