OPINION
PAUL H. ANDERSON, Justice.The following reformulated questions of law were certified to us by the United States District Court for the District of Minnesota:
I. Where a plaintiff settles with a tortfeasor for 40% of the tortfea-sor’s liability limits after giving the plaintiffs insurer notice under Schmidt v. Clothier, 338 N.W.2d 256 (Minn.1983), may the plaintiffs insurer deny an underin-sured motorist claim based on the plaintiffs failure to reach the “best settlement” with the tortfea-sor?
II. If so, does the court apply an objective or subjective test in determining whether the best settlement was reached?
III. If the plaintiff fails to satisfy the best settlement test, what is the effect on any underinsured motorist claim he may have filed?
We answer the first question in the negative and therefore it is unnecessary for us to address the other two certified questions.
Emmett Dohney was involved in a car accident with Laurel Larsen on January 25, 1995. Larsen was solely responsible for the accident. As a result of the accident, Dohney incurred medical bills of over $23,000. Dohney asserts that the accident caused ongoing headaches, wage loss, pain and suffering, and loss of future earning capacity, which he claims resulted in damages greater than $50,000. At the time of the accident, Larsen was insured for $50,000 per injury and $100,000 per occurrence through American Express Property & Casualty Insurance Company (Amex). Through his attorney, Dohney negotiated a settlement with Larsen and Amex for $20,000.
Dohney was insured for underinsured motorist (UIM) benefits through Allstate Insurance Company. Dohney’s UIM policy was for $100,000. On September 28, 1998, Dohney’s attorney sent a letter to Allstate informing Allstate about the proposed settlement with Amex. In the letter, the attorney requested that Allstate substitute its draft for Amex’s per the settlement procedure outlined in Schmidt v. Clothier, 338 N.W.2d 256 (Minn.1983). In the letter, the attorney incorrectly stated that the limit on Larsen’s Amex policy was $30,000.1 Approximately two weeks later, Allstate received a letter from Amex stating that the proposed settlement was for $20,000 on a $50,000 policy. Allstate refused to substitute its draft. Dohney then settled with Amex for $20,000.
*600In December 1999, Dohney sued Allstate in Minnesota state court, seeking UIM benefits. Allstate removed the matter to federal district court on diversity grounds. In a deposition, Dohney stated that he did not recall why he settled with Larsen for $20,000 when his claimed damages exceeded $50,000, except that he thought it was all Amex would pay. Citing attorney-client privilege, Dohney refused to answer questions about his attorney’s advice on settlement. Allstate’s subpoena to depose Dohney’s attorney on the matter was quashed by the federal district court.
Allstate moved for summary judgment or in the alternative requested that the federal court certify a question to the Minnesota Supreme Court under Minn. Stat. § 480.065 (2000). The federal court granted the request to certify and issued an order and the certification. The parties stipulated to the facts in the order. As part of the facts, the court included the following statement: “Based upon the parties’ stipulation of facts, the Court considers it probable that in the absence of [Allstate’s] underinsured motorist insurance policy, [Dohney] could have settled his case with Ms. Larsen for more than $20,000.”
We accepted the certification and reformulated the federal court’s questions.2 We also granted a motion of the Minnesota Defense Lawyers Association (MDLA) to file a brief as an amicus curiae.
I.
The certified questions presented are matters of law, which we review de novo. Foley v. Honeywell, 488 N.W.2d 268, 270 (Minn.1992). The underlying issues involve statutory interpretation and interpretation of insurance contracts, which are both legal issues that we review de novo. Hibbing Educ. Ass’n v. Pub. Employment Relations Bd., 369 N.W.2d 527, 529 (Minn.1985); State Farm Ins. Cos. v. Seefeld, 481 N.W.2d 62, 64 (Minn.1992).
Before we specifically address the best settlement certified question, it is first necessary to establish some overall understanding of Minnesota’s no-fault insurance scheme and in particular the provisions relating to UIM coverage. Minnesota Statutes §§ 65B.41-65B.71 (2000) are the no-fault automobile insurance provisions. Sections 65B.43, subd. 17 & subd. 19, and 65B.49, subd. 3a & subd. 4a (2000), govern UIM coverage. There are four general types of UIM coverage systems — difference of limits, damages less limits, limits less paid, and damages less paid — and Minnesota has utilized each of these types at some point in the last 30 years. Theodore J. Smetak, Underinsured Motorist Coverage in Minnesota: Old Precedents in a New Erct, 24 Wm. Mitchell L.Rev. 857, 865-66 (1998).
