Petersen v. Magna Corp.

We granted leave to appeal in this case to determine the parties against whom attorney fees may be prorated under MCL 418.315(1). We conclude that the term “prorate” in MCL 418.315(1) applies only to employers and their insurance carriers. Accordingly, we affirm the judgment of the Court of Appeals.

I. FACTUAL BACKGROUND AND PROCEDURAL HISTORY

This case involves a dispute over workers’ compensation benefits. Plaintiff Rick Petersen began working *304for codefendant Koleasco, a trucking company, in February 1997. In March 1997, Koleasco hired codefendant BCN Transportation Services (BCN), a human resources “employee leasing” company, to administer its employee benefits.

In November 1997, plaintiff was injured when he fell from a flatbed truck while securing a load of Christmas trees. After the accident, he underwent surgery on his right foot and applied for workers’ compensation benefits. The following year, he required treatment for back pain, which his treating physician believed was caused by the November 1977 fall.

Several questions were taken to a workers’ compensation magistrate: (1) who was plaintiffs employer at the time of his injury, BCN or Koleaseco? (2) was Midwest Employers Casualty Company the relevant insurer for workers’ compensation purposes? (3) was plaintiff disabled? and (4) if so, which injury caused his disability? The magistrate bifurcated these issues into two trials.

In the first trial, the magistrate ruled that plaintiff was a Koleaseco employee on the date of his injury despite the fact that BCN paid his wages. Thus, because BCN had stipulated that it was plaintiffs employer, the magistrate ruled that both BCN and Koleaseco were plaintiffs employers and both were liable for plaintiffs workers’ compensation benefits. On appeal, the Workers’ Compensation Appellate Commission (WCAC) affirmed that ruling.

In the second trial, the magistrate considered (1) was plaintiffs counsel entitled to an attorney fee of 30 percent of plaintiffs medical bills unpaid by defendant? and (2) who was responsible for paying plaintiffs future medical and weekly benefits? With respect to plaintiffs attorney fees, the magistrate ruled:

*305[Although Midwest... was paying the plaintiff weekly benefits, it refused to pay medical bills related to [plaintiffs] injury.... The total amount of the medical bills incurred... which defendant refused to pay is $153,448.54.1 find that plaintiffs counsel is entitled to a 30 percent attorney fee for these unpaid medical bills under Section 315(1).

With respect to whether BCN or Koleaseco was responsible for paying plaintiffs ongoing medical and weekly benefits, the magistrate ruled that BCN and its insurance carrier, Midwest, were primarily responsible.

Again, both parties appealed to the WCAC, which affirmed the award of attorney fees, observing:

While the magistrate failed to explicitly so find, in this case ... defendant knew of the medical bills in question well in advance of trial, and simply refused to pay them claiming they were not work related. Once the magistrate so found, given that prior knowledge and refusal to pay, the action in awarding attorney fees was within his discretion and hence proper.

Magna Corporation, another “employee leasing” company insured by Midwest, and Midwest sought leave to appeal both WCAC orders. The Court of Appeals initially denied the applications.1 We remanded the case for consideration as on leave granted.2

On remand, the Court of Appeals affirmed the WCAC.3 The Court found that competent evidence supported the factual findings of both the magistrate *306and the WCAC with respect to plaintiffs employment. Regarding the assessment of attorney fees, the Court held that § 315(1) is ambiguous because it does not identify the entity against which the magistrate may assess such fees.4

The Court concluded, “[W]here the remainder of [§ 315(1)] discusses the employer and/or the [insurance] carrier, it follows that the attorney fees are to be calculated or divided between those entities. The plain language of the statute does not mandate that the health care provider assume responsibility for any portion of those fees.”5 We granted leave to appeal to consider the proper interpretation of § 315(1).6

II. MCL 418.315(1)

The proper interpretation and application of a statute presents a question of law that we review de novo.7 MCL 418.315(1), part of the Worker’s Disability Compensation Act (WDCA),8 provides in pertinent part:

The employer shall furnish, or cause to be furnished, to an employee who receives a personal injury arising out of and in the course of employment, reasonable medical, surgical, and hospital services and medicines, or other attendance or treatment recognized by the laws of this state as legal, when they are needed.... After 10 days from the inception of medical care as provided in this section, the employee may treat with a physician of his or her own choice by giving to the employer the name of the physician and his or her intention to treat with the physician. The employer or the employer’s carrier may file a petition objecting to the named physician selected *307by the employee and setting forth reasons for the objection. If the employer or carrier can show cause why the employee should not continue treatment with the named physician of the employee’s choice,... the ... magistrate may order that the employee discontinue treatment with the named physician or pay for the treatment received from the physician.... If the employer fails, neglects, or refuses so to do, the employee shall be reimbursed for the reasonable expense paid by the employee, or payment may be made in behalf of the employee to persons to whom the unpaid expenses may be owing, by order of the workers’ compensation magistrate. The workers’ compensation magistrate may prorate attorney fees at the contingent fee rate paid by the employee. [Emphasis added.]

