APPEL, Justice.
In this case, we consider the duration and amount of a spousal support award in a dissolution of marriage. Based on the evidence at trial, the trial court ordered the petitioner to pay $1400 per month in spousal support, increasing to $2000 per month upon the termination of child support, for life. The trial court also divided the assets of the parties approximately equally, rejected a dissipation-of-assets claim raised by the respondent, and declined to award the respondent trial attorneys’ fees.
The petitioner appealed and the respondent cross-appealed. We transferred the case to the court of appeals, which affirmed the order of the district court. We granted further review.
On further review, we limit our review to questions arising from the award of spousal support. On the other issues raised in the petitioner’s appeal and the respondent’s cross-appeal, the order of the district court as upheld by the court of appeals is affirmed.
I. Background Facts and Proceedings.
In this case, we consider the spousal support award made by the district court in connection with the dissolution of Steven and Linda Gust’s marriage after trial in May of 2012. In its order, the district court, among other things, divided the assets and debts of the parties and required Steven to pay traditional spousal support in the amount of $1400 per month for as long as he was paying child support for a minor son, and $2000 per month thereafter.
Steven filed a posttrial motion seeking to expand the findings of the district court. Among other things, Steven asked that the spousal support begin at $1400 per month, but that it be reduced after a period of time to $1000. Steven also asked the court to place a termination date on spousal support at Steven’s retirement. The district court denied the motion.
Steven appealed and Linda cross-appealed. Steven challenged the spousal support amount as excessive in amount and duration. In her cross-appeal, Linda challenged the district court’s division of assets and sought attorneys’ fees for trial and appellate proceedings. We transferred the case to the court of appeals, which affirmed the order of the district court. We granted further review.
Based upon our review of the entire record, we make the following findings of fact. Steven and Linda Gust were married in 1985. At the time of trial, Steven and Linda had two children, aged seventeen and twenty-one. At the time of the entry of the district court’s order in this case, Steven was fifty-seven years old and Linda was fifty-two years old.
Steven received his bachelor’s degree in economics from Iowa State University in [405]*4051977. After working at several construction companies, he testified he began working at MD Construction in approximately 2005, rising to his current position of general manager. Steven testified his base salary from MD Construction was $76,000 per year. In 2011, the year prior to trial, Steven received incentive payments of about $16,000. For 2012, Steven expected to receive incentive payments of between $6000 and $8000. Through his work, Steven received partially paid health insurance, paid vacation, and paid sick leave. We, like the district court, find that Steven’s earning capacity from his position is $92,000 per year. Additionally, although Steven has type 1 diabetes, the disease does not prevent full-time employment.
Steven also testified regarding the operations of an entity called Sound Real Estate,' LLC (Sound). Linda and Steven were the sole members of Sound and under the operating agreement were entitled to equal amounts of any distributions to members. The original purpose of Sound was to flip houses. More recently, Steven restructured Sound and obtained subcontractors to engage in lead-based paint removal for MD Construction. Steven has certifications as a lead abatement contractor, a lead abatement worker, and lead abatement trainer.
Steven testified the business of Sound was a result of grants administered by the U.S. Department of Housing and Urban Development for Sioux City and Polk County. Steven’s role in Sound’s business focused on completing paperwork for lead abatement projects performed under the grants. Because of the exhaustion of grant funds, Steven’s desire not to work nights and weekends, and Steven’s concern about aggravating his diabetes, Sound ceased to be active, and Steven resigned from the entity in 2012.
During 2011 Steven withdrew $64,000 from Sound, which, combined with his compensation from MD Construction, yielded a total gross income for 2011 of approximately $156,000. The funds were largely used, however, to pay credit card debt and to provide temporary support for Linda during the pendency of the dissolution proceeding.
Steven testified he has no desire to continue Sound’s business or open a similar business at the present time. Because Sound’s business focused on completing paperwork that is no longer required in connection with lead paint abatement projects, there is no current prospect that the business could be resuscitated. Because Steven has a full-time job and because Sound has no current business viability, we do not include earnings from Sound in our calculation of Steven’s present earning capacity.
At the same time, however, we find that Steven paid $50 to the Iowa Secretary of State for filing fees on behalf of SafeCon, a business owned by his girlfriend that provides lead-based abatement services to community colleges. Steven testified he had “no idea” whether he would work for SafeCon in the future, but emphasized he was done working two jobs. We find that Steven has no plans to work for SafeCon or any other similar entity while he is employed full time by MD Construction. There is at least a possibility, however, that utilizing his lead abatement training and expertise, Steven will work with Safe-Con or a similar entity sometime in the future.
Linda testified that she was almost fifty-two at the time of the trial and lived in a rented townhouse with the parties’ minor son. She attended Des Moines Area Community College in the distant past, where she was close to obtaining a two-year degree. Between 1982 and 1986, Linda [406]*406worked as a secretary for an accounting firm. She also worked as a bookkeeper in Steven’s business, H & S Builders, Ltd., from 1992 to 1994 and had done some work for Generavivity, an assisted nursing care facility in Lake Panorama. Aside from this employment, Linda, with the agreement of Steven, took care of the house and kids until 2008 while Steven earned an income to support the family.
Beginning in 2008 as the children grew older, Linda became employed outside the home. She currently holds two part-time jobs with the Ankeny Community School District, one involving work in the media center, which pays $12 per hour, and another barcoding textbooks, which pays $9 per hour. The combination of jobs yields $15,000 in income per year. She does not, however, receive benefits from these two part-time jobs.
