This action, for damages as a result of an alleged breach of a requirements contract, raises related but distinctly separate issues as to whether the plaintiff buyer’s requirements occurred in good faith and whether those requirements were unreasonably disproportionate to the estimates stated in the contract.
In a fuel oil supply contract executed in early December, 1969, defendant Amerada Hess Corporation (Hess) agreed to supply the requirements of plaintiff Orange and Rockland Utilities, Inc. (O & R) at plaintiffs Lovett generating plant in Tompkins Cove, New York. A fixed price of $2.14 per barrel for No. 6 fuel oil, with a sulphur content of 1% or less, was to continue at least through September 30, 1974, with the price subject to renegotiation at that time. Estimates of the amounts required by plaintiff were included in the contract clause entitled "Quantity”. Insofar as those estimates are relevant to the instant controversy, they were as follows:
1970 - 1,750,000 barrels
1971 - 1,380,000 barrels
1972 - 1,500,000 barrels
1973 - 1,500,000 barrels
The estimates had been prepared by plaintiff on December 30, 1968, as part of a five-year budget projection. The estimates anticipated that gas would be the primary fuel used for generation during the period in question.1 This was a result of the lower cost of gas and of the fact that gas became readily available for power generation during the warmer months of the year as a result of decreased use by gas customers. Plaintiff expressly reserved its right to burn as much gas at it chose by the inclusion, in the "Quantity” provision of the requirements contract, of a clause to the effect that "[n]othing *112herein shall preclude the use by Buyer of * * * natural gas in such quantities as may be or become available”.
Within five months of the execution of the requirements contract, the price of fuel oil began to ascend rapidly. On April 24, 1970 the market price of the oil supplied to plaintiff stood at between $2.65 and $2.73 per barrel. On May 1, 1970 the price was in excess of $3 per barrel. The rise continued and was in excess of $3.50 per barrel by mid-August, and more than $4 per barrel by the end of October, 1970. By March, 1971 the lowest market price was $4.30 per barrel—more than double the price set forth in the subject contract.
Coincident with the earliest of these increases in the cost of oil, O & R proceeded to notify Hess, on four separate dates, of increases in the fuel oil requirements estimates for the year. By letter dated April 16, 1970, O & R notified Hess that it was expected that over 1,460,000 barrels of oil would be consumed over the period April-December, 1970. Since well over 600,000 barrels of oil had been consumed during the first three months of the year, the total increase anticipated at that time was well in excess of 300,000 barrels over the estimate given in the contract.
Eight days later, by letter dated April 24, 1970, O & R furnished Hess with a revised estimate for the period May through December, 1970. The figure given was nearly 1,580,-000 barrels which, when combined with quantities which had already been delivered or were in the process of delivery during the month of April, exceeded the contract estimate by over 700,000 barrels—a 40% increase.
The following month the estimates were again increased— this time to nearly one million barrels above the contract estimate. Hess was so notified by letter dated May 22, 1970. Finally, a letter dated June 19, 1970 indicates a revised estimate of more than one million barrels in excess of the 1,750,000 barrels mentioned in the contract; an increase of about 63%.
On May 22, 1970, the date of the third of the revised estimates, representatives of the two companies met to discuss the increased demands. At that meeting O & R’s president allegedly attributed the increased need for oil to the fact that O & R could make more money selling gas than burning it for power generation. Hess refused to meet the revised requirements, but offered to supply the amount of the contract estimate for the year 1970, plus an additional 10%.
*113The June 19, 1970 letter referred to above recited that the Hess position was "wholly unacceptable” to O & R. It attributed the vastly increased estimates to (a) an inability to burn as much natural gas as had been planned and (b) the fact that O & R had been "required” to meet higher electrical demands on its "own system” and to furnish "more electricity to interconnected systems” than had been anticipated.
Thereafter, for the remainder of 1970, Hess continued to supply the amount of the contract estimates plus 10%. A proposal by Hess, in October, 1970, to modify the existing contract by setting minimum and maximum quantities, and by setting a price keyed to market prices, was ignored by O & R. Although the proposed modification set a price 65 cents lower than the market price, it was more advantageous for Q & R to insist on delivery of the estimated amounts in the December, 1969 contract (at $2.14 per barrel) and to purchase additional amounts required at the full market price.
