[199]*199TANZER, J.
This dissolution case presents an opportunity to examine a recurrent situation in light of our observation in Grove and Grove, 280 Or 341, 344, 571 P2d 477 (1977), that support and property division provisions of a decree must often be formulated in a coordinated manner to achieve the equitable objectives of a dissolution proceeding. Here, we examine a long-term marriage, the principal asset of which is a corporate family business. Like the baby before King Solomon, this corporation cannot prudently be divided. Unlike the case facing our biblical predecessor, both of the parties urge us not to cleave the corporation but, at the same time, to divide its benefits, and the courts do not have his discretion to refuse to act.
REVIEW OF EQUITABLE DISCRETION
A threshold issue is the proper scope of appellate review of equitable remedies ordered in domestic relations cases. This being a case in equity, our permissible scope of review on the record is entire. ORS 19.125(3) and (4) provide:
“(3) Upon an appeal from a decree in a suit in equity, the Court of Appeals shall try the cause anew upon the record.
“(4) When the Court of Appeals has tried a cause anew upon the record, the Supreme Court may limit its review of the decision of the Court of Appeals to questions of law.”
Finding facts anew on the record is a relatively simple, straightforward process. See, e.g., Waterman v. Armstrong, 291 Or 551, 633 P2d 774 (1981). In this case, however, there is no reason asserted for a retrial of the facts on review. Rather, the dominant plea of the appellant in the Court of Appeals and the petitioner on review is to modify the equitable relief decreed below. Appellate review of discretionary orders in equity is a more complex process than factfinding and our initial discussion is an attempt to articulate some of the fundaments of that process which ordinarily go unspoken.
ORS 19.125(3) may promise more in the way of trial anew than the appellate courts actually deliver. Neither the Court of Appeals nor this court approaches its [200]*200appellate or review function in a vacuum as if circuit courts were mere masters and there were no trial court decree. Appeal is not regarded as simply a second trial. In recent years, most appeals of domestic relations decrees have been finally decided by the Court of Appeals. The approach of the Court of Appeals has been to promote finality of trial court decrees and to refrain from modifying decrees unless there is some good reason to do so. Most commonly, it has done so simply by concluding that an order is or is not an abuse of discretion, without further elaboration. Recently, however, that court has resorted to statements which give a somewhat more understandable explanation of its approach. In McCoy and McCoy, 28 Or App 919, 926, 562 P2d 207 (1977), the court criticized its former reliance on the term “abuse of discretion” and expressed a new standard of review for “mistake”:
“* * * It is not the policy of the law to make appellate review a matter of course where the losing party below automatically gets another chance to make his case. While de novo review may be characterized as a trial anew, the burden is on the appellant to show that the lower court made a mistake. Where there are viable alternatives available, often no two courts can agree which is the preferable. There is often no perfect remedy, but merely a choice between several ‘dull axes.’ See Wirthlin and Wirthlin, 19 Or App 256, 527 P2d 147 (1976). The role of the appellate court is not to substitute its preferences for that of the lower court. In this case the trial court has attempted to divide an array of marital assets. We may prefer reshuffling but should be reluctant to do so unless our preference is motivated with sufficient conviction to proclaim that the trial court made a mistake.”
McCoy is a valiant attempt to articulate a standard, but it has its shortcomings. The foremost difficulty of application of this standard is that it does not tell us against what rule of law, judicial policy or judge’s intuition a remedy is to be evaluated to determine if there has been a mistake.
More recently, the Court of Appeals has described its practice as reflecting judicial preference. In Pullen and Pullen, 38 Or App 137, 142, 589 P2d 1145, rev den 286 Or 449 (1979), the court said:
“Although this court’s review is de novo and we must exercise independent judgment based on our own review of [201]*201the record, our role is not to substitute our preferences for those of the lower court. We will not modify a property division unless we are convinced that we can make a significantly preferable disposition than that made by the trial court. * * *” See also Frishkoff and Frishkoff, 45 Or App 1033, 1042, 610 P2d 831 (1980).
This too has its shortcomings because, again, there is no statement of how the court determines what is “significantly preferable.” That is a question of legal substance.
