Plaintiffs appeal by right the trial court’s ruling, after a bench trial, that defendant township’s zoning regulations did not cause an unconstitutional taking. We affirm.
I. FACTS
Plaintiffs Robert, Marcia, and Dennis Leland own four parcels of land totaling approximately 220 acres (the property) in Northfield Township. Before the events that gave rise to the present dispute, the property had been zoned AR (Agriculture District), and had been farmed for over 100 years.
In January 2002, plaintiff Grand/Sakwa of North-field, LLC (or its predecessor or agent), executed an agreement to purchase the property from the Lelands for $30,000 per acre and paid a nonrefundable deposit of $25,000. On June 30, 2003, plaintiffs applied to rezone *140the property from AR to SR-1 (Single-Family Residential District One). SR-1 zoning allows up to four dwellings per acre with sewer service, or one dwelling per acre without sewer service. On November 18, 2003, the township board approved the rezoning, limited to 450 homes. Following that approval, township residents organized a successful referendum, held May 18, 2004, that overruled the board’s decision, thereby leaving the property zoned AR. After the referendum, the North-field Township Zoning Board of Appeals denied plaintiffs’ requests for use or dimensional zoning variances.
Plaintiffs filed this lawsuit on October 22, 2004. They alleged that application of any zoning classification more restrictive than SR-1 constituted a regulatory taking. Shortly after the lawsuit was filed, a new township board took office. A majority of the new board’s members were organizers or supporters of the referendum that overruled the board’s 2003 rezoning of the property to SR-1. The new board fired its planner and took action to amend the zoning ordinances, rezoning the property from AR to LR (Low Density Residential District). The LR classification itself was amended to allow only one home per two acres, instead of the previously allowed one home per acre.
At the time of the bench trial, therefore, the property was zoned LR. Plaintiffs argued that whether or not a regulatory taking had occurred should be determined by evaluating the AR zoning that existed at the time the lawsuit was filed. The township argued that whether or not there was a taking should be determined on the basis of the LR zoning that existed at the time the trial court heard the proofs and rendered a decision. Therefore, before determining whether the zoning constituted a regulatory taking, the trial court had to determine which zoning ordinance was to be tested. The trial *141court ruled that the relevant zoning ordinance was the one then in place, i.e., LR zoning. After the full trial, the court held in the township’s favor on all of plaintiffs’ claims, finding no constitutional violation. Plaintiffs appealed by right.
II. THE RELEVANT ZONING ORDINANCE
Plaintiffs first argue that the trial court erred by ruling that their challenge was to the LR zoning classification in place at the time the court made its decision rather than the AR classification in place when the lawsuit was filed.1 We disagree.
Plaintiffs’ view that the zoning classification in effect when their suit was filed should apply is contrary to the guiding caselaw. We have stated that “[t]he general rule is that the law to be applied is that which was in effect at the time of decision [by the trial court]. Thus, if a zoning ordinance has been amended [after suit was filed] ... a court will give effect to the amendment^]” Klyman v City of Troy, 40 Mich App 273, 277-278; 198 NW2d 822 (1972), citing City of Lansing v Dawley, 247 Mich 394; 225 NW 500 (1929).
This general rule is subject to two narrow exceptions. “A court will not apply an amendment to a zoning ordinance where (1) the amendment would destroy a vested property interest acquired before its enactment, or (2) the amendment was enacted in bad faith and with unjustified delay.” Lockwood v Southfield, 93 Mich App 206, 211; 286 NW2d 87 (1979) (citation omitted).
The first exception does not apply here because there is no vested property interest at issue. At the time of the *142sale, the property was zoned AR and remained so until the amendment rezoning it LR was adopted. Plaintiffs concede that the township board’s 2003 decision to rezone the property SR-1 never took effect because it was superseded by the referendum. Thus, there was never any vested right to develop the property under any zoning classification other than AR.
