This is a local property tax case wherein plaintiffs and defendant seek review of judgments of the Passaic County Board of Taxation (county board) reducing assessments on the property hereinafter described for the years 1975 through 1978. Plaintiffs seek a further reduction, while defendant counterclaims for a restoration of the original assessment. The premises in question are located at 73-89 Howe Avenue, Passaic, New Jersey, and are known and designated as Block 2153, Lot 1, on the City of Passaic tax map. The assessments and the county board action thereon were as follows:
1975-1976
Assessment County Board Judgment
Land $ 48,000 $ 48,000
Improvements 225,500 200,000
Total $273,500 $248,000
1977-1978
Assessment County Board Judgment
Land $ 48,000 $ 48,000
Improvements 225,800 200,300
Total $273,800 $248,300
The land consists of approximately 15,400 square feet, and the improvements thereon are four three-story brick apartment buildings containing 32 units, of which 30 units are rentable. The structures are about 80 years old and in poor condition. The tenant population is composed entirely of persons subsisting on transfer payments. The property is located in a declining neighborhood.
Plaintiff John Bresnowitz purchased the subject property on August 1, 1975 at a mortgage foreclosure sale for a cash *397consideration of $130,000.1 Bresnowitz bought the property to protect his own mortgage of $35,000, which ranked behind three or four other mortgages.
Bresnowitz sold the property on October 20, 1978 to Bauza Realty Corp. for $165,000. The buyer paid no cash down, the sales price being represented entirely by a purchase money mortgage taken back by the seller. That mortgage was of the direct reduction variety with interest at 9%. Amortization was predicted on a 25-year term but the maturity date was October 20, 1988, ten years after the sale.
Repairs and maintenance of the property during the years in issue were difficult as well as expensive in view of the vandalizing proclivities of the tenants.
Valuation
Plaintiff’s expert determined the true value of the property to be $165,000 for all tax years in issue. In reaching this conclusion he relied principally upon the October 20,1978 sale to Bauza Realty Corp. His analysis of the sale convinced him that the transaction was arms’ length and the stated price represented the total consideration.
Plaintiff’s expert found additional support for his true value conclusion by use of the economic approach. For the years 1975, 1976 and 1977 he calculated the gross rent potential by annualizing the October 1976 rent potential. He estimated the effective gross income, after deducting for actual vacancies for 1976, at $71,713. Utilizing stabilized and actual expenses the expert estimated total expenses at $42,683, yielding a net income of $29,030. For 1978 the expert annualized October 1977 rent in determining gross rent potential, from which he deducted actual vacancies to arrive at effective gross income of $74,783. He *398estimated expenses on the basis of actual expenses for 1977 and determined net income to be $27,215 for 1978. The expert then applied the building residual method, attributing $6,778 of net income to the land for the years 1975 through 1977 and $6,715 of net income to the land for 1978. His capitalization rate for the years 1975 through 1977 consisted of a 10% interest rate, a 5% recapture rate and an effective tax rate of 4.12%, for an overall rate of 19.12%. His capitalization rate for 1978 was composed of a 10% interest rate, a 5% recapture rate and an effective tax rate of 3.99% for an overall rate of 18.99%. The defendant’s land assessment was accepted for all years. Plaintiff’s expert capitalization income pursuant to the aforementioned rates to arrive at a true value of $164,400 for the years 1975 through 1977 and $156,000 for 1978.
The expert indicated that his basis for the return and recapture rates was an an analysis of comparable risk factors and long-term investments. He testified that he arrived at his rates of return upon the basis of such investments, his knowledge of the short and long-term money markets, the age, physical condition and tenant mix of the property, and the fact that Passaic was an area of lesser investment desirability and greater risk than many of the comparable long-term investments upon which he relied. His estimate of the rate of recapture (depreciation) was predicted upon the higher risk inherent in an old apartment property of inferior condition.
Finally, plaintiff’s expert rejected the reproduction cost approach as a valuation method because, he testified, such approach would merely have mirrored the income and market data approaches if properly used. He used no comparable sales in his market approach because he found no comparable buildings.
