(dissenting).
I agree with Division I of the majority opinion; the district court had jurisdiction to act on the application for modification. I do not agree, however, that a court has authority to modify a provision for fixed-term alimony.
The majority finds authority to modify the award in Iowa Code section 598.21(8) (1985) which provides, in relevant part, that “[t]he court may subsequently modify orders made under this section when there is a substantial change in circumstances.” I submit this is not a grant of new authority at all; for well over 100 years our Code has provided for modification under similarly-worded statutes. Compare Iowa Code § 1485 (1851) (“Subsequent changes may be made by a court in these respects when circumstances render them expedient.”) with Iowa Code § 598.21 (1979) (identical provision). The only change in the 1985 version is to authorize the modification on a “substantial change in circumstances” rather than upon expediency of the circumstances.
The majority also relies upon a 1985 amendment to section 598.21(8) which provides that, in determining whether a substantial change in circumstances has occurred, a court may consider several factors, including increased medical expenses and changes in health conditions. Reliance on this amendment, however, obscures the issue in this case. The issue is not whether such a change in circumstances has occurred, but whether the court had *398authority to order a modification under any circumstances. If a court lacks authority to order a modification, the establishment of a substantial change in circumstances is irrelevant. See Knipfer v. Knipfer, 259 Iowa 347, 356, 144 N.W.2d 140, 145 (1966); Duvall v. Duvall, 215 Iowa 24, 27, 244 N.W. 718, 719 (1932).
I submit that the actual extent of a court’s authority to order modification has never been set by statute but by our cases. These cases have imposed restrictions on the power to modify which are not apparent in the statutes. For example, the modification statutes have always appeared to authorize modification of any decree provisions, including property settlements, lump sum alimony, and even those decrees providing for no alimony at all. See, e.g., Iowa Code § 1485 (1851) (“court may make such order in relation to the ... property of the parties and the maintenance of the' wife as shall be right and proper. Subsequent changes may be made by the court in these respects when circumstances render them expedient.” (Emphasis added.)). Even present section 598.21(8) provides for modification of all orders “made under this section.” (“This section” covers orders for support, alimony, and property division. See Iowa Code § 598.21(l)-(4).)
Despite this apparently broad grant of statutory authority, our cases have uniformly denied modification of property divisions. In re Marriage of Johnson, 299 N.W.2d 466, 467 (Iowa 1980); Ash v. Ash, 172 N.W.2d 801, 803 (Iowa 1969). Our cases have also refused modification of decrees providing for lump sum alimony, Knipfer, 259 Iowa at 355-56, 144 N.W.2d at 145; Kraft v. Kraft, 193 Iowa 602, 607, 187 N.W. 449, 451 (1922); and those providing for no alimony at all. See In re Marriage of Carlson, 338 N.W.2d 136, 139 (Iowa 1983); Spain v. Spain, 177 Iowa 249, 260, 158 N.W. 529, 533 (1916).
The reason for the court’s exercise of restraint is its concern for the stability of judgments. In Ash, 172 N.W.2d at 803, we said:
Defendant urges the need for stability of judgment entries as a reason for refusal to modify the decree [in regard to property division] at this time. This is probably the principal reason for reluctance to disturb a judgment once entered and allowed to remain of record beyond the statutory period for reconsideration. While such reasoning does not rob the court of power to act where fraud or mistake has been shown, it does militate against the exercise of equitable power except where cogent reasons of justice require modification of the decree.
Accord, Johnson, 299 N.W.2d at 467.
Obviously, there are cases in which some flexibility is desirable, even necessary, in order to meet changing needs. This would be especially true in the case of child support because of the difficulty in determining in advance the future needs of children. In those cases in which flexibility is built into the decree, it does no violence to the principle of stability to allow modification. It can hardly be said that the prospect of modification can make a decree uncertain when, by its own terms, it is already uncertain.
I submit it is when a potential modification would insert doubt into an otherwise definite provision of the decree that the concept of judicial stability is violated. This is consistent with the fact our cases deny modification of property support and lump sum alimony while allowing modification of periodic alimony during a lifetime, during the time the recipient remains single, or until further order of the court. There is necessarily some uncertainty ⅛ the latter types of provisions.
I believe this is the basis on which we must distinguish the case of In re Marriage of Carlson, which the majority apparently believes is controlling here. In Carlson, the original decree provided for alimony and child support payments in set amounts “until further order of [the] court.” When the former wife became destitute, the alimony award was “terminated” by modification, in exchange for a modest lump-sum payment by her former husband. *399Later, a second modification reinstated the alimony, and this court affirmed. Carlson, 338 N.W.2d at 142. Carlson, however, is much different on its facts. In the first place, we had serious doubts whether the alimony award had effectively been terminated because of the wife’s circumstances and her lack of understanding of the original proceeding. Id. at 137-38. More significantly, the decree in Carlson was itself indefinite, providing for alimony until further order of court. Reinstating alimony in the second modification proceeding, under these circumstances, did not have the effect of destabilizing an alimony award; it was indefinite at the outset.
This is not true in the present case. Alimony was provided for until the former wife remarried, cohabitated, or died. In these respects, it was indefinite. But it was definite as to the outside limits of time it was to be paid; two years. When this court says two years does not mean two years, I believe the court creates uncertainty where there is none and offends our concept of judicial stability.
It is well known that most dissolution settlements are negotiated by the parties. I believe it is important for parties, who might well be willing to pay a premium for it in the settlement proceedings, to be able to negotiate for certainty in future payments, and they should have some assurance that a decree incorporating the concept of certainty will mean what it says.
The immediate effect of this decision will be to destabilize all decrees, past and future, which provide for fixed-term alimony and to cause alimony payers who have already satisfied such obligations to wonder when the other shoe will drop by the imposition of future judgments against them. The long-term effect will be to cast doubt on the stability of any fixed terms, because the majority’s rationale for modification of fixed-term alimony will certainly support modification of lump-sum allowances and perhaps even property division, both of which have been held to be beyond the reach of modification.
I agree with the district court that it lacked authority to enter the modification order and would therefore affirm.
SCHULTZ, CARTER and LAVORATO, JJ., join this dissent.