At trial, Steven offered expert testimony suggesting that Linda had an earning capacity of between $29,619 and $80,400 per year. Based upon our review of the record, we agree with the district court that while the expert report overstates Linda’s earning capacity somewhat, Linda’s earning capacity is $22,500 per year.
The property owned by the parties is accurately described in the appendix attached to the district court order. The district court divided the parties’ assets roughly equally, with Steven receiving a net equity of $62,249 and Linda $81,651. In the attached appendix, the district court determined Steven was to be awarded approximately $136,000 (valued in 2012) in retirement accounts, and Linda $58,000. The parties had equally divided the proceeds from the sale of the marital home with the expectation that the proceeds would pay each party’s attorneys’ fees.
With respect to maintaining the standard of living the parties were accustomed to, we find because of the inefficiencies resulting from the establishment of two households and because the parties used credit card debt to live beyond their means during the marriage, neither party will be able to maintain their predivorce lifestyle in the postdivorce world. We find Steven’s current living expenses at the time of trial were $4387 per month (assuming no reduction of principal of credit card debt). While Linda claimed $4623.99 in current living expenses, we find this amount was somewhat overstated. Making adjustments .for lower costs of health insurance, food and household expenses, cable costs, and eliminating the savings component, we find Linda’s current monthly expenses at the time of trial were $3819 per month.
II. Standard of Review.
An appeal regarding the dissolution of marriage is an equitable proceeding. Iowa Code § 598.3 (2011). Our review is therefore de novo. Iowa R.App. P. 6.907; In re Marriage of Schenkelberg, 824 N.W.2d 481, 484 (Iowa 2012). We give weight to the factual determinations made by the district court; however, their findings are not binding upon us. Iowa R.App. P. 6.904(3)(flr).
In reviewing questions related to spousal support, while our review is de novo, we have emphasized that “ ‘we accord the trial court considerable latitude.’ ” In re Marriage of Olson, 705 N.W.2d 312, 315 (Iowa 2005) (quoting In re Marriage of Spiegel, 553 N.W.2d 309, 319 (Iowa 1996)); see also In re Marriage of Schenkelberg, 824 N.W.2d at 486. We will disturb the trial court’s order “ ‘only when there has been a failure to do equity.’ ” In re Marriage of Olson, 705 N.W.2d at 315 (quoting In re Marriage of Spiegel, 553 N.W.2d at 319). We noted in In re Marriage of Benson, 545 N.W.2d 252, 257 (Iowa 1996):
[407]*407This deference to the trial court’s determination is decidedly in the public interest. When appellate courts unduly refíne these important, but often conjectural, judgment calls, they thereby foster appeals in hosts of cases, at staggering expense to the parties wholly disproportionate to any benefit they might hope to realize.
III. Discussion.
A. Overview of Iowa Law Regarding Spousal Support. We begin our discussion with an overview of Iowa law regarding spousal support. Originally, the Iowa legislature provided that in a divorce action the court could enter an order with respect to maintenance of the parties “as shall be right.” Iowa Code § 1485 (1851). In awarding spousal support under this wide-open provision, Iowa courts considered a range of facts and circumstances, which were reprised in Schantz v. Schantz, 163 N.W.2d 398, 405 (Iowa 1968). In Schantz the court laid out “a general formula” for considering the equitable determinations in divorce proceedings, including “the troublesome problem inherent in awarding alimony.” Id. at 405. Although the Schantz general formula consisted of five premarital and ten postmarital criteria distilled from prior caselaw, the court pointed out that each element is not always present or important in every case. Id.
In 1970, the legislature enacted no-fault divorce. See 1970 Iowa Acts ch. 1266 (codified at Iowa Code eh. 598 (1971)); see also In re Marriage of Williams, 199 N.W.2d 339, 344 (Iowa 1972) (noting “the overriding legislative purpose of the dissolution act is to remove fault-based standards for termination of marriages”). A decade later, the Schantz approach (except for provisions related to fault of the parties) was largely adopted by our legislature. . See 1980 Iowa Acts ch. 1175, § 3 (codified at Iowa Code § 598.21(3) (1981)). Under the current version of Iowa Code section 598.21A(1) (2011), a court
may grant an order requiring support payments ... for a limited or indefinite length of time after considering all of the following:
a. The length of the marriage.
b. The age and physical and emotional health of the parties.
c. The distribution of property made pursuant to section 598.21.
d. The educational level of each party at the time of marriage and at the time the action is commenced.
e. The earning capacity of the party seeking maintenance, including educational background, training, employment skills, work experience, length of absence from the job market, responsibilities for children under either an award of custody or physical care, and the time and expense necessary to acquire sufficient education or training to enable the party to find appropriate employment.
f. ' The feasibility of the party seeking maintenance becoming self-supporting at a standard of living reasonably comparable to that enjoyed during the marriage, and the length of time necessary to achieve this goal.
g. The tax consequences to each party-
h. Any mutual agreement made by the parties concerning financial or service contributions by one party with the expectation of future reciprocation or compensation by the other party.
i. The provisions of an antenuptial agreement.
j. Other factors the court may determine to be relevant in an individual case.
