Respondents-appellants employ members of the charging party-appellee labor unions. Pursuant to their respective collective-bargaining agreements (CBAs), respondents provide pension benefits to their employees. The CBAs provide the employees with various pension plan options, including one in which payments terminate at the death of the employee (straight-life pension) and another in which pension benefits continue until the death of both the employee and his or her spouse (joint- and-survivor pension or optional benefits plan). Since 1982, a particular mortality table was used to calculate the joint-and-survivor-pension monthly benefit. In 2006, respondents adopted a different mortality table for calculating those benefits, thereby reducing the monthly pension benefit paid under the joint-and-survivor plan. The charging parties filed a claim with the Michigan Employment Relations Commission (MERC), asserting that respondents committed an unfair labor practice (ULE) by lowering pension benefits without bargaining on the issue as required by the public employment relations act (FERA), MCL 423.201 et seq. The MERC agreed that respondents’ unilateral actions constituted a ULR ordered respondents to bargain on the issue, and held that until an agreement is reached, the joint-and-survivor pension benefits must be calculated under the mortality table adopted in 1982. We affirm.
I. UNDERLYING FACTS
The Macomb County Employees’ Retirement System Ordinance (the retirement ordinance) provides pension *154benefits for employees who are members of the system.1 Before 1982, calculation of optional joint-and-survivor pension benefits included consideration of the gender of the retiree because the average lifespans of women and men differed. In 1978, the United States Supreme Court held that the usage of separate tables constituted unlawful gender discrimination. Los Angeles Dept of Water & Power v Manhart, 435 US 702; 98 S Ct 1370; 55 L Ed 2d 657 (1978). The Michigan Attorney General then issued an opinion that public pension systems must adopt gender-neutral mortality tables. OAG, 1981-1982, No 5846, p 29 (January 22, 1981) (“[Adoption of a sexually-neutral retirement table by the [county] would comport with federal and state law.”).2
In 1982, in response to this change in the law, the Macomb County Retirement Commission asked its actuary, Gabriel, Roeder and Smith (GRS), to study the effect on the retirement system if it adopted a single mortality table for all future retirees based on a blending of male and female mortality tables into one gender-neutral, or unisex, table. GRS’s report explained that doing so would result in a range of outcomes. To the degree that the blend was weighted toward male mortality rates, the result would be “substantially lower benefits than at present for women electing a joint and survivor benefit[.]”3 To the degree that the blend was weighted toward female rates, it would result in an increase in costs because it would increase benefits to male retirees greater than the reduction in benefits to female retirees. The report went on to note that the *155only way to “make sure that no participant will receive a lesser benefit than under present procedures” was to adopt a gender-neutral table with a 100% female/0% male blend of mortality rates. The report outlined the specific additional costs to the system using this and several other blends of the male and female mortality tables and offered them as options to the retirement commission. Though cognizant of the increase in overall costs to the retirement system, the retirement commission adopted the 100% female/0% male mortality blend as its gender-neutral mortality table.
The 1982 GRS report also noted that the retirement ordinance required that the optional joint-and-survivor benefit be “the actuarial equivalent” of the standard straight-life benefit. Accordingly, the report recommended adopting a specific rule to govern the meaning of actuarial equivalence in the context of optional benefits. The report recommended adoption of a rule stating that “for purposes of determining amounts of optional benefits, the actuarial equivalent will be based upon a stipulated interest rate and unisex mortality table.” Section 15 of the retirement ordinance was thereafter amended to read:
The Retirement Commission shall from time to time adopt such mortality and other tables of experience, and a rate or rates of regular interest, as are necessary in the Retirement System on an actuarial basis. For purposes of determining actuarial equivalent Retirement Allowances, the Retirement Commission is currently using a 7V2% interest rate and a blending of male and female rates based on the 1971 group annuity mortality table projected to 1984 with ages set back 2 years .... [Emphasis added.]
The retirement commission continued to use the same mortality table for 24 years. However, in 2006, in response to another study conducted by GRS, the retirement commission adopted a new gender-neutral *156mortality table, effective July 1, 2007, which, among other things, changed the assumed ratio of retirees selecting the joint-and-survivor plan from 100% female/0% male to 60% male/40% female. This had the result of lowering the monthly retirement benefit for those under the joint-and-survivor pension. The charging parties demanded bargaining over the change. Respondents rejected the demand, and the charging parties filed ULP charges with the MERC asserting a violation of respondents’ duty under § 10(l)(e) of PERA, MCL 423.210(l)(e), to bargain over benefits.