While the Minnesota legislature has made significant changes to the UIM provisions, the definition of what is an under-insured motorist has remained essentially *601the same since the implementation of the No-Fault Act in 1975. A motorist is un-derinsured, and UIM benefits are therefore applicable, if “a bodily injury liability policy applies at the time of the accident but its limit for bodily injury liability is less than the amount needed to compensate the insured for actual damages.” Minn.Stat. § 65B.43, subd. 17. Therefore, UIM benefits only become available if the tortfeasor’s policy limits are less than the actual damages sustained by the injured UIM policyholder. In the present case, for purposes of the certified questions only, Larsen’s vehicle was underinsured.
Before the No-Fault Act became effective in 1975, insurers were required to make UIM coverage available to an insured, but insureds were required to affirmatively opt for UIM coverage or they would not receive the benefits. Minn.Stat. § 65B.25 (1971); Jacobson v. Ill. Farmers Ins. Co., 264 N.W.2d 804, 807 (Minn.1978). At that time, the UIM coverage was a “difference of limits” coverage, which meant that UIM benefits would become available only to the extent that the damages and the UIM policy limit were greater than the limit on the tortfeasor’s policy. Lick v. Dairyland Ins. Co., 258 N.W.2d 791, 793-94 (Minn.1977). Another way of explaining this type of coverage is- that the amount of UIM benefits payable were calculated by subtracting the tortfeasor’s policy limit from the injured UIM policyholder’s UIM limit. Id. .Thus, each insured controls his maximum recovery by choosing the amount of UIM coverage to carry. Id.
On January 1, 1975, the No-Fault Act became effective and as part of the act, insurers were required to offer UIM coverage, and UIM coverage shifted from “difference of limits” to “damages less limits” or “add-on” coverage. Act of April 11, 1974, ch. 408, § 9, 1974 Minn. Laws 773, codified at Minn.Stat. § 65B.49, subd. 6 (1976). As “damages less limits” coverage, the amount of UIM benefits available is calculated by subtracting the tortfeasor’s policy limit from the injured UIM policyholder’s damages, with the caveat that UIM payments could not exceed the UIM limit. Under this “damages less limits” coverage, below policy limit settlements with the tortfeasor were allowed, but the injured UIM policyholder would have to absorb the gap — UIM coverage would not be applied to cover the difference between the limit of the tortfeasor’s policy and the below-limit settlement amount. Schmidt, 338 N.W.2d at 261. In essence, UIM coverage would begin at the limit of the tort-feasor’s policy, but the UIM insurer could not require the insured to exhaust the tortfeasor’s policy limits in order to be eligible for UIM benefits as long as the insured could prove that the tortfeasor’s vehicle was actually'underinsured. Id. at 260-61.
In Schmidt, we weighed the interests of the insured in controlling the timing and terms of settlement against the interests of the UIM insurer in protecting subrogation rights against the tortfeasor. Id. at 260-63. We decided that the proper procedure would be for an insured party to notify his UIM insurer of any potential below-limit settlement. Id. at 263. The UIM insurer would then have the opportunity to substitute its payment to the insured for an amount equal to the settlement and thus preserve its subrogation rights against the tortfeasor. Id. If the UIM insurer did not substitute its draft payment, the insured could settle with the tortfeasor, the UIM insurer and the insured would then proceed to arbitrate the UIM claim, and the UIM insurer would waive any subrogation rights against the tortfeasor and her insurer. Id. If the UIM insurer did substitute its draft, it could then take action against the tortfeasor and her insurer to attempt to obtain a better *602settlement or proceed to trial. Id. In either scenario, UIM payments would be available, but the insured would be forced to absorb the gap between the below-limit settlement amount and the tortfeasor’s policy limit. Id.