The second exception applies if the trial court finds that the newer classification “was enacted for the purpose of manufacturing a defense to plaintiffs suit.” Landon Holdings, Inc v Grattan Twp, 257 Mich App 154, 161; 667 NW2d 93 (2003) (quotation marks and citation omitted). In Klyman, we defined the exception more narrowly, stating that a change in an ordinance shall be applied unless it “was ... enacted simply to manufacture a defense.” Klyman, 40 Mich App at 279 (emphasis added).
Plaintiffs have cited only one case of record, Willingham v Dearborn, 359 Mich 7; 101 NW2d 294 (1960), in support of their view on this issue. There, the plaintiff was denied a permit to construct a service garage on his property on the ground that the plans did not provide for a 160-foot setback. Id. at 8. However, no ordinance required such a setback. Id. Accordingly, the plaintiff filed suit to require the defendant city to issue a building permit. Id. While the suit was pending, the city adopted an ordinance requiring, for the first time, a 160-foot setback. Id. at 8-9. The trial court declined to consider the amended ordinance, finding that it “ ‘can place no other construction’ ” on the city’s actions other than it serving as a basis to retroactively legitimize its denial of the sought-after building permit. Id. at 9. Our Supreme Court held that the trial court properly declined to apply the zoning ordinance adopted during litigation. Id. at 10.
*143The facts in Willingham bear no resemblance to those in the instant case. In that case, the city sought to adopt an ordinance tightening its zoning requirements to bar a use that was permitted when the plaintiff initially sought the building permit.2 Here, the development sought by plaintiffs was never within the zoning classification, and the ordinance they seek to exclude from consideration is one that grants, rather than restricts, development rights.
In all the other cases addressing the issue, our courts have held that it is the postsuit ordinance that controls. Franchise Realty Interstate Corp v Detroit, 368 Mich 276, 279; 118 NW2d 258 (1962); London Holdings, 257 Mich App at 165; MacDonald Advertising Co v McIntyre, 211 Mich App 406, 410; 536 NW2d 249 (1995); Lockwood, 93 Mich App at 211; Klyman, 40 Mich App at 279.
We agree with plaintiff that the trial court wrongly characterized the relevant test as requiring application of the newer zoning ordinance unless its adoption was “done solely” to improve the municipality’s litigation posture. However, we similarly reject the notion that if improving the municipality’s litigation position plays any role in the decision to adopt the new ordinance, bad faith has been sufficiently established. None of the cases cited by the parties adopt such a standard, and there was evidence of mixed motives in several of the cases in which Michigan courts held that the newer ordinance applied. Accordingly, we will not void a municipality’s action simply because it served to strengthen its litigation position. The factual determi*144nation that must control is whether the predominant motivation for the ordinance change was improvement of the municipality’s litigation position. And, because this is a factual determination to be made by the trial court, we review it for clear error. MCR 2.613(C).3
In making their argument to the trial court, plaintiffs relied on several quotes from board meetings that demonstrated that the board was partially motivated by a desire to defend against the instant litigation. However, the trial court, after hearing the evidence, concluded that “the rezoning to LR was not done solely as an attempt to improve the Defendant’s position at trial.” The court further noted that the township’s “GMP [Growth Management Plan] was amended to reflect that the LR zoning permits ‘limited residential development while preserving significant areas of agriculture, open space, and natural features’, and preserves ‘a predominantly rural character,’ while providing ‘certain residential and public uses . . . compatible with the principal use....’” It also noted that the rezoning to LR was undertaken “pursuant to recommendations from a newly hired land use planner.” In other words, the board made a decision to allow residential development that maintained a rural character, rather than allow either more substantial development or none at all. The trial court also noted that the zoning board had previously granted plaintiffs’ request to rezone the property SR-1. Plaintiffs suggest that we *145should ignore this fact, since the membership of the township board changed after the time the request was granted and the new board was hostile to development. However, the events can fairly be read as demonstrating recognition by both boards that development was in order, though they disagreed on the degree of that development. Plaintiffs suggest that the township was opposed to all development, as demonstrated by the referendum, and only adopted the LR zoning as a litigation strategy. However, plaintiffs concede that, after the old board adopted the SR-1 zoning, it was not possible to propose a referendum that would void the SR-1 zoning and institute LR zoning in its place. The only mechanism for the residents to challenge the SR-1 zoning in a referendum was to put it to an up or down vote, i.e., SR-1 or AR.