[408]*408(Emphasis added.) Our cases applying the statute have identified three kinds of support: traditional, rehabilitative, and reimbursement. See In re Marriage of Becker, 756 N.W.2d 822, 826 (Iowa 2008); In re Marriage of Francis, 442 N.W.2d 59, 63-64 (Iowa 1989). While the categories may overlap in some cases, see, e.g., In re Marriage of Becker, 756 N.W.2d at 827 (involving spousal support award the court could not characterize as strictly rehabilitative or traditional), this case involves traditional spousal support. “The purpose of a traditional or permanent alimony award is to provide the receiving spouse with support comparable to what he or she would receive if the marriage continued.” In re Marriage of Hettinga, 574 N.W.2d 920, 922 (Iowa Ct.App.1997). Traditional support is ordinarily of unlimited or indefinite duration. See In re Marriage of Francis, 442 N.W.2d at 64.
Our cases repeatedly state that whether to award spousal support lies in the discretion of the court, that we must decide each case based upon its own particular circumstances, and that precedent may be of little value in deciding each case. See, e.g., In re Marriage of Becker, 756 N.W.2d at 825-26; In re Marriage of Fennelly, 737 N.W.2d 97, 100 (Iowa 2007); In re Marriage of Fleener, 247 N.W.2d 219, 220 (Iowa 1976). Our cases tend to emphasize the need to closely examine all the statutory factors and the entire record in each case. See, e.g., In re Marriage of Schenkelberg, 824 N.W.2d at 484-87. Further, the various factors listed in Iowa Code section 598.21A(1) cannot be considered in isolation from each other. See Iowa Code § 598.21A(1) (noting “all” factors are to be considered by trial court in awarding spousal support); In re Marriage of Schenkelberg, 824 N.W.2d at 486 (emphasizing that spousal support is calculated based on “all” the factors in 598.21A(1)).
The law of spousal support under the multifactored statutory approach has been criticized for its arbitrary nature and lack of predictability. See Robert Kirkman Collins, The Theory of Marital Residuals: Applying an Income Adjustment Calculus to the Enigma of Alimony, 24 Harv. Women’s L.J. 23, 24-25 (2001). According to the critics, the terms of the statutes embracing multifactored tests for spousal support are not well defined and the standards are so vague that just about any outcome, including those based on the personal preference of an individual judge, may be justified by citation to pliable statutory factors. See id.; David A. Hardy, Nevada Alimony: An Important Policy in Need of a Coherent Policy Purpose, 9 Nev. L.J. 325, 326 (2009) (characterizing spousal support as “judge-specific, idiosyncratic, inconsistent, and unpredictable”). Some courts have joined the fray. See, e.g., Bacon v. Bacon, 819 So.2d 950, 954 (Fla.Dist.Ct.App.2002) (Farmer, J., concurring specially) (“[B]road discretion in the award of alimony is no longer justifiable and should be discarded in favor of guidelines, if not an outright rule.”); Melzer v. Witsberger, 505 Pa. 462, 480 A.2d 991, 994 (1984) (noting under Pennsylvania law “a total lack of organization with respect to how these principles interact and how they should be applied in order to arrive at an appropriate award of support”). The criticisms are not entirely off the mark, as a multifactored legal test in which all factors are relevant and none are dispositive can be extraordinarily difficult to consistently apply.
In part in response to such criticisms, reform efforts have been undertaken as reflected in the National Conference of Commissioners on Uniform State Laws, Uniform Marriage and Divorce Act (UMDA), in 1970 (abandoning concept of fault, limiting the reliance on spousal support, and emphasizing self-support even if [409]*409at a substantially lower level than existed during the marriage), the American Law Institute (ALI), Principles of the Law of Family Dissolution: Analysis and Recommendations, in 2000 (establishing presumptions or guidelines based more on compensation for loss than upon need to provide predictability and consistency for setting spousal support award), and the guidelines of the American Academy of Matrimonial Lawyers, AAML Commission Recommendations, in 2007, see Mary Kay Kisthardt, Re-thinking Alimony: The AAML’s Considerations for Calculating Alimony, Spousal Support, or Maintenance, 21 J. Am. Acad. Matrim. Law. 61, App. A (2008) [hereinafter Kisthardt] (proposing a presumptive formula based upon duration of marriage and earning capability of the parties). As is invariably the case when a family law change is proposed, these reform efforts produced fervent supporters and bitter detractors. See, e.g., John J. Sampson, Uniform Family Laws and Model Acts, 42 Fam. L.Q. 678, 685 (2008) (noting the UMDA generated considerable controversy but limited support); Michael R. Clisham & Robin Fretwell Wilson, American Law Institute’s Principles of the Law of Family Dissolution, Eight Years After Adoption: Guiding Principles or Obligatory Footnote?, 42 Fam. L.Q. 573, 577 (2008) (noting the ALI’s Principles of the Law of Family Dissolution sparked immense commentary but have only been examined in about 100 cases). Although these reform efforts were diverse in terms of their underlying theories and substantive content, they all sought to more clearly define the law of spousal support, to provide greater structure to the application of multifactored tests, and to enhance the predictability of outcomes.
In recent years, there has been a movement to statutorily modify multifactored spousal support statutes. Colorado and Massachusetts have recently amended their law to provide more detailed guidelines for the award of spousal support. Cf. Colo.Rev.Stat. Ann. §§ 14-10-114(3)(b)(II), 14-10-122 (West, Westlaw through 2d Reg. Sess. of 69th Gen. As-semb.); Mass. Gen. Laws ch. 208, §§ 48-55 (West, Westlaw through ch. 389 of 2014 2d Ann. Sess.).