Defendant’s expert, its chief assessor, found the true value of the property to be $253,000 for all tax years in issue, with $48,000 allocated to land and the remainder to improvements. He relied principally on the economic approach to valuation. His determination of gross rent potential appears to be an annualization of the October 1976 actual rentals, stabilized for *399all four years. Unlike plaintiff’s expert, who used the actual vacancies, defendant’s expert applied a 5% vacancy and collection loss factor to arrive at effective gross income of $73,085. He testified that his section of a 5% vacancy factor was influenced by an actual vacancy rate in the subject property of less than 5%. The actual vacancies and collection losses indicate that a higher rate is appropriate. Moreover, the assessor was not aware that 15 separate summary eviction proceedings had been instituted in the Passaic County District Court during 1977. The course of action thus pursued by plaintiff involving 50% of the rentable units serves to discredit the selection of a 5% vacancy and collection loss factor.
The assessor purported to accept and stabilize claimed expenses for the years 1975 through 1977 but he ignored claimed expenses for 1978. (Expenses for the latter year were $5,000 higher than stabilized expenses for the three preceding years.) He arrived at stabilized net income for all years in issue of $31,076, to which he applied a total capitalization rate of 12.4% for all years. That rate was allocated on the basis of 7.5% return, 0.8% recapture and 4.1% effective tax rate. Using the building residual method, the assessor allocated $5,616 to land income, capitalizing the residue at 12.4% to arrive at a true value of the improvements. This exercise produced a total true value for land and building of $253,000.
While the selling price of real property involved in a judicial determination of its assessable value is a guiding “indicium” of fair value and ordinarily is merely evidential, such price under certain circumstances might become controlling. Hackensack Water Co. v. Division of Tax Appeals, 2 N.J. 157, 162, 65 A.2d 828 (1949). The statutory criterion is the consideration of the market value at a fair and bona fide sale by private contract. N.J.S.A. 54:4-23. I find that the price for which the subject premises were sold on October 20, 1978 is not indicative of the property’s true value. In reaching this conclusion I have weighted and evaluated the factors surrounding the sale and I find special circumstances which tend to increase the sales price *400in derogation of the property’s true value. See Rek Investment Co. v. Newark, 80 N.J.Super. 552, 559, 194 A.2d 368 (App.Div. 1963); McCrory Stores Corp. v. Asbury Park, 89 N.J.Super. 234, 245, 214 A.2d 526 (App.Div. 1965). The buyer paid no cash; and the plaintiff took back a purchase money mortgage for the entire price. Payments under the mortgage were calculated on the basis of a 25-year term, with maturity in ten years. On the maturity date 82.74% of the original principal amount would remain outstanding. The buyer was a corporation and the record fails to indicate the existence of personal guarantee of the mortgage debt by any of the buyer’s principals.
The indispensable component of any sale, in economic terms, is a shift in the risks and benefits of ownership. Cf. Commissioner of Internal Revenue v. Brown, 380 U.S. 563, 85 S.Ct. 1162, 14 L.Ed.2d 75 (1965). That component is missing from the sale to Bauza Realty Corp. The buyer put up no cash and thus assumed no economic risk. Plaintiff, on the other hand, retained the economic risk of ownership while parting with the legal title to the property. The mortgage debt being, for all practical purposes nonrecourse, plaintiff was solely dependent on rental income from the property for his mortgage payments. At the maturity of the debt in ten years 82.74% of mortgage principal would still be outstanding.