Given the deference we show to a trial court’s superior ability to judge the credibility of witnesses, Glen Lake-Crystal River Watershed Riparians v Glen Lake Ass’n, 264 Mich App 523, 531; 695 NW2d 508 (2004), and that affording such deference leads us to the conclusion that the evidence did not demonstrate that obtaining a litigation advantage was the predominate reason for the ordinance change, we find that the trial court did not clearly err by applying LR zoning as the law of the case.
III. REGULATORY TAKING
Plaintiffs next argue that the LR zoning constitutes an unconstitutional governmental taking.4 We disagree.
Both the United States and Michigan Constitutions prohibit the government from taking private property *146for public use without just compensation. US Const, Am V; Const 1963, art 10, § 2. A taking may be caused by overly burdensome regulations in two situations: if the regulation does not advance a legitimate state interest or if “the regulation denies an owner economically viable use of his land.” K & K Constr, Inc, v Dep’t of Natural Resources, 456 Mich 570, 576; 575 NW2d 531 (1998) (K & K I). The second type of taking may be found two ways. First, there is a “categorical taking” if a regulation denies the owner of “all economically beneficial or productive use of land.” Lucas v South Carolina Coastal Council, 505 US 1003, 1015; 112 S Ct 2886; 120 L Ed 2d 798 (1992). Second, a regulatory taking may be found on the basis of the traditional balancing test established in Penn Central Transp Co v New York City, 438 US 104; 98 S Ct 2646; 57 L Ed 2d 631 (1978).
Plaintiffs do not claim a categorical taking, arguing only that the trial court should have found a taking under the Penn Central test. Penn Central calls for the court to consider three factors: the character of the government’s action, the economic effect of the regulation on the property, and the extent by which the regulation has interfered with distinct, investment-backed expectations. Id. at 124; K & KI, 456 Mich at 577.
A. CHARACTER OP THE GOVERNMENTAL ACTION
Penn Central provides that the central question in analyzing the character of the governmental action is whether that action constituted a physical invasion. Penn Central, 438 US at 124. Where it does, the factor weighs in favor of finding a taking. Id. Here, it is undisputed that the actions of the township board did not create a physical invasion of plaintiffs’ property. Zoning regulations are not a physical invasion. Id at *147125. Indeed, the Penn Central Court cited zoning ordinances as “the classic example” of governmental action affecting land interests and stated that such regulations are generally permissible. Id.
Penn Central further provides that the “government may execute laws or programs that adversely affect recognized economic values,” and that a regulatory taking will not be found where a state tribunal reasonably concludes that the land-use limitation promotes the general welfare, even if it “destroy[s] or adversely affect[s] recognized real property interests.” Id. at 124-125.5 Accordingly, the trial court did not clearly err by finding that the first prong of the Penn Central test weighed in the township’s favor.
B. ECONOMIC EFFECT OF THE LR ZONING
Plaintiffs maintain that the LR zoning created a loss of the value that they would have received had the *148property been zoned SR-1. The township does not dispute this allegation, because it is clear that property on which 450 homes can be built has greater value than the same property on which 80 homes can be built. However, the question, contrary to plaintiffs’ suggestion, is not simply whether their preferred zoning results in an increase in the value of the land. If that were the case, virtually every zoning regulation could be successfully challenged as a regulatory taking. Plaintiffs correctly assert that a comparison of the values is relevant to the overall analysis, but it is by no means controlling.