• A major impetus to the legislation in Massachusetts was the question of the impact of retirement on spousal support (referred to as alimony in Massachusetts). Rachel Biscardi, Dispelling Alimony Myths: The Continuing Need for Alimony and the Alimony Reform Act of 2011, 36 W. New Eng. L.Rev. 1, 30-31 (2014); see also Mass. Gen. Laws ch. 208, § 49(f) (“[G]eneral term alimony orders shall terminate upon the payor attaining the full retirement age.”). In 2009, the Massachusetts Supreme Judicial Court declined to create a presumption in favor of the pay- or’s request to be relieved of alimony obligations upon retirement. See Pierce v. Pierce, 455 Mass. 286, 916 N.E.2d 330, 344-45 (2009).1 In response, the Massachusetts legislature amended its alimony statute to provide, among other things, that general (or traditional) alimony should not presumptively continue beyond the payor reaching full retirement age, absent [410]*410a showing of good cause. See Mass. Gen. Laws ch. 208, §§ 49(f), 53(e). The Alimony Reform Act of 2011 also provides that a court may consider alimony for an indefinite period only for marriages of twenty years or more and generally limits support payments to between thirty and thirty-five percent of the difference between the parties’ gross incomes at the time of the alimony order, unless there are circumstances warranting deviation. See 2011 Mass. Legis. Serv. eh. 124, § 3 (West) (codified at Mass. Gen. Laws ch. 208, §§ 49(c), 53(b), 53(d)); see also Biscardi, 36 W. New Eng. L.Rev. at 17-37; Charles P. Kindregan, Reforming Alimony: Massachusetts Reconsiders Postdivorce Spousal Support, 46 Suffolk U.L.Rev. 13, 24-41 (2013) (same).
A new statute in Colorado uses an approach similar to that in Massachusetts. The Colorado statute provides guidance regarding the duration of spousal support (referred to as spousal maintenance in Colorado), noting that in a marriage lasting more than twenty years, the court may award spousal maintenance for a specific term or an unlimited term, but in no event for less than ten years without making specific findings for ordering such reduction. Colo.Rev.Stat. § 14 — 10—114(3)(b)(II). The new statute addresses retirement situations by providing that retirement upon reaching full retirement age gives rise to a rebuttable presumption that the retirement is in good faith, thereby providing structure when the retiring payor spouses seek to terminate or reduce spousal maintenance in a modification action. See id. § 14 — 10122(2)(b). The statute presents a guideline for the amount of spousal maintenance. Id. § 14 — 10—114(3)(b)(I). Notably, the Colorado statute also preserves a previous procedure whereby a district court may retain jurisdiction to consider adjustments to spousal maintenance in the future based on specifically described events. Id. § 14 — 10—114(3)(g); In re Marriage of Caufman, 829 P.2d 501, 502 (Colo.App.1992).
Iowa was one of the very first states to adopt no-fault divorce after the promulgation of the UMDA. See In re Marriage of Williams, 199 N.W.2d at 344 (“California was the first to make the move in 1969 followed by Iowa and Michigan in 1970.”). However, the legislature did not adopt the spousal support approach of the UMDA, and there has been no legislative action as in Massachusetts and Colorado to alter our traditional multifactored statutory framework for spousal support determinations. As a result, we are compelled to follow the traditional multifactor statutory framework. A number of general principles emerge from our easelaw, however, that suggest the comparative weight or importance of various factors and provide at least a degree of structure for traditional spousal support determinations.
First, our easelaw demonstrates that duration of the marriage is an important factor for an award of traditional spousal support. Traditional spousal support is often used in long-term marriages where life patterns have been largely set and “the earning potential of both spouses can be predicted with some reliability.” In re Marriage of Francis, 442 N.W.2d at 62-63; see also In re Marriage of Kurtt, 561 N.W.2d 385, 388 (Iowa Ct.App.1997). Further, particularly in a traditional marriage, when the parties agree a spouse should stay home to raise children, the economic consequences of absence from the workplace can be substantial. See, e.g., In re Marriage of Becker, 756 N.W.2d at 827. While neither we nor the legislature have established a fixed formula, the shorter the marriage, the less likely a court is to award traditional spousal support. Generally speaking, marriages last[411]*411ing twenty or more years commonly cross the durational threshold and merit serious consideration for traditional spousal support. See, e.g., In re Marriage of Michael, 889 N.W.2d 630, 632, 635-39 (Iowa 2013) (twenty-three years); In re Marriage of Olson, 705 N.W.2d at 315-17 (twenty-three years); In re Marriage of Geil, 509 N.W.2d 738, 740, 742 (Iowa 1993) (nearly nineteen years); In re Marriage of Debler, 459 N.W.2d 267, 268-70 (Iowa 1990) (twenty-two years); In re Marriage of Muelhaupt, 439 N.W.2d 656, 661-62 (Iowa 1989) (twenty years); In re Marriage of Wegner, 434 N.W.2d 397, 397-99 (Iowa 1988) (twenty-six years); In re Marriage of Murray, 213 N.W.2d 657, 659-61 (Iowa 1973) (nineteen years); In re Marriage of Boyd, 200 N.W.2d 845, 854 (Iowa 1972) (twenty years).
