district wealth fail to take into account regional costs. School districts face significant variation [187 [187 Misc.2d Misc.2d 86] 86] in costs to deliver educational services, which in turn affects their ability to pay for various educational inputs. In 1999, SED noted the long-standing recommendations made by various blue ribbon panels to include regional cost estimates in the State aid formulas and concluded that "[t]he failure to explicitly recognize geographic cost differences within the major operating aid formulas has led to formula allocations which are inequitable." SED has quantified these differences in regional costs. New York City's regional cost ratio is the highest in the State, which means that a dollar buys fewer educational resources in New York City than anywhere else in the State. Second, the CWR's wealth equalization is undermined to a degree by the State's STAR program. Under STAR, qualified households receive a partial school property tax exemption, thereby reducing their school property tax bill. The State then reimburses local school districts for the foregone tax revenue due to property tax cuts and thereby shifts a portion of local education taxes from individual school districts to the State. The unrebutted evidence at trial indicates that New York City receives less STAR aid than localities in the rest of the State. The Independent Budget Office estimates that New York City homeowners received an average reduction in tax liability of $323 per participating household, while those in the rest of the State received an average reduction of $926. The State Comptroller has predicted that when the STAR program is fully implemented, New York City will fare even worse vis-á-vis the rest of the State. Defendants' assertion that "STAR is property tax relief, not school aid" 39 does not comport with the practices of SED or the State Division of Budget, which include STAR tax relief in their evaluations of school aid. Finally, even though State aid distribution formulas are grossly wealth equalizing, they do not appear to be particularly responsive to changes in district income. From 1988-1989 through 1997-1998, measures of New York City's property wealth varied considerably. During 1994-1998, New York City's actual value of taxable real property per "Total Wealth Pupil Units" in thousands dropped from $260.3 to $215.7, a loss in value of 17.2%. In contrast the average value in the rest of the State dropped by 1.9%. Despite these differences, during each [187 [187 Misc.2d Misc.2d 87] 87] of these years New York City received virtually the same percentage share increase in State aid. 40 For the above-stated reasons, the wealth equalization effects of the State aid formulas do not accurately reflect New York City's wealth and ability to pay for education. More importantly, defendants' other primary defense of the State aid distribution system—that it is driven in large measure by student need—is almost completely negated by the record. It is true that the State aid formulas contain factors such as TAPU that are weighted to a degree by a district's percentage of at-risk children. Additionally, some State aid formulas, such as "Extraordinary Needs Aid" and "Limited English Proficiency Aid," are purportedly designed to send more money to districts with large numbers of at-risk children. However, these formulas and weightings do not accurately account for the costs of education caused by large numbers of at-risk students in a single district. The State Education Department has long documented this misalignment of resources in its needs/resource capacity index, a measure of a district's ability to meet the needs of its students with local resources. Based on a comparison of districts according to their needs/resource capacity index ratios, SED has determined that New York City receives substantially less State aid than districts with similar needs. For example in 1996-1997 the City received $3,562 per pupil in State aid, between $1,500 and $1,800 less than districts with similar needs/ resource capacity index ratios. This analysis is substantiated by other SED analyses using different methods that demonstrate that New York City receives substantially less State aid than districts with similar numbers of at-risk students. Indeed, the City receives less State aid than many districts with substantially smaller proportions of at-risk students. [187 [187 Misc.2d Misc.2d 88] 88] Even if the State aid formulas were designed to allocate education aid more fairly, it would be difficult to change the orientation of State education funding because any annual increase in State aid has historically been divided without reference to the formulas. The evidence at trial demonstrated that the formulas and grant categories are not allowed to operate neutrally but rather are manipulated during the State's annual budget negotiations by State officials to produce consistent funding allocations of aid increases among school districts around the State. The evidence at trial demonstrated clearly what the State Comptroller has found: It is well known that the formulas are annually `worked backwards' until the politically negotiated `share' for the City schools is hit in the calculations. In this context, the data feeding into the school aid formulas for New York City is really of no practical consequence whatsoever—the City will get the negotiated share of aid regardless of what data they report. The evidence supported the Comptroller's conclusion that annual increases in State education aid are allocated pursuant to an agreement struck by the Governor and the leaders of the State Assembly and the State Senate as part of the over-all annual budget negotiations. These negotiations produce a general agreement on the over-all amount to be spent on education and how it is to be distributed across the State which is then ratified by the Legislature. This phenomenon is commonly referred to as "three men in a room." If there are "three" men in a room, and the Comptroller is not among them, one may reasonably query how the Comptroller is able to speak authoritatively regarding this process. The answer is that the process is an open secret in Albany. The means for effectuating the agreement to provide a fixed share of State aid increases are contained in State software and documents. The documents and software relied upon by the court do not reveal any deliberations of individual legislators. Rather, they simply confirm how the process of allocating increases in State school aid works. Confirmation of the process is also seen in New York City's consistent "share" of any annual State aid increase. The State has created computer software, called the "State aid modeling system," to enable governmental personnel to evaluate the distributional impact on school districts of changes to the State's public school funding formulas. The State aid [187 [187 Misc.2d Misc.2d 89] 89] modeling system includes thousands of variables that the State has devised as part of approximately 50 State aid formulas and categorical grant programs that it uses to fund public schools in New York State. The State aid modeling system is used by staff from SED, the Division of the Budget (DOB) and members and staff of both houses of the Legislature. In order to manipulate data using the State aid modeling system, personnel from SED, DOB and the Legislature fill out a form entitled "Confidential State