(dissenting). The majority holds that the Department of Health’s recalibration regulation which reduces Medicaid reimbursement rates by 3.035% is unreasonable and without rational basis, because the Department failed to consider "the actual causes of facilities’ change — if any — in CMI” (see, majority opn, at 167) and because there is no "record evidence to support [the Department’s] supposition that the CMI increase is solely attributable to 'paper optimization’ ” (see, id., at 167).1
In reaching this conclusion, the majority not only disregards the ample proof in the record that the change in average CMI resulted from "paper optimization”, it reaches well beyond the narrow limits of appropriate judicial review of the type of administrative action before us. Moreover, it ignores the very standard of review that it purports to apply. Because the *170majority’s holding does not withstand analysis and amounts to an unwarranted intrusion into matters properly left to the experience and expertise of the Commissioner of Health, we dissent.2
I
The accepted standard for reviewing an administrative regulation of the type before us is not in dispute: (1) that a regulation will be upheld "if it has a rational basis, and is not unreasonable, arbitrary or capricious” (majority opn, at 166) and (2) that if the regulation is to be nullified, "[t]he challenger must establish that '[it] is so lacking in reason for its promulgation that it is essentially arbitrary.’ ” (Id. [emphasis added], quoting Matter of Marburg v Cole, 286 NY 202, 212.) A review of the important policy considerations underlying the standard is necessary here.
The rationality standard is designed to ensure no more than that the administrative action being reviewed is reasonable and fair. It necessarily contemplates a high degree of deference to the agency’s rule-making determination, particularly where, as here, the rule involves areas within the agency’s expertise (see, 5 Davis, Administrative Law § 29:3, at 343; see also, Sunstein, Law and Administration After Chevron, 90 Colum L Rev 2071, 2117). This deference reflects both the judiciary’s recognition of its own lack of competence in the frequently complex and technical areas covered by administrative regulations and the agency’s presumed competence and proficiency in these matters (see, 5 Davis, op. cit., § 29:3, at 343). For these reasons, courts are and should be reluctant to disturb the substantive decisions of an administrative agency in the lawful exercise of its rule-making function (id.). Complying with this established doctrine, this Court has routinely sustained regulations under the rationality standard (see, e.g., Matter of New York State Health Facilities Assn. v Axelrod, 77 NY2d 340, 349-350; Molina v Games Mgt. Servs., 58 NY2d 523, 529-530; Matter of Catholic Med. Center v Department of Health, 48 NY2d 967, 968-969; Matter of Bates v Toia, 45 NY2d 460, 464; Matter of Bernstein v Toia, 43 NY2d 437, 448; Matter of St. Lukes Hosp. Center v Ingraham, 43 NY2d 771, affg on opn below at 52 AD2d 181; Ostrer v Schenck, 41 NY2d *171782, 785-788; Matter of Sigety v Ingraham, 29 NY2d 110, 115-116; Grossman v Baumgartner, 17 NY2d 345, 349-350).
To be sure, this Court has in rare instances invalidated regulations because they lacked a rational basis where the record contained no evidence to support the regulation (see, e.g., Matter of Jewish Mem. Hosp. v Whalen, 47 NY2d 331, 343; Matter of Hawley v Cuomo, 46 NY2d 990, 992).3 But such cases are entirely consistent with the rule that all that is required under the rationality standard is some evidentiary basis for the regulation.
The foregoing policy considerations are certainly applicable in reviewing the establishment of Medicaid reimbursement rates (see, Matter of Catholic Med. Center v Department of Health, supra; Matter of Sigety v Ingraham, supra; see also, Matter of New York State Health Facilities Assn. v Axelrod, supra, at 346-347). Indeed, it is difficult to imagine an area where the considerations calling for judicial deference could be more compelling than in Medicaid rate setting where the Legislature has broadly delegated the responsibility to the Commissioner of Health. In carrying out his rate-setting function, the Commissioner is required to consider various economic, geographic and health-care-related factors in setting rates, subject only to the requirement that the rates be reasonable and adequate for the efficient and economical operation of the facility (Public Health Law § 2807 [3]; § 2808 [2-a], [3]).