A. STATUTORY INTERPRETATION

This is a case of statutory interpretation. The primary goal of such interpretation is to give effect to the intent of the Legislature.9 The first step in ascertaining such intent is to focus on the language of the statute itself. If statutory language is unambiguous, the Legislature is presumed to have intended the meaning expressed in the statute.10 The words of a statute provide the most reliable evidence of the Legislature’s intent, and as far as possible, effect should be given to every phrase, clause, and word in a statute.* 11 If the statutory language is certain and unambiguous, judicial construction is neither required nor permitted, and courts must apply the statute as written.12

However, when statutory language is ambiguous, this Court has consistently held that a court construing it *308may go beyond the plain language of the statute.13 In fact, where the language leaves the statute’s meaning ambiguous, it is the duty of the courts to construe it, giving it an interpretation that is reasonable and sensible.14 Therefore, a finding of ambiguity has important interpretive ramifications.

In this case, the Court of Appeals held that the last sentence of § 315(1) is ambiguous in that it is unclear who is responsible for an injured employee’s attorney fees. Thus, the threshold question is whether, in light of the plain language of the entire provision, the last sentence of § 315(1) is ambiguous.

Section 315(1) gives magistrates the discretionary authority to prorate attorney fees at the contingent fee rate paid by the employee. “Prorate” means “to divide, distribute, or calculate proportionately.”15 In § 315(1), the term “prorate” could reasonably apply to employers, their insurance carriers, health care providers, employees seeking workers’ compensation benefits, or to any combination of them. Moreover, neither § 315 as a whole nor any other provision of the WDCA indicates the parties to whom a division or distribution of attorney fees applies.

I agree with the Court of Appeals conclusion that the final sentence of § 315(1) applies only to employers and their insurance carriers. This is because that interpretation harmonizes the final sentence with the remainder of § 315(1). The final sentence does not stand alone.16 It *309must be construed in the context of § 315(1) in its entirety and harmonized with the statute’s other provisions to satisfy the purpose intended by the Legislature.17

Here is how that construction operates. Section 315(1) concerns employer liability to pay medical benefits to workers injured in the course of employment. The statute as a whole provides a process for employers and their insurance carriers to object to the medical treatment that an injured employee seeks. Hence, “prorate” in the final sentence of § 315(1), when read with the remainder of the statute, applies to the parties who might contest the payment of medical benefits: employers and their insurance carriers.18 This interpretation unifies the last sentence of § 315(1) with the remainder of the statute.

I now consider the Legislature’s use of the word “may” with respect to prorating attorney fees. The word “may” is permissive in nature. As applied to “prorate,” “may” indicates that magistrates have discretion in determining whether to award attorney fees. Hence, magistrates are allowed to award attorney fees, but they are not required to do so.19

*310Likewise, the word “may” bears on whether a proration must occur if a magistrate does award attorney fees. I would hold that the use of the word “may” grants magistrates the discretion to prorate attorney fees among employers and their insurance carriers. Should a magistrate determine that only one of those parties is liable, the magistrate may impose attorney fees against only that party. Conversely, should the magistrate find multiple parties liable, fees may be prorated accordingly. This interpretation affords the phrase “may prorate” its full meaning.20

Nonetheless, as evidenced by the dissents in this case, competing reasonable interpretations of § 315(1) exist when its language alone is considered.21 In fact, the words of § 315(1), in light of the entire statute and the WDCA, do not clearly indicate the parties to whom a proration of attorney fees applies. As a consequence, I am unable to ascertain the intent of the Legislature based solely on the language of the statute. I turn next to the question whether § 315(1) is ambiguous.