The dissent in Schmidt agreed with the notice and substitution requirements, but disagreed that the insured should have to absorb the gap created by a below-limit settlement with the tortfeasor. Id. at 264 (Todd and Amdahl, JJ., concurring and dissenting; Scott, J., dissenting). The dissent stated that this gap rule would have “a very bad practical effect” in that it would not allow the insured to negotiate a reasonable settlement unless the insured was willing to bear the financial loss of a below-limit settlement. Id. The dissent explained that the notice and substitution requirements would correct problems resulting from any incentive the insured might have to settle for less than the tortfeasor’s policy limit created by having the UIM insurer cover the gap. Id. The position advocated by the dissent was that UIM coverage should start at the point where the settlement with the tortfeasor left off, with the notice and substitution procedures protecting the interests of the UIM insurer. Id. The dissent stated:
The injured party should negotiate the best possible settlement. The offer is communicated to the underinsured carrier. If it is not satisfied, the offer should be rejected, the underinsured carrier should immediately pay the amount of the offer to the injured party, and they should immediately proceed to arbitration. This method places in the under-insured carrier the ability to manage its role in the pending claim to its best advantage without imposing any financial burden on the injured party.
Id.
In 1980, the legislature repealed the UIM provisions; therefore, from April 1980 until October 1985, there were no UIM statutes in effect in Minnesota. Smetak, 24 Wm. Mitchell L.Rev. at 870. In 1985, the legislature made major revisions in the No-Fault Act and for the first time made minimum amounts of UIM coverage mandatory. Act of June 27, 1985, ch. 10, § 68, 1985 First Special Session Minn. Laws 1840, codified at Minn.Stat. § 65B.49, subd. 3a (1986). Additionally, the legislature changed the UIM coverage from “damages less limits” coverage to “limits less paid” coverage, in which the insured could recover the difference between his UIM insurance limit and the amount of damages recovered from the tortfeasor. Act of May 21, 1985, ch. 168, § 12, 1985 Minn. Laws 460, codified at Minn.Stat. § 65B.49, subd. 4a (1986.). Thus, UIM payments were calculated by subtracting the damages actually paid by the tortfeasor from the limit of the insured’s UIM policy. Under this system, the insured could once again control his maximum recovery by choosing the amount of UIM coverage to carry. The major difference between the “limits less paid” system and the “difference of limits” system is that the “limits less paid” system does not require the insured to absorb the gap resulting from any below-limit settlement with the tortfeasor. In Broton v. W. Nat’l Mut. Ins. Co., we characterized the 1985 legislative revisions to the No-Fault Act as codifying the dissenter’s gap position in Schmidt. Broton, 428 N.W.2d 85, 89-90 (Minn.1988).
In 1989, the legislature again modified the No-Fault Act and these changes are what govern UIM policies today. UIM coverage is once again “add on” coverage, as it was when we decided Schmidt. Act of May 19, 1989, ch. 213, § 2, 1989 Minn. Laws 648, codified at Minn.Stat. *603§ 65B.49, subd. 4a (2000). However, under the current UIM provisions, coverage is “damages less paid” coverage, meaning that UIM coverage is excess coverage, but unlike the “damages less limits” type of coverage, the “damages less paid” coverage begins where payment from the tort-feasor leaves off. As it was under the 1985 changes, the UIM insurer must cover the gap created by below-limit settlements.3 Id. Minnesota’s UIM law now is as it would have been in 1983 had the dissenters prevailed in Schmidt in that UIM insurers must cover the gap and UIM coverage is excess coverage.
With this understanding of Minnesota’s No-Fault law, we proceed to address the certified questions. Allstate and MDLA urge us to answer yes to the first question and allow UIM insurers to challenge a UIM claim on the basis that the insured faded to reach the best settlement with the tortfeasor. Allstate and MDLA argue that such a defense is the only way to prevent less expensive UIM coverage from being converted to third-party liability coverage and to balance the competing interests of the UIM insurers and their insureds.