Plaintiff’s demeanor on the witness stand revealed a property owner of advancing years and failing health, desperately anxious to rid himself of a millstone. I find that the transaction of October 20,1978 with Bauza Realty Corp. was not a usable sale for purposes of this proceeding. If the transaction be regarded as a sale in legal contemplation in the sense that title passed and legally sufficient consideration was paid (i.e., an unconditional promise to pay a specified sum, secured by a mortgage). I find that the sale was not at arms’ length. The better measure of true value of the subject property is found in the economic approach to valuation utilized by plaintiff’s expert. His capitalization of imputed net income produced values of $164,400 for the tax years 1975, 1976 and 1977 and *401$156,000 for 1978. I find the risk-oriented components of the expert’s capitalization rate reasonable and I accept them as an accurate reflection of the demands of the long-term investment market during the tax years under review. I give great weight to the opinion of plaintiff’s expert in so far as his economic approach is concerned, as I am impressed not only with his candor, intelligence, knowledge and experience but also with the facts and reasoning forming the foundation of his opinion. In re Port of New York Auth., 28 N.J.Super. 575, 101 A.2d 365 (App.Div. 1953); Passaic v. Gera Milis, 55 N.J.Super. 73, 150 A.2d 67 (App.Div. 1959), certif. den. 30 N.J. 153, 152 A.2d 171 (1959).
The testimony of defendant’s expert, on the other hand, does not withstand close scrutiny. To begin with, while his determination of gross rent potential appears reasonable, his calculation of the vacancy and collection loss factor finds no support in the record. His assumption of a 5% vacancy and collection loss factor is discredited by his concession that he was unaware of the initiation of 15 summary eviction proceedings in 1977 by the plaintiff.
In addition, the assessor claimed to utilize the expense estimates claimed by the taxpayer as of October 1, 1976. While he did in fact utilize those expenses on a stabilized basis for the years 1975, 1976 and 1977, he ignored 1978 expenses, which exceeded the stabilized expenses for the three preceding years by more than 10%.
The major flaw in the assessor’s analysis, however, lies with the development of his capitalization rate. His assumed return on investment of 7.5% is supported only by reference to the prevailing yield on short-term, low-risk investments. He made no effort to compare his return rate to yields on long-term investments prevalent during the tax years. The credible, persuasive testimony of plaintiff’s expert indicates that a real estate investor would consider the subject property to be a long-term investment of the highest possible risk.
*402Finally, the assessor’s selection of a recapture component of 0.8% is predicated upon a sinking fund approach, which is demonstrably inappropriate for the subject property. The use of a Hoskold or sinking fund method presupposes the projection of a stable income level over the remaining economic life of the improvement, i.e., the property is occupied by a high-grade tenant under a long-term lease. American Institute of Real Estate Appraisers, The Appraisal of Real Estate (7 ed. 1978), 317-318. This supposition finds no support in the record. The subject property was tenanted exclusively by high-risk welfare clients throughout the tax years under review. The actual vacancies and collection losses are extraordinarily high, as are the maintenance and repair expenses by virtue of the tenant’s vandalizing propensities. In addition, the nature of the tenant population precludes effective rent increases to match rising costs. The proposition that the property is valued as though it were demised to a high-grade tenant for a long period of time is a tenuous unwarranted hypothesis.
Although plaintiff contends that the sale of the subject property on October 20, 1978 is the best evidence of value, I find that the economic approach of plaintiff’s expert provides a better measure of true value. This court has the right and duty to apply its own expert judgment to valuation data submitted by the experts, and to ascertain true value in light of all the evidence, Samuel Hird & Sons, Inc. v. Garfield, 87 N.J.Super. 65, 73-74, 208 A.2d 153 (App.Div. 1965), even though the true value thus determined is fixed at a lesser amount than the value for which plaintiff contends. Botta v. Brunner, 26 N.J. 82, 138 A.2d 713 (1958); Perdomo v. Goldstein, 122 N.J.Super. 14, 298 A.2d 305 (Law Div. 1972); R. 4:42-6. Accordingly, I find the true value of the subject property to be $164,000 for the years 1975, 1976 and 19772 and $156,000 for the year 1978.