1. LANSING MAYOR v PUB SERVICE COMM

This Court’s most recent pronouncement on the *311proper standard for discerning whether statutory language is ambiguous was espoused in Lansing Mayor v Pub Service Comm.22 In that case, the Court examined MCL 247.183 to determine whether it required a company to obtain local government consent before beginning pipeline construction. Critical to the Court’s analysis was its discussion of the proper method for discerning statutory ambiguity.

The Court stated that statutory provisions are not ambiguous unless one “irreconcilably conflicts” with another or unless a term is “equally susceptible to more than a single meaning.”23 Applying this definition, the Lansing Mayor majority observed that “[o]nly a few provisions are truly ambiguous.”24

Lansing Mayor’s definition of “ambiguity” is unsupported by any Michigan law whatsoever, having been derived, as it were, from thin air.25 Rusinek v Schultz, Snyder & Steele Lumber Co is the only Michigan case predating Lansing Mayor that employs the language “equally susceptible.”26 However, Rusinek did not state that language is ambiguous only if it is equally susceptible to different interpretations. Instead, Rusinek simply held that statutes in derogation of common law must be strictly construed.27

*312Also unsupported as a threshold for finding ambiguity is the “irreconcilably conflicts with another provision” language found in Lansing Mayor. The Lansing Mayor majority cited Klapp v United Ins Group Agency, Inc28 for this definition of “ambiguity,” but Klapp simply states that language is ambiguous when “its provisions are capable of conflicting interpretations.”29 It neither requires an irreconcilable conflict nor that the language be equally susceptible to more than one interpretation. Thus, the two-pronged “equally susceptible” and “irreconcilably conflicts” test adopted in Lansing Mayor has no basis in Michigan law.

Furthermore, the Lansing Mayor majority made two explicit and glaring misstatements of law. First, it cited Klapp for the proposition that a finding of “ambiguity is a finding of last resort.”30 Klapp did not say this. Instead, Klapp held that the rule of contra proferentem.31 is a rule of last resort. Indeed, Klapp concluded that the language at issue in that case was ambiguous, without commenting on whether such a conclusion was a “good” or a “bad” thing.32 Wholly absent from Klapp, or other Michigan law, is any indication that consideration of whether language is ambiguous should be given only as a last resort.

Second, I note that the Lansing Mayor majority expressly rejected the “reasonable minds” standard for discerning ambiguity as applied by the dissent in that *313case. The majority stated “[t]hat is not, and has never been, the standard either for resolving cases or for ascertaining the existence of an ambiguity in the law.”33 However, the Court cited no authority for this proposition. In fact, this holding was a blatant misstatement of the law.34

Because it was based on a mythical definition of “ambiguity” and egregious misstatements of law, I reject Lansing Mayor’s standard for discerning statutory ambiguity.35

i. STARE DECISIS36

Because I conclude that Lansing Mayor’s definition *314of “ambiguous” is unsupported by Michigan law, I must now determine whether it ought to remain the controlling method for discerning ambiguity in the laws of this state. I treat the definition as governed by stare decisis for purposes of analysis.

Stare decisis is short for stare decisis et non quieta movere, which means “stand by the thing decided and do not disturb the calm.” Stare decisis attempts to balance two competing considerations: the need of the community for stability in legal rules and decisions and the need of courts to correct past errors.37 This doctrine has been part of the American legal landscape since the country’s formation.38

Alexander Hamilton wrote that, to “avoid an arbitrary discretion in the courts, it is indispensable that [courts] should be bound down by strict rules and precedents which serve to define and point out their duty in every particular case that comes before *315them ... .”39 In the early twentieth century, Justice (then-Judge) Cardozo wrote that the “labor of judges would be increased almost to the breaking point if every past decision could be reopened in every case, and one could not lay one’s own course of bricks on the secure foundation of the courses laid by others who had gone before him.”40

Although any rule is only as solid as its boundaries are clear, it was not until recently that this Court formally established a test to determine when it should depart from stare decisis. In 2000, in Robinson v Detroit,41 the Court held that the first question in deciding whether to overrule precedent is whether an earlier decision was wrongly decided.42 Next, according to Robinson, courts should review (1) whether the decision defies practical workability, (2) whether reliance interests would work an undue hardship if the decision were overturned, and (3) whether changes in the law or facts no longer justify the decision.43 Thus, Robinson enunciated a test premised on whether the questioned decision was wrongly decided, to be followed by a three-pronged analysis of whether stare decisis nonetheless counsels upholding it.