In Penn Central, 438 US at 131, the Court agreed that the regulation in question diminished the value of the plaintiffs’ property. However, it relied on the fact that the regulation “does not interfere in any way with the present uses of [the property].” Id. at 136. The same is true here. Indeed, the LR zoning classification allows a much more valuable use of the property than does AR zoning, the classification in effect when plaintiffs entered into the purchase agreement. Moreover, although the restriction imposed in Penn Central was quite significant in that it prevented the plaintiffs from building their planned structure, the Court held that this was insufficient to establish a taking because the plaintiffs were not “denied all use of even those preexisting [property] rights.” Id. at 115-117, 137. In the instant case, the only preexisting rights of use of the property were those permissible under AR zoning. No rights existing under AlR zoning are denied under the LR zoning; indeed, as noted, the LR zoning substantially expands plaintiffs’ land use rights, allowing residential development to occur.
Plaintiffs also argue that the trial court made several errors in its decision to admit or exclude certain evi*149dence regarding the value of the property. We review these evidentiary challenges for an abuse of discretion. Lewis v LeGrow, 258 Mich App 175, 200; 670 NW2d 675 (2003). First, plaintiffs contend that the trial court should not have admitted evidence of a 1998 sale of 77 acres by the Lelands for $10,000 per acre. Plaintiffs argue that because that purchase was made by the Whitmore Lake School District, it was likely a unique event and should not be considered in determining value. Second, plaintiffs object to the consideration of a 1996 appraisal that valued 120 acres of the property at $3,500 per acre. Plaintiffs note that the appraisal was made under the previous AR zoning and was, therefore, inapplicable to a current valuation. Third, plaintiffs assert that the trial court should not have considered evidence that a local church was interested in purchasing 15 acres of the property for as much as $43,000 per acre. Plaintiffs argue that this church would not have developed the land for profit, that the sale would affect only a small portion of the property, and that the church’s interest may have been a result of the anticipated residential development. We conclude, particularly given that this was a bench trial, that each of these challenges addresses the weight to be given to the evidence, not its admissibility.
Plaintiffs similarly assert that the trial court should have excluded the testimony of Robert Walworth, defendant’s expert witness, who testified regarding the economic viability of the property under the LR and AR zoning classifications. Plaintiffs argue that the method employed by Walworth was improper and that he did not provide any useful information regarding the feasibility of development. Plaintiffs’ expert, John Widmer, detailed what he perceived to be deficiencies in Walworth’s calculations. The trial court heard this testimony and appropriately held that it went to the weight *150of the evidence, rather than its admissibility. Walworth testified that he used estimated costs of development, some of which came from plaintiffs’ evidence, added in a profit margin, and calculated the average price at which each lot would need to be sold in order to be economically viable. He explained certain differences between his analysis and that of Widmer, including that Widmer’s analysis involved a prospective rate of return, whereas Walworth calculated a simple profit factor that did not discount future cash flows back to the present. Plaintiffs’ objection appears to stem from the fact that Walworth and Widmer simply tried to calculate different things. In the context of a bench trial in which the experts underwent extensive cross-examination, we find no abuse of discretion in the trial court’s admission of Walworth’s testimony.
Plaintiffs next argue that the trial court erred by excluding testimony from lay witnesses regarding the value of the property under SR-1 zoning. The court held that this evidence was only relevant to damages and deferred the testimony pending a ruling on the cause of action. We agree with plaintiffs that the trial court should have taken the testimony, given that the balancing test “requires at least a comparison of the value removed with the value that remains.” K & K I, 456 Mich at 588 (quotation marks and citations omitted). However, we cannot conclude that the court’s ruling constituted an abuse of discretion, given that the township conceded that the property would have greater value if zoned SR-1 and that the court heard extensive testimony from plaintiffs’ appraisal expert regarding the extent of the difference in value, including the admission of an economic feasibility report. It is well established that we defer to the trial court’s credibility determinations. Glen Lake, 264 Mich App at 531. Moreover, the relevant caselaw provides that even a large *151diminution of value does not, standing alone, constitute a taking. K & K Constr, Inc v Dep’t of Environmental Quality, 267 Mich App 523, 553; 705 NW2d 365 (2005) (K & K II) (noting that the United States Supreme Court has refused to find takings in cases involving 75% and even 87.5% diminutions of value). There was sufficient evidence, albeit evidence with which plaintiffs disagreed, to allow the trial court to properly conclude that the diminution in value was not so significant as to weigh the second prong of the Penn Central test in plaintiffs’ favor.