*403Discrimination
A. 1975-1977
Plaintiffs’ expert presented an analysis of the statistical data compiled and published by the State Director of the Division of Taxation (the Director). This information indicated that the overall average ratios3 for the City of Passaic for the years 1975 through 1977 were:
1975 77.73%
1976 75.34%
1977 77.15%
During the years 1975 through 1977 the ratios for Class 2 (one to four-family residential houses) and Class 4 (apartments, commercial, and industrial) properties were:
Class 2 Class 4
1976 59.16% 89.49%
1976 58.22% 90.28%
1977 56.79% 99.94%
The Director’s studies for the three years in question also indicate the following with respect to ratio ranges and clusters:
Clusters By Class
1975 1976 1977
Sales 4 Sales 4 Sales 4
Class Class Class
Ratio 2-4 2-4 2-4
100% and over 10 17 12 19 6 29
90-99 9 5 5 3 8 5
80-89 8 8 5 9 7
70-79 15 16 14 11 20
60-69 30 3 36 6 31
50-59 65 3 44 2 71
40-49 36 1 49 0 47
30-39 6 0 1 0 11
No sales below 30
179 53 166 50 201 58
*404 All Classes
1975 1976 1977
100% and over 27 32 37
14 8 13 90-99
17 14 13 80-89
32 25 28 70-79
33 42 40 60-69
68 37 46 49 73 48 50-59 40-49
6 1 11 30-39
No sales under 30
234 217 263
The Director’s published coefficients of deviation5 for the years 1975 through 1977 indicated the following with respect to Passaic:
General Coefficients
1975 27.83%
1976 31.33%
1977 34.58%
Segmented Coefficients for Class 4 Properties
1975 39.46%
1976 41.78%
1977 53.37%
*405Range of Ratios (Low and High) by Classification
Class 2 Class 4
1975 35.82-142.00 44.86-174.63
1976 39.33-212.50 53.78-260.00
1977 30.77-516.67 45.14-232.00
The subject property is included in Class 4.
The assessor’s assessing practices can only be described as random and inconsistent. He acknowledged that he changed assessments for properties that were sold, for properties for which building permits were sought and for properties that were otherwise called to his attention. He did not alter assessments for properties which fell into none of these categories. There was no consistently applied standard for increasing assessments. In some cases the assessor increased the assessment to the sales price while in other instances he calculated the assessment by applying the average ratio to the selling price. He also conceded that he would not lower an assessment on his own initiative when a property sold for less than its assessed value. In the absence of some inquiry or overture from the buyer, the assessment of such property was not disturbed.
It is now well settled in this State that “a taxpayer is entitled to ‘treatment commensurate with that given his fellow taxpayers within the municipality’ and that if it is not accorded, he is entitled to a judicial or quasi-judicial remedy.” Tri-Terminal Corp. v. Edgewater, 68 N.J. 406, 409, 346 A.2d 396, 399 (1975). Where a taxpayer’s property is assessed at a ratio to true value substantially higher than the common level of assessments, or, if there is no such common level, substantially higher than the average ratio determined by the Director, then such taxpayer is ordinarily entitled to a reduction of his assessment down to the valuation resulting from application of the Director’s ratio to the true value of the property. In re Appeal of Kents, 2124 Atlantic Ave., Inc., 34 N.J. 21, 166 A.2d 763 (1961).
*406The evidence in this case leads inescapably to the conclusion- and I so find-that no common level existed in the City of Passaic during the years 1975 through 1977. The assessor’s own testimony belies his assertion that he endeavored to maintain a common level of assessments.
Quite apart from the Assessor’s testimony, the Director’s sales ratio and coefficients of deviation studies point to random, chaotic assessing practices which reveal planlessness and the absence of any semblance of a common level. The ranges of ratios for Class 2 and Class 4 properties during the years 1975 through 1977 were extraordinarily broad, with a spread of more than 100 percentage points for each year in question. The ratio clusters do not indicate concentration within a relatively narrow band. Indeed, a band of 50 percentage points is required to cover 80% of sales of all classes in 1975 and 1976, and the same band is required to cover 78% of sales in 1977.
The coefficients of deviation, excellent statistical barometers employed in determining uniformity (or the lack thereof) in assessing practices, reinforce the conclusion of planlessness drawn from an analysis of the ratio ranges and clusters. The general coefficients of deviation ranged from 27.83% in 1975 to 34.58% in 1977. The segmented coefficients of deviation for Class 4 properties were even worse, ranging from 39.46% in 1975 to 53.37% in 1977. A coefficient of deviation in excess of 20% indicates the absence of a common level (or the erosion of a pre-existing one) and the need for a revaluation. Tri-Terminal Corp. v. Edgewater, supra at 413, 346 A.2d 396.
