650 So.2d 801 (1995)
MORTON M. GOLDBERG AUCTION GALLERIES, INC., et al.
v.
CANCO, INC., et al.
No. 94-CA-0734.
Court of Appeal of Louisiana, Fourth Circuit.
January 31, 1995.
Frank H. Walk, Jr., New Orleans, for appellant, Morton M. Goldberg Auction Galleries, Inc.
Arthur W. Landry, Plauche, Maselli & Landry, New Orleans, for appellees Canco, Inc. and American Mfrs. Mut. Ins. Co.
Robert M. Johnston, Anne Derbes Keller, Shannon Howard-Duhon, Adams and Johnston, New Orleans, for appellee Whitney Nat. Bank.
René A. Pastorek, Neal & Pastorek, Metairie, for appellee Mintz & Mintz Realty, Inc.
Before BARRY, PLOTKIN and WALTZER, JJ.
PLOTKIN, Judge.
Following a bench trial on the merits, plaintiff, Morton M. Goldberg Auction Galleries, Inc., brings this appeal seeking an increase in damages for lost profits it was awarded against defendants, Canco, Inc. ("Canco"), Whitney National Bank ("the Whitney"), and Mintz & Mintz Realty, Inc. ("Mintz & Mintz"). Based on our review of the record, we affirm.
FACTS
Plaintiff is a well-known dealer of antiques in the New Orleans area. In November of 1990, it moved its business from Magazine Street to a building located at 543-47 Baronne Street. Plaintiff leased the building from its owner, defendant Mintz & Mintz. Defendant Canco owned an adjoining piece of property fronting on Dryades Street known as the Civic Theatre. This piece of property included an arcade alleyway. Also adjoining plaintiff's building was a building owned by defendant the Whitney.
On June 10, 1991, metropolitan New Orleans suffered a serious rainstorm. According to plaintiff's petition, as a result of this deluge, rain water fell onto defendants' buildings, drained into the arcade alleyway, and ultimately overflowed into plaintiff's premises. Plaintiff alleged that the water caused damage to his business, including the costs of cleaning up the mess created by the flood and lost profits. Plaintiff brought suit against defendants on June 5, 1992.
The matter went to bench trial on February 8 and 9, 1994. On the morning of trial, defendants stipulated to their shares of liability, apportioning 90% to Canco and 5% each to the Whitney and Mintz & Mintz. At the close of the trial, the court took the matter under advisement. On February 11, 1994, the trial court rendered its decision, finding defendants at fault in causing the flooding of plaintiff's premises. It awarded plaintiff $28,650.28 in special damages and $27,157 in lost profits. Plaintiff has appealed devolutively, attacking only the trial court's award of damages for lost profits.
ISSUES
The only issue that we must resolve is whether the trial court was correct in calculating plaintiff's award for business interruption losses. However, in order to answer this inquiry, we are required to resolve four discreet sub-issues, to-wit:
1. Whether the trial court erred in subtracting fixed expenses from gross revenues in its calculation of lost profits?
2. Whether the trial court erred in relying on the regression analysis offered by defendants' expert witness to calculate plaintiff's lost profits?
3. Whether the trial court erred in awarding lost profits only for the three months following the flood?
4. Whether the trial court erred in denying an award of lost profits following plaintiff's European buying trip in August, 1991? We will address each of these questions in turn.
DISCUSSION
Deduction of Fixed Expenses
In its first assignment of error, plaintiff contends that the trial court erred in deducting fixed expenses in calculating plaintiff's lost profits. Defendants contend that both variable and fixed expenses must be deducted in calculating lost profits. In its Reasons for Judgment, the trial court noted in pertinent part:
[W]ith respect to the loss of profit or business interruption the Court finds that the approach by Dr. Wood [defendants' expert] with respect to lost revenues, cost of goods sold and variable costs is more appropriate than the competing conceptual approach and figures by Mr. Berger and Mr. Boudreaux [plaintiff's experts].
The plaintiff experts calculated loss profits by subtracting variable costs from revenues but not fixed costs. Dr. Wood subtracted both variable costs as well as fixed costs.
This latter method is the appropriate method and is in accordance with the decision in Graham v. Edwards, [614 So.2d 811, 819 (La.App. 2d Cir.1993)].
Before turning to an examination of the merits of plaintiff's claim, we must first address defendants' contention that the trial court's written reasons for judgment are not subject to appellate review. In support of this argument, defendants cite this court's statement in Cornelius v. Housing Authority of New Orleans, 539 So.2d 1250 (La.App. 4th Cir.1989), that "[w]ritten reasons for a judgment do not constitute the judgment and are not appealable. Thus our review is limited to the trial court's result, i.e., its judgment and not the reasons." Id. at 1252 (citations omitted). Defendants also point to similar language in this court's decision in Succession of Velasquez-Bain, 471 So.2d 731, 751 (La.App. 4th Cir.), writ denied, 476 So.2d 354 (La. 1985).