There is nothing to suggest that in fulfilling his rate-setting responsibilities under the Public Health Law, the Commissioner may not rely on inferences drawn from circumstantial evidence — a natural and necessary component in any decision-making process. Circumstantial proof is, of course, as probative as direct evidence and may be even more persuasive (see, 1A Wigmore, Evidence §26, at 957, 961 [Tillers rev 1983]; People v Ford, 66 NY2d 428, 442). Indeed, in Health Facilities *172(supra), we recently sustained regulations that were, in large measure, based on the Commissioner’s inferences from circumstantial evidence that nursing homes were unfairly discriminating against Medicaid patients in their admission and retention policies.
Here, the conclusion which the majority rejects as irrational —apparently because it is based largely on an inference from circumstantial evidence rather than on direct and concrete proof — is that the increase in CMI which necessitated the Medicaid rate recalibration resulted from "paper optimization” rather than from an actual change in patient condition. But the inference is compelled naturally and logically by the simple process of elimination.4
That "paper optimization” (i.e., more accurate compliance with newly imposed reporting requirements) is a generally recognized phenomenon is well documented (see, e.g., Simborg, DRG Creep: A New Hospital-Acquired Disease, N Eng J Med 304:1602-1604 [1981]; Hoke, Charting for Dollars, Am J of Nursing [June 1985]; Wennberg, Will Payment Based on Diagnosis-Related Groups Control Hospital Costs?, N Eng J Med 311:295-300 [1984]; Report and Recommendations to Secretary, US Dept of Health and Human Services, at 31-32 [1987]). The majority does not dispute this proposition or that "paper optimization” could have been a cause of the 1986 average CMI being greater than the 1985 average CMI (see, majority opn, at 167). The other possible causes for the higher 1986 CMI would, of course, be actual increases in the mix of sicker patients requiring greater care. But the Commissioner’s proof eliminates these other potential causes with far more certainty than needed to satisfy the rationality standard and compels the conclusion that "paper optimization” — the uneliminated cause — explains the increase in the 1986 CMI. The majority attempts to cast doubt on the proof eliminating these other potential causes of the comparative increase in the 1986 average CMI over the 1985 CMI and on the validity of the resultant inference that "paper optimization” is the explanation. As will be explained (see, part II, infra), this effort is ineffectual.
*173II
To discuss the Commissioner’s proof in support of the 3.035% rate recalibration, the basic relationship between the Medicaid reimbursement rates and the 1985 average CMI must be understood. One of the variables used in the computation of the reimbursement rates under the new RUG-II system is the average CMI. The reimbursement rates are an inverse function of the average CMI — i.e., if the average CMI from which the rates are computed decreases, then the resultant reimbursement rates increase. Here, the initial reimbursement rates (i.e., before recalibration) were computed from the 1985 average CMI. If, as the proof demonstrates, the 1985 CMI was understated, it follows that the corresponding reimbursement rates would have been set too high and deserved recalibration.
Concededly, the 1985 CMI was lower than the 1986 CMI; and concededly some of the difference is explained by the fact that the nursing homes in 1986 were admitting new patients requiring more care as the RUG-II system encouraged them to do. For this part of the difference, of course, the Commissioner made no recalibration. However, part of the difference between the 1985 and 1986 CMIs is not explained by the actual increase in the quantity of sicker patients in the case mix but by the more accurate reporting of patient data in 1986 for the same patients, the phenomenon known as "paper optimization”. In other words, "paper optimization”, and the inadequacy of the 1985 data reporting which it reveals, establish that the 1985 average CMI was lower than would have been the case if the 1985 data had been reported accurately. The Commissioner’s proof demonstrates clearly, in our view, the existence of "paper optimization” and a corresponding understatement of the 1985 CMI of between 3.2% and 3.9% — more than necessary to justify the Commissioner’s recalibration of 3.035% to correct the resultant overstatement in the initial reimbursement rate.