Second, the cases emphasize that in marriages of relatively long duration, “[t]he imposition and length of an award of traditional alimony is primarily predicated on need and ability.” In re Marriage of Wendell, 581 N.W.2d 197, 201 (Iowa Ct. App.1998). For over forty years, by virtue of both judicial decision and legislative provision, the yardstick for determining need has been the ability of a spouse to become self-sufficient at “a standard of living reasonably comparable to that enjoyed during the marriage.” Iowa Code § 598.21A(1)(/'); accord Schantz, 163 N.W.2d at 405; see In re Marriage of Becker, 756 N.W.2d at 827 (noting support payments needed until payee can become “self-supporting at a standard of living reasonably comparable to that she enjoyed during the marriage”); In re Marriage of Olson, 705 N.W.2d at 316 (same); In re Marriage of Geil, 509 N.W.2d at 742 (same). The standard for determining need is thus objectively and measurably based upon the predivorce experience and private decisions of the parties, not on some externally discovered and imposed approach to need, such as subsistence or adequate living standards or amorphous notions of self-sufficiency.
In determining need, we focus on the earning capability of the spouses, not necessarily on actual income. See In re Marriage of Wegner, 434 N.W.2d at 398-99 (“We have consistently examined the earning capacity [of the parties] beyond simply ascertaining present income.”). In marriages of long duration, the historical record ordinarily provides an objective starting point for determining earning capacity of persons with work experience. See, e.g., id. In order to establish earning capability for persons without work experience or who are arguably unemployed, the parties may use vocational and other experts to assist the court in making the determination. See generally Edward M. Mazze & Candace E. Mazze, Putting a Vocational Expert to Work in a Divorce Case, 36-WTR Fam. Advoc. 26, 26-30 (2014) (describing the expertise of vocational experts and the process by which they determine the earning capacity of the parties); Brett R. Turner, Earning Capacity and Spousal Support: The Uses and Abuses of Vocational Evidence in Divorce Cases, 14 No. 12 Divorce Litig. 213 (2002) (noting that “determining income capacity is a field which particularly requires experience and personal contacts-factors which trial judges are not equipped to develop”).
With respect to ability to pay, we have noted that “[fjollowing a marriage of long duration, we have affirmed awards both of alimony and substantially equal property distribution, especially where the disparity in earning capacity has been great.” In re Marriage of Geil, 509 N.W.2d at 742; accord In re Marriage of Hitchcock, 309 N.W.2d 432, 438 (Iowa 1981). Where there is a substantial dis[412]*412parity, we do not employ a mathematical formula to determine the amount of spousal support. See In re Marriage of Conley, 284 N.W.2d 220, 223 (Iowa 1979); In re Marriage of Andersen, 243 N.W.2d 562, 564 (Iowa 1976); cf In re Marriage of Huffman, 453 N.W.2d 246, 248 (Iowa Ct. App.1990) (determining an equitable property division “cannot be reduced to a precise mathematical formula”). We have, however, approved spousal support where it amounts to approximately thirty-one percent of the difference in annual income between spouses. See In re Marriage of Michael, 839 N.W.2d at 638 & n. 7. Where a spouse does not have the ability to pay traditional spousal support, however, none will be awarded. See In re Marriage of Woodward, 426 N.W.2d 668, 670 (Iowa Ct. App.1988) (noting the payor did not have the income to meet more than the parties’ children’s minimal needs).
With respect to duration, we have observed that an award of traditional spousal support is normally payable until the death of either party, the payee’s remarriage, or until the dependent is capable of self-support at the lifestyle to which the party was accustomed during the marriage. See, e.g., In re Marriage of Becker, 756 N.W.2d at 826; In re Marriage of Francis, 442 N.W.2d at 64. In order to limit or end traditional support, the evidence must establish that the payee spouse has the capacity to close the gap between income and need or show that it is fair to require him or her alone to bear the remaining gap between income and reasonable needs. See Joan M. Krauskopf, Rehabilitative Alimony: Uses and Abuses of Limited Duration Alimony, 21 Fam. L.Q. 573, 583 (1988); Joan M. Krauskopf, Maintenance: A Decade of Development, 50 Mo. L.Rev. 259, 308 (1985). Spousal support may end, however, where the record shows that a payee spouse has or will at some point reach a position where self-support at a standard of living comparable to that enjoyed in the marriage is attainable. See, e.g., In re Marriage of Becker, 756 N.W.2d at 827.
B. Impact of Retirement on Award of Traditional Spousal Support. Although traditional spousal support is generally awarded for life, the question arises how the prospective retirement of a payor or payee spouse should be considered in the spousal support analysis. Cases across the country have produced a variety of answers that have been catalogued in an extensive annotation. See Jane Massey Draper, Annotation, Retirement of Husband as Change of Circumstances Warranting Modification of Divorce Decree— Prospective Retirement, 110 A.L.R.5th 237, 258-276 (2003) [hereinafter Draper]; see also Colleen Marie Halloran, Petitioning a Court to Modify Alimony When a Client Retires, 28 U. Balt. L.Rev. 193, 207-229 (1998). The issue of continuing spousal support after retirement is particularly pressing in light of longer lifespans and the fears that persons approaching retirement may outlive their money.