Although the Robinson test was implemented as a mechanism for determining when a prior decision of the Court should be upheld, its application has proven superficial and cursory.44 In fact, an examination of *316cases applying the Robinson test demonstrates that not once has the Court cited it as a basis for upholding a prior decision.45 Robinson’s statement that a wrongly decided case should “invariably” be overruled was a chilling signal that a conclusion that precedent has been wrongly decided is sufficient justification for overruling it.46

These facts alone suffice to show that Robinson is insufficiently respectful of precedent. Therefore, I *317would modify it by shifting the balance back in favor of precedent and expanding on Robinson’s list of factors to consider in applying stare decisis.

I would hold that a stare decisis analysis should always begin with the presumption that upholding the precedent involved is the preferred course of action. The presumption should be retained until effectively rebutted by the conclusion that a compelling justification exists to overturn the precedent.

Robinson, by contrast, contained no such presumption. Moreover, the Court’s applications of Robinson suggest that such a presumption was never considered. Even if it had been initially applied, once a case was deemed to have been wrongly decided, any presumption in favor of upholding precedent disappeared.471 would reject this approach and reiterate that a rebuttable presumption exists in favor of upholding precedent.48

The question arises what deference should be paid to cases in which Robinson was used to overturn existing precedent. I believe that a lower level of deference should be accorded to these cases because they represent a departure from the traditional notions of stare decisis. In *318Adarand Constructors, Inc v Pena,49 the United States Supreme Court expressly addressed the distinction between consideration of well-established law and cases representing a recent departure from precedent:

It is worth pointing out the difference between the applications of stare decisis in this case and in Planned Parenthood of Southeastern Pa v Casey. Casey explained how considerations of stare decisis inform the decision whether to overrule a long-established precedent that has become integrated into the fabric of the law. Overruling precedent of that kind naturally may have consequences for “the ideal of the rule of law.” In addition, such precedent is likely to have engendered substantial reliance, as was true in Casey itself. (“[F]or two decades of economic and social developments, people have organized intimate relationships and made choices that define their views of themselves and their places in society, in reliance on the availability of abortion in the event that contraception should fail.”). But in this case, as we have explained, we do not face a precedent of that kind, because Metro Broadcasting itself departed from our prior cases — and did so quite recently. By refusing to follow Metro Broadcasting, then, we do not depart from the fabric of the law; we restore it. We also note that reliance on a case that has recently departed from precedent is likely to be minimal, particularly where, as here, the rule set forth in that case is unlikely to affect primary conduct in any event[50]

Furthermore, the Court stated:

Our past practice in similar situations supports our action today. In United States v Dixon, we overruled the recent case of Grady v Corbin, because Grady “lack[ed] constitutional roots” and was “wholly inconsistent with earlier Supreme Court precedent.” In Solorio v United States, we overruled O’Callahan v Parker, which had caused “confusion” and had rejected “an unbroken line of *319decisions from 1866 to 1960.” And in Continental TV, Inc v GTE Sylvania Inc, we overruled United States v Arnold, Schwinn & Co, which was “an abrupt and largely unexplained departure” from precedent, and of which “[t]he great weight of scholarly opinion ha[d] been critical.” See also, e.g., Payne v Tennessee (overruling Booth v Maryland, and South Carolina v Gathers, Monell v New York City Dept of Social Servs (partially overruling Monroe v Pape, because Monroe was a “departure from prior practice” that had not engendered substantial reliance); Swift & Co v Wickham (overruling Kesler v Department of Public Safety of Utah, to reaffirm “pre- Kesler precedent” and restore the law to the “view. . . which this Court has traditionally taken” in older cases).[51]

Thus, there is substantial support for applying a decreased presumption in favor of precedent when that precedent itself represents a recent departure from prior established caselaw.52

The next inquiry should be whether there exists a compelling justification for overruling precedent.53 A com*320pelling justification is not a mere belief that a precedential case was wrongly decided or that the Court, as currently composed, would have decided the case differently. Rather, in determining whether a compelling justification exists, the Court should consider several evaluative criteria, none of which, standing alone, is dispositive.

These criteria include, but are not limited to: (1) whether the rule has proven to be intolerable because it defies practical workability, (2) whether reliance on the rule is such that overruling it would cause a special hardship and inequity, (3) whether related principles of law have so far developed since the rule was pronounced that no more than a remnant of the rule has survived, (4) whether facts and circumstances have so changed, or come to be seen so differently, as to have robbed the old rule of significant application or justification, (5) whether other jurisdictions have decided similar issues in a different manner, (6) whether upholding the rule is likely to result in serious detriment prejudicial to public interests, and (7) whether the prior decision was an abrupt and largely unexplained departure from precedent.