Although the hearing referee and the MERC reached different rulings, they agreed on two preliminary questions. First, that under PERA, respondents have a duty to bargain over the method by which the joint-and-survivor pension benefits are determined. Second, they agreed that this duty to bargain was not eliminated by the fact that the pension plan is administered by an independent board.
The hearing referee and the MERC disagreed about whether the CBAs fully covered the issue of retirement-benefit calculations so as to satisfy the respondents’ duty to bargain. The hearing referee found that this did because the CBAs incorporated § 26 of the ordinance, which describes the optional joint-and-survivor benefits as “actuarially equivalent” to the straight-life benefits. The hearing referee found that the term “actuarially equivalent” represented a bargained benefit and that, although the meaning of the term “actuarially equivalent” as used by the parties was ambiguous, respondents’ unilateral change in the benefits paid under the optional joint-and-survivor plan did not give rise to a ULP4
*157The MERC concluded that because the term “actuarially equivalent,” as used in the CBAs, was ambiguous, the CBAs did not “contain the entirety of the parties’ agreements with respect to pension benefits.” It went on to conclude that the 24-year practice of using the 100 percent female mortality table constituted a “tacit agreement that the practice would continue.”5 It found, therefore, that the unilateral change in the mortality tables used to calculate benefits constituted a ULE
II. STANDARD OF REVIEW
The MERC’s findings of fact are conclusive if supported by competent, material, and substantial evidence on the record considered as a whole. MCL 423.216(e); Const 1963, art 6, § 28; Amalgamated Transit Union, Local 1564, AFL-CIO v Southeastern Mich Transp Auth, 437 Mich 441, 450; 473 NW2d 249 (1991). Indeed, appellate review of those findings must be undertaken with sensitivity because of the administrative expertise of the MERC. Amalgamated Transit, 437 Mich at 450; Gogebic Community College Mich Ed Support Personnel Ass’n v Gogebic Community College, 246 Mich App 342; 348-349; 632 NW2d 517 (2001). The *158MERC’s legal rulings, however, are not accorded the same deference as its factual findings. “Legal rulings of an administrative agency are set aside if they are in violation of the constitution or a statute, or affected by a substantial and material error of law.” Amalgamated Transit Union, 437 Mich at 450. Of course, whether an error of law has occurred and, if so, whether it is substantial and material are legal questions subject to review de novo. Mich Ed Ass’n v Christian Bros Institute of Mich, 267 Mich App 660, 663; 706 NW2d 423 (2005). Also subject to review de novo are issues of statutory interpretation, Kent Co Deputy Sheriffs Ass’n v Kent Co Sheriff, 463 Mich 353, 357 n 8; 616 NW2d 677 (2000), as well as whether contract language is ambiguous and the meaning of unambiguous contract language, Port Huron Ed Ass’n v Port Huron Area Sch Dist, 452 Mich 309, 323; 550 NW2d 228 (1996).
III. ANALYSIS
In this appeal, we must determine whether respondents violated their duty to bargain when they adopted new mortality tables to calculate joint-and-survivor benefits under the CBAs that had the result of reducing the monthly benefits paid under the joint-and-survivor plan. A public employer commits an unfair labor practice if it refuses to bargain in good faith regarding a mandatory subject of collective bargaining or takes unilateral action on the subject absent an impasse in the negotiations. MCL 423.210(1)(e); Detroit Police Officers Ass’n v Detroit, 391 Mich 44, 54-55; 214 NW2d 803 (1974). A public employer also “commits an unfair labor practice if, before bargaining, it unilaterally alters or modifies a term or condition of employment, unless the employer has fulfilled its statutory obligation or has been freed from it.” Port Huron, 452 Mich at 317.
*159A. MANDATORY BARGAINING
Under § 15 of PERA, MCL 423.215(1), a public employer has a duty to bargain in good faith over subjects found within the scope of the phrase “wages, hours, and other terms and conditions of employment.” See Detroit Police Officers Ass’n, 391 Mich at 54. The duty to bargain in good faith also extends to officers and agents of public employers. MCL 423.210(1).
