The opinion
STATE OF MICHIGAN
COURT OF APPEALS
TIM EDWARD BRUGGER II, FOR PUBLICATION
May 15, 2018
Plaintiff-Appellee,
v No. 337394
Midland Circuit Court
MIDLAND COUNTY BOARD OF ROAD LC No. 15-002403-NO
COMMISSIONERS,
Defendant-Appellant.
Before: SHAPIRO, P.J., and M. J. KELLY and O’BRIEN, JJ.
O’BRIEN, J. (dissenting).
“[T]he general rule is that judicial decisions are to be given complete retroactive effect.”
Hyde v Univ of Michigan Bd of Regents, 426 Mich 223, 240; 393 NW2d 847 (1986). Because I
believe that Streng v Bd of Mackinac Co Rd Comm’rs, 315 Mich App 449, 463; 890 NW2d 680
(2016), does not warrant divergence from this general rule, I respectfully dissent.
In addressing this issue, it is necessary to understand the events that led up to the Streng
decision. The following summary, although lengthy, is crucial for understanding the effects of
Streng on our jurisprudence and the reasons why it should be given retrospective application.
Our Supreme Court in Rowland v Washtenaw Co Rd Comm, 477 Mich 197; 731 NW2d
41 (2007)—the case that, as will be explained, created the issue that Streng resolved—
summarized this history as follows:
As of 1969 . . . the enforceability of notice requirements and the particular notice
requirements in governmental immunity cases was well settled and had been
enforced for almost a century. In 1970, however, there was an abrupt departure
from these holdings in the Court’s decision in Grubaugh v City of St Johns, 384
Mich 165; 180 NW2d 778 (1970)[ abrogated by Rowland, 477 Mich 197]. In
Grubaugh the Court discerned an unconstitutional due process deprivation if
plaintiffs suing governmental defendants had different rules than plaintiffs suing
private litigants. . . .
Two years later, in Reich v State Hwy Dep’t, 386 Mich 617; 194 NW2d
700 (1972)[ abrogated by Rowland, 477 Mich 197], the Court took Grubaugh one
step further and held that an earlier version of MCL 691.1404, which included a
-1-
60–day notice provision, was unconstitutional, but this time because it violated
equal protection guarantees. The analysis again was that the constitution forbids
treating those injured by governmental negligence differently from those injured
by a private party’s negligence. Leaving aside the unusual switch from one
section of the constitution to another to justify an adjudication of
unconstitutionality, this claim is simply incorrect. Private and public tortfeasors
can be treated differently in the fashion they have been treated here by the
Legislature. It does not offend the constitution to do so because with economic or
social regulation legislation, such as this statute, there can be distinctions made
between classes of persons if there is a rational basis to do so. As we explained in
Phillips v Mirac, Inc, 470 Mich 415, 431-433; 685 NW2d 174 (2004), legislation
invariably involves line drawing and social legislation involving line drawing
does not violate equal protection guarantees when it has a “rational basis,” i.e., as
long as it is rationally related to a legitimate governmental purpose. The
existence of a rational basis here is clear, as we will discuss more fully, but even
the already cited justification, that the road be repaired promptly to prevent
further injury, will suffice.
Considering the same point, Justice Brennan in his dissent in Reich pithily
pointed out the problems with the majority’s analysis:
The legislature has declared governmental immunity from tort
liability. The legislature has provided specific exceptions to that
standard. The legislature has imposed specific conditions upon the
exceptional instances of governmental liability. The legislature
has the power to make these laws. This Court far exceeds its
proper function when it declares this enactment unfair and
unenforceable. [386 Mich at 626.]
The next year, in Carver vMcKernan, 390 Mich 96; 211 NW2d 24 (1973)[
abrogated by Rowland, 477 Mich 197], the Court retreated from Grubaugh and
Reich and, in a novel ruling, held that application of the six-month notice
provision in the Motor Vehicle Accident Claims Act (MVACA), MCL 257.1118,
was constitutional, and that the provision was thus enforceable, only where the
failure to give notice resulted in prejudice to the party receiving the notice, in that
case the Motor Vehicle Accident Claims Fund (MVACF). The reasoning was
that while some notice provisions may be constitutionally permitted some may
not be, depending on the purpose the notice serves. Thus, if notice served a
permissible purpose, such as to prevent prejudice, it passed constitutional muster.