During the remainder of the contract period Hess continued to deliver quantities approximately equal to the estimates stated in the subject contract. O & R purchased additional oil for its Lovett plant from other suppliers. The contract between Hess and O & R terminated one year prematurely by reason of an environmental regulation which took effect on October 1, 1973 and which necessarily curtailed the use of No. 6 fuel oil with a sulphur content as high as 1%. During the period 1971 through September, 1973 O & R consistently used more than double its contract estimates of oil at Lovett.2
This action was commenced in mid-1972. O & R’s complaint seeks damages consisting of the difference between its costs for fuel oil during the period in question and the cost it would have incurred had Hess delivered the total amount used by O & R at the fixed contract price of $2.14 per barrel. The trial was conducted in September, 1975 before Mr. Justice Donohoe, sitting without a jury. In an opinion dated March 8, 1976, Trial Term held that plaintiff should be denied any recovery on the ground that its requirements were not incurred in good *114faith. Specifically, Trial Term found that plaintiffs greatly increased oil consumption was due primarily to (a) increases in sales of electricity to other utilities and (b) a net shift from other fuels, primarily gas, to oil. The former factor was condemned on the premise that "[i]ndirectly, O & R called upon Hess to supply the demands for electricity to the members of the [New York Power] Pool. O & R then shared the savings in the cost of fuel with the other members of the Pool”. The latter factor was not elaborated on to any great degree. Trial Term did, however, infer that O & R seized "the opportunity to release its reserve commitment of gas” and thereby reaped very substantial profits.
Although Trial Term stated in its opinion that one of the questions before it was whether plaintiffs demands were unreasonably disproportionate to the estimates set forth in the contract, it failed to reach this question in the light of its conclusion that plaintiff had failed to act in good faith. Plaintiff contends on this appeal (1) that Trial Term’s finding of an absence of good faith is unsupported by the record and (2) that since its requirements for the entire term of the contract were less than twice total contract estimates, its demands were not "unreasonably disproportionate” as a matter of law. We reject both contentions upon the facts of this case and affirm Trial Term’s dismissal of the complaint.
It is noted at the outset that the parties agreed, pursuant to their contract, that New Jersey law should apply. The governing statute is subdivision (1) of section 2-306 of the Uniform Commercial Code, which provides, in relevant part: "A term which measures the quantity [to be supplied by a seller to a purchaser of goods] by the * * * requirements of the buyer means such actual * * * requirements as may occur in good faith, except that no quantity unreasonably disproportionate to any stated estimate or in the absence of a stated estimate to any normal or otherwise comparable prior * * * requirements may be * * * demanded” (NJ Stat Ann, 12A: 2-306, subd [1]; matter in brackets added).
There is, as Trial Term observed, a good deal of pre-code case law on the requirement of "good faith”. It is well settled that a buyer in a rising market cannot use a fixed price in a requirements contract for speculation (New York Cent Iron Works Co. v United States Radiator Co., 174 NY 331, 335-336; Asahel Wheeler Co. v Mendleson, 180 App Div 9, 12; see Moore v American Molasses Co., 106 Misc 263). Nor can a *115buyer arbitrarily and unilaterally change certain conditions prevailing at the time of the contract so as to take advantage of market conditions at the seller’s expense (Andrews Coal Co. v Board of Directors of Public Schools, Parish of Orleans, 151 La 695).
There is no judicial precedent with respect to the meaning of the term "unreasonably disproportionate” which appears in subdivision (1) of section 2-306 of the Uniform Commercial Code. Obviously this language is not the equivalent of "lack of good faith”—it is an elementary rule of construction that effect must be given, if possible, to every word, clause and sentence of a statute (2A Sutherland, Statutory Construction [4th ed], § 46.06; Rainbow Inn v Clayton Nat. Bank, 86 NJ Super 13; see McKinney’s Cons Laws of NY, Book 1, Statutes, § 231). The phrase is keyed to stated estimates or, if there be none, to "normal or otherwise comparable prior” requirements. While "reasonable elasticity” is contemplated by the section (see Official Comment, par 2 to Uniform Commercial Code, § 2-306), an agreed estimate shows a clear limit on the intended elasticity, similar to that found in a contract containing minimum and maximum requirements (see Official Comment, par 2 to Uniform Commercial Code, § 2-306). The estimate "is to be regarded as a center around which the parties intend the variation to occur” (supra).
The limitation imposed by the term "unreasonably disproportionate” represents a departure from prior case law, wherein estimates were generally treated as having been made simply for the convenience of the parties and of no operative significance (Note, Requirements Contracts under the Uniform Commercial Code, 102 U Pa L Rev 654, 660-661; Note, Requirements Contracts: Problems of Drafting and Construction, 78 Harv L Rev 1212, 1218; cf. Shader Contrs. v United States, 276 F2d 1). It is salutary in that it insures that the expectations of the parties will be more fully realized in spite of unexpected and fortuitous market conditions (see Note, Requirements Contracts under the Uniform Commercial Code, 102 U Pa L Rev 654, 666-667, supra). Thus, even where one party acts with complete good faith, the section limits the other party’s risk in accordance with the reasonable expectations of the parties.
