*490OUTLINE OP OPINION
I Outline of Major Issues 497
II “Pair Share” and “Region” — Preliminary Considerations 498
III Madison — Its Growth and Development 500
IV The Zoning Ordinances 503
V “Least Cost” versus “Law and Moderate Income” Housing, etc. ■ 510
VI Incapacity of 1973 Ordinance to EfEect Lower Income Housing 514
VII The “Pair Share” Approach of the Defendant 524
VIII “Pair Share” and “Region” — General Considerations 531
IX Environmental Considerations 544
X “Affirmative Action” for Lower Income Housing 546
XI The Validity of the Zoning Statute 547
XII Relief for Corporate Plaintiffs 548
XIII Remedy and Remand 552
*491The opinion of the court was delivered by
Conrord, P. J. A. D.,Temporarily Assigned. We today review the decision of Judge Eurman invalidating the 1973 amendatory zoning ordinance of defendant Township of Madison.1 128 N. J. Super. 438 (Law Div. 1974). That determination culminated an action instituted by plaintiffs in November 1970 challenging the validity of a zoning ordinance adopted by the township in September 1970 to replace a previous one in effect since 1964.2 Judge Furman had invalidated the 1970 ordinance in Oakwood at Madison, Inc. v. Tp. of Madison, 117 N. J. Super. 11, 21 (Law Div. 1971), but at the same time rejected an attack by plaintiffs on the constitutionality of the enabling zoning statute, N. J. S. A. 40:55-30 et seq. Id. at 16.
Defendant obtained a stay of judgment pending its appeal to the Appellate Division, and plaintiffs filed a cross-appeal as to that part of the judgment sustaining the validity of the statute. On plaintiffs’ motion, and because of the importance of the case, we certified the appeals pending unheard in the Appellate Division pursuant to R. 2:12-2. 62 N. J. 185 (1972).
Oral argument was originally heard by the court on March 5, 1973, and additional argument was requested for January 8, 1974. However, on October 1, 1973 Madison Township adopted a major amendment to the 1970 ordinance. Conse*492quently, on January 8, 1974, while retaining jurisdiction, we remanded the action to the trial court for a trial and ruling on the ordinance as amended, with the result stated above.
Oral argument on the present phase of the appeal has been had twice, emphasis being placed on the effect on the issues herein of our intervening decision in So. Burl. Cty. N.A.A.C.P. v. Tp. of Mt. Laurel, 67 N. J. 151, app. dism. and cert. den. 423 U. S. 808, 96 S. Ct. 18, 46 L. Ed. 2d 2028 (1975) (“Mount Laurel" hereinafter). We have received and considered supplemental briefs and materials.
Plaintiffs herein comprise two groups. Oakwood at Madison, Inc. and Beren Corporation (hereinafter “corporate plaintiffs”), both New Jersey corporations, were developers owning a tract of vacant developable land of some 400 acres, the disputed Oakwood-Beren tract. Six individuals were low income persons acknowledged by the trial judge as “representing as a class those who reside outside the township and have sought housing there unsuccessfully.” Oakwood at Madison, Inc. v. Tp. of Madison, supra (117 N. J. Super. at 14). Plaintiffs alleged, inter alia, (a) that the exclusionary nature of the ordinance rendered it unconstitutional; (b) that the enabling legislation was unconstitutional in its failure to provide adequate standards for municipal exercise of the zoning power; and (c) that the restrictive effect of the ordinance as applied to corporate plaintiffs’ property rendered it confiscatory.
The trial court invalidated the 1970 ordinance, primarily on the grounds that in zoning massive areas of vacant developable land for one-and two-acre single family residences, beyond the reach of 90% of the population, and in allocating only “miniscule” acreage for multi-family dwelling units, it ignored the housing needs of the township and the region, and failed “to promote reasonably a balanced community in accordance with the general welfare.” 117 N. J. Super. at 20-21. The court upheld the constitutionality of the enabling legislation; it did not reach the issue of confiscation, appar*493ently regarding the invalidation of the entire ordinance as rendering that matter moot.
While the 1973 amendatory ordinance transferred substantial areas from large lot to smaller lot zoning, made more land available for multi-family development and provided for planned unit development (PUD) and “cluster” zones, the evidence in the case convinced the court that the mu-' nieipality still was not satisfying its obligation to “provide its fair share of the housing needs of its region”, particularly in relation to the low-income and moderate-income population. 128 N. J. Super. at 447. The amended ordinance was therefore again struck down in its entirety. Ibid.
The main lines of the Law Division opinion striking down the 1973 ordinance may be summarized as follows. A crisis in housing needs continues, most serious for those of low and moderate income. The region, whose housing needs must reasonably be provided for by such municipalities as Madison, is not necessarily coextensive with Middlesex County. “Rather, it is the area from which, in view of available employment and transportation, the population of the township would be drawn, absent invalidly exclusionary zoning”. 128 N. J. Super. at 441. Almost all of Madison’s employed residents work outside the township, 50% in the county, 15% in New York City, 10% in Essex County, and the remainder in nearby counties, including 7% in Monmouth County. After an analysis of the testimony concerning the number of housing units which could be expected, under the amended ordinance, to be produced and to be' affordable by low and moderate income households, the court said:
Of the total 20,000 to 30,000 housing units which may be built in Madison Township under the 1970 zoning ordinance as amended, about 3500 [12% to 17%] at most would be within the reach of households with incomes of $10,000 a year, the upper limit of moderate incomes, and virtually none within the reach of households with incomes of $9,000 a year or less. This contrasts with the present township population, approximately 12% low income and 19% moderate income. Id. at 446.
*494The court assessed Madison Township’s obligation to provide its fair share of regional housing needs as follows:
Without the rigidity of a mathematical formula this court holds that Madison Township’s obligation to provide its fair share of the housing needs of its region is not met unless its zoning ordinance approximates in additional housing unit capacity the same proportion of low-income housing as its present low-income population, about 12%, and the same proportion of moderate-income housing as its present moderate-income population, about 19%. The amended zoning ordinance under review falls palpably short and must be struck down in its entirety. Id. at 447.
The court did not specify any absolute numerical quota of low and moderate income units the ordinance would be expected to render possible, but found that annual needs “into the 1980’s were 750 to 1000 units, 500 to 600 of those low and moderate income.” Id. at 442.
The court dealt with the defendants’ argument that ecological and environmental factors justified the RP, R-80 (2 acre lot minimum) and R-40 (1 acre lot minimum) zones by pointing out that such problems had “no bearing” except in specified limited areas and that “ample land outside these areas is available” with which the township could meet its obligation to provide its fair share of needed housing. 128 N. J. Super. at 447.
It should be stated at the outset that the basic rationale embraced by Judge Furman in both of his opinions in the case is substantially that adopted by this court in Mount Laurel, with the qualification that our determination there rested on the state constitutional ground that due process and equal protection are denied if “substantial segments of the population” are improperly precluded from residing within the municipality because of local zoning regulations. 67 N. J. at 175. The “substantial” segments thus identified were those low and moderate income people of the region economically unable to afford suitable housing in developing municipalities of the region because of their highly cost-generating zoning restrictions.
*495General guidelines toward eliminating undue cost-generating restrictions were stated in Mount Laurel (67 N. J. at 187) :
By way of summary, what we have said comes down to this. As a developing municipality, Mount Laurel must, by its land use regulations, make realistically possible the opportunity for an appropriate variety and choice of housing for all categories of people who may desire to live there, of course including those of low and moderate income. It must permit multi-family housing, without bedroom or similar restrictions, as well as small dwellings on very small lots, low cost housing of other types and, in general, high density zoning, without artificial and unjustifiable minimum requirements as to lot size, building size and the like, to meet the full panoply of these needs.
In the absence of legislation providing for regional zoning authorities, while municipalities are empowered to zone individually, they must nevertheless (if of the “developing” category described in Mount Laurel, 67 N. J. at 160), by their zoning regulations serve and not impede the general welfare represented by satisfaction of the housing needs of lower income people throughout the region. 67 N. J. at 188 — 190;3 id. at 194 (Pashman, J., concurring).
*496After the last argument in this matter the Legislature revised the zoning statutes of this State by enactment of the “Municipal Land Use Law”, L. 1975, c. 291, which by its terms became operative August 1, 1976. We invited and have received from counsel supplemental comment as to any effect of the new law on the issues herein, particularly as to the continued viability of Mount Laurel. We find nothing in the statute inconsistent with the doctrine there laid down. (The decision would control, in any event, in view of its constitutional underpinning.) Without here indulging in any comprehensive evaluation of the new law, certain of the purposes stated in Section 2 seem particularly pertinent.
d. To ensure that the development of individual municipalities does not conflict with the development and general welfare of neighboring municipalities, the county and the State as a whole;
e. To promote the establishment of appropriate population densities and concentrations that will contribute to the well-being of persons, neighborhoods, communities and regions and preservation of the environment ;
g. To provide sufficient space in appropriate locations for a variety of agricultural, residential, recreational, commercial and industrial uses and open spaces, both public and private, according to their respective environmental requirements in order to meet the needs of all New Jersey citizens.
*497At the same time, the new law reminds ns, as we emphasized in Mount Laurel, that out of a proper concern for adequate housing there should not and need not be over-intensive and too sudden development, future suburban sprawl and slums, or sacrifice of open space and local beauty. 67 N. J. at 191. Thus, the newly articulated purposes of Section 2 of the statute include:
c. To provide adequate light, air and open space.
j. To promote the conservation of open space and valuable natural resources and to prevent urban sprawl and degradation of the environment through improper use of land.
I
Outline of Major Issues
The judgment of the trial court, the intervention of our decision in Mount Laurel and the nature of the record and briefs before us combine to cast the issues for determination as follows:
1. Is the Madison 1973 zoning ordinance exclusionary, i. e., whether or not so intended, does it operate in fact to preclude the opportunity to supply any substantial amounts of new housing for low and moderate income households now and prospectively .needed in the municipality and in the appropriate region of which it forms a part?
2. If, as we have concluded, the affirmative response to the foregoing question by the trial court should be sustained, is it incumbent upon the courts, pursuant to Mount Laurel, to demarcate a pertinent region and to fix a specific number of lower-cost housing units as the “fair share” of the regional need therefor to be made possible by the Madison ordinance ?
3. If, as we have concluded, the foregoing question should be answered in the negative, what kind of an order should be made to assure Madison’s compliance, as a developing municipality, with Mount Laurel’s mandate that its zoning ordinance “afford the opportunity” for at least “the mu*498nicipality’s fair share of the present and prospective regional need” for “decent and adequate low and moderate income housing”? 67 N. J. at 188.
II
“Fair Share” and “Region”— Preliminary Considerations
As noted above, the prime question before us, in Mount Laurel terms, is whether the trial court has correctly found that Madison’s zoning ordinance does not provide the opportunity to meet a fair share of the regional burden for low and moderate income housing needs. We have seen that the trial court did not specify the precise boundaries of the applicable region nor fix an absolute number of appropriate housing units to be provided. It merely described the pertinent region as the area from which the population of the township would be drawn, absent exclusionary zoning.
A substantial body of evidence was adduced by the defendant below purporting to specify Madison’s fair share of Middlesex County’s unmet need for low and moderate income housing as of 1975. Moreover, the record before us, the briefs and the literature in the field supply abundant data concerning- methods' and techniques for estimating a municipality’s fair share of a regional housing need. We propose to comment on these matters hereinafter, for three purposes: (a) to explain our conclusion that the evidence concerning fair share adduced by defendant does not refute the trial court determination that the Madison ordinance is deficient in the respects noted; (b) to elucidate the considerations relating to the appropriate “region” whose housing needs are relevant to this action; and (c) to furnish guidance to courts, counsel and expert witnesses in this area in applying the principles of Mount Laurel to litigated controversies generally.
However, we deem it well to establish at the outset that we do not regard it as mandatory for developing municipalities whose ordinances are challenged as exclusionary to devise specific formulae for estimating their precise fair share of the lower income 4 housing needs of a specifi*499eally demarcated region. Nor do we conceive it as necessary for a trial court to make findings of that nature in a contested case. Firstly, numerical housing goals are not realistically translatable into specific substantive changes in a zoning ordinance by any technique revealed to us by our study of the data before us. There are too many imponderables between a zone change and the actual production of housing on sites as zoned, not to mention the production of a specific number of lower cost units in a given period of time. Municipalities do not themselves have the duty to build or subsidize housing. Secondly, the breadth of approach by the experts to the factor of the appropriate region and to the criteria for allocation of regional housing goals to municipal “subregions” is so great and the pertinent economic and sociological considerations so diverse as to preclude judicial dictation or acceptance of any one solution as authoritative. For the same reasons, we would not mandate the formula approach as obligatory on any municipality seeking to correct a fair share deficiency.
We are convinced from the record and data before us that attention by those concerned, whether courts or local governing bodies, to the substance of a zoning ordinance under challenge and to bona fide efforts toward the elimination or minimization of undue cost-generating requirements in respect of reasonable areas of a developing municipality represents the best promise for adequate productiveness without resort to formulaic estimates of specific unit “fair shares” of lower cost housing by any of the complex arid controversial allocation “models” now coming into vogue.5
It is desirable that administrative agencies acting under legislative authorization assume the regulation of the housing distribution problem. Until then, in the current post-Mount Laurel period judicial emphasis on approaches such as those just outlined, and exemplified in the remedial sec*500tion of this opinion, will, it is hoped, suffice to move the State toward the objective of “available housing- in the developing municipalities for a goodly number of the various categories of people of low and moderate income who desire to live therein and now cannot.” Mount Laurel, 67 N. J., at 188, n. 21.
Ill
Madison — Its Growth and Development
Madison Township consists of approximately 42 square miles, or 25,000 acres, in the southeast corner of Middlesex County, of which almost 40% is vacant developable land. Its location within the gap between the metropolitan centers of New York and Philadelphia is a strategic one: this “Atlantic urban region” gap is expected to be bridged within the next 25 years, with a concomitant population increase of 75%. The Tri-State Regional Planning Association (covering counties in New York, New Jersey and Connecticut and including Middlesex) predicts that Middlesex will be one of four counties to experience the greatest rates of growth in the tri-state area from 1970 to 2000.