But, if it served some other purpose (the Court could not even imagine any other)
then the notice required by the statute became an unconstitutional legislative
requirement. Thus, the Court concluded that in order to save the statute from
being held unconstitutional, it had to allow notice to be given after six months and
still be effective unless the governmental agency, there the MVACF, could show
prejudice. Whatever a court may do to save a statute from being held to be
unconstitutional, it surely cannot engraft an amendment to the statute, as was done
in Carver. See, e.g., North Ottawa Community Hosp v Kieft, 457 Mich 394, 408 n
-2-
14; 578 NW2d 267 (1998). Notwithstanding these problems, they went unnoticed
and the rule now was “only upon a showing of prejudice by failure to give such
notice, may the claim against the fund be dismissed.” Carver, 390 Mich at 100.
Returning to the Carver approach in 1976, this Court in [Hobbs v
Michigan State Hwy Dep’t, 398 Mich 90, 96; 247 NW2d 754 (1976), overruled by
Rowland, 477 Mich 197], held regarding the notice requirement in the defective
highway exception to governmental immunity:
The rationale of Carver is equally applicable to cases brought
under the governmental liability act. Because actual prejudice to
the state due to lack of notice within 120 days is the only
legitimate purpose we can posit for this notice provision, absent a
showing of such prejudice the notice provision contained in [MCL
691.1404] is not a bar to claims filed pursuant to [MCL 691.1402].
Finally, in 1996, in [Brown v Manistee Co Rd Comm, 452 Mich 354; 550
NW2d 215 (1996), overruled by Rowland, 477 Mich 197], this Court reassessed
the propriety of the Hobbs decision and declined to overrule it on the basis of
stare decisis and legislative acquiescence. [Rowland, 477 Mich at 206-209 (some
alterations in original).]
Relevant to the current appeal, this Court in Crook v Patterson, 42 Mich App 241, 242; 201
NW2d 676 (1972), held—in a half-page decision that relied exclusively on Reich—that MCL
224.21 violated the equal protection clause and was, therefore, unconstitutional and void. In
1996, the Michigan Supreme Court in Brown also held that MCL 224.21 was unconstitutional on
equal protection grounds, but noted that it was “not the same equal protection issue raised in
Reich” because “[t]his Court is no longer persuaded that notice requirements are unconstitutional
per se.” Brown, 452 Mich at 361 n 12. Instead, the Brown Court held that MCL 224.21 violated
equal protection because the 60-day notice provision had no rational basis to “[t]he only
purpose . . . for a notice requirement,” which was “to prevent prejudice to the government.” Id.
362-364.
In 2007, the Michigan Supreme Court in Rowland corrected this long line of cases that
impermissibly engrafted an “actual prejudice” requirement into statutory notice requirements to
avoid governmental immunity. In our Supreme Court’s words,
The simple fact is that Hobbs and Brown were wrong because they were built on
an argument that governmental immunity notice statutes are unconstitutional or at
least sometimes unconstitutional if the government was not prejudiced. This
reasoning has no claim to being defensible constitutional theory and is not rescued
by musings to the effect that the justices “ ‘look askance’ ” at devices such as
notice requirements, Hobbs, 398 Mich at 96, quoting Carver, 390 Mich at 99, or
the pronouncement that other reasons that could supply a rational basis were not
to be considered because in the Court’s eyes the “only legitimate purpose” of the
notice provisions was to protect from “actual prejudice.” Hobbs, 398 Mich at 96.
[Rowland, 477 Mich at 210.]
-3-
The Rowland Court went on to cite a number of purposes for notice provisions, thereby expelling
the long-held notion that the only purpose of a notice requirement in governmental immunity
cases was to prevent prejudice. The Rowland Court concluded that “[t]he notice provision
passes constitutional muster” and rejected “the hybrid constitutionality of the sort Carver,
Hobbs, and Brown engrafted into our law.” Id. at 213.