C. INTERFERENCE WITH DISTINCT INVESTMENT-BACKED EXPECTATIONS
The role of investment-backed expectations was discussed at length in K & K IT. In that case, we did not wholly foreclose a taking claim based on a regulation in effect at the time the land was purchased; however, we held that “[a] key factor is notice of the applicable regulatory regime[.]” Id. at 555. A claimant who purchases land that is subject to zoning limitations with the intent to seek a modification of those limitations accepts the business risk that the limitations will remain in place or be only partially modified. As Justice O’Connor noted in her concurrence in Palazzolo v Rhode Island, 533 US 606, 633; 121 S Ct 2448; 150 L Ed 2d 592 (2001), “the regulatory regime in place at the time the claimant acquires the property at issue helps to shape the reasonableness of those expectations.” Notice of the regulation cannot absolutely bar a taking claim, but should be taken into account. When plaintiffs entered into the purchase agreement, they were aware that the property was zoned AR. Plaintiffs argue that they had a reasonable expectation that the zoning classification would change, but they do not refer us to any evidence in support of that proposition. Instead, *152they rely on the fact that the previous board agreed to change the zoning to SR-1 in 2003. However, as plaintiffs conceded at oral argument, they understood that the zoning modification adopted by the board remained subject to a timely referendum challenge6 and that, when a timely challenge is made, the new zoning classification would not take effect at all, unless approved by the voters. Thus, plaintiffs’ implicit suggestion that the property was for some time subject to the SR-1 zoning and that the SR-1 classification was taken away from them, after they spent money on the project, fails because of the fundamental fact that the property was never actually zoned SR-1. Moreover, any funds expended by plaintiffs once the petition was filed cannot be said to have been expended with a reasonable expectation that the proposed development could be built, in light of the referendum challenge.
In sum, Grand/Sakwa chose to purchase AR-zoned property upon which, according to its own arguments and expert testimony, it could not build an economically viable development. It made efforts to get the zoning changed and failed. Contrary to plaintiffs’ claim, we are unaware of any caselaw that provides that monies expended in pursuit of a zoning change are, themselves, grounds to claim a taking.
The trial court did not clearly err by holding that the third Penn Central factor favored the township. Accord*153ingly, because each of the Penn Central factors weighed in the township’s favor, we find that the trial court did not err by finding that the rezoning of the property to LR did not constitute an unconstitutional regulatory taking.
IV DUE PROCESS AND EQUAL PROTECTION
Finally, plaintiffs argue that rezoning the property to LR violated their due process and equal protection rights because it rendered the property not economically viable.7 We disagree.
To show a violation of substantive due process, “a plaintiff must prove (1) that there is no reasonable governmental interest being advanced by the present zoning classification, or (2) that the ordinance is unreasonable because of the purely arbitrary, capricious and unfounded exclusion of other types of legitimate land use from the area under consideration.” A & B Enterprises v Madison Twp, 197 Mich App 160, 162; 494 NW2d 761 (1992). Plaintiffs argue that AR and LR zoning render the property dead land. However, as already discussed, the trial court did not clearly err when it found that this was not the case. Accepting the trial court’s finding on that point, plaintiffs cannot show that it was a due process violation for the township to zone the property LR.