In essence, Allstate and MDLA argue that the Schmidt notice and substitution procedure provide inadequate protection of a UIM insurer’s subrogation rights. They argue that substitution requires the insurer to bear the risk that the below-limit settlement was the best settlement, or in other words, that the actual liability of the tortfeasor was below the settlement amount. In support of its position that we require more from a settlement than just an insured’s best settlement, Allstate cites to a number of cases in which we disallowed uninsured motorist (UM) or UIM benefits after a settlement due to concerns about overcompensation to the insured. Employers Mut. Cos. v. Nordstrom, 495 N.W.2d 855, 857 n. 3 (Minn.1993) (characterizing the court of appeals’ decision in Royal-Milbank Ins. Co. v. Busse, 474 N.W.2d 441 (Minn.App.1991)); Dairyland Ins. Co. v. Starkey, 535 N.W.2d 363 (Minn.1995); Gusk v. Farm Bureau Mut. Ins. Co., 559 N.W.2d 421 (Minn.1997). However, contrary to Allstate’s assertion, none of these cases stand for the proposition that a UIM insurer may challenge an insured’s below-limit settlement with the tortfeasor as not the best settlement.4 Here, Dohney *604will not be overeompensated if Allstate has to pay his UIM benefits. Nor is it the case generally that injured UIM policyholders are overcompensated if UIM insurers cover the gap between below-limit settlements and the tortfeasor’s policy limits.5 The issue is who pays what, not whether the insured recovers more than his damages. Finally, the dissent asserts that at present we require more from a best settlement than the insured’s best settlement, but it makes this assertion without citing any authority and overlooks what we said in Schmidt.
We used the best settlement language in Schmidt when we explained why an insured party does not need to exhaust the liability limits of the tortfeasor’s policy in order to be eligible for UIM benefits. Id. We stated that “[t]he insured has the right to accept what he or she considers the best settlement available and to proceed to arbitrate the underinsurance claim * ⅜ Id. at 261 (emphasis added). Here, the dissent fails to appreciate the fact that the best settlement language as included in Schmidt was the insured’s best settlement, not a UIM insurer’s best settlement or even a court-determined best settlement. Id. However, in Schmidt we did echo some of the concerns of Allstate and the MDLA that if the UIM insurer was forced to cover the gap created by a below-limit settlement, the insured “would have no incentive to obtain the best settlement.” Id. at 261.
Allstate and MDLA expand on these concerns, and Allstate asserts that without an incentive to make a best settlement, an insured “would be able to convert inexpensive underinsured motorist vehicle coverage into additional, not-bargained-for, third-party liability coverage, * * *.” They further argue that this process results in a shift of the cost of defense from insureds and liability insurers to UIM insurers. In order to stop this shift, Allstate and the MDLA suggest that the words best settlement as we used them in Schmidt should mean more than just the insured’s best settlement and that failure to meet an objective best settlement standard should provide the UIM insurer with a defense to the entire UIM claim or at least to covering the gap.
In asking us to allow a best settlement defense, Allstate and the MDLA outline a number of potential tests we could adopt for determining whether an insured reached the best settlement. The first is a subjective best settlement test that would require a court to inquire into the reasons why an insured settled below the tortfea-sor’s liability limit. Based on our language in Nordstrom, this inquiry would focus on whether the insured settled “in good faith.” 495 N.W.2d at 858. Both Allstate and Dohney argue that a subjective test is unworkable and would create more litigation and delay.
A second alternative test would be an objective/bright line test to determine whether an insured has made the best settlement. Allstate proposes a set percentage rule of 90%, which would establish that in order for a settlement to be a best settlement, the insured must settle at or above 90% of the tortfeasor’s policy limit. Allstate stated that the 90% figure was not critical, but that the important feature of *605their proposal is the establishment of a set percentage, not what the precise percentage is. Allstate argues that this proposal provides insureds with an incentive to negotiate a better settlement, but still would allow below-limit settlements.
Dohney counters this argument by pointing out that a 90% set percentage rule is an arbitrary rule, which frustrates the purposes of the No-Fault Act by requiring insureds to litigate to final judgment just as the exhaustion requirement did. Dohney also argues that the policy reasons we endorsed in Schmidt when declaring exhaustion clauses void apply with equal force to a set percentage. Finally, Dohney argues that a set percentage rule of 90% will eliminate any incentive insurers have for offering settlements over 90% of the policy limits.
Dohney correctly points out some of the problems with a set percentage rule. Further, a 90% rule, or any rule establishing a set percentage, is troublesome because it fails to accommodate those situations in which an insured may be justified in accepting a settlement significantly below the policy limits. Broton, 428 N.W.2d at 89 (recognizing reasons why an insured may accept a below-limit settlement).