Parts of the township lie within 20 miles of the. highly urbanized areas of Elizabeth and Newark. Easy access to and from the municipality is provided by several major highways which traverse it. The Garden State Parkway and State Highways 34 and 35 cut across the eastern portion of the township, connecting- Madison to both the south Jersey shore area and the Newark-New York metropolitan area. State Highway 18 and U. S. Route, 9 run through the center of the township, and there are three major county roads, 527, 516 and 520. The New York-Long- Branch Railroad runs through the eastern portion of the municipality, and although there are no commuter stations in the township itself there are in nearby communities. The accessibility of the township is readily illustrated by the status of the com*501nmnity as a commutershed. Only 1% of Madison’s employed residents work within the township. As found by the trial court, 50% of the work force are employed in Middlesex County, 15% in New York City, 10% in Essex County, 9-12% in Union County and 7% in Monmouth County.
Madison is an archetypal “developing” municipality within the contemplation of the Mount Laurel specifications. 67 N. J. at 173, 187. During the past 25 years, it has experienced explosive growth. Its population increased over two decades by 561%, from 7,366 in 1950 to 48,715 in 1970. This boom has continued, with the population climbing to 50,000 by the time of the first trial and 55,000 by the second in 1974. With the growth and concomitant municipal problems came a steady rise in tax rates.
Even in light of this period of great expansion, Madison still has large potential for further growth.6 Among the twenty-five municipalities in Middlesex County, Madison in 1970 ranked 20th lowest both in population density and housing density. Vacant acreage is plentiful; of the township’s 25,000 acres, between 8,143 and 11,000 are vacant and developable. The township is a sprawling municipality marked by little continuity and spotty development. The area is laced by a network of streams and rivers eventually feeding into South River to the north. Cheescquake State Park occupies a sizeable portion of land in the eastern part of the town.
The older residential development is concentrated to some extent in the Old Bridge and Browntown areas and on the Raritan Bay, which forms the eastern boundary of the municipality. Lawrence Harbor and Cliffwood Beach, the two major developments located on the Raritan Bay, consist mainly of bungalows on 50-foot lots originally built, in the *5023.920’s as summer cottages and since converted into year-round residences. Newer single family developments and apartment complexes are peppered over the township, except for the southwestern third of the township which is largely undeveloped. Most single family homes are on lots less than an acre. It does not appear that prior to the first trial any one-or two-acre housing developments had been constructed except for some two-acre homes built in the 1920’s. Commercial land uses are scattered, generally following some of the major highway routes. Industrial usage is slight.
Construction within the township fell off from 1970 to 1973. Comparing Madison with four nearby municipalities of generally similar characteristics (with large undeveloped areas) for the said three year period, Madison issued an average of only 53 dwelling unit building permits per year as against 368 in East Brunswick, 309 in Monroe, 89 in Sayreville and 212 in South Brunswick. Although Madison contains 20% of the county’s vacant residentially zoned land, and from 3960 to 1970 issued 15% of all the building permits, from 3970 to 1972 its percentage of county building starts fell to 6%.
From 1950 to 1970 the housing growth in the township was characterized by construction of single family homes built on lots of 15,000 square feet or less and of a number of multifamily garden apartment developments. Virtually all the apartment units in the township, however, were constructed after 1963, and by April 1969 they comprised 3,700 or 27.4% of the total of 33,499 housing units in the township. In 1970, 56% of the single family dwellings in the township were valued at $25,000 or less. Figures from the 1970 census show that, in terms of statewide quintiles (20%) of income category, 12% of the township’s households had incomes below $6,627, 19% between $6,627 and $9,936, 24% between $13,088 and $19,236, and 18% above $19,236. As of 1970, existing land uses were predominantly residential: 68% of all realty taxes were paid by single family homeowners, 16% by apartments and condominiums and the rest by commercial users, farms and industry.
Thus the overall pattern of land use confronting Madison Township planners and officials in 1970 was one of substantial *503but scattered residential growth, with little industrial and commercial development. The 1970 ordinance was a hurried effort to slow population growth and the accompanying rise in the tax rate and largely to confine new population to designated areas. See 117 N.J. Super. at 14.
IV
The Zoning Ordinances
A. The 1970 Ordinance
For present purposes the salient provisions of the 1970 ordinance are adequately summarized in the first opinion of the trial court. 117 N.J. Super. 16-17. The patent intent and effect of the ordinance was to prevent construction of a substantial number of homes or apartments, particularly at low cost. Most of the land area was zoned for one- or two-acre single family homes — uses not only beyond the reach of 90% of the general population but also responsive to little if any existing market. Ibid,. It goes without saying that the ordinance was clearly violative of the principles later enunciated in Mount Laurel. Judge Furman properly condemned it as pure “fiscal zoning”, not taking into consideration “[h]ousing needs of the region” and failing to promote “reasonably a balanced and well ordered plan for the entire municipality.” 117 N.J. Super. at 18.
B. The 1973 Ordinance
The 1973 ordinance extensively revised the land use restrictions of the prior ordinance. The amount of land zoned nonresidential (commercial, office and industrial) was decreased by 760 acres from 19.80% to 16.70% of the total. A new “open space” zone — RP or Recreation-Preservation — was created. This encompassed the areas deemed environmentally sensitive by the township: Cheesequake State Park, the Old Bridge sands area, Burnt Fly Bog, the meadowlands adjacent to Deep Run (the latter three containing underground water resources areas), and the Raritan beachfront. The RP and the RR zone (also an open space area devoted to substantial preservation in a natural condition) were permitted to be developed as R-80 on two-acre lots until acquired by the municipality.
*504Although the fact was not stressed at trial, Madison has placed more than 4,000 acres in zones restricted to industrial and office uses despite the fact that only some 600 acres have ever been devoted to that use. By comparison, we criticized Mount Laurel for zoning 4,100 acres industrial although only 100 acres had ever been so used. 67 N. J. at 162-163, 184.
The 1973 ordinance considerably increased the facial housing potential of the prior ordinance. It enlarged the total acreage available therefor by 800 acres and the potential housing capacity, inclusive of existing housing, by 16,000 units or 46,000 persons. These figures, supplied by the Madison Township Housing Study, may be misleading, as they assume all acreage zoned residential is or will be developed to its maximum permissible density whereas some of it is already developed, either non-rcsidential or below permissible density, or is undevelopable.
Under the 1973 ordinance, there; are five single-family zones, accounting for 72% of the total vacant residential area.7 The most restrictive zone, the R-80, with a minimum lot size of two acres, was reduced from 9,134 to 3,040 acres. The R-40 zone (one acre minimum lot size) was increased from 5,557 to 7,511 acres. Together, however, these two zones account for 42% of the total acreage within the township, 58% of its vacant developable acreage, 70% of the total acreage zoned single-family, and 80% of vacant developable single family acreage.8 The R-20 zone, 1,977 total *505or 1,285 vacant developable acres, requires a 20,000 square foot, minimum lot size.
These three zones (R-20, R-40 and R-80) may be compared with the zones considered exclusionary in Mount Laurel. There more than half the township was zoned R-3, requiring single family homes on half acre lots; in the instant ease, over 50% of the township is zoned for half acre lots or larger, and 42% for one-or two-acre lots. Considering only vacant developable acreage, the total for the three zones is over 65%, 58% comprising R-40 and R-80.9
The R-15 zone10 and R-l 0 zones, requiring 15,000 and 10,000 square foot lots respectively, account for another 5% of the land. Both are more restrictive than the R-l zone (9,375 square foot, 75' wide lots) involved in Mount Laurel. Calling for some “very small lot” zoning in a developing municipality, 67 N. J. at 170, n. 8, 187, Justice Hall noted that minimum size lots of 9,375 to 20,000 square feet “cannot be called small lots and amounts to low density zoning.” 67 N. J. at 183. Yet almost 70% of Madison Township is zoned at such or lower densities (including the RP and RR zones).
Only the R-7 zone allows residential single family development on lots smaller than those found in Mount Laurel to constitute low density zoning. It allows 7,500 square foot lots and two-family dwellings. However it accounts for only 5.8% of the total acreage and 2% of the vacant developable acreage in the township.
*506The AF or multi-family apartment zone was enlarged by 150 acres. The prescribed bedroom ratio of the prior ordinance (80% one bedroom, 20% two bedroom) was deleted and replaced by a floor area ratio (FAR) limiting construction to a maximum of 10,000 square feet per acre. Although the ordinance presumably allows any size units in any combination, up to the maximum FAR, the impact of the building area ratio combined with the profit incentive which motivates developers is such that, according to the proofs, only small units (efficiencies and one bedrooms) will be constructed.11
Of the 676 acres zoned AF (2.7% of the total township acreage) at most only 193 acres are vacant and developable (2.3% of township total). The true figure, however, as indicated by Judge Furman, is more likely to approximate 120 acres (or 1.5% of township total).12 The AF zone is limited to the development of parcels of six or more acres. The parties agreed that the AF zone could hold at least 800 housing units, but defendant maintains the maximum capacity *507is 1,700. However, as against township planner Abeles’s estimate of 15,600 to 20,700 total additional units of all kinds possible under the ordinance, the potential AP units constitute only 5.1% to 8.2% thereof.13
Madison Township relies heavily on provisions in the 1973 ordinance for PUDs (planned unit developments) and clustering to satisfy its obligation with respect to low and moderate income housing. On the evidence, that reliance is illusory.
The PUD, an overlay zone, is a modern planning modality, introduced by the 1973 ordinance pursuant to N. J. S. A. 40:55 — 54 et seq. See Mount Laurel, 67 N. J. at 166; L. 1975, c. 291, Sees. 2k, 29.1b. Three areas are zoned for PUD — two of which are on remote sites unserviced by water and sewer utilities. PUD requirements vary, depending upon the amount of land in the developer’s tract.14 A Class I PUD, between 150 and 300 acres, has a maximum density of 3.5 units per acre. Of all the units constructed in a Class I PUD, a minimum of 30% must be detached single-family units,15 and the remainder may be medium density multi-family.16 A Class II PUD, between 300 and 500 acres, has a maximum density of 4.25 units per acre; a minimum of 17.5% of the units must be single-family, a maximum of 12.5% may be *508high density,17 and the remainder medium density. Class III PUDs (over 500 acres) are the most favored, with an allowable density of 5.0, 12.5% minimum single-family detached and maximum 17.5% high density. The densities allowed in the PUD zones are 20% lower than those originally proposed by the municipal planners. Moreover, it is unlikely that the highest density (5.0) will ever be utilized as there are within the PUD zones no 500-acre parcels owned by a single entity, and accumulation of the necessary number of acres is, according to the credible evidence, neither “possible nor probable.”18
Eor every PUD project the developer is required to maintain at least 12.5% of the gross project area as undeveloped open space and 7.5% as developed open space; at least 5% of a Class I and 10% of a Class II or III tract must be nonresidential development; at least 3% in all the classes must be commercial and 2% of Class II special development. There is the additional requirement that the developer build a school large enough to accommodate .5 children per dwelling unit and dedicate the land to the township. Streets and utility hookups must also be provided by the developer. There is a lengthy three-stage approval process, which, according to the estimate of the township planner, may take as long as a year but plaintiffs’ expert pegs at 18 months to two years.
The proofs render it patent that the PUD requirements have the potential for greatly increasing the costs of housing units in that zone. The school requirement alone adds a cost of $2.2 million (or 66% of all the central improvement costs involved in a PUD project) to the project budget of a Class IT PUD.
*509The cluster provisions (a PUD variation, not including commercial uses) apply to any lands in the R-40 or R-80 zones not alternatively designated PUD. Under the cluster provisions, a developer is allowed to build at increased densities if he preserves a proportion of his land as open space, public purpose space or donated public purpose space. If 20% of the gross project area is devoted to open space, the allowable densities are increased from normal R-40 and R-80 densities of 1 unit and y2 unit per acre to 1-1/3 and 5/6 units per acre, respectively. If an additional 20% of the land is devoted to public purpose space, a further increase in density to 1-2/3 unit and 1.0 units per acre is allowed. Furthermore, if the public purpose land (in an R-40 zone) is dedicated to the township, the developer may build 20% of all his units in attached buildings of four units each. A cluster development must be a minimum of 25 and maximum of 150 acres.
The credible proofs indicate that the clustering provisions are unlikely to have a significant impact on the cost of housing as the low densities and limits on numbers of attached units make significant economies of scale unlikely, and therefore, according to plaintiffs’ experts, cluster development may not occur at all. Yet even under the cluster provisions costs would be prohibitive to most lower income families. According to the township planner, an R-40 townhouse would cost a minimum of $29,000; plaintiffs’ expert places the average townhouse figure at $52,000. Moreover, the single family detached units which constitute 80% of an R-40 cluster would according to plaintiffs’ expert cost $64,000. The allowable densities are still too low to create significant cost savings. By comparison, Mount Laurel had a similar cluster provision allowing a density of 2.25 units per acre upon dedication of 15% of the total acreage to the municipality and reservation of 25% for public use.
The distribution of vacant and developable acreage (and total acreage) among the various zones under the ordinance shows that low density, middle and high income residential *510uses are strongly favored. Only a maximum of 2.37% of the town’s vacant developable residential acreage is zoned for multi-family apartments (AF), and the correct figure may be as low as 1.02% or 0.84%. An additional 2% is zoned R-7 for small lot attached double houses. Though 9.9% is zoned for PUD development, the location of two of the three PUD tracts makes their development highly unlikely. Using the township planner’s estimates of the potential future building capacity under the 1973 ordinance, the R-7 and AF zones account for a maximum of 16% of all housing units. By contrast, the R-80, R-40 and open space zones account for over 71% of the vacant developable residential acreage and over 41% of the "housing units. If the R-15 and R-20 zones are counted, the large-lot single-family acreage figure increases to 82% and the unit figure to almost 50%.
