(dissenting).
I respectfully dissent. The majority overrules Lindahl by concluding it failed to properly apply the rules of statutory construction to accurately discern the legislative intent of sections 516A.1 and 516A.2. I not only disagree with this conclusion, but find it problematic that the majority would overrule a case involving statutory interpretation that has been accepted as a part of our law for well over a decade.
Within the construct of our three branches of government, there exists a fundamental principle that it is the prerogative of the legislative branch to declare the law, and the obligation of the judicial branch to interpret the law. See Slockett v. Iowa Valley Community Sch. Dist., 359 N.W.2d 446, 448 (Iowa 1984). There is an equally fundamental principle that the goal of judicial statutory construction is to ascertain and give effect to the intention of the legislature. Wellsburg-Steamboat Rock Community Sch. Dist. v. Iowa Dep’t of Educ., 523 N.W.2d 749, 751 (Iowa 1994). A judicial body which does anything more risks infringing upon the important balance of governmental power that lies at the heart of our democratic society. Courts, indeed, have broad powers but these powers must not intrude into matters properly left to the legislative or political arena. Charles D. Breitel, The Lawmakers,. 65 Colum. L.Rev. 749, 777 (1965). In overruling Lindahl, I believe the majority has improperly assumed the legislative lawmaking role.
Chapter 516A was originally enacted'in 1967. It required motor vehicle liability insurers in this state to provide coverage “for the protection of persons insured” under the policy for injuries caused by uninsured or hit and run motorists, unless such coverage was specifically rejected by the insured, Iowa Code § 516A.1. It further permitted the insurer to place exclusions from such coverage in the policy “which are designed to avoid duplication of insurance or other benefits.” Id. § 516A.2. Our legislature amended section 516A.1 in 1980 to expand the required coverage to include underinsured motorist coverage.
■ Lindahl was decided in 1984. In Lin-dahl, we held that an owned-but-not-insured exclusion in an automobile policy violated the mandatory coverage requirement in section 516A.1 when the insured under an automobile policy was injured while occupying a motorcycle which was owned by the insured but not insured under the policy. We recognized this holding to be the majority rule-of courts in other jurisdictions that had addressed the question under similar statutes. Lindahl, 345 N.W.2d at 550.
During the fifteen years that have come and gone since Lindahl, the statutory language interpreted in Lindahl has remained unchanged. During the same time, however, other language in the statute was amended and rewritten by our legislature following our interpretation. In 1990, for example, we refused to enforce antistacking provisions in a policy of insurance as violative of the protection giv*308en to insureds under section 516A.1. Hernandez v. Farmers Ins. Co., 460 N.W.2d 842, 845 (Iowa 1990). In direct response to our interpretation of section 516A.1 in Hernandez, our legislature promptly rewrote section 516A.2 to specifically abrogate our interpretation of the statute in Hernandez. It declared that “enforcement of the antistacking provisions contained in a motor vehicle insurance policy does not frustrate the protection given to an insured under section 516A.1.” See Iowa Code § 616A.2C1).
Our history is replete with examples, like Hernandez, of the legislative response to judicial interpretation of statutes, Outside a constitutional challenge, the legislative branch is not required to live with a judicial interpretation of its statute. If the legislature disagrees with our interpretation of law it enacted, it has the prerogative to change or rewrite the law to conform to its intention. Consequently, we recognize that judicial construction given to a statute which is left undisturbed by the legislature over a period of years gives rise to the inference that the legislative branch has accepted our judicial interpretation. State v. Anderson, 517 N.W.2d 208, 214 (Iowa 1994). Furthermore, we assume that when the legislature re-enacts a statute, as was done to section 516A.2 following Hernandez, it was aware of our prior decisions interpreting the statute and would have specifically altered our interpreted intention if it did not agree with it. State v. Stueve, 260 Iowa 1023, 1027, 150 N.W.2d 597, 599-600 (1967); see also Young v. City of Des Moines, 262 N.W.2d 612, 615 (Iowa 1978), overruled on other grounds by Parks v. City of Marshalltown, 440 N.W.2d 377, 379 (Iowa 1989) (legislature’s re-enactment of statute, without pertinent change, following judicial interpretation may be presumed adopted or approved by the legislature).