That "paper optimization” would be expected to occur under the circumstances here cannot be disputed. The RUG system had just been instituted. The patient data from which the 1985 average CMI was computed came from newly required RUG-II reporting documents ("Patient Review Instruments” [PRIs]) which the facilities had never before used. The PRI is a complex and detailed reporting form consisting of 37 separate items and 15 pages of instructions. The report requires the collation and analysis of extensive physical and medical infor*174motion in making the sometimes difficult' evaluation of a patient for placement in the proper RUG-II category. A PRI form must be submitted on behalf of each patient in a facility twice a year. In view of the complexity of this form, it is hardly surprising that the facilities reported the data with greater thoroughness and accuracy when they made their second filing in 1986. In fact, the Commissioner’s informational bulletin to the facilities prior to the implementation of the RUG-II system advised them that just such a recalibration for "paper optimization” should be anticipated.
To establish the extent of the understatement in the 1985 average CMI due to inaccurate data reporting, the Commissioner compared the 1985 PRI data with the 1986 data for the same patients in the same facilities. New patients were excluded from the comparison. Since the comparison was limited to patients residing in the facilities in both years, any patient admitted in 1986 was not considered. Thus, the increase in CMI could not have resulted from the admission of new patients in 1986 requiring more intensive medical care.
Nor do the other possible explanations suggested by plaintiff for the 1986 CMI increase provide a reason for rejecting the Commissioner’s conclusion. For example, the 1986 increase cannot be explained by the release of patients requiring less intensive care, as plaintiff surmises. On the contrary, the evidence establishes that patients requiring more intensive care in the higher CMI categories were discharged at a greater rate than those in the lower categories requiring less care. Additionally, the record evidence refutes plaintiffs suggestion that the increase was due to the over-all deterioration in patient condition or temporary changes in that condition. The statistical proof — comparing data pertaining to all of the 1985 and 1986 patients (including new admittees) — shows the same general over-all increase in CMI for the new short-term patients as for the previously admitted long-term patients.
Moreover, again totally refuting plaintiffs suggestion to the contrary, the 3.2% "paper optimization” increase in the 1986 average CMI — adopted by the Commissioner as the basis for his initially proposed 3.2% recalibration — cannot be attributed to the facilities providing more rehabilitation services to patients in 1986. The Department’s analysis showing the 3.2% increase was based on a 1985-1986 comparison from which patients in the rehabilitation categories were excluded.
The fact that the 3.035% recalibration is a reduced figure *175agreed to by the Commissioner after discussion and compromise with the facilities is certainly not a ground for holding that figure unreasonable. Initially, the Department sought a 3.2% reduction in reimbursement rates, corresponding to the 3.2% increase in average CMI for patients in groups other than rehabilitation. Thus, the negotiation and compromise to which the majority takes exception (see, majority opn, at 168) actually contributed to the adoption of a recalibration that was more favorable to nursing homes than the one the Department initially sought to impose (based on the 3.9% CMI increase when rehabilitation patients were included and 3.2% when they were excluded).
As noted, that recalibration would be necessary to correct the reimbursement rates for "paper optimization” was a probable eventuality foreseen from the very inception of the RUG-II system — a probability of which the facility operators were made fully aware. Thus, it can hardly be thought that the correction for "paper optimization” was adopted as an afterthought only because the department "did not like some of what [the] new [RUG-II] process had wrought” (majority opn, at 167).
Finally, there can be no merit to plaintiffs suggestion— espoused, it seems, by the majority (majority opn, at 168)— that because the recalibration effected an across-the-board reduction in the RUG-II reimbursement rates, a facility having more patients requiring intensive care in the higher RUG-II categories would somehow be deprived of its full compensation based on its actual CMI. The recalibration, of course, in no way altered any particular facility’s CMI but effected only the reimbursement rates — the constant multipliers used in conjunction with a facility’s CMI to compute the amount of its reimbursement.