Nevertheless, these Court of Appeals holdings represent a move away from the “abuse of discretion” standard, under which a modification was tantamount to a condemnation of the trial judge for wrongdoing, to appellate review based on reasoned preference of sufficient degree to justify disturbing the circuit court decree. Their approach at the jurisprudential level is at once an invitation to principled arguments and a statement of self-restraint in favor of stability. ORS 19.125(3) requires the Court of Appeals to try every case anew upon proper invocation of that statutory right, but it leaves the Court of Appeals with discretion as to whether it will modify decrees. McCoy and Pullen represent policies which are permissible under ORS 19.125(3).
The earlier pronouncements in this court have been similarly addressed primarily to a policy of appellate restraint rather than substantive law. For example, in Sandner v. Sandner, 243 Or 349, 350, 413 P2d 424 (1966), this court repeated its intention to disturb only those decree provisions which are “clearly erroneous,” but it did not express a legal rule, policy or principle by which to determine the existence of clear error. More typically, in Hofer v. Hofer, 247 Or 82, 87, 427 P2d 411 (1967), we reviewed for an “abuse of discretion,” and found none, citing Smith v. Smith, 212 Or 654, 656, 320 P2d 1111 (1958), wherein we reviewed for “clear abuse” of discretion. Again, these are statements of restraint, not statements of substantive law. They leave open the question of what equitable policies are to be applied by trial courts in the exercise of their discretion.
The policy favoring finality of decrees suggests that we modify discretionary acts rarely, and that has been [202]*202our practice since 1969 when we became a court of review. We recognize that any review of trial court discretion must be founded on a healthy respect for the wisdom of our modern chancellors in the trial courts who see the people, decide the cases and develop a feel for the best solution in what are often difficult circumstances. On the other hand, judicial discretion is not a license for judges simply to exercise their individual will, for we are judges, not kings. Discretion is a device for reason, not will. Thus, equitable discretion should be exercised in a principled manner, based on experience and learning. Professor Re has described the process by which reasoned decisions of past cases become, by replication in similar cases, equitable principles:
“Judicial discretion implies a prudence and correctness of judgment that is just and fair under all of the circumstances. It does not mean a decision or conclusion based upon unknown or questionable assumptions and idiosynchratic [sic] factors. Rather, judicial discretion implies the exercise of a judicial value judgment based upon pertinent and relevant factual data. This value judgment has also taken into consideration the existing precedents in like cases. Hence, the importance of recognizing and presenting the factors or equities that affect that judgment-i.e., the exercise of judicial discretion. An appeal to ‘equity,’ therefore, requires a knowledge of these factors or equities, and not a tugging at the heartstrings and an appeal to emotions. * * *” (Original emphasis.) Edward D. Re, Cases and Material on Equity and Equitable Remedies 222-23, 11-14 (5th ed 1975).
Equitable principles, developed by repeated judicial application, are seldom articulated. It seems that they live mostly as unstated empirical intuition, embodied largely in the feel of lawyers for what the courts (or any particular judge) will do when faced with any particular set of facts. A problem with this is that individual decisions apply policies and, as long as equitable principles remain intuitive instead of articulated, adjoining circuit courts may choose to apply different policies to similar cases. Particularly in an era of social change (such as contemporary changes in the roles of women) and in a legal context of general statutes, unpredictability and inconsistency in decrees are likely if the policies or principles which guide [203]*203and confine the exercise of equitable discretion are left unstated by appellate courts.
Accordingly, in our more recent cases we have attempted to rationalize and articulate the equitable policies that guide us. Intimations of this notion are found in Abraham v. Abraham, 248 Or 163, 165, 432 P2d 797 (1967), where, after a customary acknowledgment of the “clear abuse” standard, we said:
“* * * ‘Discretion’ as used in this context does not mean the power of free decision unreviewable by an appellate tribunal, nor does it mean a latitude of judgment where there is an absence of any rules. It means the quality of sound and non-arbitrary judgment in accordance with certain limiting guidelines.”
In Grove we described our function on review:
“* * * [Although each case must be considered on its own facts, it is proper to develop general principles, in addition to and consistent with those provided by the legislature, to the end that similar cases will be treated similarly. * * *” 280 Or at 353.