Y
“Least Cost” versus “Low and Moderate Income” Housing
A key consideration in this particular case as well as a factor integral to the entire problem, generally, is the well-known fact, amply corroborated by this record, that private enterprise will not in the current and prospective economy without subsidization or external incentive of some kind construct new housing affordable by the low income population and by a large proportion of those of moderate income.19 We recognized this fact in Mount Laurel. 67 N. J. *511at 170, n. 8; 188, n. 21. The amount and kind of governmental subsidies available for housing has always been fragmentary, and federal sources have recently been restricted.20 What can legally be required of municipalities by way of initiation of public housing programs and provision of zoning incentives for production of lower income housing will be discussed infra. But it will be apparent that sources extraneous to the unaided private building industry cannot be depended upon to produce any substantial proportion of *512the housing needed and affordable by most of the lower income population.
In view of the foregoing, defendant implies that the mandate of Mount Laurel is impracticable in the current economy and that litigation to enforce it is futile. Thus defendant flatly asserts in a supplemental brief: “We do not believe that substantial low and moderate income housing can be created by zoning.” However, it goes on to make an observation which appears to us to provide the clue to the only acceptable alternative recourse if in fact private enterprise cannot economically construct the housing needed for lower income families. It states:
Planned Unit Development can help by providing large amounts of additional housing some of which is in the moderate income range. The effect of new construction is also to create filtering whereby families in the moderate income group move into new housing created in the PUD zone making available existing housing for lower income families who cannot afford the new. Without subsidization, this is undoubtedly the most reasonable and certain method of creating housing opportunities for low income families.
To the extent that the builders of housing in a developing municipality like Madison cannot through publicly assisted means or appropriately legislated incentives (as to which, see infra) provide the municipality’s fair share of the regional need for lower income housing, it is incumbent on the governing body to adjust its zoning regulations so as to render possible and feasible the “least cost” housing, consistent with minimum- standards of health and safety, which private industry will undertake, and in amounts sufficient to satisfy the deficit in the hypothesized fair share. As the matter was put in a supplemental amicus brief of The Public Advocate:
* * * for now, and in the foreseeable future, it is absolutely essential to build a substantial amount of housing units at the lowest cost feasible and consistent with health and safety. Builders now must be given the opportunity to build as inexpensively as possible in order to accommodate the low, moderate-subsidized and, especially, moder*513ate-conventional population. -Thus, in one sense, future disparities in the increases in housing cost and median income are not relevant; that is, we should be building at the lowest cost feasible now.
This sentiment has also been expressed by a léading commentator who responded thus to the argument that the inability of lower income persons to afford unsubsidized housing rendered njinimal the impact of exclusionary ordinances:
As a factual matter, the situation may often be otherwise; particularly where the ordinance is enacted by an undeveloped suburb attempting to stem an urban tide, the availability of low cost housing a decade hence may be very much a function of zoning requirements today.
Sager, “Tight Little Islands: Exclusionary Zoning, Equal Protection and the Indigent”, 21 Stanford L. J. 767, 792 (1969).
Nothing less than zoning for least cost housing21 will, in the indicated circumstances, satisfy the mandate of Mount Laurel. While compliance with that direction may not provide newly constructed housing for all in the lower income categories mentioned, it will nevertheless through the “filtering down” process referred to by defendant tend to augment the total supply of available housing in such manner *514as will indirectly provide additional and better housing for the insufficiently and inadequately housed of the region’s lower income population. See also Mount Laurel, 67 N. J. at 205 (Pashman, J., concurring).22
It will be apparent from our survey of the facts and the discussion hereinafter that the 1973 ordinance under review not only fails to provide directly for Madison’s fair share of the region’s low and moderate income' housing needs but also is not geared to satisfy such a share in terms of “least cost” housing in the sense just described. The failure will be seen to be both quantitative and qualitative. Insufficient areas are zoned to permit such housing, and the zoning restrictions are such as to prevent production of units at least cost consistent with health and safety requirements.
VI
Incapacity \of the 1973 Ordinance to Affect Adequate Lower Income Housing
As we shall indicate hereinafter, Madison’s planners have, for purposes of this litigation, formulated a study purporting *515to demonstrate an unmet need in Madison for low and moderate income housing, as its share of a larger Middlesex County regional need therefor, of some 1800 housing units as of 19 75.23 Assuming, for present purposes, the legitimacy of that estimate — a matter for discussion in a later section of this opinion — the evidence is convincing that the 1973 zoning ordinance does not hold the promise of an opportunity to meet that need and at the same time satisfy the prospective continuing need in the foreseeable period following 1975. This, in effect, was the substance of Judge Furman’s holding, quoted earlier herein. From his conclusions,24 amply supported by the record, it appears that no new housing is feasible under the ordinance for persons in the bottom income third of the population (under $9,000); that at most 12% (3500/30,000) to 17% (3500/20,000) of all new housing units are attainable by persons earning $10,00025 a year; so that 83% to 88% of the feasible future units would be zoned out of the reach of the lowest 40% of the population.
The anatomy of these failures is apparent.
In our analysis of the minimum lot single-family and multi-family zones in IVB, supra, the prima fade disproportion of land zoned for high cost residences vis-ovis that *516zoned for lower cost residences and multi-family units was fully canvassed. According to the proofs here adduced, there is little or no present market for the R-80 and R-40 zoning as such. Cf. Schere v. Township of Freehold, 119 N. J. Super. 433 (App. Div.), certif. den. 62 N. J. 69, cert. den. 410 U. S. 931, 93 S. Ct. 1374, 35 L. Ed. 2d 593 (1972). While the R-7 zone may permit a marginal amount of moderate income housing, new home ownership in that zone is precluded for the low income population. While the latter condition may be economically unavoidable (see V, supra), Madison has provided for no home ownership at all on “very small lots”, as mandated by Mount Laurel. 67 N. J. at 187. Clearly no effort was made to permit “least cost” single family homes — and certainly not in reasonable numbers.
We have further seen that the multi-family zoning regm lations are not only substantially deficient in areas of developable vacant land made available therefor, but also defective in their susceptibility to entrepreneurial concentration in one-and two-bedroom configurations.
Mount Laurel requires that a municipality must allow for “an appropriate variety and choice of housing.” 67 N. J. at 174. This obligation extends to all types of housing and includes moderate and large sized multi-family units. In MiOunt Lmrel, bedroom limitations were considered “so clearly contrary to the general welfare as not to require further discussion.” 67 N. J. at 183. Although the express bedroom restrictions of the 1970 Madison ordinance were' excised in 1973, the maximum bulk and density regulations in the AR and PUD zones (the sole sites of multi-family units),26 when combined with the economics of building, effectively dictate development on an 80% one bedroom, 20% two bedroom mix, and such a combination was within the contemplation of the township planners. This is not an *517inevitable result of zoning and economics, for a municipality through the zoning power can and should affirmatively act to encourage a reasonable supply of multi-bedroom units affordable by at least some of the lower income population. Such action should include a combination of bulk and density restrictions, utilization of density bonuses,27 minimum bedroom provisions and expansion of the EAR ratio in the AE zone to encourage and permit larger units.
Although the validity of a zoning provision for density bonuses in the sense stated in note 27, supra (as distinguished from unit bonuses for rental concessions — i. e., “rent skewing”), has not been argued in this case, we see no objection to it in principle. Comparable bonuses are expressly permitted by the statute in relation to PUDs and clustering. N. J. S. A. 40:55-57(b) (2) and (3). While there is no express statutory sanction for a density bonus provision outside the PUD context, this type of regulation is directly tied to the physical use of the property and is thus within the recognized ambit of the zoning power. There was unanimity of opinion among the experts at the trial herein that such a device is a vital weapon in the armament of affirmative zoning for adequate housing of families in all income categories. Recognizing that the objectives of Mount Laurel are essential to the effectuation of the general welfare, and are within the broad legislative delegation to municipalities of both the zoning and the general police power, see Ward v. Scott, 11 N. J. 117 (1953); 16 N. J. 16 (1954), and Inganamort et al. v. Bor. of Fort Lee, et al., 62 N. J. 521 (1973), we hold that provision for density bonuses in the sense indicated is within the municipal zon*518ing power, and, in situations such as that here presented, is a necessary implement in the encouragement of builders to provide multi-family housing for those of lower income. Cf. Mount Laurel, 67 N. J. at 170, n. 8.
We are constrained to take a more reserved position as to the validity of zoning provisions for “rent skewing”, or the allowance of greater density in either sale or rental accommodations in exchange for special concessions by the developer of rental or sale price of a limited number of units. Although this is also a widely recommended zoning technique for handling the problem of encouraging private construction of lower income housing,28 we discern serious problems with the exercise of local zoning power in such a manner without express legislative authorization. See Board of Supervisors v. De Gnoff Enterprises Inc., 214 Va. 235, *519198 S. E. 2d 600 (Sup. Ct. 1973); Annot. 62 A. L. R. 3d 880 (1975). We will not here resolve the issue in the absence of adequate argument on the matter. However, we are not to be understood as discouraging local initiative in this area; the question, moreover, deserves legislative study and attention.
It seems useful to point out, in connection with the revision of the ordinance which will be required by our judgment herein, that sound planning calls for providing for a reasonable cushion over the number of contemplated least cost units deemed necessary and believed theoretically possible under a particular revision. Plaintiff adduced testimony that a reasonable margin over any formulaic quota was necessary in order to produce any likelihood of achievement of the quota. The reasons are evident. Many owners of land zoned for least cost housing may not choose to use it for that purpose. And developers of least cost housing may not select all of the zoned land available therefor, or at least not within the anticipated period of need. Thus over-zoning for the category desired tends to solve the problem.29
The PUD provisions do not contribute to the desideratum of Mount Laurel. The provisions were analyzed *520in IVB, supra. We assume, for present purposes, that PUD zoning is valid. It has not been attacked in this case, and the relevant legislation, N. J. S. A. 40:55-54 et seq.; L. 1975, c. 291, Secs. 2 k, 29.1 b, as in the case of all legislation, is presumptively valid. Cf. Mount Laurel, 67 N. J. at 166-167, n. 5. In any event, it is a corollary of Mount Laurel that when municipal exactions from developers .reach such proportions as to exert an exclusionary influence, whether in a PUD or any other context, they offend the constitutional precept of Mount Laurel and must be remedied.
As pointed out by Heyman and Gilhool in their penetrating study of the rationale for upholding subdivision requirements: “But such exactions raise the spectre of exclusion: arguably they will add so to the cost of suburban housing as to exclude an even larger portion of lower income and nonwhite population than is presently relegated to life in the central cities by the higher suburban costs.” “The Constitutionality of Imposing Community Costs of New Suburban Residents Through Subdivision Exactions”, 73 Yale L. J. 1119, 1155 (1964). The authors conclude, however, that the exclusionary impact of such exactions “will be strikingly slight because legislative and judicial pressures will tend to require the establishment of reasonable ceilings.” Ibid. Cf. Berger, Land, Ownership and Use 786, 787 (2d ed. 1975).
Heyman and Gilhool found subdivision exactions ranging from $37.50 to $325 per lot to he reasonable. Id. at 1156. In contrast, a $2.2 million expenditure for schools in a Class II 300 acre PUD with a maximum capacity of 1,725 units yields a per unit exaction of $127530 which does *521not seem to be reasonable even at the high price levels of 1974.31 This requirement must be omitted in the revision of the ordinance.
The requirement for the provision by the PUD developer of r-oads, water and sewage facilities presents a different situation. Such statutorily authorized municipal requirements for approval of subdivision plats have generally been held to be valid exercises of the police power. Brazer v. Borough of Mountainside, 55 N. J. 456 (1970). Nevertheless, as already noted, application of the police power through zoning cannot be had in a manner contrary to the general welfare. Mount Laurel, 67 N. J. at 175.
Other limitations on this statutory power have previously been recognized. Cf. Divan Builders v. Planning Bd. Tp. of Wayne, 66 N. J. 582 (1975). There this court considered a prerequisite to subdivision approval that the developer contribute toward the cost of an off-site drainage facility without allocating any part of the cost to other properties specially benefitted by the improvement, and found it invalid. Similarly, in Longridge Builders, Inc. v. Planning Bd. of Princeton Tp., 52 N. J. 348 (1968), the court condemned a condition to subdivision approval that plaintiff pave an off-site right of way, resulting in the imposition of the entire cost upon' plaintiff, when other lands would benefit from the improved road and there were inadequate standards and procedures for the allocation of costs.
In this light, the $300,000 and $600,000 costs involved in bringing roads and utilities to the removed PUD sites (a distance of up to two miles, compared to the 361-foot distance of the improvement involved in Longridge, *522supra) bear examination. To the extent that these costs do not prohibit development, they nonetheless add sufficiently to final costs as to tend to have an exclusionary impact. Further considerations highlight the questionable validity of these requirements.
Only a limited area has been denoted PUD (9% of the township’s vacant developable residential acreage), and 2/3 of it is in remote areas. This is not a case of a substantial category of land being reasonably zoned as a whole, with only a minor portion excessively burdened. Quite the opposite. The municipality could well have located these PUDS in more accessible areas of the town had it been motivated to render housing for lower income families more available. However, the record clearly shows that the sites were deliberately chosen in order to force the PUD developers and their customers to carry the burden of developing these remote areas. The township planner testified that the “decision was made that two PUDs would be the incentive to complete and bring around the water system * * * to provide a main system that would rationally serve 1/3 of the township.” This kind of motivation is echoed in the minutes of the planning board and township council where it was noted that because of the locations chosen the developers would bo required to build portions of the trans-Madison highway as well as to widen and improve Union Hill Road. Such expenditures by PUD developers would directly benefit the owners of other property advantaged by the added facilities and paved roads yet not required to contribute a proportionate amount of the cost.
The potential impact of the water and sewer line requirements is shown by the conclusion of the Middlesex County Planning Board, in reviewing the PUD ordinance, that the two remote PUD areas would probably not be developed at all within the next ten years. Under the totality of the stated circumstances, it must be concluded that a prima facie, case of exclusion has been made out with respect to the road and facility requirements, and the burden shifts *523to the municipality to justify those provisions of the ordinance. Cf. Mount Laurel, supra, 67 N. J. at 181. As the municipality has not met its burden, the municipality will be directed on remand to do one or more of the following in the course of revision of the ordinance (if it continues in its position that the PUD provisions partly meet its obligation to zone for least cost housing) : (1) eliminate these requirements or revise them to render them not exclusionary; (2) require proportionate donation by other property holders; or (3) relocate these or other PUD tracts nearer to utility hookups.