Ill
The question here is not whether the Department can "document and substantiate its conclusion” (majority opn, at 168) by direct, concrete proof, as the majority holding clearly seems to require. The question is whether plaintiff can meet its burden of showing that the conclusion is unreasonable and unsupported by some evidence (see, Matter of Catholic Med. Center v Department of Health, 48 NY2d 967, 968, supra). The fixing of Medicaid rates under the broad authority delegated to the Commissioner (Public Health Law § 2807 [3]; § 2808 [2-*176a], [3]) — like any rate-making determination (see, e.g., Public Service Law §§ 65, 66; Matter of Abrams v Public Serv. Commn., 67 NY2d 205, 211-215, 217-218) — requires an exercise of judgment (see, Matter of Catholic Med. Center v Department of Health, supra, at 968-969). Here, the criteria established by the Legislature to guide the Commissioner in setting Medicaid rates are only that the rates be "reasonable and adequate to meet the costs which must be incurred by efficiently and economically operated facilities” (see, Public Health Law § 2807 [3]). It was in the exercise of this broad, rate-making authority that the Commissioner determined that the over-all Medicaid reimbursement rate should in fairness be recalibrated.
That "paper optimization” explains some of the 1986 CMI increase and thus justifies some recalibration of the reimbursement rates cannot be challenged. The extent of the justified recalibration is something that, by its very nature, cannot be established with mathematical certainty. Of necessity it requires proof by inference and an exercise of judgment in evaluating that proof. The ultimate question is whether the recalibration is reasonable and fair and whether there is not some evidentiary basis for it. To require more and, in effect, to impose a far stricter scrutiny than the rationality standard calls for is to engage in the sort of judicial second-guessing that the Legislature in its delegation of power to the Commissioner could not have contemplated (see, Public Health Law § 2807 [3]; § 2808 [2-a], [3]). In employing a strict scrutiny standard under the rubric of a test characterized as its "narrow holding of irrationality in this particular case” (majority opn, at 167),5 the majority frustrates the fundamental policies underlying the concept of judicial deference and creates a significant and, we suspect, potentially troublesome precedent in what has been a well-established area of administrative law. For these reasons, we dissent.
Judges Simons, Kaye and Alexander concur with Judge *177Bellacosa; Judge Hancock, Jr., dissents and votes to affirm in a separate opinion in which Judge Titone concurs; Chief Judge Wachtler taking no part.
Order reversed, with costs, and judgment of Supreme Court, Albany County, reinstated.
. The plaintiff here represents the interest of 42 counties. But, the regulation relates to reimbursement rates applicable to public and private facilities alike (see, 10 NYCRR 86-2.3; see generally, Public Health Law §§ 2807, 2808).
. We agree with the majority’s conclusion that the action was timely commenced within the four-month limitations period, measured from the facilities receipt of the Department’s rate calculation notice on June 2,1987.
. In Matter of Jewish Mem. Hosp. v Whalen (47 NY2d 331), for example, the Court invalidated a regulation which disqualified 10% of facilities’ educational costs from possible insurance reimbursement as costs not directly related to patient services. There, the Department had conceded that it never conducted a study of educational costs to justify its conclusion that 10% was the proper percentage to exclude from reimbursement as unrelated to patient care. Similarly, in Matter of Hawley v Cuomo (46 NY2d 990), the Court invalidated a two-county nonsolicitation order issued by the Secretary of State, where the record was devoid of any evidence of racial blockbusting tactics on which the regulation was based.
. The majority’s suggestion that the dissent is based on " 'simple processes of elimination’ at the appellate review stage” (majority opn, at 169) misses the point. It is not the Court, but the Commissioner, who has drawn the perfectly natural and proper inference from circumstantial evidence that the change in average CMI resulted from "paper optimization”, the only potential cause which was not eliminated.
. Labeling the decision as "narrow” and other similar efforts to limit the breadth and effect of its holding (see, e.g., majority opn, at 169 [terming its decision a "straightforward application of the traditional judicial review function to a particular record and regulation”]), of course, cannot change what it does and what it means as a precedent. That the holding is not "narrow” and goes well beyond its purported application of "traditional review” will be seen if the entirely reasonable methods employed by the Commissioner in exercising his rule-making function are studied in this record.