Not all of these factors will be applicable in every case. Nor is there a magic number of factors that must favor overruling a case in order to establish the requisite compelling justification. Rather, I believe that the conclusion about whether these factors support finding a compelling justification should be reached on a case-by-case basis.

ii. APPLICATION OF STARE DECISIS TO LANSING MAYOR

As stated above, I begin my stare decisis analysis with a presumption in favor of upholding precedent. *321Only if a compelling justification exists should the Court overrule the prior decision. Although Lansing Mayor was wrongly decided with respect to the definition of “ambiguity,” this fact does not constitute the requisite compelling justification to overrule it. Instead, we must examine additional factors to determine whether there exists a compelling justification to overrule it.

First, I consider whether the method for discerning ambiguity in Lansing Mayor has proven intolerable because it defies practical workability. I believe that it does. Intrinsically, an analytical approach to interpreting statutes on the basis of their “plain meaning,” where reasonable minds disagree on what that meaning is, is unworkable. This standard gives judges unfettered discretion to pick and choose among available “plain meanings” or dictionary definitions, and thus sheds little light on what the Legislature intended statutory language to mean.54 It also potentially leads to arbitrary outcomes and injects instability into the law.

Moreover, the mere fact that different justices of this Court, judges of the Court of Appeals, and trial judges disagree on the meaning of statutory language suggests that ambiguity exists. Allowing a judge to pick one meaning among severál equally plausible meanings without using the rules of statutory construction is quite simply an exercise in speculation. As Justice Stevens of the United States Supreme Court stated:

[T]he “minimalist” judge who holds that the purpose of [a] statute may be learned only from its language retains greater discretion than the judge who will seek guidance *322from every reliable source. A method of statutory interpretation that is deliberately uninformed, and hence unconstrained, increases the risk that the judge’s own policy preferences will affect the decisional process.[55]

I share Justice Stevens’s concerns and believe that Lansing Mayor’s definition of “ambiguity” is inherently unworkable.

I also note that application of Lansing Mayor’s definition of “ambiguity” has never once led the Court to find statutory language ambiguous.56 If a rule of statutory interpretation inevitably leads to the same result in each case in which it is applied, such a rule is *323innately unworkable. Accordingly, this factor weighs strongly in favor of abrogating the Lansing Mayor definition.

Second, I examine whether reliance on the rule is such that overruling it would cause a special hardship and inequity. I believe that litigants have reasonably relied on Lansing Mayor’s definition of “ambiguity.”57 Its application potentially leads to a different statutory interpretation than one based on alternative definitions. Hence, I conclude that rejecting the definition may prejudice current litigants who have relied on it.

However, I also recognize that litigants have continued to rely on previous standards for discerning ambiguity. This reliance is also reasonable given the inconsistent application of Lansing Mayor’s definition,58 as well as the disagreement over the proper definition of *324“ambiguity” among the justices of this Court.59 Such reliance is not surprising given that Lansing Mayor’s definition is so vastly different from the definitions of “ambiguity” that preceded it for more than 150 years.60 In fact, despite its explicit holding, Lansing Mayor did not overrule or cite any previous cases that addressed the proper method for discerning ambiguity.

Nevertheless, because Lansing Mayor stands as the most recent declaration of how to discern statutory ambiguity, litigants reasonably relied on it for this point. I therefore conclude that this factor weighs moderately in favor of upholding the definition.

Third, I consider whether related principles of law have so far developed since the Lansing Mayor definition of “ambiguity” was pronounced that only a remnant of the definition has survived. This factor is inapplicable to the stare decisis analysis in this case. The definition of “ambiguity” is a tool of judicial construction. Its only relevance is to discern ambiguity or the lack of it in a given statute. Thus, the definition of “ambiguity” stands alone and is not inherently related to other principles of law. Accordingly, this factor weighs neither in favor of nor against replacing the Lansing Mayor definition.

Fourth, I examine whether facts and circumstances have so changed, or have come to be seen so differently, as to rob the old definition of significant justification. This factor focuses on real-world practicalities.61 How*325ever, the definition of “ambiguity” is not beholden to the underlying facts and circumstances of a given case. Accordingly, because the definition of “ambiguity” does not implicate practical concerns, I do not believe that this factor weighs in favor of or against replacing the Lansing Mayor definition.