A third potentially cost-generating requirement in the PUD provisions is the approval process, which allegedly results in high carrying charges and has the potential for delay effectively barring the project. The approval process is a three-stage procedure, which, according to the township’s witness, may be completed within a year, but which plaintiffs’ experts testify would take eighteen months to two years.
Admittedly, a protracted approval process will add greatly to the cost of any project and hence may tend to render development prohibitive to lower income users. The evidence as to the cost impact of the Madison PUD provisions was contradictory. As noted, they were adopted pursuant to the PUD enabling legislation, N. J. S. A. 40:55-54 to 67. The statute undertakes to replace existing multifarious procedures with “an expeditious method of processing a plan for a planned unit development * * * to avoid the delay and uncertainty” inherent in the other procedures. N. J. S. A. 40:55-59. It then proceeds to lay out a two-stage approval process which the Madison ordinance follows with one exception. The ordinance adds a third “informal preliminary application” stage which must be completed prior to the filing of an application for tentative approval. This “informal” stage adds approximately 40 days. Without it, *524the entire process on its face takes 150 days.32 Consequently the township’s testimony that the procedure may be completed within nine months is supportable on the record. Except for the “informal preliminary application” stage, the procedure is valid. The latter, being unduly cost-generating, and not contemplated by the statute, should be eliminated.
We need add no more to the discussion above in IY B of the inadequacy for fair share purposes of the cluster provisions of the ordinance.
In summation of this point, the 1973 Ordinance is shown not to provide the opportunity for a substantial amount of new housing which could be available to the lower income segments of the population. This failure arises from both (a) the inadequacy or non-existence of areas zoned for homes on very small lots or for multi-family housing; and (b) the undue cost-generating features inherent in the ordinance which raise the expense of purchasing or renting new housing units above the reach of the great majority of the lower income population.
YII
The “Fair Share” Approach of the Defendant
We made the preliminary observation in II, supra, that although we would neither make nor require a finding of *525fact as to a given number of lower income bousing units to be made possible by the Madison Zoning Ordinance, we would, for 'the purposes there enunciated, nevertheless, discuss the evidence herein .concerning Madison’s fair share of a regional need for such housing. By way of further preliminary, we adhere to the broad principle of Mount Laurel that each developing municipality must by its zoning ordinance provide the opportunity for a fair share of the lower income housing needs of its region. We intend that our judgment herein shall subserve that principle notwithstanding that we do not propose to, nor require that the trial court shall demarcate specific boundaries for a pertinent region or fix a specific unit goal as defendants’ fair share of such housing needs.
Defendant undertook at the trial to establish what would constitute a fair share of the regional need for low and moderate income housing for Madison’s zoning ordinance to render possible. It relied upon two allocation studies, one made in 1972 on behalf of the Middlesex County Planning Board by a planner named Kim (“Kim study”); the other made on behalf of the Madison Township Planning Board in 1974, during the litigation below, by its planning adviser Abeles and his firm (“Abeles study”). The Abeles study, in turn, partly relied on an estimate of Madison’s unmet lower income housing needs as of 1975, made in 1970, by the Middlesex County Planning Board (“County study”), updated by Abeles in 1974. Both the Kim and the Abeles studies took Middlesex County as- the relevant region, making allocations of “fair shares” of lower income housing for each of the 25 municipalities in the county. The allocations of the two studies for Madison were relatively close, the Kim study arriving at about 1600 lower income units, and the Abeles study, about 1800 units, both as of 1975. Neither, however, ventured a prognostication as to Madison’s continuing fair share of the region’s needs on an annual or *526other periodic basis for the foreseeable future beyond 1975.33 In the latter regard, however, Abeles foretold that Madison’s housing unit growth (total) until 1975 would be at the annual rate of 600-800 units and from 1975 to 1980 at 800-1000. This was not broken down between lower income and other categories.
By comparison with the foregoing, Davidoff, plaintiffs’ planning expert, estimated an unmet need in Madison as of 1975 of 3,000 lower income units, and a continuing need thereafter into the 1980’s of 1000-1200 units annually (total) of which 500-600 should be lower income.34
The county study indicated an unmet county need for lower income housing units as of 1975 of 23,600 units. Significantly, the study stated:
“To meet this deficit, increases in public and private resources will be required. For the county and its municipalities these resources would include the freeing up of more land for residential development of a particular density and location commensurate with socio-economic function of units required * * *” (emphasis added).
The plan of the Kim study was to estimate what the demand (need) for low and moderate income housing in the county would be by 1975 and to compare it with the prospective supply of such units by that date. The pertinent lower income category used was those households with less than $12,000 annual income as of 1970. Kim estimated a 1975 county population of 766,946 (583,000 in 1970) and *527the number of persons employed as 390,700 (341,000 in 1970). The estimates were based on expected movements of people and jobs from large urban centers to the suburbs. In general there would be an “expansion” from the New York metropolitian region to the outer boundaries of Middlesex County and to the south and west thereof.
Kim estimated a demand for lower income housing in 1975 of 136,374 households or units (111,301 in 1970) to which he added a factor of 6.5% for normal vacancy rate, or a total 1975 demand of 134,589.
The Kim approach to the estimate of anticipated housing deficiency was to balance prospective employment in the county with housing, following the approach of the TriState Regional Planning Commission. His study disclosed a 1975 housing supply of 115,791 lower income units, which, subtracted from demand, left a 1975 deficit of 18,798 lower income units.
The Kim allocation (fair share) of the deficit among the 35 municipalities of the county was based on a complex mathematical formula reflecting the following factors: (a) housing location in relation to work place; (b) housing location in relation to housing construction costs, i. e., land costs per unit; (c) other “constraints”, e. g., the subregion’s capacity to absorb more housing. The application of the formula to Madison was to allocate to it a “fair share” of 8.4%, or about 1600 units.
The Abeles study avowed as a goal “increasing the housing supply for all income groups.” The new housing “should be provided in a mixture of housing types and costs at various locations in the township.” The job dispersion outside Middlesex County of Madison’s resident work force as compared with the work force resident in the county as a whole is indicated by the percentages of 50% for the township and 36% for the county. Correlatively, Madison has 8% of the population of the county but only 0.8% of the jobs.
*528The Abeles study averred that important to a fair share-regional plan is the consideration of the relevant “housing market area”. This is defined as “the geographic area in which housing units are in competition for the people who are seeking housing”. The study conceded that “Madison Township is a relatively small part of a market area which could encompass virtually all of Central New Jersey,” but stated that its purposes would be met by assuming a housing market area confined to Middlesex County.35
The Abeles study projected a growth of 18,000 households in Middlesex County from’ 1970-1975 and another 15,000 between 1970 and 1980. This growth will take place primarily in municipalities with an ample supply of vacant land. The existing supply of housing in the county is very limited, as indicated by the abnormally low vacancy rates (1% in the county in 1970; 1.5% in Madison). Sixty minutes is a “commonly acceptable limit for commutation.” The highway network allows for considerable mobility and creates substantial housing demand in Madison. *529Unlike the Kim study, the Abeles study did not approach the concept of housing needs from a demand and supply relationship, but rather from that of the number of obsolete housing units in existence. Obsolescence was defined as (a) undesirable physical conditions; (b) overcrowding of occupants of unit; and (e) excessive share of income being paid as rent. The total thus determined was found to concur generally with a 1974 revision of the 1970 estimate by the county study of an unmet need as of 1975 for lower income housing in the county. The revision increased the original estimate of 33,600 to 39,351 units.
The county need was allocated as “fair shares” to municipalities in accordance with a formula generally resembling that devised by the Metropolitan Washington Council of Governments (COG). This involves housing need factors and housing supply factors. To those the Abeles study added certain modifiers. The need factors were generally those aforementioned related to obsolescence. The supply factors comprise vacant residential land as zoned and vacant housing units. The modifiers were (1) per capita financing resources; (3) the existing ratio of low and moderate income housing; (3) a multi-family housing index based on growth in multi-family units from 1960-1970; and (4) a housing density index, i. e., the ratio of housing to total land area.
Application of the resulting formula to the 39,351 county units needed ascribed to Madison a “fair share” of 6.1%, or about 1800 lower income units. As already noted, the Abeles study made no prognostication as to a fair share of the continuing housing need subsequent to 1975.
The post-litem, motivation of the Abeles study is apparent. It was approved by the township planning board April 33, 1974, two days before the end of the second trial and over 18 months after adoption of the 1973 ordinance. The analysis of “need” is largely weighted in favor of indigenous need, i. e., of those persons already in the township, as compared with potential aspirants for housing from *530outside the municipality. Of the 1784 units which the Abeles study attributes as Madison’s fair share, 1394 units will represent the indigenous need and only 390 the need of the region (Middlesex) .36
In several respects, over and above the apparent limitation of the region to the county, the formula presented by the Abeles study appears to be a self-serving one. First, the original COG formula contained an accessibility factor as one of the modifiers which Abeles omitted, ostensibly on the ground that a substantial proportion of all employment opportunity in Middlesex County lies within 45 minutes commuting time from any municipality therein. However, this unwarrantably favors Madison. As it is nearer to heavy employment centers such as Perth Amboy and Woodbridge than are places like Monroe and Cranbury, relatively more people would be likely to want to live in Madison, and correspondingly Madison’s comparative fair share would be larger if accessibility were taken into account.
The inclusion of a multi-family index and a housing density index are two additional self-serving features of the Abeles study. These factors did not appear in the COG formula, and their use results in a substantial credit in Madison’s favor. While the housing concentration factor defensibly affords some recognition to those subregions which “have already contributed to the supply of rental housing for low and moderate income families,” Abeles nevertheless takes duplicative credits for both Madison’s increase over the past ten years in multi-family units and for its existing housing density. As a result of the interplay of these factors, Madison, with 11,142 vacant and developable acres, ends up with a 6.1% allocation compared to towns of similar amounts of vacant residential acres, Monroe (12,067 acres), and *531South Brunswick (10,778 acres), which, under the Abeles formula, receive fair share allocations of 16.1% and 12.6% respectively. Towns with closely corresponding fair share allocations, East Brunswick (7.5%), North Brunswick (6.0%), and Woodbridge (5.0%), have vacant residential acres of 4,722, 986, and 230 respectively.
VIII
“Fair Share” and “Region”
— General Considerations
The probative value of the Kim and Abeles fair share studies should be appraised against the background of the substantial body of experience that has been developed by governmental planning bodies in recent years in devising fair share plans for voluntary housing planning purposes as distinguished from litigation. All of them involve realistic housing market areas larger and functionally more appropriate, in Mount Laurel terms, than the small Middlesex County region. Before discussing those specific plans, some preliminary observations as to the concepts of “fair share” and “region” seem appropriate by way of background.
Of primary significance is the difference between the situation of an administrative planning agency functioning under authorizing legislation and that of a court dealing with an attack by litigation on the adequacy of the zoning ordinance of an isolated municipality. The former is dealing with a comprehensive, predetermined region and can render or delegate the making of allocations with relative fairness to all of the constituent municipalities or other subregions within its jurisdiction.37 Moreover, it presumably has ex*533pertise suited to the task. The correlative disadvantages of a court adjudicating an individual dispute are obvious.38
The formulation of a plan for the fixing of the fair share of the regional need for lower income housing attributable to a particular developing municipality, although clearly envisaged in Mount Laurel, 67 N. J. at 162, 189-190, involves highly controversial economic, sociological and policy questions of innate difficulty and complexity. Where predictive responses are called for they are apt to be speculative or conjectural.39 These observations are supported not only by the published literature40 but by the proofs and comprehensive briefs supplied us by the parties and amici.
*534Some of the problems catalogued above were touched upon in Mount Laurel, e. g., “region”, 67 N. J. at 162, 189— 190; incidence of subsidized construction in contemplation, id. at 170, n. 8, 188, n. 21; sources of reliance fox “fair share” guidance, id. at 190; quantity of needed housing reasonably expectable under proper zoning, id. at 188, n. 21. We take this occasion to make explicit what we adumbrated in Mount Laurel and have intimated above — that the governmental-sociological-economic enterprise of seeing to the provision and allocation throughout appropriate. regions of adequate and suitable housing for all categories of the population is much more appropriately a legislative and administrative function rather than a judicial function to be exercised in the disposition of isolated cases.41 Cf. 67 N. J. at 189, n. 22, 190.
*535Fortunately, the other branches of government are giving the matter their attention.42 But unless and until other ap*536propriate governmental machinery is effectively brought to bear the courts have no choice, when an ordinance is challenged on Mount Laurel grounds, but to deal with this vital public welfare matter as effectively as is consistent with the limitations of the judicial process.
We address the question, implicated by defendant’s evidential studies, of the appropriate concept of a “region” in the context of a litigation challenging the housing adequacy of a particular zoning ordinance. Defendant purports to justify its fair share allocation on the basis of a single county as a region. However, both the Kim and the Abeles studies, in estimating anticipatorily the need for and supply of housing as of 1975 in Middlesex County, apparently recognized the influence of growth of population and jobs emanating from the broader region of northeastern New Jersey and the New York metropolitan region. Thus, while it is questionable whether the functionally relevant regional housing need was adequately realized in the Kim and Abeles studies, it does appear that in effect they envisaged a need emanating beyond the county boundaries.
The technical details of the basis for fair-share allocations of regional goals among municipalities, pertaining as they do to an area of considerable complexity and theoretical diversity, are not as important to a reviewing court concerned with effectuating Mount Laurel objectives as the consideration that the gross regional goal shared by the constituent municipalities be large enough fairly to reflect the full needs of the housing market area of which the subject municipality forms a part.
*537In broad principle, we believe Judge Furman was correct in conceiving the appropriate region for Madison Township as "the area from which, in view of available employment and transportation, the population of the township would be drawn, absent invalidly exclusionary zoning”. 128 N. J. Super. at 441. This is essentially like the housing market area concept espoused in the Abeles report as sound in principle, although not directly employed in the Abeles fair share study.