We have considered the impact of future retirement on support awards in only a handful of cases. Our most recent case dealing with retirement benefits in the context of traditional spousal support is In re Marriage of Michael, 839 N.W.2d at 632, 638. In this case, a sixty-three-year-old payor spouse sought modification of a prior spousal support order. Id. at 632. Among other things, the district court modified the order to provide that spousal support should terminate when the payor spouse reached the age of sixty-seven. Id. The court of appeals reversed. Id. at 634. We sided with the court of appeals on this issue. Id. at 634, 640. We held that the question of whether the payor spouse’s obligation should terminate at his retire[413]*413ment “will depend on the circumstances of the parties prevailing at that time.” Id. at 689 n. 8. Thus, the payor spouse in In re Marriage of Michael was required to wait until actual retirement to seek a reduction in his spousal support payments through a modification action based upon the impact of his retirement on his ability to pay. See id.; see also In re Marriage of Maro, No. 04-1043, 2005 WL 427720, at *8 (Iowa Ct.App. Feb. 24, 2005) (reversing trial court order which found that if a fifty-five-year-old payor spouse retired before the payee spouse reaches the age of sixty-five, this would constitute a significant change of circumstance meriting modification, reasoning that such a determination was premature and must await circumstances that will prevail at the time of trial of any future modification action).
We took a somewhat different path several decades ago in Locke v. Locke, 246 N.W.2d 246 (Iowa 1976). In Locke, we noted that the future receipt of social security benefits was a factor to consider under the Schantz support formulation. Id. at 254. Yet, we held the record was inadequate to calibrate the appropriate amount of spousal support because there was an absence of important testimony regarding the value of marital assets. Id. As a result, we remanded the case to the district court to develop an adequate record. Id. The implication of Locke is that although the record on appeal was inadequate, the trial court’s consideration of the impact of future retirement on support obligations when the initial support order was being crafted was appropriate in that case. See id. The inadequacy in the record, however, was not based upon lack of evidence regarding future events, but present valuations. See id.
In another dated case, Craft v. Craft, 226 N.W.2d 6, 8-9 (Iowa 1975), we considered a support order where support of $200 per month continued until the payor applied for social security benefits and filed proof that payee was receiving benefits as a divorced spouse. Upon such a showing, the amount of social security benefits paid to the payee spouse would be a credit against the original support obligation. Id. at 9; see also In re Marriage of Helmle, 514 N.W.2d 461, 464 (Iowa Ct. App.1994) (reducing spousal support payments when payee is entitled to social security benefits by amount of such benefits). The case does not suggest a reevaluation of the need for spousal support so much as merely providing a credit against preestablished spousal support for future retirement benefits.
Other Iowa appellate court cases have addressed the issue of the impact of future retirement on support obligations in a variety of ways. Sometimes the initial support obligation has been adjusted based upon the future receipt of retirement benefits by the payee spouse. For example, in In re Marriage of Ennenga, No. 04-1641, 2005 WL 2756723, at *3 (Iowa Ct.App. Oct. 26, 2005), the court of appeals held that because the disparity in the parties’ income will be significantly reduced when the payee spouse retired, support should terminate upon her receipt of social security benefits. Similarly, in In re Marriage of Schriner, No. 03-1131, 2004 WL 1898484, at *3-4 (Iowa Ct.App. Aug. 26, 2004), opinion vacated in part on other grounds by 695 N.W.2d 493, 495 (Iowa 2005), the court of appeals held a reduction of support payments from $1400 to $700 per month based upon the payee’s retirement was equitable in light of the expert testimony adduced in the case.
The court of appeals has also held support obligations are affected by the date the payor retires. For example, in In re Marriage of Markham, No. 02-1134, 2003 WL 21074445, at *1-2 (Iowa Ct.App. May [414]*41414, 2003), the court of appeals approved an order where support terminated when the payor retired or reached sixty-five. See also In re Marriage of Miller, 524 N.W.2d 442, 444-45 (Iowa Ct.App.1994) (involving support order in which payor’s obligation to make spousal support payments terminated when payee retired, either party died, or when the payor took social security benefits). On the other hand, in In re Marriage of Bell, 576 N.W.2d 618, 623 (Iowa Ct.App.1998), abrogated on other grounds by In re Marriage of Wendell, 581 N.W.2d at 200, the court of appeals held that under all the facts and circumstances a support obligation should not be affected by retirement and that if the pay- or was unable to continue support he could petition for modification of the decree.
Our caselaw also reveals other instances where spousal support orders take future retirement into account in reducing but not eliminating traditional spousal support. In In re Marriage of McCumin, 681 N.W.2d 322, 326, 331 (Iowa 2004), we modified the trial court award of spousal support and ordered a more appropriate award of $1500 per month until the payor retired or the payee reached sixty-five, whichever occurred last, and then reduced the payments to $1500 minus the payee’s social security payments. Another payments-reduction case is In re Marriage of Bell, 576 N.W.2d at 621, 623, where the district court reduced support payments by fifty percent on the date of retirement. In In re Marriage of Ask, 551 N.W.2d 643, 644 (Iowa 1996), the parties stipulated to an arrangement whereby the $1000 per month support payments would be reduced upon the payor’s retirement to one-half of the payor’s monthly IPERS payment. Further, “[t]he parties would combine their social security benefits and each would receive one-half of the combined amount.” Id.