Discrimination is shown by the ratio between the assessed value of the subject property and its true value, as herein determined. That ratio is 165%, substantially in excess of the Director’s average ratios of 77.73%, 75.34% and 77.15% for the years 1975, 1976 and 1977, respectively.
In view of the foregoing I find that plaintiffs are entitled to discrimination relief for the years 1975, 1976 and 1977 and that the assessable value of the subject property shall be the true value thereof, as hereinabove determined, to which I will *407apply, in the interest of relative stability in assessments, the average of the Director’s average ratios for the 3 years in question. New Brunswick v. Division of Tax Appeals, 39 N.J. 537, 189 A.2d 702 (1963); Feder v. Passaic, 105 N.J.Super. 157, 251 A.2d 457 (1969). That figure is 76.74%.
B. 1978
Discrimination relief from the assessments on the subject property is appropriate for 1978. N.J.S.A. 54:2-40.4, effective for 1978, provides in pertinent part:
Whenever the [Tax Court] is satisfied by the proofs that the ratio of the assessed valuation of the subject property to its true value exceeds the upper limit or falls below the lower limit of the common level range, it shall revise the taxable value of the property by applying the average ratio to the true value of the property . . .
The “common level range,” as defined in N.J.S.A. 54:l-35a, is “that range which is plus or minus 15% of the average ratio for that district.” The average ratio is defined in the same statute (for 1978) as the “unweighted, unclassified arithmetic average as determined by the Director of the Division of Taxation from the latest 1-year study data compiled by the director ... as of October 1 of the year preceding the tax year . . .” N.J.S.A. 54:l-35b instructs the Director, on or before April 1 in each year, to determine the average ratio and the common level range and to certify the same to the county boards, the assessors, and the municipal clerks. The average ratio so certified by the Director for Passaic is 74% and the common level range, also certified, is 63% (lower) and 86% (upper).
Defendant argues that the average ratio is 80.51% for 1978, as that is the figure which appears in the Table of Equalized Valuations promulgated by the Director pursuant to N.J.S.A. 54:1-35.1 for school aid purposes. This argument overlooks the fact that the latter Table is weighted. See Willingboro Tp. v. Burlington Cty. Bd. of Tax., 62 N.J. 203, 212, 300 A.2d 129 (1973). N.J.S.A. 54:l-35a, prior to its amendment for 1979 and later years, defined the average ratio in terms of the unweighted, unclassified arithmetic average as determined by the Director. This unweighted figure, reflected in the Director’s certification, is, as stated earlier, 74% for 1978.
*408As the subject property was assessed at 175% of its true value, which I have found to be $156,000 for 1978, I find that the assessment lies beyond the common level range and, accordingly, the assessment will be reduced for 1978 to the property’s true value multiplied by the average ratio of 74% pursuant to N.J. S.A. 54:2-40.4.
In view of the foregoing judgment will be entered determining the assessments to be as follows:
1975,1976,1977
Land $48,000
Improvements 77,850
Totals $125,850
1978
Land $48,000
Improvements 67,440
Totals $115,440
Hereafter, all references to a singular plaintiff are to John Bresnowitz, unless otherwise indicated. Plaintiff Niktan Realty owned the property for only eight months of the first year under review. Bresnowitz was the owner for the balance of the period involved.
piaintiff expert’s valuation adjusted by using effective tax rate of 4.16 derived from application of average ratios as corrected.
The ratio of assessed values to market values, determined by usable sales.
Sales of Class 1 (vacant land) properties were too few in number to be statistically significant.
Coefficients of deviation are useful statistical tools in the measurement of uniformity of assessment practices. A high coefficient denotes lack of uniformity, whereas a low coefficient signifies reasonable equality of assessment. The general coefficient is the average deviation from the average assessment ratio in the taxing district, expressed as a percentage of average assessment ratio. The segmented coefficient is the average deviation of assessment ratios of each property class from the average assessment for all properties of all classes expressed as a percentage of average assessment ratios for all properties of all classes.