Respondents assert that they have no duty to bargain over actuarial assumptions because actuarial assumptions are the sole fiduciary responsibility of the retirement commission. Although actuarial assumptions used to determine whether the retirement system is receiving sufficient contributions to maintain adequate funding may not be a subject of bargaining, Bd of Trustees of the Policemen and Firemen Retirement Sys of Detroit v Detroit, 270 Mich App 74, 82-85; 714 NW2d 658 (2006), the purpose of those assumptions is wholly different from those used to calculate joint-and-survivor pension benefits, which, as the hearing referee concluded, are subject to mandatory bargaining:
[T]he mortality table at issue here, although an “actuarial assumption,” is used to calculate the benefits received by retirees from the system. While Respondents’ contributions are affected by the choice of the mortality table used for this purpose, they are also affected by the methods used to calculate final average compensation. In general, if benefits rise, so must Respondents’ contributions. I find that like the methods used to calculate final average compensation, the mortality table used to calculate joint and survivor pension benefits is a matter properly within Respondents’ control and is a mandatory subject of bargaining under PERA.
It is well settled that “ ‘[a]n employer is responsible for its bargaining obligations regardless of whatever *160actions are taken by an independent pension board.’ ” Detroit Police Officers Ass’n v Detroit, 212 Mich App 383, 390; 538 NW2d 37 (1995), aff'd 452 Mich 339 (1996) (citation omitted). Moreover, “[i]t is improper for an employer to remove a subject of mandatory bargaining from the scope of PERA by assigning its management to a body insulated from PERA.” Detroit Police Officers Ass’n, 212 Mich App at 389. Our Supreme Court explained that the basis for this principle originates from the supremacy of state law over local ordinances:
The enactment of an ordinance, however, despite its validity and compelling purpose, cannot remove the duty to bargain under PERA if the subject of the ordinance concerns the “wages, hours or other terms and conditions of employment” of public employees. If the [relevant] ordinance were to be read to remove a mandatory subject of bargaining from the scope of collective bargaining negotiations, the ordinance would be in direct conflict with state law and consequently invalid. Therefore, if. .. [the subject of the ordinance] is a mandatory subject of bargaining, a city ordinance cannot foreclose collective bargaining on the subject. [Detroit Police Officers Ass’n, 391 Mich at 58 (citations omitted).]
Retirement or pension benefits and methods of calculating them are mandatory subjects of collective bargaining. Id. at 63-64; Lieutenants & Sergeants Ass’n v City of Riverview, 111 Mich App 158, 161; 314 NW2d 463 (1981). Accordingly, the existence of an ordinance that grants the retirement commission the authority to adopt the actuarial assumptions used to calculate retirement benefits does not foreclose collective bargaining on the issue, and respondents’ lack of control over the retirement commission cannot excuse avoiding mandatory bargaining under PERA. Moreover, respondent Macomb County has the authority to amend the *161ordinance to comply with any term of a bargained agreement and, as just noted, whatever the ordinance may provide, it cannot foreclose statutorily mandated collective bargaining.6 With regard to the other respondents, we note the ordinance already provides that union members’ retirement benefits are controlled by the terms of the members’ pertinent CBA. Macomb County Employees’ Retirement System Ordinance, § 53b.
Accordingly, we hold, as did the hearing referee and the MERC, that the actuarial assumptions used to calculate the optional forms of benefit payments under the terms of the CBAs are subject to mandatory bargaining under PERA.
*162B. WHETHER THE CBAs ARE AMBIGUOUS IN THE USE OF THE TERM “ACTUARIAL EQUIVALENCE”
Whether the term “actuarial equivalence” is ambiguous or unambiguous determines the standard by which the past actions of the parties may be seen to establish a term or condition of employment. If the term is ambiguous, then a “tacit agreement” that “past practice will continue” renders that practice a term or condition of employment that cannot be unilaterally altered. If the term is unambiguous, then a past practice may constitute a term or condition of employment only if it “is so widely acknowledged and mutually accepted that it amends the contract,” i.e., that “the parties had a meeting of the minds with respect to the new terms or conditions so that there was an agreement to modify the contract.” Port Huron, 452 Mich at 312.
