Use of Statistical Sampling Procedures To Verify Contractor's Invoices

B-170479, JAN 15, 1976

GAO LIMITATION, PURSUANT TO 31 U.S.C.A. SEC. 82B-1(A), ON USE OF STATISTICAL SAMPLING PROCEDURES TO PREPAYMENT EXAMINATION OF DISBURSEMENT VOUCHERS FOR AMOUNTS NOT EXCEEDING $300 IS APPLICABLE TO ICC PROPOSAL FOR RANDOM SAMPLING TO VERIFY CONTRACTOR'S INVOICES UNDER STENOGRAPHIC REPORTING SERVICES CONTRACT. HOWEVER, IF CONTRACT WERE TO PROVIDE FOR SEPARATE INVOICING OF TRANSCRIPTS INDIVIDUALLY UNDER $300, STATISTICAL SAMPLING OF THOSE TRANSCRIPTS WOULD BE PERMISSIBLE.

STATISTICAL SAMPLING PROCEDURES:

THIS DECISION RESPONDS TO A REQUEST BY THE MANAGING DIRECTOR OF THE INTERSTATE COMMERCE COMMISSION (ICC) FOR ADVICE AS TO THE EFFECT OF OUR LIMITATION ON THE USE OF STATISTICAL SAMPLING PROCEDURES TO THE PREPAYMENT EXAMINATION OF DISBURSEMENT VOUCHERS FOR AMOUNTS NOT IN EXCESS OF $300, ON AN ICC PROPOSAL TO INSTITUTE A RANDOM SAMPLING PROCEDURE TO VERIFY CONTRACTOR'S INVOICES UNDER A STENOGRAPHIC REPORTING SERVICES CONTRACT. THE MANAGING DIRECTOR EXPRESSES CONCERN THAT APPLICATION OF THE $300 LIMITATION COULD HINDER THE EFFECTIVENESS OF THE PROPOSED SAMPLING SYSTEM, SINCE A SIGNIFICANT NUMBER OF THE COMMISSION'S HEARINGS RESULT IN TRANSCRIPTS WITH INDIVIDUAL COSTS EXCEEDING THE MAXIMUM.

THE ICC CONTRACTS WITH A PRIVATE CONCERN FOR STENOGRAPHIC HEARING TRANSCRIPTS BY THE PAGE, AND HAS BEEN COUNTING THE PAGES IN EACH TRANSCRIPT TO DETERMINE THE NUMBER OF "ALLOWABLE" PAGES UNDER THE CONTRACT TERMS. THIS INFORMATION IS FURNISHED THE CONTRACTOR WHO THEN SUBMITS AN INVOICE. THE INVOICE IS CHECKED AGAINST ICC RECORDS AS TO NUMBER OF TRANSCRIPTS AND NUMBER OF PAGES WITHIN EACH TRANSCRIPT BEFORE BEING APPROVED FOR PAYMENT. IT IS PROPOSED THAT THE ICC NO LONGER COUNT PAGES, BUT RATHER HAVE THE CONTRACTOR DO SO AND SUBMIT PAGE COUNTS WITH EACH TRANSCRIPT. THE CONTRACTOR WOULD SUBMIT A MONTHLY INVOICE FOR ALL TRANSCRIPTS PREPARED DURING THE MONTH, USING HIS PAGE COUNTS. THE ICC WOULD THEN, ON A STATISTICAL SAMPLING BASIS, VERIFY THE CONTRACTOR'S SUBMITTED PAGE COUNTS. SAMPLE SIZE WOULD BE DETERMINED ON THE BASIS OF A 99 PERCENT CONFIDENCE FACTOR. THE MANAGING DIRECTOR STATES WITH RESPECT TO THIS PROPOSAL:

"WHILE STATISTICIANS IN THE COMMISSION BELIEVE THAT A SAMPLING SYSTEM CAN BE DESIGNED FOR THIS FUNCTION THAT WILL ACHIEVE THE DESIRED ACCURACY IN PAYMENTS, I AM CONCERNED ABOUT THE APPLICABILITY OF THE $300 SAMPLING LIMITATION. UNDER OUR 1975 CONTRACT, WE ARE PAYING A MINIMUM OF $1.89 AND $2.98 A PAGE FOR HEARINGS HELD IN WASHINGTON, D.C., AND ALL OTHER POINTS, RESPECTIVELY. IF, AS AN EXAMPLE, THE $300 LIMITATION WERE APPLIED TO THE SAMPLE INVOICES ENCLOSED, 7 OF THE 32 WASHINGTON, D.C., HEARINGS AND 45 OF THE 76 FIELD HEARINGS WOULD BE EXCLUDED FROM THE SAMPLING SYSTEM. THE LIMITATION HAS AN EVEN GREATER IMPACT IN THIS APPLICATION BECAUSE FIELD HEARINGS, WHICH ARE MORE LIKELY TO BE EXCLUDED, ACCOUNT FOR ABOUT 75-80 PERCENT OF THE TOTAL VOLUME, AND HIGHER PAGE RATES ARE LIKELY IN 1976. THE LIMITATION IS APPLICABLE, IT COULD HINDER THE EFFECTIVENESS OF A SAMPLING SYSTEM TO SUCH A DEGREE THAT THE PROPOSAL WOULD NOT BE PRACTICAL. ON THE OTHER HAND, INDIVIDUALLY CHECKING 273,000 PAGES AND COUNTING EACH LINE SEEMS UNDULY BURDENSOME AND NOT COST EFFECTIVE."

OUR MEMORANDUM TO HEADS OF DEPARTMENTS AND AGENCIES DATED FEBRUARY 25, 1975, B-153509, CAPTIONED "STATISTICAL SAMPLING PROCEDURES IN THE EXAMINATION OF DISBURSEMENT VOUCHERS," PROVIDES IN RELEVANT PART:

"BASED ON THE AUTHORITY CONTAINED IN TITLE 1 OF PUBLIC LAW 93-604 APPROVED JANUARY 2, 1975, STATISICAL SAMPLING PROCEDURES AUTHORIZED BY PUBLIC LAW 88-521 (31 U.S.C. 82B-1(A)) MAY BE USED IN THE PREPAYMENT EXAMINATION OF DISBURSEMENT VOUCHERS FOR AMOUNTS NOT IN EXCESS OF $300.

"THE HEAD OF EACH DEPARTMENT AND AGENCY WILL SET HIS OWN LIMIT WITHIN THE $300 MAXIMUM. HE WILL DEMONSTRATE BY COST-BENEFIT ANALYSIS THAT ECONOMIES WILL RESULT BY USE OF THE LIMIT HE SELECTS. IN OUR REVIEWS OF YOUR ACCOUNTING SYSTEMS, WE WILL EVALUATE THE ADEQUACY AND EFFECTIVENESS OF YOUR STATISTICAL SAMPLING PROCEDURES."

THE $300 LIMITATION, AS SET FORTH ABOVE, IS BY ITS TERMS CLEARLY APPLICABLE TO THE ICC PROPOSAL. HOWEVER, SECTION 44 OF TITLE 3 OF THE GENERAL ACCOUNTING OFFICE POLICY AND PROCEDURES MANUAL FOR GUIDANCE OF FEDERAL AGENCIES PROVIDES FOR THE USE OF SAMPLING IN THE EXAMINATION OF SINGLE VOUCHERS IN LARGE AMOUNTS REPRESENTING NUMEROUS TRANSACTIONS OF A SIMILAR KIND. ACCORDINGLY, IF ICC'S CONTRACTUAL ARRANGEMENTS WERE TO PROVIDE FOR SEPARATE INVOICING OF THOSE TRANSCRIPTS WHICH ARE INDIVIDUALLY UNDER $300 FROM THOSE WHICH ARE $300 OR MORE, STATISTICAL SAMPLING OF THOSE UNDER $300 WOULD BE AN ACCEPTABLE METHOD OF VERIFYING THE INVOICE COVERING THOSE TRANSCRIPTS.