Similarly, in Smith v. Smith, 290 Or 675, 680, 626 P2d 342 (1981), we said:
“Although each case must be considered on its own facts, it is proper to develop general principles to the end that similar cases will be treated similarly. * * *”
In Grove and Smith we attempted to actually formulate principles of general application to discretionary decisions in cases involving similar problems.
It is consistent with a proper relationship of trial and appellate courts that the latter “develop general principles” which state equitable policies and goals so that the trial courts can be guided by them in fashioning each decree. If appellate courts perform their job of stating the principles which are to guide trial court discretion, then the trial courts can properly perform their job of applying those principles to varying situations in individual cases. Professor Rosenberg has described the role to be performed by trial courts in such cases when appellate courts properly perform their law-stating function:
“* * * The element of flexibility and choice in the process of adjudicating is precisely what justice requires in [204]*204many cases. Flexibility permits more compassionate and more sensitive responses to differences which ought to count in applying legal norms, but which get buried in the gross and rounded-off language of rules that are directed at wholesale problems instead of particular disputes. Discretion in this sense allows the individualization of law and permits justice at times to be hand-made instead of mass-produced.” Rosenberg, M., Judicial Discretion of the Trial Court, Viewed from Above, 22 Syr L Rev 635, 642 (1971).
Moreover, when the principles of equity are stated, appellate review can also be principled rather than ad hoc.1
Both the circuit court and the Court of Appeals went about the task of formulating a decree with intelligence and sensitivity. We will review each provision and attempt to articulate certain equitable principles which guide us in our review and which should guide both the trial and appellate courts in future like cases.
THE FACTS AND PROCEEDINGS
We take as the facts those found by the Court of Appeals. ORS 19.125(4). The parties are in their mid-forties. They began their marriage with little more than a paid-up Chevrolet. They were married for 24 years. Their four children are grown. Following the recent separation of the parties, the wife moved to Portland and enrolled in a beauty school. For medical reasons, her prospective work [205]*205week will be limited to four days. She anticipates net income of approximately $500 per month from this vocation.
The principal marital asset is the family business, a wheat farming operation purchased during the marriage, in 1963, from the husband’s mother for $127,000. The husband has run the business; the wife’s participation in its management has been minimal. The tillable acreage was doubled in 1977 by the purchase of adjoining land for $225,000.
In 1977, the family business was incorporated. It was capitalized at $500,000, or $500 per share for 1,000 shares of stock issued. Ownership of the shares was initially allocated based upon the source of the funds and tax considerations; the wife did not participate in that determination. At the time of trial, the wife owned 81 shares, the children 40, and the husband 879. The present value is $500 per share.
The trial court divided the property and awarded spousal support with the express purpose of keeping the farming corporation intact insofar as possible and providing funds for the near future and long-term financial security of the wife. It awarded to the wife vacation property and a new automobile together worth $100,000. The corporation was required to redeem her stock valued at $40,500 over a four-year period commencing June 1, 1980. All other property was awarded to the husband. That omnibus award included his stock in the corporation and, according to the testimony, these items: husband’s retirement account of $35,000, bank stock worth $4,000, life insurance worth $5,000 and $6,000 in a checking account, totalling $50,000. The wife was also awarded $1,500 monthly spousal support until November 1, 1980, and $1,000 per month thereafter. She was also awarded costs and attorney fees.
The Court of Appeals affirmed the award of property and modified the attorney fees award to $6,600. It was concerned that an award of spousal support did not adequately provide for the wife’s future because it was subject to modification or termination in the event of the husband’s death. It therefore modified the decree by [206]*206eliminating spousal support entirely and substituting in its place a requirement that the husband pay to the wife $235,500 composed of $40,500 for her stock interest and $195,000 to balance the greater award to the husband, payable in annual installments of $12,000 plus interest. That, together with the $100,000 in property awarded to the wife, equals one-half of the value of the property of the parties.
PROPERTY DIVISION AND SUPPORT
ORS 107.105(1)(e) requires that the court divide the property of the parties in such manner “as may be just and proper in all the circumstances.” Like the support provisions considered in Smith and unlike those in Grove, there is not an extensive statement of statutory objectives to guide us. ORS 107.105(1)(e) requires that we presume a homemaker spouse to have contributed equally to the marital assets which suggests an equal division of assets if the division is based on contribution.2 It does not require division based on contribution and does not exclude other considerations.