The MERC found that because the retirement ordinance does not provide a definition of “actuarial equivalence,” the term is ambiguous. It further found that the past practice of the use of the 100% female/0% male table had become a term or condition of employment under Port Huron and therefore could not be unilaterally altered. We agree with the MERC’s conclusion that the term “actuarial equivalence” is ambiguous. We also conclude that its findings regarding the past practice were supported by competent, material and substantial evidence. Amalgamated Transit Union, 437 Mich at 450.
The only expert testimony in the record regarding actuarial equivalence was provided by a witness for the charging parties. The expert testified that as long as the same assumptions are used for everyone, they are actuarially equivalent, irrespective of whether the benefits themselves are equal in value:
*163Q. So that if I’m 62 and I retire and [someone else] is 65 and he retires, if we live to our life expectancy, we should get the exact or close as can be calculated the same amount of pension benefits from the social security system, right?
A. That’s the goal, correct.
Q. And that goal is known as actuarial equivalence, isn’t it?
A. I’m not sure. I don’t think so.
Q. ... Is it your understanding that [the GRS report is] saying, look, if you want the surviving spouse options to be the actuarial equivalent of the single life, here’s what we do.
A. Well, I’m not entirely sure what they mean by actuarially equivalent because they are by plan definition actuarially equivalent.
Q. But the value—
A. The value is different than-I’m sorry. Go ahead.
Q. Correct. We talked about that social security example. The value of the benefits were different, right?
A. Yes.
Q. Those were actuarially equivalent?
A. Right.
A. Well, by plan definition actuarial equivalence is using a set of factors, they are equivalent. The plan, when they value these particular benefits, there’s a greater value to somebody who elects a life option — I’m sorry, a joint and survivor option than a life benefit.
*164A.... Individually they’re using factors that are neutral and by definition they’re actuarially equivalent because that’s what they have to be.
A. Well, in terms of valuating a life benefit and a joint and survivorship, you can have a stipulated set of factors that you use in a mortality interest rate and to equate one form to another using those assumptions, they will be equal, in other words, they will be actuarially equivalent based on those assumptions. Like for example, they use the 71 female with a two-year set back. Every option using those tables is equivalent to every other option.
Q. And even though they’re not going to end up being equal?
A. Well, now we’re talking about how do we value these benefits.
Q. Yes.
A. The value of those benefits will not be equal because you’re using a different set of assumptions to value these benefits.
Q. That’s what I’m getting at. There’s a difference between being of equal value and being actuarially equivalent?
A. Right. I mean actuarially equivalent is usually a term used in a plan document to set the optional forms to another optional form. The valuation of those optional forms is a different matter, whole different assumption set. They don’t have to be different. In most cases they are not the same. They’re always different.
Q. In most cases the actual valuation is going to be different than what you came up with to be actuarially equivalent to start with?
A. Right.
*165Q. To be actuarially equivalent is really a matter of choosing what factors that you decide will make it actuarially equivalent, is that right?
A. Right.
Q. Valuation would be I wait and see how long you live. I see how many dollars. I pay it off. I multiply the years times the dollars or times the month and that’s a value and they may not be the same for the same people at all?
A. They may not be the same.
Q. They may be different for male and female?
A. Right.
Q. So let me guess. You could end up having benefits having different values and still be perfectly actuarially equivalent and plans run that way every day and have for the last 25 years of your experience?
A. Yes. [Emphasis added.]
It is abundantly clear from this expert testimony that the MERC had substantial evidence from which to conclude that the term “actuarial equivalence” as used in this case did not unambiguously mean “equal in value.”7
*166C. PAST PRACTICE
The MERC concluded that the parties had engaged in a past practice of accepting a 100% female/0% male mortality table for calculation of the optional joint-and-survivor pension and that this practice constituted a “tacit agreement” with respect to the application of the term “actuarial equivalence” as used in the contract that cannot be unilaterally altered. We agree with this conclusion.8
We further conclude that, even if “actuarial equivalence” had the unambiguous meaning of “equal in value,” the parties’ practices over the subsequent 24 years would have constituted a modification of the contract that could not be unilaterally altered.
*167[T]he unambiguous contract language controls unless the past practice is so widely acknowledged and mutually accepted that it amends the contract. The party seeking to supplant the contract language must show the parties had a meeting of the minds with respect to the new terms or conditions so that there was an agreement to modify the contract. [Port Huron, 452 Mich at 312.]