It would be unwise to attempt to define the phrase "unreasonably disproportionate” in terms of rigid quantities. In order that the limitation contemplated by the section take effect, it *116is not enough that a demand for requirements be disproportionate to the stated estimate; it must be unreasonably so in view of the expectation of the parties. A number of factors should be taken into account in the event a buyer’s requirements greatly exceed the contract estimate. These include the following: (1) the amount by which the requirements exceed the contract estimate; (2) whether the seller had any reasonable basis on which to forecast or anticipate the requested increase (see City of Lakeland, Florida v Union Oil Co. of Cal., 352 F Supp 758; but, cf. Waddell v Phillips, 133 Md 497); (3) the amount, if any, by which the market price of the goods in question exceeded the contract price; (4) whether such an increase in market price was itself fortuitous; and (5) the reason for the increase in requirements.
Turning once again to the facts of the instant case, we conclude that, at least as to the year in which this controversy first arose, there was ample evidence to justify a finding of lack of good faith on plaintiff’s part. Even through the thicket of divergent and contrasting figures entered into exhibit at trial, the following picture emerges: nonfirm sales3 from plaintiff’s Lovett plant, presumably in large part to the New York Power Pool, increased nearly sixfold from 67,867 megawatt hours in 1969 to 390,017 megawatt hours in 1970. The significance of that increase in nonfirm sales lies in the fact that such sales did not enter into the budget calculations which formed the basis of the estimates included in the contract. Even assuming that a prudent seller of oil could anticipate some additional requirements generated by nonfirm sales, an increase of the magnitude which occurred in 1970 is unforeseeable. That increase, of 322,150 megawatt hours, translates into the equivalent of over 500,000 barrels of oil.4 The conclusion is inescapable that this dramatic change in plaintiff’s relationship with the New York Power Pool came about as a result of the subject requirements contract, which insured it a steady flow of cheap oil despite swiftly rising prices.5 O & R’s *117use of the subject contract to suddenly and dramatically propel itself into the position of a large seller of power to other utilities evidences a lack of good faith dealing.6
In addition to this massive increase in sales of power to other utilities, the evidence indicates that at about the time O & R was demanding roughly one million barrels of oil in excess of the 1970 contract estimate, there was an internal O & R proposal to release gas to a supplier which represented the equivalent of 542,000 barrels of oil. An internal O & R memorandum dated May 26, 1970 (four days after the meeting at which Hess refused to supply the one million additional barrels demanded) recommended that in view of the Hess position, the proposed release be canceled. Significantly, O & R never did burn as much oil as had been demanded in May and June, 1970. Its total usage for the year was 2,294,845 barrels—471,155 barrels less than its maximum demand. This was explained, by O & R officials, in part, on the ground that their "gas department” had made a "pessimistic estimate” which did not turn out to be quite true.
Thus it appears that in May, 1970 Hess refused an O & R demand of roughly one million barrels in excess of the contract estimate, which demand was occasioned by greatly increased sales to other utilities and a proposed release of gas which might otherwise normally have been burned for power generation.7 The former factor is tantamount to making the other utilities in the State silent partners to the contract (cf. City of Lakeland, Florida v Union Oil Co. of Cal., 352 F Supp *118758, supra), while the latter factor amounts to a unilateral and arbitrary change in the conditions prevailing at the time of the contract so as to take advantage of market conditions at the seller’s expense (Andrews Coal Co. v Board of Directors of Public Schools, Parish of Orleans, 151 La 695, supra). Hess was therefore justified in 1970 in refusing to meet plaintiff’s demands, by reason of the fact that plaintiff’s "requirements” were not incurred in good faith.