Dissolution of a marriage is analogous in many ways to dissolution of a partnership or other joint financial venture. Note: The Implied Partnership: Equitable Alternative to Contemporary Methods of Postmarital Property Distribution, 26 U of Fla L Rev 221, 226-27 (1974). The analogy is not complete, however. Unlike a business dissolution, a marital dissolution often requires the achievement of certain social as well as financial objectives which may be unique to the parties. See, e.g., Clark, Law of Domestic Relations, 450-51 (1968). For example, the parties [207]*207to a long-term marriage should be awarded the resources for self-sufficient, post-dissolution life apart insofar as possible within the limitations of the capabilities and property of the parties. Often, as here, that objective cannot be well served by a simple division of assets. As the husband argues, we can expect no more golden eggs if the decree kills the goose that lays them.
In formulating a decree, the Court of Appeals relied on an earlier statement of guiding principles which we also find to be apt:
“* * * [I]n dividing the property the dissolution decree should seek to disentangle the parties’ financial affairs and make them free from each other’s interference. The friction resulting from the unsuccessful marriage partnership almost inevitably makes continued business association untenable. While it is common in marriages of this long duration to attempt an equal division of property, there is ‘no hard and fast formula.’ Property division and spousal support should be considered together in attempting to put the two parties in a position so that they may leave the marriage in a self-sufficient status. * * *” (Citations omitted.) Slauson and Slauson, 29 Or App 177, 183-84, 562 P2d 604 (1977).
The award by the courts below to the husband of his shares in the corporation and the requirement that the wife’s shares be purchased from her is sound. The corporation is the principal source of capital and income upon which the future financial security of the parties must be based. As in. McCoy and McCoy, supra, the wife’s participation in the management and development of the corporate interest has not been substantial. Rather, the corporation reflects the industry and operational control of the husband and its profitability is dependent upon his continued participation. It would be legally possible to order the sale of the corporation and divide the proceeds, thus providing to each party a base upon which to rebuild financially. This solution is not desirable in this case, however, because evidence establishes that the assets of the parties would be substantially reduced by taxes on the proceeds of sale. The value of the corporation to both parties is greatest as an operating entity in the continued operational control of the husband, potentially generating cash, profits and appreciated value, free from the influence of minority ownership in a former spouse. Compare: Malloy and Malloy, 31 [208]*208Or App 1359, 572 P2d 672 (1977), and Ford and Ford, 26 Or App 353, 552 P2d 563 (1976). Where division of the principal asset of the marriage would unnecessarily dissipate its value and where alternative means can be found for dividing the financial benefit of the asset, the asset should be awarded intact to the spouse best able to manage it and other forms of balancing awards of property or support should be employed.
We agree with the division of noncorporate assets by the courts below. We uphold the award of $100,000 in noncorporate assets to the wife. This promotes her self-sufficiency by allowing her to use them immediately for income or to liquidate them, depending upon market conditions and her ambitions, for her short-range needs or for investment. Also, the award to the wife of these assets minimizes the financial drain on the corporation which will be necessary in order to equalize awards to the parties. We also agree that the interest of the wife in the corporation should be purchased to unify the husband’s ownership. The award to the husband of his retirement account and his life insurance worth $40,000 recognizes that their value cannot be readily severed from his ownership. The award to him of $10,000 in cash and securities will give him a degree of fiscal flexibility in operating the farm and meeting his immediate financial obligations to the wife. We conclude that these assets were divided in a way consistent with the statement in Slauson which will tend to preserve the value of the assets and best enable them to begin life anew with maximum self-sufficiency.
The trial court and the Court of Appeals differed on how wife’s minority interest is to be purchased. The trial court ordered that the corporation redeem her stock for $40,500 over a period of four years. The Court of Appeals added $40,500 as a component of the judgment to be paid by husband. In attempting to avoid unnecessary dissipation of the marital assets under the circumstances of this case, it is proper for the court to consider evidence of the tax consequences of its decree. ORS 107.105(2); Cook and Cook, 280 Or 589, 572 P2d 312 (1977); see also Engle and Engle, 52 Or App 561, 629 P2d 397, rev allowed (1981). In determining which manner of purchase should be decreed, we are considering such testimony.