Fifth, I consider whether other jurisdictions have decided similar issues in a different manner. My review indicates that Lansing Mayor’s definition of “ambiguity” is unsupported by any other jurisdiction. In fact, not a single jurisdiction, state or federal, requires an irreconcilable conflict between provisions or that language be equally susceptible to more than one meaning before finding statutory ambiguity.62 Accordingly, this factor weighs strongly in favor of rejecting the Lansing Mayor definition.

I note that the sixth factor, whether upholding the Lansing Mayor definition is likely to result in serious detriment prejudicial to public interests, has no bearing on this case. As previously stated, the definition of “ambiguity” is merely a tool of statutory interpretation. As such, its application has no relevance to public interests because the definition of “ambiguity,” standing alone, does not dictate any course of action in a particular case.

Finally, I consider whether the prior decision was an abrupt and largely unexplained departure from precedent. As noted earlier, Lansing Mayor’s definition of “ambiguity” is nothing more than a legal fiction, wholly unsupported by any law. Furthermore, it represented a decisive and abrupt shift from precedent without recognition of the prior standards for discerning statutory ambiguity. Accordingly, this factor weighs heavily in favor of abrogating Lansing Mayor’s definition of “ambiguity.”

*326Under my analysis of the aforementioned factors, Lansing Mayor’s definition of “ambiguity” has proven unworkable, is unsupported by other jurisdictions, and represents an abrupt and unexplained departure from precedent. Although it has been relied on in Michigan, reliance has been uneven and short-lived. Accordingly, I conclude that a compelling justification exists for replacing it.63

*3272. THE PROPER METHOD OF DISCERNING AMBIGUITY

Having rejected the Lansing Mayor definition of “ambiguity,” I now iterate the proper method of discerning statutory ambiguity. Before Lansing Mayor was decided, Michigan courts used several analogous variations of statutory interpretation. For example, in In re MCI Telecom Complaint,64 we held that “[s]hould a statute be ambiguous on its face ... so that reasonable minds could differ with respect to its meaning, judicial construction is appropriate to determine the meaning.”65 This Court has applied the “reasonable minds” standard on numerous other occasions.66

*328This Court has also employed a “doubtful” standard in deciding whether ambiguity exists. In Smith v Grand Rapids City Comm,67 we held that “[w]here . . . the language of a statute is of doubtful meaning, the court should give it a reasonable construction looking to the purpose to be subserved thereby, and the object sought to be accomplished and its occasion and necessity.”68

Finally, this Court has used a “susceptible” standard. Applying this method of interpretation, we held that “[i]t is only where a statute is unclear and susceptible to more than one interpretation that judicial construction is allowed.”69 These corresponding approaches to evaluating statutes for ambiguity have endured in Michigan caselaw throughout the nineteenth and twentieth centuries.70

These historical standards for discerning ambiguity are easily reconcilable and analogous. The crux of the “reasonable minds” standard is that, when two persons reasonably afford different meanings to statutory language, it is ambiguous. As for the “doubtful” standard, it suggests that a statute is ambiguous when its language is of questionable or unclear meaning. The “susceptible” standard is self-*329explanatory in that, if a statute is susceptible to more than one interpretation, it is ambiguous.

I adopt a definition of “ambiguity” that encompasses all three of the aforementioned well-established standards for determining ambiguity.71 Specifically, I would hold:

[W]hen there can be reasonable disagreement over a statute’s meaning, or, as others have put it, when a statute is capable of being understood by reasonably well-informed persons in two or more different senses, [a] statute is ambiguous. For example, this Court has concluded that statutes [are] ambiguous when one word in the statute has an unclear meaning, when a statute’s interaction with another statute has rendered its meaning unclear, or when application of the statute to facts has rendered the correct application of the statute uncertain.[72]

This standard gleans the fundamental principles from the “reasonable minds,” “doubtful,” and “susceptible” *330tests. It has been applied, in some variation, by every other state in the country,73 all the federal circuit *331courts,74 and the United States Supreme Court.75