There is also authority for the proposition that traditional spousal support may terminate upon reaching retirement. In In re Marriage of Anliker, 694 N.W.2d 535, 539 (Iowa 2005), the district court imposed an award of traditional spousal support until the payee attained the age of sixty-five or either party dies. While the limitation was not challenged on appeal, we found the traditional spousal support so limited was “equitable and [therefore, it] should not be disturbed.” Id. at 541; accord In re Marriage of Soloski, No. 05-0310, 2006 WL 623583, at *2, 3-4 (Iowa Ct.App. March 15, 2006) (holding spousal support which terminated when payor reached age sixty-five equitable as to amount and duration); In re Marriage of Aronson, No. 05-0373, 2006 WL 334250, at *3-5 (Iowa Ct.App. Feb. 15, 2006) (holding termination of spousal support when payor reached age sixty-five equitable where fifty-one-year-old wife received one-half of payor’s retirement benefits at that time).
On the other hand, some appellate court cases have held that retirement has no effect on traditional spousal support under the facts and circumstances of the case. See In re Marriage of Wiedemann, 402 N.W.2d 744, 749 (Iowa 1987) (holding trial court did not err in awarding traditional spousal support without decrease at time of retirement where there was disparity in parties’ earning capabilities, husband would receive much higher rate of social security, and husband was awarded greater share of assets); In re Marriage of Hayne, 334 N.W.2d 347, 353 (Iowa Ct.App. 1983) (holding trial court did not err in ordering former husband to continue to pay spousal support after his retirement).
C. Analysis.
1. Introduction. In analyzing the questions posed in this case, we consider two related but separate issues. The first [415]*415issue is the general question of whether traditional spousal support of unlimited duration in the amount of $2000 per month was fair in this case. We next consider how we should treat the issue of the potential impact of Steven’s future retirement on the spousal support obligation.
2. Initial spousal support obligation. We begin our analysis here by canvassing the traditional factors that have had a substantial bearing on the determination of spousal support. We note the marriage was of long duration, nearly twenty-seven years. As is often the case where traditional spousal support is awarded, Linda spent many years as a stay-at-home mom. The length of the marriage is comfortably within our caselaw where a spouse may be considered for indefinite spousal support. See, e.g., In re Marriage of Michael, 839 N.W.2d at 632, 635-39; In re Marriage of Olson, 705 N.W.2d at 315-17; In re Marriage of Geil, 509 N.W.2d at 740, 742.
Further, the record establishes Linda will not be able to be self-supporting in a lifestyle to which she was accustomed during the marriage. Although she has now returned to the workforce, her economic prospects are limited. While she may be expected to earn $22,500 per year, spousal support would be necessary for her to live in a fashion approaching her lifestyle during the marriage. See, e.g., In re Marriage of Becker, 756 N.W.2d at 827; In re Marriage of Olson, 705 N.W.2d at 316; In re Marriage of Geil, 509 N.W.2d at 742.
Steven, on the other hand, earns $92,000 per year. There is thus considerable disparity between the annual income Linda can reasonably be expected to earn, $22,500, and Steven’s expected income, $92,000. See, e.g., In re Marriage of Geil, 509 N.W.2d at 742; In re Marriage of Hitchcock, 309 N.W.2d at 438. We further note that under the trial court’s order, Linda will live off income and support of approximately $46,500 per year ($24,000 in spousal support a year, plus $22,500 in salary), while Steven will have approximately $68,000 per year ($92,000 in salary, minus $24,000 per year in spousal support payments) at his disposal.
We recognize it may be that neither party will be able to maintain their marital lifestyle, as the parties at times lived beyond their means using credit card debt, and two households are inevitably more expensive to maintain than one. However, the spousal support award allows Linda to live in a fashion that approaches the lifestyle to which she was accustomed in the marriage without undermining Steven’s ability to do the same.
In considering duration of the award, we note that traditional spousal support is ordinarily unlimited in duration except upon the remarriage of the payee spouse, or death of either party. See, e.g., In re Marriage of Cooper, 451 N.W.2d 507, 509 (Iowa Ct.App.1989); In re Marriage of Bomstein, 359 N.W.2d 500, 503 (Iowa Ct. App.1984); In re Marriage of Hayne, 334 N.W.2d at 351. There can, however, be exceptions to the general rule. For example, in In re Marriage of Becker, 756 N.W.2d at 827, we limited spousal support in a twenty-two year marriage where the record showed that after a period of rehabilitation and retraining, the income of the payee spouse “should allow her to become self-supporting at a standard of living reasonably comparable to the standard of living she enjoyed during the marriage.” In contrast, in this case, there is no reason to believe Linda will ever be able to generate enough income to support herself at the standard of living she enjoyed during the marriage. Under the record of this case, Linda’s need for support to maintain the lifestyle she has grown accustomed to and Steven’s ability to pay should remain unchanged for the indefinite future. See [416]*416Schroeder v. Schroeder, 924 S.W.2d 22, 25-27 (Mo.Ct.App.1996) (holding where evidence indicates that dependent spouse’s financial prospects will not improve materially in the future and the means of the spouse providing maintenance are not likely to decrease, original maintenance should stay in place).
We also recognize the trial court was in the best position to balance the parties’ needs, and we should intervene on appeal only where there is a failure to do equity. See In re Marriage of Olson, 705 N.W.2d at 315; In re Marriage of Benson, 545 N.W.2d at 257. Setting aside the retirement issue, we do not find the award of $2000 per month fails to do equity in this case.2
3. Treatment of future retirement. The fighting issue in this case is whether the district court erred in not fashioning a spousal support order that took into account the future retirement of Steven. In this regard, we have several options. First, we could follow the approach in In re Marriage of Michael, 839 N.W.2d at 639 n. 8, and hold that the question is not ripe for review and must await Steven’s actual retirement. Second, we could consider a flexible approach which allows the district court wide latitude to consider the issue when there is an adequate factual record, but to defer the question to a modification action where there is a paucity of information. Third, we could simply do our best to consider the likely impact of retirement on both the parties in light of their ongoing needs as well as the reduced ability to pay that ordinarily flows from retirement.