[209]*209The tax consequences of the approaches of the trial court and the Court of Appeals are substantially different. The corporation is the sole source of funds with which the husband can buy out the wife’s interest. If he does so by paying off the judgment as required by the Court of Appeals, he must withdraw from the corporation sufficient funds to do so. His withdrawals will be considered salary, dividends or return of capital and taxable as such. He will therefore be required to withdraw sufficient additional funds to pay taxes thereon at regular income or capital gains rates, plus more to pay the taxes thereon, and so on.
The result of the circuit court approach may prove to be more prudent. If the corporation redeems the wife’s shares, there will be no tax consequences to the corporation or to the husband. The tax consequences to the wife are the same whichever way the purchase is made. We therefore reinstate the provision of the circuit court decree requiring that the husband is to cause the corporation to redeem wife’s interest in the sum of $40,500 prorated over a four-year period, commencing June 1, 1980.
The Court of Appeals also modified the decree by eliminating the spousal support order. It concluded that spousal support was too impermanent an arrangement to secure the wife’s long-term financial interest where the payments are essentially a compensating device for the imbalance in the property division resulting from the award of the corporation to the husband. It therefore ordered a judgment for $195,000 to the wife payable in annual installments of $12,000 plus nine percent interest.
We calculate a judgment of $174,500 will balance the awards to the parties, excluding consideration of attorney fees. The equation is:
To Husband:
$ 50,000 439,000 (174,500) $ 315,000 stock, insurance and accounts stock in farming corporation judgment to wife
[210]*210To Wife:
$100,000 vacation property and automobile 40,500 stock and redemption payments 174,500 judgment to wife $315,000
We therefore modify the judgment amount to $174,500.
In circumstances where periodic payments are desirable as a continuation of the obligation of one spouse to support the other, it is generally appropriate that they be in the form of spousal support, subject to various statutory objectives such as self-sufficiency, and further subject to the risks of modification or termination similar to those which exist in a continuing marriage. See Grove and Grove, supra. Here, however, although the payments are in the form of support, they are actually awarded due to the practical inability to equally divide the farm. Therefore, we agree with the Court of Appeals that they should have a more permanent character than that of support payments. Also, property division in the form of periodic payments defers the ability of the recipient to use or profit from the unpaid principal amount. We therefore agree with the Court of Appeals that payments of the judgment amount should bear interest at the rate of nine percent per annum, ORS 82.010(3).
The husband’s petition for review objecting to this modification can best be summarized as expressing the maxim that even a court of equity cannot get blood from a turnip. Also, we have already concluded that the corporate turnip should be neither divided, destroyed nor sold. He points out that the payments for the first several years will be in excess of $30,000 each, that his source of funds will be the corporation, and that in addition to withdrawals of funds for principal and interest, he will have to withdraw sufficient funds to cover his own taxes payable upon the withdrawals. He must also have some money available for his own support. The only way to accomplish all this will be to sell off assets of the corporation, thus diminishing its value and its income-producing capacity, and reducing the likelihood that he can make future payments. In other [211]*211words, husband describes a “downward spiral” in which the value and capacity of the corporation is steadily depleted.
Husband suggests an alternative means to accomplish the same purpose. He urges that we retain the spousal support provision and guarantee its permanence by requiring husband to maintain adequate life insurance to reflect the declining balance. This will protect the wife’s entitlement to financial security even if the husband dies. Husband represents that this is a solution commonly employed by domestic relations trial courts. Assuring permanence of support by requiring life insurance may often be a desirable device in fashioning an equitable decree, but we do not believe it is the most desirable solution in this case. The husband presently has life insurance in an unspecified amount worth $5,000 and term insurance for $5,000. There is no way for us to assure ourselves that the husband is insurable for more or will remain so. As he grows older, the insurance premiums may increase to amounts which he will find difficult or impossible to pay, and the same arguments about depletion of the corporation may become applicable later. Under the circumstances of this case, we reject husband’s solution.
Although we have rejected his proposed solution, the objections he urges are nevertheless reasonable. Unless circumstances change greatly, and we do not see that likelihood from the evidence, husband will be unable to make payments without losing the very asset in which he is buying out the wife’s marital interest. An equitable division must take account of this practical reality.