The record contains the GRS report entitled “Study Regarding Unisex Mortality Tables,” which was presented to the retirement commission on July 26, 1982. The report included discussions regarding the effect of various actions the commission might take and included sample male-female blended morality tables labeled U1 (90% male/10% female), U2 (75% male/25% female), U3 (50% male/50% female), and U4 (0% male/100% female). The added costs of these options to the retirement system as a percentage of payroll were estimated as follows: U1 (0.01 percent); U2 (0.19 percent); U3 (0.47 percent); U4 (1.01 percent). The retirement commission adopted the U4 (100 percent female) mortality table for determining optional joint-and-survivor benefits under the retirement ordinance. The retirement commission continued using the 100 percent female mortality table until 2006.
Respondents claim that the adoption of the 100 percent female table “unknowingly” created unequal payments and that “it took a team of actuaries to discover the overpayments in 2006.” However, this is not borne out by the record. The initial GRS study that resulted in the adoption of the 100 percent female table specifically indicated that there would be an increased cost to the system. In fact, the 100 percent female table created the highest effect on costs of the four options. In spite of that, the retirement commission elected to adopt that table. Further, as conceded by Macomb County’s finance director, whatever “inequality” was occurring in 2006 had been *168occurring over the past 24 years. Moreover, the record indicates that GRS performed an experience study in 1993 to review the actuarial assumptions the system was using to calculate benefits and did not recommend any changes to the system.9
In any event, § 15 of the retirement ordinance was amended shortly after the adoption of the 100 percent female mortality table to read: “For purposes of determining actuarial equivalent Retirement Allowances, the Retirement Commission is currently using a lxk% interest rate and a blending of male and female rates based on the 1971 group annuity mortality table projected to 1984 with ages set back two years.” This amendment was clearly adopted to set the optional joint-and-survivor benefits at values that were not strictly “equal in value” to those provided in the straight-life benefit. The original GRS plan expressly stated:
COMMENT C: The Retirement System Ordinance provides that an optional benefit will be “the actuarial equivalent” of the standard benefit. The Retirement Commission could adopt a rule stating that for purposes of determining *169amounts of optional benefits, the actuarial equivalent will be based upon a stipulated interest rate and unisex mortality table. This could eliminate the need for an ordinance change.
Indeed, GRS’s 1982 report went on to provide: “[A] unisex approach subsidizes optional elections for men. If, in recognition of this, more men elect joint and survivor benefits than in the past, cost to the system will be greater than is shown.” Thus, the initial GRS study both recognized and explicitly informed the county that adoption of any of the unisex tables could result in optional joint-and-survivor benefits that were not equal in value to those of the straight-life benefits. Rather, every unisex table would, to some degree, create an approach that resulted in greater benefits for men making one of the optional elections. The report went further, suggesting that the retirement commission adopt language designed to circumvent the equivalence requirement by providing an open-ended formulaic definition for “actuarial equivalent” that would be based on an interest rate and unisex mortality table.
Consequently, assuming respondents’ definition of “actuarially equivalent” is correct and unambiguous, the retirement commission’s selection of the 100 percent female table, in conjunction with respondents’ adoption of the suggested language and the continued use of the 100 percent female table for 24 years, even after actuarial review in 1993, represented a “definite, certain, and intentional” modification of the actuarial equivalent requirement that was “unequivocal.”10 See *170Port Huron, 452 Mich at 329. The acceptance of this provision is clear from the language in the controlling CBAs, which provide that retirement benefits are to be continued “as presently constituted,” i.e., in accordance with the 100% female/0% male mortality table.
The evidence presented showed that, despite respondents’ claim that the clear terms of the CBAs required that the joint-and-survivor pension be “equal in value” to the straight-life pension, for 24 years — from adoption until 2006 — the parties continuously used the 100 percent female mortality table without regard to whether it would create equal-in-value pensions. Accordingly, even were respondents correct that the term “actuarial equivalence” as used by the parties was unambiguous, we would still find for the charging parties because the usage of the 100 percent female mortality table was “so widely acknowledged and mutually accepted that it creat[ed] an amendment to the contract.”11 Port Huron, 452 Mich at 329.