After Rowland abrogated Reich, Crook’s holding that MCL 224.21 violated equal
protection was no longer good law. But even before Rowland, it is debatable whether Crook was
good law; Brown decided that MCL 224.21 was unconstitutional but expressly rejected reliance
on Reich—upon which Crook was exclusively decided—because our Supreme Court was “no
longer persuaded” by those reasons. Brown, 452 Mich at 361 n 12. In contrast to Crook, Brown
held that MCL 224.21 violated equal protection because it was not rationally related to “the only
purpose” of a notice statute: “to prevent prejudice to the governmental agency.” Id. 362.
Rowland expressly overruled Brown and its “reading an ‘actual prejudice’ requirement into”
notice statutes. Rowland, 477 Mich at 213. Rowland also rejected the idea that the sole purpose
of a notice statute was to prevent prejudice. See id. at 211-213. In so doing, it rejected Brown’s
reasoning that MCL 224.21 was unconstitutional. See Brown, 452 Mich at 362.
It was in this context that this Court, in 2016, addressed Streng. As explained, after
Rowland was decided, the notice requirements in MCL 224.21 were no longer unconstitutional.
This created the question of whether the notice requirements in either MCL 224.21 or the GTLA
applied to injuries due to a highway defect on county roads. No published opinion addressed this
issue until Streng, which held that the notice requirements in MCL 224.21 controlled. Streng,
315 Mich App at 463.
The question now before us is whether Streng should be given retroactive effect. The
Michigan Supreme Court in Pohutski v City of Allen Park, 465 Mich 675, 696; 641 NW2d 219
(2002), provided guidance for a court faced with a decision of this type:
This Court adopted from Linkletter v Walker, 381 US 618; 85 S Ct 1731; 14 L Ed
2d 601 (1965), three factors to be weighed in determining when a decision should
not have retroactive application. Those factors are: (1) the purpose to be served
by the new rule, (2) the extent of reliance on the old rule, and (3) the effect of
retroactivity on the administration of justice. People v Hampton, 384 Mich 669,
674; 187 NW2d 404 (1971). In the civil context, a plurality of this Court noted
that Chevron Oil v Huson, 404 US 97, 106-107; 92 S Ct 349; 30 L Ed 2d 296
(1971), recognized an additional threshold question whether the decision clearly
established a new principle of law. Riley v Northland Geriatric Center (After
Remand), 431 Mich 632, 645-646, 433 NW2d 787 (1988) (GRIFFIN, J.).
Guiding this analysis are the principles that prospective-only application is an “extreme
measure,” Co of Wayne v Hathcock, 471 Mich 445, 484 n 98; 684 NW2d 765 (2004), and
decisions are generally given retrospective application, Hyde, 426 Mich at 240.
Initially, I question the majority’s conclusion that Streng established new law. Streng did
not overrule any caselaw, nor did it introduce a novel interpretation of a statute. Instead, it
resolved a dispute between two conflicting statutes. The majority is correct that this dispute laid
-4-
dormant since this Court’s decision in Crook in 1972. However, as stated, Brown, in 1996,
rejected the basis for the Crook decision. More pointedly, Rowland, in 2007, overruled Brown
and abrogated Reich—on which Crook exclusively relied—making the holdings of both Crook
and Brown no longer binding on the interpretation of MCL 224.21.1 Thus, Streng did not clearly
establish a new principle of law in 2016; the only new principles of law were established by
Rowland in 2007, and Streng simply resolved the ensuing conflict between two statutes—MCL
224.21 and the GTLA notice provision—in the post-Rowland legal landscape.
Further, as observed in Devillers v Auto Club Ins Ass’n, 473 Mich 562, 587; 702 NW2d
539 (2005), “prospective-only application of our decisions is generally limited to decisions
which overrule clear and uncontradicted case law.” (Quotation marks and citation omitted). As
explained, Rowland—not Streng—upended over 30 years of caselaw governing notice
requirements. Streng merely interpreted the pertinent statutes post-Rowland and did not, itself,
“overrule” any caselaw. Moreover, as a result of Rowland, the caselaw governing the applicable
notice requirements at the time that Streng was decided was not “clear and uncontradicted”; by
abrogating the reasoning employed by the relevant cases, Rowland, at the very least,
“contradicted” the applicable case law.2
Even assuming that this Court’s resolution of the highly unusual situation faced in Streng
created new law, I do not believe that the remaining factors weigh in favor of retroactivity. The
purpose of the Streng holding was to resolve a conflict between two conflicting statutes. The
Streng Court decided that, of those two statutes, the Legislature intended for the 60-day notice
1
To the extent that Rowland did not explicitly overrule Brown’s holding that MCL 224.21 was
unconstitutional, Rowland clearly rejected Brown’s reasoning with regard to that issue by
explaining that there were numerous reasons, besides preventing prejudice, to find a rational
basis for a notice requirement.