We therefore agree with the Court of Appeals that a judgment should be awarded to the wife, payable by the husband together with nine percent interest. The judgment shall be in the amount of $174,500. Instead of requiring that the payments of principal be the same each year, however, we order that the husband be required to make monthly payments of principal plus nine percent interest per annum on an amortized basis over a 25-year period. Twenty-five years correlates roughly to the husband’s life expectancy, is sufficient time for the wife to establish independent financial means, and will take her into her period of social security eligibility. Therefore, we order that [212]*212the judgment be payable in monthly installments of $1,464.41 payable on the first of each month commencing upon the first day of the month following the entry of the mandate. The husband has been paying $2,000 monthly support during the litigation. This figure is within his means, although he will be hard pressed to make redemption payments. For the wife, this payment plus her earnings will roughly maintain her standard of living.
ATTORNEY FEES AND COSTS
Last, we come to the question of attorney fees and costs. The husband cross-appealed, challenging the award to the wife of attorney fees. The Court of Appeals modified the award to eliminate payment for her attorney’s travel from Portland to Pendleton.
ORS 107.105(l)(h) authorizes the trial court to award “such further sums as additional attorney fees or additional costs and expenses of suit or defense as the court finds reasonably and necessarily incurred by such party,” and similar authority is given to us on appeal by subsection (5). The statute grants judicial discretion, but does not direct when or how that discretion is to be exercised.
We find only one case in which this court undertook to clarify how the statutory discretion to award attorney fees was to be equitably exercised. In Turner v. Turner, 237 Or 39, 40-41, 390 P2d 360 (1964), we held without elaboration that an award of attorney fees and costs
“* * * js a matter largely within the trial court’s discretion, considering, among other factors, the financial resources of the parties, the property division made by the decree, and the fault of the parties. Blake v. Blake, 147 Or 43, 53, 31 P2d 768 (1934). The trial court acted within its discretion.”
The first part of the formula is general and the last part obsolete, see ORS 107.036. Also, the gender of the spouse is no longer a controlling factor. This case presents an opportunity to be more specific.
First, we observe that the statute is not intended to compensate the prevailing party. Indeed, it is often difficult to determine which party prevails in a dissolution suit. If the statute were so construed, the husband, as [213]*213successful petitioner, would be entitled to attorney fees and costs. He makes no such claim and the statute makes no such restriction.
We regard the statute as a legislative recognition that dissolution affects interests of the greatest personal importance and that neither spouse should be denied the opportunity to sue or defend due to lack of equal access to marital resources which may be available for that purpose. The “financial resources of the parties, the property division made by the decree,” and the support orders, if any, are all relevant to a trial court’s determination of attorney fees under the statute. If neither party has the resources, whether in the form of assets or earning power, to pay the costs of litigation, it would be inequitable to require either to pay the costs of the other. On the other hand, if the parties are equally able to bear the costs of litigation, an order for one to pay the other’s costs would not be equitable. The more difficult situation occurs where the cash, liquid assets or income-producing capability are in the hands of one party as, for example, when a custodial parent receives a house, young children and a support order, and the other spouse receives his or her professional practice or the family business. In such a case, provision for the spouse without resources at hand should be made and reflected in the property division. We need not and should not attempt to outline the variables of such cases here, for this is a simpler case.
Here, the trial court properly awarded attorney fees to wife as a component of an overall decree. The decree has been substantially modified on appeal and review. Therefore, the issue of costs and attorney fees should be reconsidered in light of the overall provisions of the present form of decree. Property and payments have been awarded to the parties evenly. Substantial costs of litigation for both sides have been reasonably incurred. If the husband is to pay the wife’s cost of litigation, then the property division should be adjusted accordingly to reflect his assumption of her obligation. That is not necessary, however, because both parties are financially able to pay their attorneys with resources now or soon to be available to them: the husband may look to the business and the property awarded him as a source of funds; the wife will receive payments from the [214]*214corporation and on the judgment and she can sell some property. In other words, both parties will have equal estates and access to resources with which to pay for the litigation. There is no reason to require either to bear the other’s cost.
The award of attorney fees and costs is vacated. The same considerations are applicable on appeal and review. Costs of appeal and review were reasonably incurred (unlike some dissolution appeals) and the parties are equally able to bear their own.
Affirmed as modified. No costs to either party.