In the alternative, Allstate proposes a test based on a concurring opinion of Minnesota Court of Appeals Judge Jack Davies that would address some of the deficiencies of the 90% proposal. See United States Automobile Ass’n v. Morgan, C1-96-1333, 1997 WL 360595 at *3 (Minn.App. July 1, 1997) (Davies, J., concurring specially), rev. denied (Minn. Oct. 1, 1997). Judge Davies concluded that an insurer should be able to challenge an “improvident” below-limit settlement. Id. Judge Davies explained the gap issue in terms of “coverage gaps” and “liability-based gaps.” Id.
Gaps created because of exhausted coverage or insurer insolvency are legitimate “coverage gaps” resulting from a best settlement. But gaps created by settlements founded on uncertain liability or comparative fault are “liability-based gaps” and represent settlements that do not meet the best-settlement standard.
Id. Judge Davies concluded that UIM insurers should be required to pay coverage gaps, but should not be required to pay liability-based gaps. Id. Under this proposal, only below-limit settlements resulting from coverage gaps would be best settlements.
Allstate acknowledges that Judge Davies’ approach would result in additional litigation because there would be a need to determine the reasons for each below-limit settlement. In addition to the increased litigation that may result, this approach ignores other justifications why — aside from coverage-based reasons — an injured insured should be able to accept a below-limit settlement. For example, in Broton, we specifically referenced uncertain liability as one reason why an insured would settle for less than the limit of the tortfea-sor’s policy. 428 N.W.2d at 89. In Schmidt we recognized time and litigation costs as other reasons why an insured may choose to accept a below-limit settlement. 338 N.W.2d at 260.
The MDLA offers yet a fourth test to determine whether a best settlement was reached — a multi-factor test. MDLA suggests that “[ajbsent other factors justifying * * * a substantial discount, ⅜ * * if a claimant agrees to take less than 50% of a tortfeasor’s available liability limits, it strongly suggests that the tortfeasor is not underinsured as a matter of law.” In essence, the MDLA suggests a rebuttable presumption that the settlement was not a best settlement if it was less than 50% of *606the tortfeasor’s policy limit unless there were multiple claimants. Other factors that the court could then consider would be whether there were concerns of possible exhaustion of coverage or whether there was insurer insolvency.
The MDLA test has the appeal of incorporating many of the positive aspects of the other tests. It utilizes a bright line rule, but is not as arbitrary as the 90% rule. It also recognizes the coverage gap issues that Judge Davies raised and, as a multi-factor test, would allow for consideration of any other factors the court would consider important in a specific case such as time delay and uncertainty. However, the MDLA test would require a court to make the best settlement determination in every case. Additionally, even though MDLA has labeled its test as an objective test, under the test each court would need to not only look at objective factors, but determine subjectively how to weigh competing findings. Finally, while this test is comprehensive, it is quite uncertain and could spawn significant litigation.
Each of the best settlement tests proposed provides some benefit over our current procedures. However, each test presents its own consequences that will operate to undermine the purposes of the No-Fault Act, such as easing the burden of litigation and encouraging prompt payment of claims.6 Schmidt, 338 N.W.2d at 260. While Allstate is correct that we have stepped in and altered UIM procedures in the past, we are hesitant to judicially alter the UIM procedures established by the legislature and in the process possibly adversely affect the delicate balance between the rights of UIM insureds and UIM insurers.
We have been balancing the interests of UIM insurers and insureds in light of the purposes of Minnesota’s No-Fault law ever since the Act’s enactment in 1975. In Schmidt, when holding that exhaustion clauses violated the policies of the No-Fault Act and were therefore void, we adopted the procedure for notice and substitution to preserve the UIM insurer’s subrogation rights and the insured’s right to settle. 338 N.W.2d at 261. In Nord-strom, we held that an insured must first recover from a tortfeasor before proceeding to arbitrate a UIM claim, because this was “the fairest solution” to both UIM insurers and insureds. 495 N.W.2d at 858. In Oanes v. Allstate Ins. Co., we reaffirmed our support of the Schmidt rule as an accommodation of the interests of both the UIM insurer and the insured and stated that the notice and substitution process “protects both the UIM claimant’s ability to settle with the tortfeasor and also the UIM insurer’s subrogation interest.” 617 N.W.2d 401, 407 n. 3 (Minn.2000).