2
Plaintiff’s strongest argument that Streng created new law is that the Rowland Court applied the
120-day notice provision from the GTLA rather than the 60-day notice provision from MCL
224.21. See Rowland, 477 Mich at 219. Perhaps this was because, under either standard, the
plaintiff’s claim in Rowland was barred because she served notice 140 days after her injury. Id.
at 201. But regardless of the Supreme Court’s reasoning, as recognized in Streng,
The Rowland Court made no mention of MCL 224.21, nor did it discuss the
reasoning in Brown . . . regarding the notice period. . . . Rowland expressed
neither approval nor disapproval regarding that choice but simply focused on the
lack of statutory language in MCL 691.1404 allowing exceptions to the time limit.
[Streng, 315 Mich App at 459-460.]
Thus, the Rowland decision provides no help to plaintiff because MCL 224.21 “was not
discussed by the Supreme Court and implicit conclusions are not binding precedent.” Galea v
FCA US LLC, ___ Mich App ___, ___; ___ NW2d ___ (Docket No. 334576); slip op at 6; see
also People v Heflin, 434 Mich 482, 499 n 13; 456 NW2d 10 (1990) (“”[J]ust as obiter dictum
does not constitute binding precedent, we reject the dissent’s contention that ‘implicit
conclusions’ do so.”).
-5-
requirement in MCL 224.21 to control. This purpose is not served by applying the notice
requirements of the GTLA—the statute that the Streng Court held that the Legislature did not
intend to apply—to control.3
With respect to the second factor, I do not believe that it is proper to look back at the
entire history of reliance on the GTLA notice provision as the majority does. As discussed,
Rowland abrogated precedent establishing that MCL 224.21 was unconstitutional, which in turn
created the question of whether the notice provisions of MCL 224.21 or the GTLA applied in
cases such as the one before us. Rowland was decided in 2007, and I believe that the proper
inquiry is the extent of reliance on the GTLA notice provision following Rowland. Orders by the
Supreme Court following Rowland did not apply MCL 224.21, see Mauer v Topping, 480 Mich
912 (2007); Leech v Kramer, 479 Mich 858 (2007); Ells v Eaton Co Rd Comm, 480 Mich 902,
903 (2007), but none of those opinions addressed whether MCL 224.21 was applicable. Instead,
each case dismissed the plaintiff’s claim for failure to file notice within 120 days. Therefore,
none of these cases established that a case filed after 60 days but before 120 days of the injury
satisfied the applicable notice requirement; the claims would have failed under either the GTLA
or MCL 224.21. The majority has not cited a single binding case decided after Rowland that
allowed a claim noticed after 60 days of the injury but before 120 days to proceed. Therefore, in
the relevant post-Rowland timeframe, there does not appear to be extensive reliance on the
120-day GTLA notice provision.
The last factor, however, weighs in favor of plaintiff. Plaintiff attempted to comply with
what he believed was the proper statute and filed notice within 120 days of his injury. However,
plaintiff was injured six years after Rowland was released. At that time, MCL 224.21 was again
constitutional and, as later decided by Streng, applied to claims such as plaintiff’s. At the very
least, when plaintiff was injured, there was a question of whether the notice requirements in
MCL 224.21 or the GTLA applied to his claims. Ultimately, in light of the other factors—and
guided by the principles that retrospective application is the general rule and prospective-only
application is an extreme measure—I would hold that retrospective application is appropriate in
this case.