With respect to subsequent years however, the record is ambiguous as to the cause of plaintiff’s drastically increased requirements. Nonfirm sales from Lovett actually declined slightly in 1971 and 1972 although they were still greatly in excess of 1969 sales.8 If one takes 1969 as a base year, increased nonfirm sales from Lovett in 1971 amounted to the equivalent of about one-half million barrels of oil,9 while in 1972 they amounted to the equivalent of just over 300,000 barrels.10 Comparable figures for 1973 are impossible to arrive at with any degree of confidence because sales figures in the record are for the full year, while defendant’s obligation to supply oil extended through only three quarters of the year. *119In any event, it is apparent that O & R’s tremendously expanded use of oil during the period subsequent to 1970 cannot be explained solely by reference to increased sales to other utilities. In 1971 oil use exceeded the contract estimate by over 1,750,000 barrels; the 1972 figure was in excess of 1,825,000 barrels; and for the first nine months of 1973 the increase was more than 1,275,000 barrels.11
It appears that a large portion of the difference between actual use and contract estimates during this period can be attributed to a rather large decline in plaintiff’s "actual take” of gas as opposed to the estimates of gas availability which were made in 1968 (and which were used in the computation of the December 30, 1968 budget). This decline, with the equivalent figure in barrels of oil,12 was as follows:
ESTIMATE ACTUAL TAKE EQUIV. BARRELS
(MCE) (MCE) DECREASE OF OIL
1971 - 40,615,000 34,518,000 6.097.000 1,016,167
1972 - 43,661,000 36,274,000 7.387.000 1,231,167
1973
(9 mos) 34,034,700 25,783,000 8,251,700 1,375,283
Even allowing for the fact that O & R’s actual system requirements were slightly lower during this period than the estimated system requirements (thus theoretically leaving more gas available for electric generation), it is clear that the decline from the estimates in gas received by O & R was a very major factor in plaintiff’s increased use of oil during this period.
The record is unclear as to why this decline came about. Plaintiff introduced into evidence a Public Service Commission memorandum which indicates that gas supplies available to interstate transmission companies had become extremely tight. However plaintiff failed to call one witness who was expert in its gas operations and who could testify as to the link, if any, between this general shortage and plaintiff’s operations. While an unfavorable inference may be drawn when a party fails to produce evidence which is within his control and which he is naturally expected to produce (2 Wigmore, Evidence [3d ed], § 285; Bayer v Farrell, 69 NJ Super 347; see, also, Richardson, Evidence [Prince, 10th ed], *120§ 92), we decline to speculate as to causes of the decline in gas received by plaintiff. In any event, such speculation is not necessary for resolution of this appeal.
We hold that under the circumstances of this case, any demand13 by plaintiff for more than double its contract estimates, was, as a matter of law, "unreasonably disproportionate’ (Uniform Commercial Code, § 2-306, subd [1]) to those estimates. We do not adopt the factor of more than double the contract estimates as any sort of an inflexible yardstick.14 Rather, we apply those standards set forth earlier in this opinion, which are calculated to limit a party’s risk in accordance with the reasonable expectations of the parties.
Here, as noted, plaintiff’s requirements during the period 1971 through September, 1973, were more than double the contract estimates. Defendant had no reasonable basis on which to forecast or anticipate an increase of this magnitude. Indeed the contract suggests the parties contemplated that any variations from the estimate would be on the downside— else why did plaintiff expressly reserve for itself the right to burn as much as it chose? The market price of the grade of oil supplied had more than doubled by March, 1971. It stayed at or above $4 per barrel for the rest of the applicable period and had reached nearly $5 per barrel by the end of September, 1973. The record is silent as to whether defendant had any reason to anticipate this enormous increase in oil prices. Finally, the increase in requirements was due in part to plaintiff’s increased sales to other utilities and also due to a significant decline in anticipated deliveries of gas, the cause of which was inadequately explained by plaintiff. The quantities of oil utilized by plaintiff during the period subsequent to 1970 were not within the reasonable expectations of the parties when the contract was executed, and accordingly we hold that those "requirements” were unreasonably disproportionate to the contract estimates (see Uniform Commercial Code, § 2-306, subd [1]).
*121Hopkins, J. P., Martuscello and O’Connor, JJ., concur.
Judgment of the Supreme Court, Rockland County, entered June 4, 1976, affirmed, with costs.
. For example, it was projected that in 1970 gas would generate 14,047,545,000,000 BTU while oil would be used to generate 10,810,740,000,000 BTU. The comparable ratios for the following years are: 1971—15/8; 1972—16/9; 1973—9/5.
. The Hess contract estimated 1971 usage at 1,380,000 barrels. In fact, 1,301,045 barrels were supplied that year by Hess and 1,844,947 barrels were supplied by other companies, for a total of 3,145,993 barrels. The contract estimate for 1972 was 1,500,000 barrels; a total of 3,325,037 barrels was purchased by O & R. The contract estimate for the first nine months of 1973 was 1,125,000 barrels (75% of the 1,500,000 listed in the contract as an estimate for all of 1973); 2,401,979 barrels were received by O & R during that nine-month period.