Following these modifications, the legislature weighed the interests of insureds and insurers against the purposes of the No-Fault Act and made significant revisions to the Act. Despite the concerns we expressed in Schmidt, the legislature amended the UIM provisions to require UIM insurers to cover the gap created by below-limit settlements. In 1989, after we noted that insurers were now required to cover the gap, the legislature again amended the act, but chose to retain the gap rule. Minn.Stat. § 65B.49, subd. 4a (1990). Thus, while Allstate and the MDLA have voiced valid concerns about a possible imbalance in favor of the insureds in the current UIM settlement procedures, we decline their invitation to create a new *607UIM settlement rule. Instead, we maintain the status quo — a best settlement with a tortfeasor for purposes of a UIM claim is an insured’s best settlement. We conclude that an insurer may not deny a UIM claim based on the insured’s failure to reach the best settlement with the tortfeasor.7 Therefore, we answer the first certified question in the negative.
Because we answered the first certified question in the negative, it is unnecessary to address the other certified questions.
First Certified Question answered in the negative.
. Based on this fact, there is some question as to whether Dohney properly notified Allstate. However, the sufficiency of the Schmidt notice is not before us.
. Under Minn.Stat. § 480.065, subd. 4 (2000), we may reformulate a question of law certified to us. The federal district court certified the questions to us as follows:
When a plaintiff settles with a tortfeasor for 40% of the tortfeasor's liability limits, does the plaintiff satisfy the requirement of Schmidt v. Clothier, 338 N.W.2d 256 (Minn.1983), and Employers Mut. Cos. v. Nordstrom, 495 N.W.2d 855 (Minn.1993), that he reach the best settlement with the tortfeasor before submitting his claim for underin-sured motorist benefits?
A. Does the Court apply an objective or subjective test in analyzing the best settlement test?
B. If the plaintiff fails to satisfy the best settlement test, what is the effect on any underinsured motorist case he may have filed?
. Minnesota Statutes § 65B.49, subd. 4a, provides:
Liability on underinsured motor vehicles. With respect to underinsured motorist coverage, the maximum liability of an insurer is the amount of damages sustained but not recovered from the insurance policy of the driver or owner of any underinsured at fault vehicle. * * * However, in no event shall the underinsured motorist carrier have to pay more than the amount of its underinsured motorist limits.
(Emphasis removed.)
. Our characterization of the settlement in Busse as not the best settlement was based on the fact that Busse’s damages were well below the liability limit of the tortfeasor's policy. Nordstrom, 495 N.W.2d at 857 n. 3; Busse, 474 N.W.2d at 442-43. Thus, he did not meet the threshold requirement of being underin-sured and was not eligible for UIM benefits. Busse, 474 N.W.2d at 443.
In Starkey, the claimant’s case establishing damages and fault went to trial and fault was apportioned between the two other drivers, one of whom was insured and one of whom was uninsured. The total damages as found by the jury did not exceed the insured driver’s policy limits. 535 N.W.2d at 363-64. Before judgment was entered, the claimant settled with the insured driver for slightly under the jury verdict and attempted to collect UM benefits for the full amount of the uninsured driver’s liability as apportioned by the jury. Id. We did not allow the claimant to collect more than the jury’s verdict and in denying any UM claim, distinguished this case from settlement cases because in Starkey the damages and fault had already been established by a jury. Id. at 364-65.
*604Finally, while Allstate is correct that we did reaffirm “our clear policy against overcompensation” in Gusk, we did so while also reaffirming the very procedures that Allstate is arguing against here. Gusk, 559 N.W.2d at 425.
. In fact, because UIM coverage is excess coverage, below-limit settlements may sometimes result in insureds receiving less compensation because the payments made by the UIM insurer to cover the gap count toward the UIM limit.
. The dissent does not specify explicitly how a court should determine a best settlement and discounts the problems such an approach may spawn — despite the fact that both Allstate and the MDLA acknowledge that increased litigation may result.
. Our conclusion does not, as the dissent asserts, place "in the sole and exclusive control of the UIM insured [the ability] to declare the tortfeasor underinsured, * * * ” because the insured still must prove damages greater than the tortfeasor’s policy limit in order to be eligible for UIM benefits. Minn.Stat. § 65B.43, subd. 17 (2000).