Regarding the equal protection challenge, it is true that the rezoning to LR affected only plaintiffs’ property. However, it is not the case that the rest of the AR-zoned land in the township was rezoned to SR-1, with only plaintiffs left behind. Rather, after the referendum, the township acted to give plaintiffs at least *154some of the relief they sought without completely abandoning the traditionally rural character of the area. It was appropriate to rezone only plaintiffs’ property when it was the only property for which a change in zoning was sought. Moreover, the amendment of the LR zoning classification itself affected all LR-zoned properties, not just plaintiffs’ property.
The township’s goals of controlling growth and maintaining open space were legitimate, the method chosen was not arbitrary or capricious, and plaintiffs’ property was not improperly singled out under the circumstances. Moreover, following the referendum denial of the SR-1 zoning, the township acted to provide plaintiffs with a more economically viable zoning classification than AR.
Accordingly, we find that the trial court did not err by ruling for the township on plaintiffs’ due process and equal protection claims.
Affirmed.
SHAPIRO, EJ., and HOEKSTRA and WHITBECK, JJ., concurred.Following a bench trial, we review the trial court’s findings of fact for clear error and review de novo its conclusions of law. City of Flint v Chrisdom Props, Ltd, 283 Mich App 494, 498; 770 NW2d 888 (2009).
The Willingham Court also criticized the city for unduly delaying its change of classifications. Id. at 10. Plaintiffs here do not claim that the township unduly delayed action, complaining instead that the township acted too quickly.
The township urges us to apply an abuse of discretion standard, citing Landon Holdings, 257 Mich App at 161, for the view that “a trial court’s decision to admit or exclude evidence of ordinance amendments during litigation” is reviewed “for an abuse of discretion!.]” We decline to do so because the trial court was not asked to decide whether evidence of an ordinance change was admissible, but rather to make a factual finding that would determine which of the ordinances was to undergo constitutional review.
We review de novo constitutional questions. Great Lakes Society v Georgetown Charter Twp, 281 Mich App 396, 425; 761 NW2d 371 (2008).
Essentially ignoring the question of physical invasion, plaintiffs argue that this prong of the Penn Central test should weigh in their favor, relying on Pulte Land Co LLC v Alpine Twp, unpublished opinion per curiam of the Court of Appeals, issued September 12, 2006 (Docket Nos. 259759 and 261199). Because Pulte is unpublished, it is not binding. MCR 7.215(C)(1). Moreover, in Pulte, while the zoning ordinance restricted the plaintiffs’ land to agricultural uses, the master plan showed that the area was planned for medium-density residential development in the future. Pulte, unpub op at 2. Because the zoning ordinance and master plan were in conflict, this Court found that, in the long term, retaining the agricultural zoning of the plaintiffs’ property would harm the public interest, rather than serve it. Id. at 5-6. On that basis, this Court held that the character of the governmental action favored the plaintiffs. Id. at p 6. By contrast, in the instant case, plaintiffs assert that the township’s growth management plan called for the residential development of plaintiffs’ property. However, plaintiffs fad to note that the plan, according to a report commissioned by plaintiffs, specified a density of one dwelling unit per five acres, i.e., the exact density permitted under AR zoning. The excerpts of the plan in the record do not demonstrate an intent to adow high-density residential development of plaintiffs’ property.
MCL 125.3402(2) provides that a petition to overrule the zoning ordinance may be submitted within 30 days of its publication. The petition must contain the signatures of “not less than 15% of the total vote cast within the zoning jurisdiction for all candidates for governor at the last” gubernatorial election. The statute provides that if such a petition is filed and determined to contain the requisite number of signatures, “the zoning ordinance adopted by the legislative body shall not take effect until... the ordinance is approved by a majority of the registered electors residing in the zoning jurisdiction!.]” MCL 125.3402(3)(c).
We review de novo constitutional questions. Great Lakes Society, 281 Mich App at 425.