B. APPLICATION TO § 315(1)

I believe that § 315(1) is ambiguous because the statute is capable of being understood by reasonably well-informed persons in two or more different senses. As mentioned before, the term “prorate” could reasonably be understood to apply to employers, their insurance providers, health care providers, and to employees seeking workers’ compensation benefits. Indeed, § 315 and the entire WDCA are entirely silent as to who, among these parties, is subject to a proration of attorney fees. Absent extra-textual sources, any determination of the Legislature’s intent becomes a mere exercise in speculation. Thus, having found that § 315(1) is ambiguous, I may now look to extra-textual sources to aid in its interpretation.76

*3321. HARMONIZATION WITH THE WDCA

I seek an interpretation of § 315(1) that is in accord with the principles underlying the WDCA as a whole. The act is remedial in nature.77 Thus, where ambiguity exists and judicial construction of the act is necessary, we construe the act’s terms liberally to grant rather than deny benefits to injured workers.78 This canon of statutory construction is deeply embedded in both American and Michigan jurisprudence.79

The WCAC has explicitly recognized the importance of holding employers and their insurance carriers responsible for a proration of attorney fees pursuant to § 315(1). In Harvlie v Jack Post Corp, the WCAC stated, “[T]he purpose of [the] attorney fee provisions ... is not merely to assure that a claimant’s attorney is paid, but also to deter employers from breaching their statutory duty to provide medical treatment to injured workers.”80

I find this reasoning persuasive and applicable to this case. If employers and their insurance carriers are not held accountable for a prorated share of attorney fees, they will have an incentive to deny medical benefits.81 *333They will be tempted to deny an injured employee’s request for coverage in the hope that, in subsequent litigation, they will not be found liable. And even if they were found liable, they would be responsible only for actual medical expenses but not for attorney fees, despite their wrongful denial of valid claims.

Likewise, if injured workers were forced to pay a prorated portion of their own attorney fees, their ultimate recovery could be reduced below their actual costs of securing medical treatment. Such a result would violate the remedial goal of the WDCA.

2. CASELAW HAS EMPLOYED SIMILAR REASONING

Caselaw also supports my interpretation of § 315(1). I find persuasive the fact that every previous case that has analyzed § 315(1) with respect to the proration of attorney fees has employed reasoning similar to mine.

The Court of Appeals first considered the proration of attorney fees pursuant to § 315(1) in Boyce v Grand *334Rapids Asphalt Paving Co 82 There, the plaintiff argued that his health care provider should be held responsible for a portion of his attorney fees. The Court rejected this argument, noting that a party does not become liable for attorney fees merely by accepting the benefits of an attorney’s services.83 The Court also ruled that § 315(1) could be construed to require either the employer or its insurance carrier to pay a plaintiffs attorney fees.

Finally, the Court noted that Administrative Rule 14 of the Bureau of Workers’ Compensation, which was in effect when the plaintiff was injured, precluded attorneys from recovering a percentage fee for accrued medical services.84 The Court questioned the soundness of Rule 14, noting that requiring employers or insurance carriers to pay attorney fees when they refuse to pay mandatory medical benefits would serve justice.85

*335In Watkins v Chrysler Corp,86 the Court of Appeals reaffirmed the principle expounded in Boyce that the term “prorate” in the final sentence of § 315(1) applies to employers and their insurance carriers. Watkins involved an injured plaintiff who sought and was awarded workers’ compensation benefits. He subsequently requested a hearing regarding his right to attorney fees related to medical expenses ultimately paid by Blue Cross Blue Shield of Michigan. The magistrate ruled that he was not entitled to such fees. The Workers’ Compensation Appeal Board (WCAB) reversed the magistrate’s decision and awarded attorney fees to be paid by the plaintiffs employer.

On appeal, the Court of Appeals reversed, holding that the imposition of attorney fees under § 315(1) would be unconscionable. It noted that the plaintiffs medical expenses had been timely paid and that there had been no neglect, breach of duty, or failure to provide medical care.87 Nonetheless, the Court reiterated that the WCAB had improperly ignored the policy aspect of Boyce that an employer and its insurer should bear attorney fees when medical expenses are not timely paid.88

In contrast to Watkins, defendants in this case did not timely pay plaintiffs medical expenses such that there was no neglect, breach of duty, or failure to provide *336medical care. Midwest failed to pay medical expenses despite the fact that it was paying other workers’ compensation benefits to plaintiff. As the WCAC noted, Midwest knew of plaintiff’s medical bills well in advance of trial, yet simply refused to pay them. Thus, the employer “fail[ed], neglect[ed], or refus[ed]” to furnish reasonable medical expenses for plaintiff’s injuries under § 315(1).89

