§275.3. Federal monitoring. — Inbound Citations

(2) As specified in § 275.3(c), FNS will validate each State agency's estimated payment error rate by rereviewing the State agency's active case sample and ensuring that its sampling, estimation, and data management procedures are correct.

(i) Once the Federal case reviews have been completed and all differences with the State agency have been identified, FNS shall calculate regressed error rates using the following linear regression equations.

(A) 1′ = y1 + b1 (X1 −x1), where y1′ is the average value of allotments overissued to eligible and ineligible households; y1 is the average value of allotments overissued to eligible and ineligible households in the rereview sample according to the Federal finding, b1 is the estimate of the regression coefficient regressing the Federal findings of allotments overissued to eligible and ineligible households on the corresponding State agency findings, x1 is the average value of allotments overissued to eligible and ineligible households in the rereview sample according to State agency findings, and X1 is the average value of allotments overissued to eligible and ineligible households in the full quality control sample according to State agency's findings. In stratified sample designs Y1, X1, and x1 are weighted averages and b1 is a combined regression coefficient in which stratum weights sum to 1.0 and are proportional to the estimated stratum caseloads subject to review.

(B) 2′ = y2 + b2(X2−x2, where y2′ is the average value of allotments underissued to households included in the active error rate, y2 is the average value of allotments underissued to participating households in the rereview sample according to the Federal finding, b2 is the estimate of the regression coefficient regressing the Federal findings of allotments underissued to participating households on the corresponding State agency findings, x2 is the average value of allotments underissued to participating households in the rereview sample according to State agency findings, and X2 is the average value of allotments underissued to participating households in the full quality control sample according to the State agency's findings. In stratified sample designs y2, X2, and x2 are weighted averages and b2 is a combined regression coefficient in which stratum weights sum to 1.0 and are proportional to the estimated stratum caseloads subject to review.

(C) The regressed error rates are given by r1′ = y1′/u, yielding the regressed overpayment error rate, and r2′ = y2′/u, yielding the regressed underpayment error rate, where u is the average value of allotments issued to participating households in the State agency sample.

(D) After application of the adjustment provisions of paragraph (b)(2)(iii) of this section, the adjusted regressed payment error rate shall be calculated to yield the State agency's payment error rate. The adjusted regressed payment error rate is given by r1″ + r2″.

(ii) If FNS determines that a State agency has sampled incorrectly, estimated improperly, or has deficiencies in its QC data management system, FNS will correct the State agency's payment and negative case error rates based upon a correction to that aspect of the State agency's QC system which is deficient. If FNS cannot accurately correct the State agency's deficiency, FNS will assign the State agency a payment error rate or negative case error rate based upon the best information available. After consultation with the State agency, the assigned payment error rate will then be used in the liability determination. After consultation with the State agency, the assigned negative case error rate will be the official State negative case error rate for any purpose. State agencies shall have the right to appeal assessment of an error rate in this situation in accordance with the procedures of Part 283 of this chapter.

(iii) Should a State agency fail to complete 98 percent of its required sample size, FNS shall adjust the State agency's regressed error rates using the following equations:

(A) 1″ = r1′ + 2(1−C)S1, where r1″ is the adjusted regressed overpayment error rate, r1′ is the regressed overpayment error rate computed from the formula in paragraph (b)(2)(i)(C) of this section, C is the State agency's rate of completion of its required sample size expressed as a decimal value, and S1 is the standard error of the State agency sample overpayment error rate. If a State agency completes all of its required sample size, then r1″ = r1′.

(B) 2″ = r2′ + 2(1−C)S2, where r2″ is the adjusted regressed underpayment error rate, r2′ is the regressed underpayment error rate computed from the formula in paragraph (b)(2)(i)(C) of this section, C is the State agency's rate of completion of its required sample size expressed as a decimal value, and S2 is the standard error of the State agency sample underpayment error rate. If a State agency completes all of its required sample size, then r2″ = r2′.

(155) The provision at § 275.23(e)(5)(iii) is effective and is to be implemented on July 16, 1999. The following provisions are effective on October 1, 1999 and are to be implemented on October 1, 2000, with the start of the Fiscal Year 2001 quality control review period: § 271.2; § 275.3(c)(3)(ii); § 275.10(a); § 275.11(c)(1); § 275.11(e)(2); § 275.11(f)(2); § 275.13(a); § 275.13(b); § 275.13(c)(1); § 275.13(c)(2); § 275.13(f)(2) and § 275.23(c)(4). The remaining provisions of this rule are effective and are to be implemented October 1, 1999, with the start of the Fiscal Year 2000 quality control review period, which begins with the October 1999 sample month.

(i) With the exception of the changes to § 275.3(c)(4) [Arbitration], § 275.23(e)(5) [State agencies' liabilities for payment error-Fiscal Year 1992 and beyond], § 275.23(e)(7) [Good Cause], and § 275.23(e)(9) [timeframes], all quality control changes that are made by Amendment No. 366 shall be implemented July 2, 1997.

(2) In addition, the household shall be determined ineligible if it refuses to cooperate in any subsequent review of its eligibility as a part of a quality control review. If a household is terminated for refusal to cooperate with a quality control reviewer, in accordance with §§ 275.3(c)(5) and 275.12(g)(1)(ii) of this chapter, the household may reapply, but shall not be determined eligible until it cooperates with the quality control reviewer. If a household terminated for refusal to cooperate with a State quality control reviewer reapplies after 125 days from the end of the annual review period, the household shall not be determined ineligible for its refusal to cooperate with a State quality control reviewer during the completed review period, but must provide verification in accordance with paragraph (f)(1)(ix) of this section. If a household terminated for refusal to cooperate with a Federal quality control reviewer reapplies after nine months from the end of the annual review period, the household shall not be determined ineligible for its refusal to cooperate with a Federal quality control reviewer during the completed review period, but must provide verification in accordance with paragraph (f)(1)(ix) of this section. In the event that one or more household members no longer resides with a household terminated for refusal to cooperate, the penalty for refusal to cooperate will attach to household of the person(s) who refused to cooperate. If the State agency is unable to determine which household member(s) refused to cooperate, the State agency shall determine the household to which the penalty shall apply.