. Nonfirm sales are to be contrasted with "firm” sales, which are synonymous with predictable sales. Firm sales include predictable sales to O & R’s own customers and sales pursuant to contract with other utilities.
. A figure of 515,400 barrels of oil is arrived at by multiplying the increase in sales of nonfirm megawatt hours (322,150) by a very conservative conversion factor of 1.6.
. Notwithstanding the fact that Hess failed to supply all that was demanded by O & R, it is apparent that even with a mix of Hess and non-Hess oil, plaintiff’s costs of producing electricity with oil would have been lower than those of any companies purchasing oil on the open market or with "floating” contracts. This observation is *117underscored by the fact that O & R found it more advantageous to purchase the balance of its "requirements” on the open market (over and above the contract estimates which Hess was supplying) rather than to renegotiate the contract at a price 65 cents lower than posted market prices.
. "[A] sudden expansion of the plant by which requirements are to be measured would not be included within the scope of the contract as made but normal expansion undertaken in good faith would be within the scope of this section.” (Official Comment, par 2 to Uniform Commercial Code, § 2-306.)
. While there was testimony by O & R officials to the effect that this gas, supplied by the Home Gas Company, could not be burned at the Lovett plant because that gas was only transmitted to another geographical sector of O & R’s territory, it strains credulity to believe that the interstate pipeline system is so parochial that releases in one sector would not affect operations in another sector. Home Gas Company was, in fact, a subsidiary of the Columbia Gas Transmission Company at the time, and Columbia Gas was burned at the Lovett plant. Furthermore, plaintiffs contention is belied by (a) the May 26, 1970 memo which linked the proposed release to oil deliveries at Lovett and (b) a revised calculation from O & R’s gas department, dated May 19, 1970, which allocated Home Gas between O & R’s two geographical divisions for the purpose of showing gas availability for electrical generation.
. In 1971 total sales from Lovett amount to 387,874 MWH; firm sales were 17,294 MWH. Assuming these firm sales were all generated at Lovett, total nonfirm sales would have been 370,580 MWH—down slightly from the 390,017 MWH sold on a nonfirm basis in 1970. In 1972 total sales from Bowline (a new generating plant), from Lovett, and from gas turbines amounted to 722,329 MWH. Firm sales for the year were 423,734, of which 374,327 MWH were generated at Bowline. Assuming that the remaining 49,407 MWH in firm sales were generated at Lovett (and not by the gas turbines), that plant’s nonfirm sales for the year would have amounted to 256,320 MWH.
We have focused on nonfirm sales in our analysis because the record does not reveal whether firm sales during this period were taken into account in the preparation of the December 30, 1968 budget projection which formed the basis of the contract estimates. If they were, they would be meaningless in attempting to explain the tremendous increase in plaintiff’s oil requirements.
While there was evidence that electrical power is "fungible” and that allocation of sales to various plants is an "accounting function”, the testimony at trial did establish that the allocation is performed in a sufficiently rational manner. We therefore decline to attribute all O & B sales to the Lovett plant.
. We choose the increase over 1969 rather than the absolute figure for sales because a certain amount of power pool sales, based on O & B’s past experience, would be reasonably foreseeable. In 1971, nonfirm sales from Lovett amounted to 370,580 MWH (see n 8), an increase of 303,007 MWH over the 67,867 MWH sold on a nonfirm basis in 1969. If we multiply this increase by the very conservative conversion factor of 1.6, we arrive at a barrel equivalent of 484,811.
. This calculation is made in exactly the same manner as the one for 1971. The increase in sales of nonfirm power amounts to 188,453 MWH (256,320 MWH less 67,867 MWH) and application of the conversion factor results in a barrel equivalent of 301,525.
. This figure is arrived at by subtracting three quarters of the contract estimate from total oil deliveries during the first nine months of 1973.
. One barrel of oil is equivalent in heating value to 6 Mcf of gas.
. There is no indication in the record that O & R continued to supply Hess with updated "requirements” demands after June 19, 1970. In fact, it can be inferred from a fair reading of the record that O & R entered into a contract with another supplier to furnish amounts required in excess of the amounts Hess was willing to supply. Nevertheless, we assume for the sake of argument that there was a continuing "demand” by plaintiff for its total requirements.
. Interestingly, plaintiff would have us hold as a matter of law that where an actual requirement varies "only by a factor of two from a stated estimate”, it is not unreasonably disproportionate to such estimate.