The concept of a county per se as the appropriate region was thought not to be "realistic” by Justice Hall in writing Mount Laurel. He there said (67 N. J. at 189-190):
The composition of the applicable “region” will necessarily vary from situation to situation and probably no hard and fast rule will serve to furnish the answer in every case. Confinement to or within a certain county appears not to be realistic, but restriction within the boundaries of the state seems practical and advisable. (This is not to say that a developing municipality can ignore a demand for housing within its boundaries on the part of people who commute to work in another state.)
Justice Hall defined the region applicable there as “the outer ring of the South Jersey metropolitan area, which area we define as those portions of Camden, Burlington and Gloucester Counties within a semicircle having a radius of 20 miles or so from the heart of Camden city.” 67 N. J. at 162, 190. What was material to that determination was the proximity of Mount Laurel to the highly urbanized Camden area, its residential development due to the influx of new residents from nearby central cities, existing and projected employment patterns, the "highway network” linking Mount Laurel with all parts of the Camden area and the contrast of its vacant acreage (65%) with the land supply situation in those nearby central cities. See 67 N. J. at 161-162.
Eor purposes of our present problem, we distinguish the situation with which we would be confronted if the municipality whose ordinance was under attack had been the subject of an official fair share housing study of a group of *538counties or municipalities conducted under such auspices as the DVPRC or the planning boards of a county or group of counties functioning under Executive Order No. 35 (see note 42, supra,). We conceivably might regard a “region” so constructed, and the dependent fair share allocations thereby arrived at, as meriting prima facie judicial acceptance. The Kim and Abeles “studies” before us do not have that authority or stature, and we have accorded them only such weight as they deserve on their merits, as analyzed above.43
For examples of regions large enough and sufficiently integrated economically to form legitimately functional housing market areas, we turn to some of the pioneering fair *539share allocation, plans executed under official or quasi-official auspices. The Miami Yalley (Dayton, Ohio) Regional Planning Commission includes five counties and 31 municipalities as far as 60 miles from the center of Dayton. The Metropolitan Washington GOG (see supra p. 529) covers 15 counties and local governmental jurisdictions, including the District of Columbia, San Bernardino County, California, although a county, occupies 20,000 square miles. The Metropolitan Council of the Twin Cities (Minneapolis-St. Paul) covers 7 counties, including almost 300 jurisdictions, with a total population of 1.9 million. The DVRPC, as already shown, comprises nine counties in Pennsylvania and New Jersey. The present significance of the cited plans is that their regions are of such size that it is difficult to conceive of a substantial demand for housing therein coming from any one locality outside the jurisdictional region, even absent exclusionary zoning. The essence of the cited plans is “to provide families in those economic categories [low and moderate] a choice of location.” 16 Trends on Housing, No. 2 p. 2 (1972).
We thus proceed to formulation of our position as to the concept of region in the context of an ad hoc application of Mount Laurel principles to a single litigated ordinance, having in mind our determination in II, supra, that it would not generally be serviceable to employ a formulaic approach to determination of a particular municipality’s fair share. We conclude that, in general, there is no specific geographical area which is necessarily the authoritative region as to any single municipality in litigation. Different experts may quite reasonably differ in their concepts of the pertinent region. See Lindbloom, “Defining 'Fair Share’ of 'Regional Need’”, 98 N. J. L. J. 633-644 (July 24, 1975). But in evaluating any expert testimony in terms of the Mount Laurel rationale, weight should be given to the degree to which the expert gives consideration to the areas from which the lower income population of the municipality would substantially be drawn absent exclusionary zoning. (Evidence of the historical sources *540of a municipality's population, among other indicia, is releven! thereto.) This is broadly comparable to the concept of the relevant housing market area, to which there has been prior reference herein.
The factors which draw most candidates for residence to a municipality include not only, for employed persons and those seeking employment, reasonable proximity thereto of jobs and availability of transportation to jobs, as mentioned by Judge Eurman and stressed by most of the experts,44 but *541proximity to and convenience of shopping, schools and other amenities. Retired people, who represent a substantial part of the lower to moderate income population, might be attracted from a greater distance than employed people.
Einally, we submit general observations as to the techniques of “fair share” allocation to municipalities within an assumedly valid region. There is much greater diversity among the experts in this regard than in relation to determining pertinent regions. Moreover, as already noted herein, harm to the objective of securing adequate opportunity for lower income housing is less likely from imperfect allocation models than from undue restriction of the pertinent region. The essential thing from that standpoint is that the true regional need be adequately quantified.
The trial court specified that for Madison to meet its fair share of the housing needs of the region its zoning ordinance must approximate “in additional housing unit capacity the same proportion of low-income housing as its present low-income population, about 12%, and the same proportion of moderate-income housing as its moderate-income population, about 19%”. The 1973 ordinance was held “palpably short” of these requirements. 128 N. J. Super. at 447.
Mount Laurel devised no formula for estimating “fair share”, but the matter was left for the municipality to apply the expertise of the “municipal planning adviser, the county planning boards and the state planning agency.” 67 N. J. at 190.
The number and variety of considerations which have been deemed relevant in the formulation of fair share plans is such as to underscore our earlier observation that the entire *542problem involved is essentially and functionally a legislative and administrative, not a judicial one.45
*543The formula specified by the trial court would not necessarily be properly utilizable in other contexts. Some municipalities might have such very high or very low existing proportions of lower-income families in their population makeup as to render such a formula patently unfair. If the existing municipal proportions correspond at least roughly with the proportions of the appropriate region the formula would appear prima facie fair. The evidence herein is that the stated municipal proportions approximate those of the county of Middlesex. We are without data as to the comparative proportions of such a larger area as would include the more urban counties in the northeast New Jersey region.
Harking back to our statement in II as to why we proposed in this opinion to discuss the concepts of fair share and region notwithstanding that we would not, nor would we require the trial court to specify a pertinent region or fix a fair share housing quota for Madison, we summarize the observations in YII and VIII as follows:
1. Based upon our analysis and findings in IV and VI, the 1973 ordinance is clearly deficient in meeting Madison’s obligation to share in providing the opportunity for lower cost housing needed in the region, whether or not the specific fair share estimates submitted by defendant are acceptable. Those estimates are, in any event, defective at least in not including prospective need beyond 1975.
3. The objective of a court before which a zoning ordinance is challenged on Mount Laurel grounds is to determine whether it realistically permits the opportunity to provide a fair and reasonable share of the region’s need for housing for the lower income population.
3. The region referred to in 3 is that general area which constitutes, more or less, the housing market area of which the subject municipality is a part, and from which the prospective population of the municipality would substantially be drawn, in the absence of exclusionary zoning.
4. Fair share allocation studies submitted in evidence may be given such weight as they appear to merit in the *544light of statements 2 and 3 above. But the court is not required, in the determination of the matter, itself to adopt fair share housing quotas for the municipality in question or to make findings in reference thereto.
IX
Environmental Considerations
As noted above, a considerable amount of vacant acreage in Madison borders certain streams or comprises important aquifer storage and discharge areas. Depositions and counter-depositions were taken by defendant and plaintiffs, respectively, bearing upon the effect of development of varying kinds on such areas as Burnt Ply Bog, the Old Bridge Sands, Raritan Bay beachfront, the salt marshes behind Raritan Bay and the four streams flowing into South River. Defendant offered the depositions at the trial to establish that certain areas zoned R-80, R-40 and RP were so sensitive to flood, water contamination and related problems that they should be kept from development at all or restricted to very low residential density. The trial judge declined to consider this evidence on the ground that considerable other land, free from such ecological considerations, and amenable to higher density development, was available within the township with which it could meet its fair share obligation for its own and the region’s housing needs. 128 N. J. Super. at 447.
Plaintiffs’ experts testified on depositions that the answer to the ecological problems posed was not prohibition or regulation of the density of development per se but careful use of the land, with adequate controls in respect of construction, sewerage, water control and treatment, sufficient open space per structure and other services.
Ecological and environmental considerations were also advanced by the municipality in Mount Laurel to justify large lot zoning throughout the township. We pointed out there that while such factors and problems were always to be given consideration in zoning (see 3 Williams, American *545Land Planning Law (1975) § 66.12, pp. 30, 34-35), “the danger and impact must be substantial and very real (the construction of every building or the improvement of every plot has some environmental impact) — not simply a makeweight to support exclusionary housing measures or preclude growth * * 67 N. J. at 187.
Notwithstanding the foregoing, we conclude the trial court erred in not receiving in evidence and giving consideration to the environmental depositions mentioned. It is not an answer to say there is ample other land capable of being deployed for lower income housing. The municipality has the option of zoning areas for such housing anywhere within its borders consistent with all relevant considerations as to suitability. There are proponents of scattering lower income housing widely throughout a municipality as well as adherents of segregating such housing in limited areas. Since the municipal fathers should have the widest latitude of judgment in that regard, it is in the interests both of the municipality and the plaintiffs that the parties have the benefit of findings by the court, from the proofs, as to exactly which of the allegedly environmentally sensitive areas, if any, are in fact not susceptible of housing development at all; which, of only low density development; and which are free of any environmental constraints in respect of density or type of housing.
We shall, in the remedial portion of this opinion, be directing amendment of the ordinance to add substantial areas to districts zoned for multi-family housing and for single-family homes on very small lots. It therefore will be necessary for the governing body of defendant to be apprised, from the findings of the court in the respects just noted, what part of the areas claimed by defendant to be environmentally unsuited for such zoning need not be resorted to for that purpose.
The court may, in its discretion, permit the depositions to be amplified at a hearing on remand, but any such hear*546ings and the added findings here directed to be made shall be expedited.
In concluding this point, however, we find no basis in the record for determining that, in any view of the environmental proofs, defendant does not have sufficient vacant developable land free from disabling ecological considerations to enable it to create the zoning opportunity for its fair share of the region’s need for least cost housing.46
X
“Affirmative Action” for Lower Income Housing
Plaintiffs and supporting amici press for a judicial mandate that developing municipalities be required affirmatively to act for creation of additional lower income housing in more ways than by eliminating zoning restrictions militating against that objective. Of the devices which have been suggested to this end, tax concessions and mandatory sponsorship of or membership in public housing projects must be summarily rejected. Tax concessions would unquestionably require enabling legislation and perhaps constitutional amendment. While we have described the sponsorship of public housing projects as a moral obligation of the municipality in certain specified circumstances, Mount Laurel, 67 N. J. at 192, we have no lawful basis for imposing such action as obligatory. It goes without saying, however, that the zoning in every developing municipality must erect no bar or impediment to the creation and administration of public housing projects in appropriate districts. See also id. at 211-212 (Pashman, J., .concurring).
We have hereinabove indicated that provision by zoning for density bonuses keyed to quantitative or bulk concessions *547by the builder (e. g., added bedrooms) is both valid and mandatory where necessary to achieve sufficient suitable least-cost housing, but that we are not -prepared presently to pass upon the validity of zoning for bonuses keyed to rental or sale price concessions.
Various additional suggestions for encouraging the proliferation of lower cost housing on municipal initiative are set forth in the supplemental amicus brief of The Public Advocate but are not deemed to require comment here as none warrant mandatory imposition in any revised Madison ordinance. See also Mount Laurel, 67 N. J. 209-213 (Pashman, J., concurring).
XI
The Validity of the Zoning Statute
Plaintiffs’ original cross-appeal from the trial court’s first determination assailed that portion of the court’s decision upholding the constitutionality of the zoning enabling act, N. J. S. A. 40:55-30 et seq., and their second brief to this court incorporates the supporting arguments. Plaintiffs’ novel contention is that the general zoning purposes stated in N. J. S. A. 40:55-32, although adequate when enacted in 1928, today fail to provide detailed standards to guide municipalities in their exercise of the zoning power, to wit, they fail to direct that a municipality must be racially and economically “inclusionary” rather than exclusionary.
The stated argument was formulated before we decided Mount Laurel and is basically mooted by our holding there,, in effect, that the zoning statute is to be construed to conform with state due process and equal protection so as to compel zoning in developing municipalities to affirmatively combat exclusion of the lower income population needing housing.47
*548In any event, Ave find the contention to lack intrinsic merit. The statute expressly sets out as a standard for the exercise of the zoning power the promotion of health, morals and the general welfare as well as other subordinate criteria. N. J. S. A. 40:55-32. As noted, a zoning ordinance contrary to the general welfare is invalid, Mount Laurel, supra, 67 N. J. at 175, and the term “general welfare” requires the consideration of regional housing needs. Ibid. Cf. Ward v. Scott, 11 N. J. 117 (1952), where a similar-attack rvas made on N. J. S. A. 40:55-39(d), providing for the issuance of use variances, on the grounds that specific standards did not accompany the delegation of power. Pinding that this section incorporated the standard of N. J. S. A. 40:55-32, and hence was valid, Justice Jacobs noted that although “the Legislature may not vest unbridled or arbitrary poAver in the administrative agency but must firrnish a reasonably adequate standard to guide it * * * [nonetheless] the exigencies of modern government have increasingly dictated the use of general rather than minutely detailed standards in regulatory enactments under the police power.” Id. at 123-4. Provision for local and regional housing needs, although not expressly enumerated under N. J. S. A. 40:55-32, is required in the promotion of the general welfare. Mount Laurel, supra.
It goes Avithout saying that the statutory and constitutional prohibition, by judicial construction, of zoning to exclude, encompasses exclusion by race as well as by economic circumstances.
The statute is valid.