Although we follow a host of rules to aid in our efforts to define the legislative intent behind a statute, the best gauge of our attempt to accomplish this goal, in the final analysis, is the reaction to our interpretation by those who write the law. A statute that has remained unchanged for well more than a decade after our interpretation of the statute reveals our legislature has concurred with our interpretation. If we had missed the mark in Lindahl, as the majority now claims, our legislature would have told us long ago. Consequently, we have no business under the framework of our powers and the principles which guide the exercise of those powers to change our interpretation now. It does an injustice to the venerable and essential concept of stare decisis, and violates those corresponding principles that give stability, strength, and confidence to our judicial process.
The majority attempts to shield itself from this criticism by rallying behind another legal principle that courts have an obligation to abandon past judicial decisions when they are found to be wrong. Yet, this principle is better applied to cases involving judge-made law, not those which involve statutory interpretation. See Kersten Co. v. Dep’t of Social Servs., 207 N.W.2d 117, 118 (Iowa 1973). Moreover, it is a principle that only applies when our past decision is found to be clearly erroneous, not when an old debate is revived without an accompanying change in the grounds or circumstances which justified the original decision. Id. at 121; see also Young, 262 N.W.2d at 615 (stare decisis is no bar where error is manifest); Stuart v. Pilgrim, 247 Iowa 709, 714, 74 N.W.2d 212, 215-16 (1956) (interpretation, once made, should be overturned only for most cogent reasons). Lindahl involved a case of statutory interpretation, and cannot fairly be characterized as clearly erroneous. Not only is it the established law of this state and the majority rule nationwide, but it has been followed by other jurisdictions since the decision was made. See Hillman v. Nationwide Mut. Fire Ins. Co., 758 P.2d 1248, 1251 (Alaska 1988) (upholding exclusion violates public policy of protective *309UM/UIM statutory scheme); Calvert v. Farmers Ins. Co., 144 Ariz. 291, 697 P.2d 684, 687 (Ariz.1985) (same); Bray v. North Carolina Farm Bureau Mut. Ins. Co., 341 N.C. 678, 462 S.E.2d 650, 653 (N.C.1995) (same); Dullenty v. Rocky Mountain Fire & Cas. Co., 111 Idaho 98, 721 P.2d 198, 208 (Idaho 1986) (Bistline, J., dissenting) (same).
The majority has simply beefed-up the dissent originally voiced in Lindahl. It adds nothing new to the debate, and can identify no change in conditions to justify a departure from our precedent. Furthermore, the criticism lodged by the majority to the framework of Lindahl is unfair.
Lindahl did not read the word “designed” out of section 516A.2, nor did it mischaraeterize the policy of section 516A.1 by describing it as a “broad mandate of coverage.” Lindahl properly observed that the legislature had a purpose in enacting section 516A.1 which would be lost if insurers could use exclusions in the manner now articulated by the majority. Thus, Lindahl properly observed that there must be some actual, not fanciful, potential for duplication of insurance or benefits, which the exclusion is designed to avoid. See Kluiter v. State Farm Mut. Auto. Ins. Co., 417 N.W.2d 74, 76-77 (Iowa 1987).
The majority also declares Lindahl was built on a faulty premise that section 516A.1 mandated uninsured coverage. The majority claims that insurers must only “offer” coverage under section 516A.1, which does not constitute a mandate. This, in my view, is only a matter of semantics and simply an artful diversion from the primary goal of statutory interpretation.
A mandate is a requirement, and section 516A.1 unmistakably requires motor vehicle insurance policies delivered or issued for delivery in this state to contain uninsured motorist protection. In the legalese of contract law, this may be viewed as ■a term of an offer, but it is a mandatory term. Furthermore, insurance policies are contracts of adhesion in which the terms of the policy are established by the insurer in conjunction with certain legislative requirements. The statute makes it clear that uninsured motorist coverage is required unless specifically rejected by the insured. There is no dispute about this, and it is wrong for the majority to transform it into a faulty premise. The statute mandates the insurer to include the coverage, and only the insured has the ability to elect not to accept it. This is what is meant by mandatory coverage, and Lin-dahl meant nothing more.