Lastly, the majority contends that “[t]he role of the government in creating confusion
concerning a legal standard weighs strongly against sanctioning a party for acting in good faith
based upon the apparent law.” In support of this assertion, the majority points to Bryant v
Oakpoint Villa Nursing Ctr, 471 Mich 411; 684 NW2d 864 (2004). Put simply, Bryant is
inapplicable to this case; it does not address whether a case should apply retroactively and, as
will be explained, neither supports nor contradicts the majority’s argument.
At issue in Bryant was whether the plaintiff’s claims sounded in medical malpractice or
ordinary negligence. Id. at 414. This determination was significant because if the plaintiff’s
3
The majority states that the purpose of Streng “was to correct an apparent error in interpreting a
provision of the GTLA.” I do not believe that Streng resolved any error in the interpretation of
the GTLA because, both before and after Streng, the notice provision of the GTLA has been
interpreted to be a 120-day notice requirement.
-6-
claims sounded in medical malpractice, then they would have been filed after the statute of
limitations had run. Id. at 418-419. Our Supreme Court, after significant analysis, concluded
that two of the plaintiff’s four claims sounded in medical malpractice, and then it addressed
“whether [the] plaintiff’s medical malpractice claims [were] time-barred.” Id. at 432. Our
Supreme Court stated that, normally, the plaintiff’s medical malpractice claims would be
time-barred, but the “equities in” the case compelled “a different result.” Id. The Bryant Court
explained as follows:
The distinction between actions sounding in medical malpractice and those
sounding in ordinary negligence is one that has troubled the bench and bar in
Michigan . . . . Plaintiff’s failure to comply with the applicable statute of
limitations is the product of an understandable confusion about the legal nature of
her claim, rather than a negligent failure to preserve her rights. [Id.]
Had the plaintiff proceeded under the correct understanding of her legal claims, her first
complaint would have been filed within the medical-malpractice statute of limitations, see id. at
418-419, and the Supreme Court ultimately allowed her claims to go forward, id. at 432.
Contrary to the majority’s reading of Bryant, the “understandable confusion” identified in
that case had nothing to do with the Legislature or the Judiciary. Rather, Bryant was simply
recognizing that it is difficult to distinguish a medical malpractice claim from an ordinary
negligence claim, and, therefore, the plaintiff’s confusion with classifying her claims was
understandable. Indeed, the general difficultly of determining whether a claim sounded in
medical malpractice or ordinary negligence was on full display in Bryant: the first judge at trial
decided that the plaintiff’s claims sounded in ordinary negligence; after the first judge recused
herself, the second judge decided that the plaintiff’s claims sounded in medical malpractice; on
appeal, two judges on a panel of this Court held that the plaintiff’s claims sounded in ordinary
negligence, while a dissenting judge believed that the plaintiff’s claims sounded in medical
malpractice; then, at our Supreme Court, five justices held that two of the plaintiff’s four claims
sounded in medical malpractice, while two justices dissented and would have held that all of the
plaintiff’s claims sounded in ordinary negligence. Bryant did not ascribe this difficulty—and the
resulting “understandable confusion”—to either the courts or the Legislature. Therefore,
Bryant’s holding simply does not support the majority’s contention that the role of the
government in creating confusion weighs in favor of prospective-only application.
Because Bryant does not support the majority’s contention that “the role of the
government in creating confusion” supports prospective application, and because the majority
does not otherwise support this assertion, I question whether “the role of the government in
creating confusion” is a valid consideration for prospective-only application. If it were, it would
“strongly” weigh in favor of prospectively applying virtually all cases that deal with the
interpretation of an ambiguous statute. When the Legislature enacts an ambiguous statute, it
creates confusion in the statute’s interpretation, which is ultimately resolved by the courts.
Under the majority’s reasoning, if a party attempted to comply with an ambiguous statute in
good faith but ultimately failed to do so, the well-intentioned-plaintiff’s actions would “strongly”
weigh in favor of prospective application of the court’s interpretation of the ambiguous statute.
As such, I do not believe that “[t]he role of government in creating confusion concerning a legal
standard” has any application to whether a decision should apply retrospectively.