XII
Belief for Corporate Plaintiffs
The corporate plaintiffs contend that-the 1973 ordinance is invalid not only generally, as exclusionary of lower in*549come housing, but specifically as to their own tract of land because of zoning restrictions which are confiscatory. They therefore ask that the court specifically order the township to place their property in an appropriate multi-family or PUD zone to be created — in effect, to grant them a permit to build the kind of moderate-to-middle income housing they have in mind.48
Plaintiffs’ expert witnesses testified that the restrictions upon their land (originally encompassing about 400 acres, but later reduced by a “Green Acres” taking to about 200 acres in the R-40 zone) were such that even using the clustering device the residences to be produced would have to sell for about $63,000, for which there was no feasible market in the number producible on the property. Plaintiffs’ witness Chester conceded that if only 30 or 40 units were involved they might be marketable, but not the hundreds contemplated by plaintiffs’ project. We cannot render an assured determination that the zoning is confiscatory against corporate plaintiffs on such proofs. Plaintiffs are not'necessarily entitled to zoning feasible for their holdings as an entirety if they are reasonably utilizable as divided in separate ownerships. Nor are they entitled to zoning permitting the most profitable development of the property. Cobble Close Farm v. Bd. of Adjustment, Middletown Tp., 10 N. J. 442, 452 (1952). Yet it appears that even if divided into smaller ownerships it would be difficult to market the property for residential uses under the zoning restrictions as they stand, having in mind the historical absence of any large lot development in the 'area. Cf. Schere v. Township of Freehold, supra (119 N. J. Super. 433).
A consideration pertinent to the interests of justice in this situation, however, is the fact that corporate *550plaintiffs have borne the stress and expense of this public-interest litgation, albeit for private purposes, for six years and have prevailed in two trials and on this extended appeal, yet stand in danger of having won but a pyrrhic victory. A mere invalidation of the ordinance, if followed only by more zoning for multi-family or lower income housing elsewhere in the township, could well leave corporate plaintiffs unable to execute their project. There is a respectable point of view that in such circumstances a successful litigant like the corporate plaintiffs should be awarded specific relief. Mytelka and Mytelka, “Exclusionary Zoning: A Consideration of Remedies”, 7 Seton Hall L. Rev. 1, 26-29 (1975); Rubinowitz, “Exclusionary Zoning: A Wrong in Search of a Remedy”, 6 Mich. J. L. Ref. 625, 668 (1973); Hartman, “Beyond Invalidation — The Judicial Power to Zone”, 9 Urban L. An. 159, 162-168 (1975); cf. Mallach, “Do Lawsuits Build Housing?: The Implications of Exclusionary Zoning Litigation”, 6 Rutqers-Camden L. J. 653, 675, 677 (1975).
There is also judicial precedent for such action. In Appeal of Girsh, 437 Pa. 237, 263 A. 2d 395 (Sup. Ct. 1970), a builder succeeded in obtaining an adjudication of the invalidity of an ordinance precluding apartment development. The municipality rezoned to create an apartment district but did not include plaintiff’s land. The plaintiff then sought to compel issuance of a permit, but the town announced the property would be condemned for a park. When plaintiff sued to enjoin condemnation, the Supreme Court ordered issuance of a permit. Order No. MP-12, 271 (August 29, 1972). Hartman, op. cit supra, 9 Urban L. An. at 161—162. The same court took similar action in Township of Williston v. Chesterdale Farms, Inc., 462 Pa. 445, 341 A. 2d 466 (Sup. Ct. 1975). See also Franklin v. Village of Franklin Park, 19 Ill. 2d 381, 167 N. E. 2d 195 (Sup. Ct. 1960).
Such judicial action, moreover, creates an incentive for the institution of socially beneficial but costly litigation such *551as this and Mount Laurel, and serves the utilitarian purpose of getting on with the provision of needed housing for at least some portion of the moderate income elements of the population. We have hereinabove referred to the indirect housing benefits to low income families from the ample provision of new moderate and middle income housing. Point V.
The foregoing considerations have persuaded us of the appropriateness in this case of directing the issuance to the corporate plaintiffs, subject to the conditions stated infra, of a permit for the development on their property of the housing project they proposed to the township prior to or during the pendency of the action, pursuant to plans which, as they originally represented, will guarantee the allocation of at least 20% of the units to low or moderate income families.49 This direction will be executed under the enforcement and supervision of the trial judge in such manner as to assure compliance with reasonable building code, site-plan, water, sewerage and other requirements and considerations of health and safety. Cf. Township of Williston v. Chesterdale Farms, Inc., supra (341 A. 2d at 468-469).
An express condition of this holding, moreover, is that the trial court, after consideration of the ecological and environmental proofs referred to in IX, supra, determine that the plaintiff’s land is environmentally suited to the degree of density and type of development plaintiffs propose. Subject to these conditions it is our purpose to assure the issuance of a building permit to corporate plaintiffs within the very early future.50
*552XIII
Remedy and Remand
We herewith modify the judgment entered in the Law Division to hold, as we did in Mount Laurel as to the ordinance there involved, that the 1973 zoning ordinance is invalid, not 'in loto, but only “to the extent and in the particulars set forth in this opinion”. Mount Laurel, 67 N. J. at 191. Eor the reasons elaborated above the ordinance is presumptively contrary to the general welfare and beyond the scope of the zoning power in the particulars mentioned. Mount Laurel, 67 N. J. at 185. The municipality has not borne its consequent burden of establishing valid reasons for the deficiencies of the ordinance. Id. at 185. It is obvious that a revision of the residential provisions of the ordinance is called for in order to provide the opportunity for that amount of least-cost housing in the township which will comply- with the directions contained in this opinion.
In Mount Laurel we elected not to impose direct judicial supervision of compliance with the judgment “in view of the advanced view of zoning law as applied to housing laid down-by [the] opinion”. 67 N. J. at 193. The-present case is different. The basic law is by now settled. Further, the defendant -was correctly advised by the trial court as to its responsibilities in respect of regional housing needs in October 1971, over five years ago. 117 N. J. Super. 11. It came forth with an amended ordinance which has been found to fall short of its obligation. Considerations bearing upon the public interest, justice to plaintiffs and efficient judicial administration preclude another generalized remand for another unsupervised effort by the defendant to produce a satisfactory ordinance. The focus of the judicial effort after six years of litigation must now be transferred from theorizing over zon*553mg to assurance of the zoning opportunity for production of least cost housing. See Mytelka and Mytelka, “Exclusionary Zoning: A Consideration of Remedies”, 7 Seton Hall L. Rev. 1, 18, 33 (1975).
The trial court on remand shall execute the directions in IX and XII above and render its findings thereon with the reasonable dispatch appropriate to the age of this litigation. It shall become the obligation of the defendant, within 90 days thereafter, unless more time is allowed by the trial court, to submit to the trial court for its approval a revised ordinance.
The revision shall zone, in the manner specified in this opinion, to create the opportunity for a fair and reasonable share of the least cost housing needs of Madison’s region, the concept of “region” to be understood as generally set forth in II and VIII hereinabove. While no formulaic determination or numerical specification of such a fair and reasonable share is required, we do not preclude it if the municipal planning advisors deem it useful. The revision shall, as minima: (a) allocate substantial areas for single-family dwellings on very small lots; (b) substantially enlarge the areas for dwellings on moderate sized lots; (c) substantially enlarge the AE district or create other enlarged multi-family zones; (d) reduce the RP, R-80 and R-40 zones to the extent necessary to effect the foregoing, subject to the directions in IX, supra; (e) modify the restrictions in the AE zones and PUD areas discussed hereinabove which discourage the construction of apartments of more than two bedrooms; (f) modify the PUD regulations to eliminate the undue cost-generating requirements specified above; and (g) generally eliminate and reduce undue cost-generating restrictions in the zones allocated to the achievement of lower income housing in accordance with the principles of least cost zoning set forth in V hereinabove.
The trial court shall have discretion, in the event of undue delay in compliance with this opinion or of a finding by the court that any zoning revision submitted by defendant fails *554to comply with, this opinion, to appoint an impartial zoning and planning expert or experts. Such expert may be directed to file a report or to testify, as the court may deem appropriate, as to a recommedation for the achievement by defendant of compliance with this opinion or with any further directions by the court pursuant thereto. See Mount Laurel, 67 N. J. at 216 (Pashman, J., concurring); Pascack Assoc. v. Mayor, Coun. Tp. of Washington, 131 N. J. Super. 195 (Law Div. 1974); Mytelka and Mytelka, “Exclusionary Zoning: A Consideration of Remedies”, 7 Seton Hall L. J. 1, 31 (1975); Hartman, “Beyond Invalidation: The Judicial Power to Zone”, 9 Urban L. An. 159, 171-173 (1975); Rubinowitz, “Exclusionary Zoning: A Wrong in Search of a Remedy”, 6 Mich. J. L. Ref. 625, 656-657 (1973).
Judgment modified, and affirmed as modified; no costs.
Since the last oral argument the name of the municipality was changed to Old Bridge. For convenience and conformity with the record we use the former name, Madison, in this opinion.
In 1969, prior to adoption of the ,1970 ordinance, the township declared a moratorium on all residential construction other than owner-occupied single family dwellings pending the adoption of a new master plan and zoning ordinance. As a result of litigation challenging the moratorium, the Superior Court in Middlesex County, in Verterre Corp. v. Township Comm. of Twp. of Madison (Docket No. L-13820-68 P. W., 7/30/69), ordered the submission to the Township of a new master plan by January 1, 1970 and adoption of a new zoning ordinance by July 1, 1970.
Since Mount Laurel is based on State constitutional grounds, its requirements are not affected by the less restrictive federal concept of equal protection in this area. See Arlington Heights v. Metropolitan Housing Development Corp., — U. S. —, 97 S. Ct. 555, 50 L. Ed. 2d 450 (1977).
Mo^lnt Laurel has been the subject of extensive discussion in the literature. See Ackerman, “The Mount Laurel Decision: Expanding the Boundaries of Zoning Reform”, 1976 U. of Ill. Load Forum 1; Payne, “Delegation Doctrine in the Reform of Local Government Law: The Case of Exclusionary Zoning”, 29 Rutgers L. Rev. 803, 805-819, 859-866 (1976) ; Williams, American Land Planning Law (1975) Addendum Ch. 66; Rose, “The Mount Laurel Decision: Is It Based on Wishful Thinking?”, 4 Real Estate L. J. 61 (1975); Mytelka and Mytelka, “Exclusionary Zoning: A Consideration of Remedies,” 7 Seton Hall L. Rev. 1, 3-4 (1975); Kushner, “Land Use Litigation and Low Income Housing: Mandating Regional Eair Share Plans”, 9 Clearinghouse Rev. 10 (1975) (terming Mount Laurel the “Magna Carta of suburban low and moderate income housing”) ; Rohan, “Property Planning and the Search for a Comprehensive *496Housing Policy — The View from Mount Laurel”, 49 St. Johns L. Rev. 653 (1975); Williams and Doughty, “Studies on Legal Realism: Mount Laurel, Belle Terre and Berman”, 29 Rutgers L. Rev. 73 (1975) (calling Mount Laurel a “major turnaround on a major current problem”); Mallach, “Do Law Suits Build Housing? The Implications of Exclusionary Zoning Litigation”, 6 Rutgers-Camden L. J. 653 (1975); Rose, “Exclusionary Zoning and Managed Growth: Some Unresolved Issues”, 6 Rutgers-Camden L. J. 689 (1975); 6 Powell, Real Property, § 872.1[2][g] (1975); Rose and Levin, “What is a ‘Developing Municipality’ Within the Meaning of the Mount Laurel Decision?”, 4 Real Estate L. J. 359 (1976). See also Berger, Land Ownership and Use 790-799 (2d ed. 1975). For a journalistic appraisal, see “U. S. Journal: Mount Laurel, N. J. — Some Thoughts on Where Lines are Drawn”, New Yorher, 69 (Feb. 2, 1976). See also Note, “The Inadequacy of Judicial Remedies in Cases of Exclusionary Zoning”, 74 Mich. L. Rev. 760 (1976).
“Lower income” is intended as a generic reference to low and moderate income, collectively.
See notes 39, 40, infra. We of course do not deprecate regional “fair share” studies by official or quasi-official governmental agencies or commissions such as those discussed hereinafter. Indeed, as will he emphasized, the basic underlying social problem is far better addressed by administrative action than litigation. In any case, the setting of numerical housing goals is only an incidental phase of the solution. *500The effective substantive revision of restrictive ordinances and governmental desire to implement such goals affirmatively are the essential prerequisites for bousing relief.
We recognize, moreover, that fair share studies by expert witnesses may be of substantial evidential value to a trial court confronted with a litigated issue like the present one.
Our discussion of the facts is not to be taken as laudation of growth per se. The fact and extent of anticipated growth are circumstances material to the need for housing all segments of the population. The control of growth is attracting widespread attention as vital to the maintenance of an acceptable environment. See. Federal Environmental Law (1974) at pp. 1420-1426. See also 3 Williams, American Land Planning Law (1975), Ch. 73, “Timing of Development”, p. 345 et seq.; Mount Laurel, 67 N. J. at 191; id. at 213 (Pashman, J., concurring).
Trial figures on the amount of acreage in each zone, both total and vacant developable, were often in conflict. However, unless otherwise stated, the statistics cited herein for total acreage, total units and total population are those submitted by the township’s planning firm. The vacant developable acreage figures are those of defendant’s expert witness Dennis Lanning, and, with the exception of the AF zone, were stipulated by the parties.
The minimum floor space limitations in the R-40 and R-80 zones were eliminated, and minimum floor space limitations per room were established in all residential zones. With respect to the R--40 and R-80 zones, however, these changes have no effect on provision of low cost housing.
We have no intent to impugn large lot zoning per se. If a developing municipality adequately provides by zoning for lower income housing it may zone otherwise for large lots to the extent that the owners of property so zoned have no other legitimate grievance therewith. Cf. Berger, Land Ownership and Use, 735, 756 (2d ed. 1975).
The R-15 is a new zone on the outskirts of existing high density development and contains a little over 500 acres. The 1964 ordinance had a major residential area similar to the R-15 in which there were more than 3,000 vacant and developable acres compared to 168 acres under the 1973 ordinance. Thus although the R-15 has been restored, it is reduced to a small fraction of its original size.
Taken in entirety, the record herein justifies the following conclusions. Using the allowable 10,000 square foot figure, it would be possible to build 16 efficiencies per acre, which, at an average rental of $180 per month, would yield a total rental income per acre of $2880 per month. An acre can alternatively accommodate twelve one- and two-bedroom apartments on an 80-20 ratio, which, at rentals of $200 and $240 per month respectively, yields a monthly rental income per acre of $2496. If three-bedroom units were built, only 7 to 8 units would be possible, which, at a monthly rent of $280 per month, gross only $1960 or $2,240 per acre per month. Obviously, given equal marketability of all units, it is more profitable to develop efficiencies. Consequently we find justified Judge Furman’s acceptance of plaintiffs’ witness’ testimony that under the restricted FAR provision, and without a maximum density per acre, efficiencies and one bedroom apartments will predominate. 128 N. J. Super. at 443.