*171IV CONCLUSION
In sum, we agree with the MERC that the term “actuarial equivalence” is ambiguous and that a past practice of accepting a 100% female/0% male mortality table constituted a tacit agreement by the parties that that table would continue to be used. We further conclude that, even if “actuarial equivalence” had the unambiguous meaning of “equal in value,” there was sufficient evidence of a meeting of the minds that the 100% female/0% male table was accepted for calculating pension benefits in lieu of any “equal in value” requirement that the table could not be unilaterally changed. Accordingly, we agree with the MERC that “Respondents violated their duty to bargain when, without bargaining, they changed the method used to calculate joint and survivor benefits under the parties’ collective bargaining agreements.”12
Affirmed.
Fitzgerald, J., concurred with Shapiro, J.This includes employees who are not represented by the unions.
The opinion also included a discussion of “actuarially equivalent,” although in a different context than at issue here, but noted that the term was undefined.
This would also result in slightly higher benefits being paid to men than had been paid to that date.
The hearing referee apparently concluded that this ambiguity in contract language could be resolved through grievance arbitration and *157that, therefore, a unilateral change in optional retirement benefits did not give rise to a ULE However, even if the hearing referee was correct that the matter could give rise to a grievance, this would not eliminate the presence of a ULP under PERA. In Bay City Sch Dist v Bay City Ed Ass’n, Inc, 425 Mich 426, 436-437; 390 NW2d 159 (1986), our Supreme Court held that “[w]here a controversy gives rise to both contractual and statutory claims .. ., grievants have been allowed to pursue different avenues of relief in different fora.” If contractual and statutory claims arise out of the same controversy, parallel proceedings are permitted. Id. at 437-440.
The hearing referee did not address whether the use of a particular mortality table for 24 years represented a past practice rising to the level of a term or condition of employment.
The dissent agrees with us that employers cannot avoid their duty to bargain over actuarial assumptions that govern pension-benefit amounts and that the assumptions in this case must therefore be subject to bargaining. Curiously, despite its explicit rejection of respondent’s view that the duty to bargain does not apply, the dissent also observes that respondent’s position “has merit.” However, other than a brief discussion of the terms “trustee” and “agent,” the dissent fails to explain what it finds meritorious in the argument. Indeed, the distinction between the terms “trustee” and “agent” is fully consistent with the notion that the retirement commission, as a trustee, may act to secure and manage reserves in order to ensure the county’s ability to pay the benefits, the amount of which is exclusively a matter for bargaining between the county and its employees.
As for the dissent’s concern regarding MCL 46.12a(l)(b), we note that the language cited by the dissent requires that pension or retirement benefits be granted “according to a uniform scale for all persons in the same general class or classification.” It is clear that a uniform scale is being used here. Each retiree is receiving benefits under the same 100% female/0% male table. The dissent is concerned that the amounts of benefits ultimately received are equal, but that is entirely different from whether a uniform scale is used to calculate the benefits. A nonuniform scale would, for example, require that the benefits received by men he calculated using a different mortality table than for those received by women. It was precisely to adopt a uniform scale that the 100% female/0% male table was adopted for all employees, even though separate scales for male and female retirees likely resulted in greater consistency in the actual dollar amount of benefits received by each retiree.
The dissent ignores the testimony and documentary evidence of the parties’ use of the term “actuarial equivalence” and instead elects to impose its own definition using a dictionary to define the term despite the fact that the term “actuarial equivalence” cannot be found in the cited dictionary. Although undefined contract terms are generally interpreted in accordance with their “commonly used meaning,” Frankenmuth Mut Ins Co v Masters, 460 Mich 105, 113-114; 595 NW2d 832 (1999), courts are not to resort to a lay dictionary for specialized terms of art, particularly in technical fields in which professionals obtain advanced degrees, see People v Thompson, 477 Mich 146, 151-152; 730 NW2d 708 (2007). For example, we give tax terms their specialized meanings. Prod Credit Ass’n of Lansing v Dep’t of Treasury, 404 Mich 301, 312; 273 NW2d 10 (1978) (stating that “terms of art” should be interpreted "in *166accordance with the experience and understanding of those who would be expected to use and interpret the act”). Similarly, we routinely accept testimony about the meaning of medical and engineering terms of art.