-7-
Turning to the concurring opinion, I disagree that Streng rested exclusively “upon the
principle of statutory interpretation that between a general and specific statute the more specific
statute controls.” Rather, Streng also interpreted MCL 224.21 and the GTLA in pari materia.
Specifically, Streng pointed to language from MCL 224.21(2) that provides that liability was
governed by the GTLA and language from the GTLA that provides the “ ‘liability, procedure,
and remedy as to county roads under the jurisdiction of a county road commission shall be as
provided in . . . MCL 224.21.’ ” Streng, 315 Mich App at 463, quoting MCL 691.1402(1)
(alteration in original). Streng concluded that “[a] close reading of the language of MCL
224.21(2) dictates that only those GTLA provisions of law that deal with ‘liability’ apply to
counties, while under MCL 691.1402(1), procedural and remedial provisions for counties should
be those of MCL 224.21.” Id. at 462-463. Accordingly, Streng concluded that the procedural
notice requirements in MCL 224.21 controlled.
I also disagree with the concurring opinion’s conclusion that Streng “could of course
have reached the opposite conclusion by following the interpretive principle that a later-adopted
statute controls over an earlier-adopted conflicting statute.” The current version of MCL
691.1402 became effective March 13, 2012, see 2012 PA 50, which is after MCL 691.1404
became effective. MCL 691.1402(1) contains the language that Streng relied upon to conclude
that the “procedural and remedial provisions for counties should be those of MCL 224.21” rather
than of the GTLA. Streng, 315 Mich App at 463. Therefore, if the later-adopted statute
controlled, the GTLA’s notice requirements were subject to MCL 224.21 for “county roads
under the jurisdiction of a county road commission.” MCL 691.1402(1).
Further, the concurring opinion misapplies the holding of Apsey v Mem Hosp, 477 Mich
120; 730 NW2d 695 (2007). At issue in Aspey were two statutes that provided conflicting
requirements for notarizing an affidavit of merit in medical-malpractice cases. However, one of
the statutes at issue provided that it was “an additional method of proving notarial acts.” MCL
565.268. The Supreme Court explained that this
sentence of MCL 565.268 indicates that the URAA is an additional or alternative
method of proving notarial acts. As an “additional” method, the URAA does not
replace the prior method. Instead, it is intended to stand as a coequal with it.
Because the two methods are alternative and coequal, the URAA does not
diminish or invalidate “the recognition accorded to notarial acts by other laws of
this state.” MCL 565.268. Simply, MCL 600.2102(4) is not invalidated by the
URAA. It remains an additional method of attestation of out-of-state affidavits.
Because the two methods exist as alternatives, a party may use either to validate
an affidavit. [Apsey, 477 Mich at 130.]
Clearly, the Aspey Court did not conclude “that in passing two statutes designating proper
procedure, the Legislature had provided ‘alternative method[s]’ to accomplish this task,” as the
concurring opinion in this case asserts. Rather, the Aspey Court relied on language from MCL
565.268, which explicitly stated that it was “an alternative method,” to conclude that the
Legislature intended to provide an alternative method.
In contrast to Aspey, there is no language in either MCL 224.21 or the GTLA providing
that the statute is “an additional method” of providing notice for purposes of governmental
-8-
immunity. Without some indication that the Legislature intended for these statutes to be
alternative methods for providing notice, Aspey simply has no bearing on whether Streng was
wrongly decided. See Mich Ed Ass’n v Secretary of State, 489 Mich 194, 218; 801 NW2d 35
(2011) (“[N]othing may be read into a statute that is not within the manifest intent of the
Legislature as derived from the act itself.”) (quotation marks and citation omitted).4
Ultimately, however, any disagreement I have with the concurring opinion will be
resolved another day. With regard to the issue before us, because I would apply Streng
retrospectively, I respectfully dissent to the majority opinion.
/s/ Colleen A. O'Brien
4
Also notable, the concurring opinion of Justice Young in Aspey, which the concurring opinion
in this case cites to, concurred in result only. Five justices agreed with the majority, and one
wrote a dissenting opinion It is unclear why the concurring opinion in this case takes the
position that the reasoning of one justice, which was not adopted by a single other justice,
“unmistakably leads to” any conclusion grounded in the jurisprudence of this state.
-9-