All income, sale or rental figures set forth in this opinion, unless otherwise indicated, reflect economic conditions as of 1974 when the case was tried.
The township tax assessor gave two figures for the vacant developable acreage in the AF — 112 and 125 acres. Plaintiffs contend the figure is closer to 67 acres.
Based on other figures adduced at trial which were more favorable to the township, Judge Furman found a total additional capacity under the ordinance of 20,000 to 30,000 units. 128 N. J. Super. at 446. Using this capacity range, the percentage range is 4% to 5.7%.
PUD tracts may be enlarged by up to 15% from neighboring (non-PUD) tracts if approval is obtained. On the other hand, if the site is inadequately serviced by water, sewage or traffic facilities, the maximum allowable density may be reduced.
With an average lot size of 15,000 square feet.
Medium density (garden apartments, townhouses, attached single family) means maximum of two stories, 8 dwelling units per building, and 10,000 gross square feet of development per acre.
High density means a maximum height of 7 stories or 75 feet and 35,000 gross square feet per acre.
In its density regulations, the PUD in the instant case is more restrictive than the PUD involved in Mount Laurel where allowable densities were as high as 6 and 7 units per acre.
Kleven, “Inclusionary Ordinances — Policy and Legal Issues in Requiring Private Developers to Build Low Cost Housing”, 21 U. C. L. A. L. Rev. 1432, 1434, 1451, 1456, 1466 (1974); Rose, “The Mount Laurel Decision: Is It Based on Wishful Thinking?”, 4 Real Estate L. J. 61, 68 (1975); Mallach, “Exclusionary Zoning and Managed Growth: Some Unresolved Issues”, 6 Rutgers-Camden L. J. 653, 660 (1975). See generally, HUD, Housing m the Seventies (1974), for a detailed discussion of housing costs, housing consumption by income groups, and the impact of subsidy programs.
In January 1973, President Nixon announced the imposition of a moratorium on spending under the principal subsidized housing programs pending a reassessment of all federal efforts in the housing area and development of new housing policies. Among those programs suspended were the Section 235 homeownership assistance program established in 1968, the Section 236 cooperative housing program, the rent supplement program (Sections 221(d)(3), 236 and 231), the low rent public housing program originating in 1937, and the Farmers Home Administration Sections 502 and 504 Programs. See HUD, Housing in the Seventies: A Report of the National Housing Policy Review (1974) for a discussion and evaluation of these past programs.
However, there is hope that these programs will be extensively revitalized. When President Ford signed the Housing and Community Development Act of 1974, a number of the suspended programs were replaced. The § 235 homeownership subsidy program was reinstated, 12 U. S. C. § 1715z, and will provide monies to moderate income families for the purchase of dwelling units. A Section 8 Housing Assistance Payment Program, 42 U. S. C. § 1437f, provides housing assistance rent payments on behalf of lower income families occupying newly constructed, substantially rehabilitated or existing housing. A multi-family housing program for the elderly, Section 202 (42 U. S. C. § 1437 a note), was also instituted. A specific plan for the subsidization of 25,000 units in New Jersey under the Section 8 program was recently announced by HUD officials in New Jersey. Neioarh Star Ledger, April 25, 1976.
Limited State sources also exist for the financing of low and moderate income housing. During the year 1976, approximately $126 million were raised from the sale of bonds for loans to moderate income developers under the Housing Finance Agency Law of 1967. N. J. S. A. 55:14j-1 et seq. Additional monies are available under the Department of Community Affairs Demonstration Grant Act, N. J. S. A. 52:27D-59 et seq. and the Mortgage Finance Agency Law. N. J. S. A. 17 :1B-4 et seq.
The concept of least cost housing is not to be understood as contemplating construction which could readily deteriorate into slums. We have emphasized the necessity for consistency of such housing with official health and safety requirements. The recently enacted State Uniform Construction Code Act, L. 1975, c. 217 (N. J. S. A. 52:27D-119 et seq.) states among its purposes “to encourage innovation and economy in construction * * *” and “to eliminate *• * * construction regulations that tend to unnecessarily increase construction costs * * yet be “consistent with reasonable requirements for the health, safety, and welfare of occupants or users of buildings and structures”. Sec. 2.
We envisage zoning provisions which will permit construction of housing, in reasonable amounts, at the least cost consistent with sueh standards. Observation in many areas of the State confirms that low cost housing can be maintained without becoming a slum. See also Mount Laurel, 67 N. J. at 191.
See Lansing, Clifton and Morgan, New Homes and Poor People: A Study of Chains of Moves (1969) (a study of the construction of over 1,000 new units in 17 metropolitan areas and its effects, especially on lowest income groups) ; Grigsby, Housing Markets and Public Policy 84-130 (1963); Fisher & Winnick, “A Reformulation of the Filtering Concept,” 1951 Journal of Social Issues 47-58; Mallach, “Do Lawsuits Build Housing? The Implications of Exclusionary Zoning Litigation”, 6 Rutgers-Camden L. J. 653, 666 (1975) (emphasizing filtering process may take a lifetime to occur).
The “filtering down” or “trickling down” theory has also been advocated in support of subsidies for middle income housing. See HUD, Housing in the Seventies 172-173 (1974).
Added support for this “filtering down” theory was adduced at the trial by Peter Abeles, township planner, who acknowledged that the movement of upper moderate or middle income families to newly constructed housing would leave their former housing available for families lower in the income scale. This movement can comprise a chain of families “moving up”. The shorter the chain, the sooner the needs of the lowest income families are met and presumably the *515better the facilities made available to them. The shortness of the chain obviously depends on the inexpensiveness of the most recently-constructed housing. Lansing et al., supra, at pp. 5, 65.
This is without regard to a share of the prospective need for the foreseeable future thereafter.
Judge Furman’s finding that 80% of vacant developable residential land is zoned R-40 and R-80 and 4% is zoned R-7 and R-10 is based on acreage in the single family zones only. See supra (p. 505).
This was the figure used at trial for the upper limit of moderate income families, based on the 1972 median income in New Jersey of $11,600. Low income was set at below $7,000. Based on 1974 statistics of the Newark area, a low income family of four earns below $8,150 and a moderate income family earns between $8,150 and $13,050.
Except for the AR zone, a small zone restricted to senior citizens housing.
Tke density bonus indicated in this context is the bonus of, for example, an additional single-bedroom or efficiency (in addition to those densities generally permitted) for every three-or four-bedroom unit constructed. Compare this to density bonuses as incentives for construction of subsidized or lower income housing discussed infra, n. 28 and accompanying text.
“Rent skewing” is a generic term referring to the imposition of a greater proportion of land, construction or other costs on one group of units in a development in order to lower the eventual rental or sale price of another group of units therein. Rent skewing can be encouraged by a municipality in two ways: requiring that a mandatory percentage .of moderately priced dwellings be constructed (this is often referred to as an MPMPD ordinance) or allowing a developer a density bonus enabling him to build, for example, one conventional unit for every two low or moderate income units constructed. See Kleven, “Inclusionary Ordinances — Policy and Legal Issues in Requiring Private Developers to Build Low Cost Housing”, 21 U. C. L. A. Rev. 1432 (1974).
Various alternatives have been suggested for satisfying the low and moderate income requirement: constructing federally subsidized housing, renting to low income families under a rent subsidy program, constructing units selling or renting at or below máximums fixed in the ordinance, conveying land to the county or its designee, selling or leasing units to a redevelopment or housing authority or giving the authority first refusal to rent or buy. See Kleven, supra, at 139-147.
On density bonuses or MPMPD’s generally, see Rose, “The Mandatory Percentage of Moderately Priced Dwelling Ordinance (MPMPD) Is the Latest Technique of Inclusionary Zoning”, 3 Real Estate L. J. 176 (1974) ; Rose, “The Mount Laurel Decision: Is It Based on Wishful Thinking?”, 4 Real Estate L. J. 61, 68-9 (1975) ; Brooks, Lower Income Housing: The Planner’s Response, ASPO Report No. 282 (Am. Socy. of Planning Officials, July-August 1972).
Of further significance, there is the possibility that low cost units actually built will not be utilized by persons needing low cost housing, but will be inhabited instead by higher income persons wishing to economize. According to the township’s housing study statistics, 31% of its households are low or moderate income, yet 62.4% of its housing stock can be categorized as low and moderate income housing. These figures seem to support the inference that since not all inexpensive dwellings will be inhabited by households economically requiring such accommodations, a municipality should overzone to meet the requirements of those who do. See Mallaeh, “Exclusionary Zoning and Managed Growth: Some Unresolved Issues,” 6 Rutgers-Camden L. J. 660, 668 (1975).
Finally, in this connection, it is obvious that a zoning ordinance may be revised periodically if experience shows that the allocations for a particular type of housing are excessive or impracticable. Note that the new statutory zoning revision, see supra (p. 496) calls for a general reexamination at least every six years of a master plan and development regulations. L. 1975, c. 291, Sec. 76.
Plaintiffs’ expert witness allocated costs of $1380 and $4340 to Class II PUD units (depending upon type of unit, the lowest figure representing a mid-rise and the highest a single family unit) as the central improvement costs of the project. Of the total $3,383,200 cost of such improvements, $2,229,500 or 66% is directly attributable to the school requirement. This yields a per.- unit exaction for schools, alone, of between $911 and $2864. •
The educational exaction demanded in the Mount Laurel PUDs was not as onerous as that here involved. A Mount Laurel developer was required to pay the costs of educating the additional school children in a PUD if the total number of. school children in the PUD exceeded .3 per multi-family unit; in Madison, the developer must actually build all necessary schools and deed them to the township.
The tentative approval stage requires a public hearing within 45 days of the filing of the application and a grant or denial of tentative approval within 60 days thereafter. Once an application for final approval is filed, unless the plan is substantially different from that receiving tentative approval, a final determination must.be made within 45 days. It must be recognized, however, that preparation and planning required in the development of the application pursuant to the ordinance can add greatly to this 150 day period. However, we deem. the evidence adduced at trial to be- insufficient to permit us to hold the application requirements invalid as exclusionary as a matter of law.
As seen above, the Mount Laurel description of a municipality’s fair share is that “of the present and prospective regional need * *• 67 N. J. at 188 (emphasis added). The recent preliminary housing allocation plan of the State Division of State and Regional Planning (see note 37 infra) properly places considerable emphasis on prospective housing needs for the foreseeable future.
Judge Furman’s finding, as noted above, was that Madison’s need into the 1980’s would be for 750-1000 new units annually, 500-600 of those to be low or moderate income units. 128 N. J. Super. at 442.
To the extent that “housing market area” is identifiable with “region”, in the Mount Laurel sense, the great predominance of the proofs in this record is that the area pertinent to Madison includes at least the seven northeastern counties of New Jersey, and is sometimes referred to as the New York Metropolitan Region, which is generally inclusive of those counties. (Note that 15% of Madison’s resident work force is employed in New York City.) Indeed, another section of the Abeles study, in projecting Madison’s future growth, said: “As part of the Middlesex County region, which is itself part of the larger New' York metropolitan region, Madison Township is unalterably affected by regional population forecasts because of its relationship to housing and employment changes. Middlesex County estimates are influenced by the wider regional context in respect to land use, housing allocations, transport systems, employment locations and other factors.”
The preliminary draft of “A Statewide Housing Allocation Plan for New Jersey” (Nov. 1976) by the State Division of State and Regional Planning places Middlesex County in an eight-county region of which the other counties are Bergen, Essex, Hudson, Morris, Passaic, Somerset and Union, pp. 10-11. See note 37, infra.
See the criticism of the policy emphasis under the COG formula on housing replacement for substandard indigenous housing, to the subordination of regional need satisfaction, in NCDH, “ ‘Fair Share’ Evolves” in Trends in Housing, Vol. 16, No. 3, pp. 1-2 (Fall 1972).
A preliminary but comprehensive housing allocation plan for the entire State has just been published by the State Division of State and Regional Planning. “A Statewide Housing Allocation Plan for New Jersey”, op. oit. supra, note 35. This undertaking, which is subject to public hearings and further review, was directed by Gov*532ernor Byrne’s Executive Order No. 35, dated April 2, 1976. That order acknowledged the impetus of our decision in Mount Laurel, and it directed that the State Division of State and Regional Planning prepare State housing goals “to guide municipalities in adjusting their municipal land-use regulations in order to provide a reasonable opportunity for the development of an appropriate variety and choice of housing to meet the needs of the residents of New Jersey.” County planning boards were permitted to be enlisted in the studies for regions, which could consist of counties or groups of counties. The order states the “law of the State of New Jersey” in language tracking the fair share-regional concept set forth in Mount Laurel.
In allocating regional goals the Division was directed to take into account: (1) the extent of housing need in the region; (2) the extent of employment growth or decline; (3) fiscal capacity to absorb the housing goal; (4) availability of appropriate sites for the housing goal; (5) “other factors as may be necessary and appropriate.”
The resulting State Division study is presented in four sections: (1) present housing needs: 1970; (2) prospective housing needs: 1970-1990; (3) substate regions for housing allocation; and (4) housing allocation methodology. The study culminates in tables setting forth a “fair share” number of housing units for low and moderate income households to meet present and prospective housing needs allocated to every municipality in the State
Present housing needs are estimated by the criteria of (1) dilapidated units; (2) overcrowded units and (3) necessary vacant units, as of 1970. p. 13.
Prospective housing needs are based upon estimates of growth of housing units from 1970 to 1990. pp. 13-15.
The criteria for fixing regions were: (1) sharing housing needs; (2) socio-economic interdependence; (3) data availability; (4) the intent of Executive Order No. 35. pp. 7-8. There are 12 resulting regions: one from the eight northeastern counties (including Middle-sex) ; one for Camden, Gloucester and Burlington; and one for each of the other ten counties of the State, pp. 10-11.