Ignoring this rule, the dissent attempts to craft a definition by breaking the term into component parts, finding definitions for each word, and then rejoining them. The absurdity of attempting this with a technical term is evident when one considers medical terms such as “gall bladder.” The dictionary defines “gall” as “[bjitterness of feeling; rancor” and “[ojutrageous insolence,” and defines “bladder” as “[a]ny of various distensible membranous sacs ... found in most animals and that serve as receptacles for fluid or gas.” The American Heritage Dictionary of the English Language (2001). Adopting the dissent’s approach to defining terms of art, the legally binding definition of “gall bladder” would be “a distensible membranous sac found in animals that serves as a receptacle for bitterness of feeling, rancor, and outrageous insolence.”
The dissent concludes that the MERC’s decision is not supported by competent, material and substantial evidence. However, we reiterate that this Court must be extremely deferential when reviewing the MERC’s factual findings. “ ‘Review of factual findings of the commission must be undertaken with sensitivity, and due deference must be accorded to administrative expertise. Reviewing courts should not invade the exclusive fact-finding province of administrative agencies by displacing an agency’s choice between two reasonably differing views of the evidence.’ ” St Clair Intermediate Sch Dist v Intermediate Ed Ass’n/MEA, 458 Mich 540, 553; 581 NW2d 707 (1998) (citation omitted).
The report stated that it would be reasonable to expect that using a merged gender table would result in “substantially lower benefits than at present for women electing a joint and survivor benefit, and slightly higher benefits than at present for men.” On the other hand, the report stated that using all female factors for future retirees would “make sure that no participant will receive a lesser benefit than under present procedures.” However, this option “would necessarily entail a cost for the plan since men electing optional forms of payment would be subject to a smaller reduction in benefits than required on an actuarial basis.” Thus, the record evidence is that the retirement commission selected the 100 percent female table, even though it resulted in the highest costs to the system, because it left the female benefits the same and increased the male benefits, as opposed to adopting some other merged table, which would have left male benefits the same or better, but reduced female benefits. This suggests that the selection was made without regard to cost or whether the options were equal in value, but was instead based on how the adopted table would affect those receiving the benefits.
The dissent contends that the parties’ knowledge that the benefits were unequal was insufficient to amend the parties’ agreement, citing Port Huron, 452 Mich at 332. We agree. That is why we have relied not simply on their knowledge, but their actions to find evidence of an unequivocal modification. We also note that in Port Huron, the Court held that there was no evidence that the district had intentionally taken *170actions contrary to the agreements; rather, they happened by happenstances or oversight. Id. at 332 n 22. Such is clearly not the case here, where there is no happenstance or oversight, but deliberate acceptance based on a clear understanding of the implications. We reject the dissent’s conclusion that the GRS reports cannot establish respondents’ intent because it was respondents that incorporated language identical to that from the GRS report in the ordinance establishing the adoption of the 100 percent female table. Ironically, the dissent asserts that the 1982 GRS report cannot establish respondents’ intent in the same footnote that it acknowledges respondents’ amendment of the retirement ordinance in accordance with the report’s recommendation. The amendment is necessarily tied to the report. Accordingly, it is appropriate to use it in determining their intent. In any event, the dissent fails to provide or cite any facts in the record that are inconsistent with the MERC’s findings.
The dissent alludes to the fact that not permitting the retirement commission to change the actuarial tables used to calculate the retirement benefits could potentially destabilize the retirement funds. However, there is no evidence in the record to support such an assertion, nor has the commission concluded that additional funding is either necessary or unavailable. Thus, the dissent’s expression of concern about the financial stability of the retirement system appears to be intended to inflame rather than clarify. Moreover, our conclusion that the parties had *171a past practice of using the 100 percent female table does not prevent the parties from selecting a new table. It merely requires that they do so at the bargaining table rather than by a unilateral change.
Finally, the dissent’s claim that the use of a gender-neutral table based on a 100 percent female assumption results in inequities constitutes a criticism of using a gender-neutral table at all. Any gender-neutral table will invariably result in some difference in payments given that women in fact do generally live longer than men and that this reality cannot, by definition, be reflected in a gender-neutral table. Moreover, the dissent’s assertion that the commission “did not accept a sex-blended mortality table until 2006” is simply wrong. A100 percent female table is still a sex-blended mortality table; the blending just assumes zero percent men. What makes it a sex-blended table is that the same assumptions are used for everyone, as opposed to having different tables to calculate benefits for men and women.
In light of our conclusion, we do not address the charging parties’ claim that respondents had a separate duty to bargain over the effects of implementing the new mortality table.