The statutory language at issue in this case is found in section 516A.2. This section allows insurers to exclude uninsured motorist coverage by including exclusions in the policy “which are designed to avoid duplication of insurance or other benefits.” The majority concludes this not only means exclusions designed to prevent the insured from obtaining duplicative uninsured motorist benefits, but also exclusions that deny coverage to the insured when there would have been duplicative coverage if the insured had purchased duplica-tive coverage. This interpretation guts the very policy behind section 516A.1 and permits the insurer to exclude the mandatory coverage if the insured is injured while operating an owned-bub-not-insured vehicle. The statute, however, does not contain such a limitation. To the contrary, section 516A.1 mandates coverage “for the protection of persons insured.” It does not limit this coverage depending on whether the insured person was occupying an insured or uninsured vehicle. The uninsured coverage mandated by section 516A.1 follows the person, not the covered vehicle. Thus, it is the majority, not Lin-dahl, which takes liberties with the rules of statutory interpretation.
The “owned-but-not-insured” exclusion in the insurance policy in this case is designed to exclude uninsured motorist coverage for injury to the insured while occupying an owned-but-not-insured vehicle. *310Such an exclusion is invalid under section 516A.1 unless “designed to avoid the duplication of insurance or other benefits” under section 516A.2.
On its face, the exclusion has nothing to do with avoiding the duplication of insurance or benefits. However, a duplication of insurance benefits could occur if the owned-but-not-insured vehicle was insured for uninsured motorist coverage under another policy of insurance. Under this situation, the design of the exclusion would include avoiding the “duplication of insurance or other benefits.” However, a duplication of uninsured motorist insurance could not arise if the owned-but-not-insured vehicle was not insured for uninsured motorist coverage.
Notwithstanding, the majority satisfies the statutory duplication of insurance requirement by concluding duplication would have occurred if the insured had purchased uninsured motorist coverage. In doing so, the statutory requirement is no longer about permitting an insurer to use an exclusion to avoid the duplication of insurance, but permitting an insurer to use the exclusion to create a duplication of insurance based on the failure of the insured to purchase duplicative insurance. This was not the design of the legislature. 'The legislature knew multiple insurance policies often come into play, and permitted the exclusion to avoid the duplication of uninsured motorist benefits that would otherwise occur when more than one policy of insurance provides uninsured motorist coverage. The majority has now not only eliminated the need for actual duplication of uninsured motorist coverage, but the very need for any duplication at all. This is in direct contravention of the legislative mandate for uninsured motorist coverage.
Stare decisis provides the needed stability in the law, which business and individuals alike rely upon to guide their decisions. See Pilgrim, 247 Iowa at 714, 74 N.W.2d at 216 (“Legal authority must be respected; not because it is venerable with age, but because it is important that courts, and lawyers and their clients, may know what the law is and order their affairs accordingly.”) I am confident the insurance industry, over the last fifteen years, has adjusted its premiums to accommodate the articulated policy of section 516A.1, as interpreted by Lindahl, and the premiums paid by insured persons reflect the possibility that an insured may be injured while occupying an owned-but-not-insured vehicle. Likewise, I must assume, as we do in other areas of the law, that the insured in this ease understood the law as interpreted by Lindahl. Miller obtained uninsured motorist coverage when he purchased the Westfield policy. This coverage, under Lindahl, followed him whether he was occupying an insured or uninsured vehicle. Thus, it would not have made economic sense for him to purchase duplicative uninsured motorist coverage, and pay an additional premium, when he subsequently obtained insurance from Midwest. Miller had a right to rely on the law in purchasing insurance, and we have done a disservice to him by changing that law. We have also done a disservice to the important principles which guide the very exercise of our powers. I am therefore compelled to depart from the majority, and voice my dissent.
LARSON and SNELL, JJ., join this dissent.