Allocation of fair shares to municipalities was developed separately for present housing needs and prospective housing needs. The method used for allocating present housing needs is to take the percentage of present housing needs of the region to the region’s total housing stock and then apply that percentage to each municipality’s housing stock, p. 13. The method used for allocating prospective housing needs was to apply an average of four indexes, being: (1) vacant ■developable land; (2) employment growth; (3) municipal fiscal •capability (in terms of growth of non-residential ratables) ; and (4) personal income per capita, p. líG-15.
Each municipality’s fair share is thus the sum of the two fair ■shares arrived at as aforesaid, p. 15.
By Executive Order No. 46, dated December 8, 1976, Governor Byrne has ordered postponement of hearings and final review of the preliminary study until after November 1977.
*533See also the nine-county housing allocation formula and projection published in 1973 by the Delaware Valley Regional Planning Commission (DVRPC). This agency monitors all federal funding for housing in its five Pennsylvania and four New Jersey counties (Mercer, Burlington, Camden and Gloucester). It also does regional planning for that area. The DVRPC determined total housing need in the region through the year 2000 and allocated the need by units among the nine counties. Each of the counties was then assigned the task of making municipal sub-allocations pursuant to DVRPC guidelines, and this has now been done in the four New Jersey counties named.
We are not here confronted with litigation joining all the municipalities in a region or county. See Mount Laurel, 67 N. J. at 216-217 (Pashman, J., concurring) ; cf. Urban League of Greater New Brunswick, et al. v. The Mayor and Council of the Borough of Carteret, et al., 142 N. J. Super. 11 (Ch. Div. 1976).
One commentator has specifically warned that because of the conjectural nature of such calculations, utilization of the court as the forum for determining a municipality’s fair share may result in “statistical warfare” between the litigants. Rose, “The Mount Laurel Decision: Is It Based on Wishful Thinking?”, 4 Real Estate L. J. 61, 67 (1975).
Tke recent literature on “fair share” methodologies is considerable. The leading theoretical analysis of fair share plans is found in Brooks, Lower Income Housing: The Planner’s Response, ASPO Report No. 282 (Am. Socy. of Planning Officials, July-August 1972). In addition to Brooks, and the commentaries on Mount Laurel cited in note 3 supra, a representative sampling of fair share literature includes: Kelly, “Will the Housing Market Evaluation Model Be *534The Solution to Exclusionary Zoning?” 3 Real Estate L. J. 373 (1975); Listokin, “Fair-Share Housing Distribution: Will It Open the Suburbs to Apartment Development?” 2 Real Estate L. J. 739 (1974) ; Listokin, “Fair Share Housing Distribution: An Idea Whose Time Has Come?” in New Jersey Trends 353 (T. Norman, ed., 1974) ; Lindbloom, “Defining ‘Fair Share’ of ‘Regional Need’: A Planner’s Application of Mount Laurel”, 98 N. J. L. J. 633 (1975) ; Moskowitz, “Regional Housing Allocation Plans: A Case History of the Delaware Valley Regional Plan,” 7 J. Urban Law 292 (1975) ; National Committee Against Discrimination in Housing (NCDH), “ ‘Fair Share’ Idea Begins to Spread”, Trends in Housing, vol. 16, no. 2, pp. 2-3 (July 1972) ; NCD, “ ‘Fair Share Evolves’ ”, Trends in Housing, vol. 16, no. 3, pp. 1-2 (Fall 1972) ; Holmgren & Erber, “Fair Share Formulas”, 4 HUD Challenge 22 (April 1973).
See also, Erber & Prior, Housing Allocation Planning: An Annotated Bibliography, Council of Planning Librarians Exchange Bibliography No. 547 (March 1974).
Professor Charles M. I-Iaar, in his pioneering work “Regionalism and Realism in Land Use Planning”, 105 U. Pa. L. Rev. 515, 530-531 (1957), noted:
The limitations of the adversary process and the specialization of courts evoke serious doubts as to judicial competence in deciding the proper allocation of regional resources * ’h * Unless the courts are far more adept and skillful than claimed, or the adversary system lends itself to such analysis, or regionalism is not a job that requires scientific planning and engineering techniques, there is a patent need for further state legislation as to who should be the *535ultimate resolver of regional disputes. For the fact remains that no presently constituted agency can adequately meet this issue. The court lacks the staff, the time or the ability to prepare a rational plan. This kind of decision seems eminently suited for the administrative process. If, however, the other governmental agencies default, certainty as well as the need to come to a final decision may he as important as the merits of the particular decision, [Emphasis supplied].
Although Professor Haar was addressing himself to all types of regional planning, his remarks are. particularly relevant to the problem of fair share allocations. Jerome Rose, in dealing specifically with this matter, prophetically observed that: •
The courts seem to have undertaken the delicate and difficult task of maintaining community balance that the legislatures have been reluctant to accept. This is an unfortunate development because the nature of the problem is one that is best resolved by the kinds of political negotiation and compromise upon which a democratic legislative process is based. The legislative process has defaulted and as Governor Cahill predicted, if the legislature fails to meet its responsibility, there will be no alternative but for the courts to step in.
Rose, “The Courts and the Balanced Community: Recent Trends in New Jersey Zoning Law,” 1973 AIP Journal 265, 274 (1973). See also Mytelka & Mytelka, “Exclusionary Zoning: A Consideration of Remedies”, 7 Seton Hall L. Rev. 1, 5, 13 (1975) ; Note, “A Wrong Without a Remedy: Judicial Approaches to Exclusionary Zoning”, 6 Rutgers-Camden L. J. 727, 741 (1975).
See Payne, “Delegation Doctrine in the Reform of Local Government Law: The Case of Exclusionary Zoning”, 29 Rutgers L. Rev. 803 (1976), where not only is essentially the same point made as in the text, but a highly novel idea is proposed for solution of the problem without “judicialization” of it. See also Note, “The Inadequacy of Judicial Remedies in Cases of Exclusionary Zoning”, 71 Mich. L. Rev. 760, 773-779 (1976).
On March 24, 1975, State Senator Greenberg introduced a bill entitled, The Comprehensive Balanced Housing Plan, Senate Bill 3100, providing a new planning framework for the allocation of regional housing needs. Regrettably, this bill died in committee, thus following the path of past legislative efforts to remedy exclusionary zoning. For a discussion of these efforts, see Mallach, supra note 3, 6 Rutgers-Camden L. J. at 677-683.
However, a strong impetus to statewide allocation of goals for provision of low and moderate income housing was provided by Governor Byrne’s Executive Order No. 35, dated April 2, 1976, men*536tioned in note 37, supra, and the study which followed, there discussed.
The Department of Community Affairs had been well along in developing a housing allocation for the State when Executive Order No. 35 was announced.
As noted above, n. 37, separate fair share allocations have been promulgated in Mercer, Burlington, Camden and Gloucester Counties under the auspices of the DVKPC.
We do not fully accord such status to the preliminary statewide housing allocation plan of the State Division of State and Regional Planning as it is only tentative and subject to further public hearings and review. See notes 37 and 42 supra.
To date a number of fair share plans have been proposed or implemented by governmental or regional bodies. The plans vary, depending upon the type of body selected to formulate or implement the plan, the geographical areas encompassed, and the type of housing being allocated (e. g., subsidized or total low income). See Listokin, “Pair Share Housing Distribution: Will It Open the Suburbs to Apartment Development?”, 2 Real Estate L. J. 739, 743 (1974). The most noted of those plans which have been actually implemented include the following: Miami Valley (Dayton, Ohio area) Regional Planning Commission, The Miami Valley Region’s Housing Plan (1973) (updating its 1970 housing allocation plan) ; Metropolitan Washington (D. O. area) Council of Governments, Fair Share Housing Formula (1972) ; San Bernardino County (California) Planning Department, Government Subsidized Distribution Model for Valley Portion of San Bernardino County (1972); Metropolitan Council of the Twin Cities Area (Minneapolis-St. Paul); Housing: Plan, Policy, Program (1973) ; Delaware Valley Regional Plan, see supra note 37.
The utility of almost all these plans is that they are intended to subserve the actual construction or subsidization of low cost housing. By contrast, a plan for a Mount Laurel type litigation, as the present, is not capable of direct utilization by the affected municipality or by the court.
Por a list of twenty-five additional plans, see Erber and Prior, Housing Allocation Planning: An Annotated Bibliography (Council of Planning Librarians Exchange Bibliography #547, 1974).
The criteria relevant for determining “region” have not received the same attention as those used to measure fair share, see note 45 infra, because of the fact that most of the fair share plans under discussion have accepted the geographic area within the jurisdiction of the planning agency as the appropriate region. Moreover, urban literature on derivation of regional boundaries often pertains to subject matter other than housing allocation, e. g., water supply, environment, transportation, etc. However, suggestions do emerge from fair share discussions as to the criteria for determining the appropriate region. The most mentioned is that of journey to work. Rose, “Exclusionary Zoning and Managed Growth: Some Unresolved Issues”, 6 Rutgers-Camden L. J. 689, 717-720 (1975) ; Lustig and Pack, “A Standard for Residential Zoning Based Upon the Location of Jobs”, 1974 AIP Journal 338 (1974) ; Burchell, Listokin and James, “Exclusionary Zoning: Pitfalls of the- Regional Remedy”, 7 Urban Lawyer 262, 271 (1975). This implicates existing job and transportation patterns. Burchell, et al., supra.
The Federal Housing Authority (FHA) has defined a housing market region as the geographic entity within which non-farm dwelling units are in mutual competition, HUD, FHA Economic and Market Analysis Division, FHA Techniques of Housing Market Analysis 12, and hence the location of actual and prospective business centers and the availability of transportation facilities are important. Id.
Rubinowitz has suggested that the relevant region should be the area in which development and movement are or will be taking place, places where middle income families have already exercised the option to move and which would be desirable to low and moderate income groups if housing were available. “Exclusionary Zoning: A Wrong in Search of a Remedy”, 6 Mich. J. L. Reform 625, 654-5 (1973).
The Department of Community Affairs in its November 26, 1975 status report (see note 42 supra) rejected journey to work as the sole criterion for delineation of regions in northern New Jersey, suggesting demarcation of regions “which are large enough or within *541which a burden may be shared.” Plaintiffs’ witness Paul Davidoff suggested three factors for determining the relevant region: (1) the volume of transactions (communications, trade, employment) between component sections of the region; (2) the demand exerted for housing within the region; and (3) the area within which a satisfactory solution of the housing need may he found.
The most frequently mentioned fair share criteria have been grouped under four headings. We list them for informational purposes without necessarily implying our approval.
“Equal share” criteria have as their objective equal distribution of housing, e. g., by establishing a minimum percentage of low and moderate income housing units to be contained within each community.
“Need” criteria allocate housing to regions where there is the greatest need, but have been criticized as perpetuating slums.
“Distribution” criteria allocate low and moderate income units to areas lacking the same in order to achieve a greater income and racial mix.
“Suitability” criteria select areas containing the most suitable housing sites based on physical and fiscal capacity. Kelly, “Will the Housing Market Evaluation Model Be the Solution to Exclusionary Zoning?”, 3 Real Estate L. J. 373 (975); Brooks, Lower Income Housing: The Planner’s Response, op. cit. supra, n. 40; Listokin, “Fair-Share Housing Distribution: Will It Open the Suburbs to Apartment Development?”, 2 Real Estate L. J. 739, 746 (1974).
The most important single criterion emerging from fair share literature is the amount of vacant developable land, as “access to land is the basic issue in exclusionary zoning.” Rubinowitz, “Exclusionary Zoning: A Wrong in Search of a Remedy”, 6 Mich. J. L. Reform 625, 661 (1973). Other basic criteria include employment opportunity, fiscal measures (including per capita income, equalized assessed valuation per pupil, degree of underutilization of classrooms) and existing housing or population density. See generally, Brooks, supra; Listokin, supra; Kelly, “Will the Housing Market Evaluation Model be the Solution to Exclusionary Zoning?”, 3 Real Estate L. J. 373 (1975) ; Rubinowitz, supra; authorities cited supra note 39.
It has been emphasized that many of the potential fair share criteria measure the same factors, Rubinowitz, supra, 6 Mich. J. L. Reform at 660-661, and the effort should be made to keep the formula factors simple to avoid duplication and the “statistical warfare” which may otherwise result from over-sophisticated formulae. Cf. Rose, “The Mount Laurel Decision: Is It Based on Wishful Thinking?”, 4 Real Estate L. J. 61, 67 (1975).
The Delaware Valley Regional Planning Board adopted a formula equally weighing only three criteria: relative wealth (based upon the market value of all taxable real estate in the county compared to the region total) ; equalization criteria (would give each county the same proportion of income groups) ; and projected employment opportunities. See Moskowitz, “Regional Housing Allocation Plans: A Case History of the Delaware Valley Regional Plan”, 7 Urban Lawyer 292 (1975).
See generally Ackerman, “The Mount Laurel Decision: Expanding the Boundaries of Zoning Reform,” 1976 U. of Ill. Law Forum 1, 43-71 (1976).
In addition, it should be noted that on January 14, 1976 a new zoning statute, L. 1975, e. 291, was enacted to supersede the one under consideration, and took effect August 1, 1976. Under this *548statute, housing needs, regional needs and low cost development are specified as among the concerns of zoning:. L. 1975, c. 291, § 2(a), CU, (e), (g), (m).
Plaintiffs offered purported proof, and contend that if their land were appropriately zoned they could develop 20% of their proposed housing for families of moderate income.
The income standards specified are to accord with data applicable as of the date of the completion of the buildings.
The “Statewide Housing Allocation Plan for New Jersey” study, op. cit. supra, n. 35, sets 1970 ranges for low income households as up to $5,568/year and moderate income household as $5,569-$8,567/ year. As of 1976 the estimate for low and moderate income households is stated as “up to approximately $13,000”. p. 492.
This determination is not to be taken as a precedent for an automatic right to a permit on the part of any builder-plaintiff who *552is successful in having a zoning ordinance declared unconstitutional. Such relief will ordinarily be rare, and will generally rest in the discretion of the court, to be exercised in the light of all attendant circumstances.