Finally, in Harvlie v Jack Post Corp,90 the Court of Appeals found a unity of purpose for § 315(1) by holding that its last sentence authorizes a magistrate to prorate attorney fees among an employer and its insurance carrier. The Court held:

Here, the WCAC majority’s construction of § 315(1) is consistent with a harmonious reading of the last two sentences of § 315(1). The third sentence of § 315(1) provides that “the... magistrate may prorate attorney fees at the contingent fee rate paid by the employee.” Standing alone, this sentence contains ambiguity because it fails to identify whom the magistrate may order to pay the attorney fees. This sentence is not to be construed in isolation, however, but instead must be read in the context of the whole statute and harmonized with the statute’s other provisions in a manner that effectuates the purpose intended by the Legislature. The second sentence of § 315(1) addresses the consequences of an employer’s failure to pay medical expenses and authorizes a magistrate to order the employer to reimburse either the injured claimant or the claimant’s medical insurance provider for the reasonable medical expenses incurred. This second sentence addresses the consequences to a nonpaying employer that “fails, neglects, or refuses” to provide reasonable medical services. The WCAC properly construed the final two sentences of § 315(1) and provided a unity of purpose for this statute.[91]

*337The Court noted that its holding was consistent with prior interpretations of 315(1).92

In sum, I agree with the magistrate, the WCAC, the Court of Appeals, and established caselaw that employers and their insurance carriers are the only parties subject to a proration of attorney fees under § 315(1). This interpretation harmonizes the last sentence of § 315(1) with the preceding sentences of the provision as well as with the remedial goals of the WDCA. Workers’ compensation magistrates may thus prorate an employee’s attorney fees incurred in procuring payment for medical expenses against an employer, or against its insurance carrier, or against both.

Here, the magistrate’s proration of attorney fees against defendants was appropriate, given defendants’ failure to pay plaintiffs medical services as mandated by § 315(1). Accordingly, I conclude that the magistrate properly prorated plaintiffs attorney fees under § 315(1) against Midwest and Magna.

III. THE INAPPLICABILITY OF THE AMERICAN RULE

Finally, I note that Michigan courts follow the so-called “American rule” with respect to the payment of attorney fees.93 We have held that, “[u]nder the American rule, attorney fees are not recoverable from the losing party as costs in the absence of an exception set forth in a statute or court rule expressly authorizing such an award.”94 This rule is codified at MCL 600.2405(6), which provides that among items that may be taxed as *338costs are “[a]ny attorney fees authorized by statute or court rule.”95

Here, § 315(1) explicitly grants magistrates the discretionary authority to prorate attorney fees related to an employer’s failure to provide services in accordance with the statute. Thus, the statute contemplates that a party other than a plaintiff could, in the discretion of the magistrate, be ordered to pay attorney fees in connection with a plaintiffs suit. Accordingly, § 315(1) constitutes an express statutory authorization of attorney fees, which trumps the otherwise applicable American rule.

IV CONCLUSION

I recommend that the Court modify and expand the principles set forth in Robinson v Detroit governing when the Court should depart from the principle of stare decisis. A stare decisis analysis should always begin with a presumption that upholding precedent is the preferred course of action. Next, the Court should determine whether a compelling justification exists to overturn the precedent. A compelling justification is not a mere belief that the precedential case was wrongly decided or that the Court as currently composed would have decided the case differently. The factors listed in this opinion should be used on a case-by-case basis to determine whether a compelling justification exists to overrule an existing precedent.

However, consistent with United States Supreme Court precedent, I would accord a lower level of deference to cases that represent a recent departure from the traditional notions of stare decisis.

*339I also reject as unworkable the definition of statutory ambiguity espoused in Lansing Mayor. I conclude that MCL 418.315(1) is ambiguous because it is capable of being understood by reasonably well-informed persons in two or more different senses. As a consequence, the statute fails to clearly indicate the parties among whom attorney fees may be prorated.

Finally, we would hold that the Legislature intended that the term “prorate” in the last sentence apply only to employers and their insurance carriers. This construction is consistent with well-established principles of statutory interpretation, caselaw, the remedial nature of the WDCA, and the purpose of § 315(1). I would also hold that the American Rule of attorney fees does not apply to § 315(1).

Plaintiffs motion to dismiss defendants’ application for leave to appeal is considered, and it is denied. We affirm the judgment of the Court of Appeals.