We think the best course in this case is to follow In re Marriage of Michael. Under this approach, future retirement will ordinarily be considered to raise too many speculative issues to be considered in the initial spousal support award. In this case, as in In re Marriage of Michael, we simply do not know important facts. For starters, we do not know when Steven will actually retire. He may retire at his normal social security retirement [417]*417age, but he may also continue to work at a time when Linda still has significant need. We do not know what the relative financial position of the parties will be at the time of Steven’s eventual retirement. When Steven does retire, he may maintain consulting relationships or other arrangements that enhance his retirement income. We do not know whether there will be health considerations that would impact the equities. We do not know what the impact of Linda’s retirement will have on the party’s relative financial posture. And, we do not know whether the retirement will be motivated by a desire to avoid or reduce support obligations, a factor that normally would not support a reduction in support benefits. See Smith v. Smith, 419 A.2d 1035, 1037-38 (Me.1980) (noting “retirement of the payor spouse for the primary purpose of avoiding alimony does not of itself bring about the substantial change in the payor’s circumstances needed to justify a reduction in alimony”).
Our approach in In re Marriage of Michael is not an outlier. A number of courts in other jurisdictions appear to have taken a similar approach and deferred consideration of the impact of retirement upon spousal support obligations. See, e.g., De-Shazo v. DeShazo, 582 So.2d 564, 565 (Ala. Civ.App.1991) (affirming award of alimony because ability of husband to pay alimony in future is speculative); Chaney v. Chaney, 145 Ariz. 23, 699 P.2d 398, 401-02 (Ct.App.1985) (holding that though future retirement was contemplated by parties at time decree was issued, precise date and what payor’s benefits' would amount to were speculative and therefore did not bar later modification); Frost v. Frost, 203 Ark. 1147,155 S.W.2d 895, 896 (1941) (indicating payor spouse retiring in the next few years will have opportunity at time of retirement to show that he is required to pay support in a sum beyond his means); In re Marriage of Kuppinger, 48 Cal. App.3d 628, 639, 120 CaLRptr. '654 (Ct. App.1975) (holding mere eligibility to retire at age sixty-five not a basis for adjustment in spousal support); Ryan v. Ryan, 697 A.2d 60, 61-62 (Me.1997) (holding parties should be allowed to reopen the relative economic issues at a more “propitious time,” as payor’s retirement was speculative and payee’s future earning ability was unknown); Sommer v. Sommer, 636 N.W.2d 423, 430-31 (N.D.2001) (declining to reduce support in the future based upon payor’s retirement noting exact date of retirement is unknown and there was no evidence the reduction of income that would occur); Mottice v. Mottiee, 118 Ohio App.3d 731, 693 N.E.2d 1179, 1183 (1997) (holding because payor spouse submitted he was considering retirement, but provided no details as to retirement income or when he was going to retire, therefore, his anticipated retirement did not constitute a basis for reduction or termination of support payments); In re Marriage of Wilson, 186 Or.App. 515, 63 P.3d 1244, 1249 (2003) (“[E]ven if husband’s retirement was foreseeable when the parties’ marriage was dissolved, the timing of its occurrence was speculative and, thus, it could not have affected his support obligations at that time.”); Browder v. Brow-der, 382 S.C. 512, 675 S.E.2d 820, 824 (Ct.App.2009) (“Any' change in circumstances regarding Husband’s retirement may warrant a modification of alimony when that event occurs; however, consideration of this anticipated but speculative occurrence at this time was inappropriate.”); Lambertz v. Lambertz, 375 N.W.2d 645, 646-47 (S.D.1985) (per curiam) (holding trial court was aware that husband might retire after decree issued; however, evidence was speculative and did not bar modification based on substantial reduction in income); Adamson v. Adamson, No. 20010516-CA, 2002 WL 31770882, at *1 [418]*418(Utah Ct.App. Dec. 12, 2002) (finding pay- or’s speculative retirement was not ripe for decision). Lengthy annotations outline various situations in which state courts have considered whether a spousal support order should be modified based upon retirement. See, e.g., Draper at 258-76.
Further, the legislature has expressly directed that in order to modify spousal support awards, a court must consider a number of specific factors. Iowa Code § 598.21C(1). Some of these factors cannot be properly considered at the time the initial spousal support award is determined (changes in resources, changes in medical expenses, changes in health, possible support of a party by another person). See id. The most consistent approach with the statutory scheme is that unless all of the factors in Iowa Code section 598.21C(1) can be presently assessed, future retirement is a question that can be raised only in a modification action subsequent to the initial spousal support order.
Based upon In re Marriage of Michael, the authorities in other jurisdictions, and our desire to vindicate the statutory scheme established by the legislature, we conclude the question of whether Steven’s spousal support should be modified upon his retirement must be made in a modification action when retirement is imminent or has actually occurred.
IV. Conclusion.
For all the above reasons, we affirm the decision of the court of appeals.
AFFIRMED.
All justices concur except WIGGINS, WATERMAN, and MANSFIELD, JJ., who concur in part and dissent in part.