Defendants' argument is misplaced. In both Cornelius and Velasquez-Bain, this court was addressing cases in which the trial court had entered a single, combined judgment and reasons for judgment. Such a document contravenes the requirements of Code of Civil Procedure article 1918, which states that written reasons for judgment "shall be set out in an opinion separate from the judgment." La.C.C.P. art. 1918. Thus, in cases in which the trial judge has failed to comply with article 1918, only the "judgment" portion of the document in question is appealable. However, this does not mean that the appellate court may not consider the trial court's reasons for judgment when it reviews the record and applies the appropriate standard of review. In fact, the entire reason for having the trial court provide written reasons for judgment is to furnish the appellate court with guidance in understanding how the lower court reached its decision. As we noted in Velasquez-Bain, "it is the preference of the appellate court to receive written reasons for judgment." 471 So.2d at 751 n. 16. Defendants' contention that we may not examine the trial court's reasons for judgment in this case is without merit.
In order to facilitate resolution of the question we now confront, a brief statement of economic principles relied on by the experts in this case is necessary. All businesses have two categories of expenses: fixed costs and variable costs. As explained by plaintiff's expert, Mr. Kent Berger, fixed costs "are those costs that the business must expend regardless of the sales level. For example, something like rent." In contrast, variable expenses are those that change in relation to the level of sales by a business. Naquin v. Department of Trans. & Dev., 604 So.2d 62, 68 (La.App. 1st Cir.), writ denied, 608 So.2d 169 (La.1992). The court in David Sloane, Inc. v. Stanley G. House & Assocs., Inc., 311 Md. 36, 532 A.2d 694 (1987), offers this explanation of the distinction between fixed and variable costs:
In the broad sense, all of the seller's costs are necessary to the production of revenues. Each expense plays a role in the present and continuing operations of the business, and the central object of the business is to make sales in order to produce revenues and ultimately profits. Nevertheless, some costs are more directly related to the performance of individual contracts than are others, and their amount will vary with the volume of sales output. The other expenses will remain relatively stable through a given range of output because they bear no direct relationship to the individual contracts. Accountants and economists refer to these two distinct kinds of costs as variable and fixed; the latter is also commonly referred to as overhead. The variable element represents those items of costs which may be identified as belonging to a specific contract or product of sale. For example, the cost of materials and labor which go directly into the production of contract goods are considered variable coststhey vary with the number of contracts which are performed. If the contract is not performed they presumably are not incurred. An example of a variable cost for a nonmanufacturing seller is the cost to him of purchasing from his supplier a product intended for resale. In contrast to items of cost which vary directly with the volume of output, there are costs which have only an indirect relation to output. These costs are necessary to the maintenance of the business but do not change except with significant changes in output. For example, rental payments or depreciation on a plant, warehouse, store or office may be the same whether few products are produced and sold or many are. Executive salaries are usually fixed costs and are set before the precise output has been determined. Clerical and administrative salaries, and advertising and insurance costs may or may not be considered fixed costs in a particular period, but if they are, they are attributed to that period's total output. It has been said that fixed costs are those which "continue if the firm is temporarily shut down, producing nothing at all."
Id. 532 A.2d at 698 n. 4.
When added together, fixed costs and variable costs account for 100% of the firm's total expenses. However, not all costs of a business are entirely fixed or entirely variable. Rather, a particular cost may have both a variable and a fixed component. This is known as a mixed cost. In point of fact, according to Dr. J. Stuart Wood, defendants' expert economist, most costs of a business are mixed.
Expert witnesses for both sides agreed that calculation of business interruption losses such as those at issue in this appeal is a multi-step process. The first step is to take the firm's gross sales for the interruption period and deduct the cost of goods sold. This leaves gross profits. The second is to subtract from this figure the firm's variable costs. This leaves what was referred to at trial as "lost profits before fixed expenses." These initial steps were agreed to by all three expert witnesses that testified at trial. However, it is after these two steps are performed that the witnesses disagreed. Plaintiff's experts contend that "lost profits before fixed expenses" is the amount that should be awarded as damages for lost profits. Defendants' expert contends that it is necessary to deduct the firm's fixed expenses from "lost profits before fixed expenses" prior to making an award of lost profits.
The jurisprudence addressing business interruption losses has established two rules, application of which depends on whether the business continues to operate following the incident in question. In cases where the business ceases operation, courts generally do not include the business' fixed costs in the award of damages. Conversely, in cases where the business continues to operate, the award for loss of profits will include fixed costs. Peacock's, Inc. v. Shreveport Alarm Co., 510 So.2d 387, 407 (La.App. 2d Cir.1987) ("Loss of projected gross profits ... instead of net profits, may be employed to measure tort damages when the business and its normal overhead continues after the damage."); see White v. Rimmer & Garrett, Inc., 340 So.2d 283, 286 (La.1976); see also Radiofone, Inc. v. Pricelluar Corp., Civ.A. No. 91-4306, 1993 WL 133830, at *7 (E.D.La. Apr. 16, 1993); Jetz Serv. Co. v. Salina Properties, 19 Kan.App.2d 144, 865 P.2d 1051, 1057 (1993).
The rationale for this distinction is straightforward. Deduction for fixed costs is appropriate when the firm ceases operation following the delict because it has not incurred those costs. If the court were not to deduct fixed costs, the plaintiff would receive a double recovery. In contrast, when the firm continues to operate following the delict, fixed costs persist and must be paid by the firm. Accordingly, deduction of fixed costs in such a case would, in effect, be punishing for the firm for continuing to operate. Thus, the net result of deducting fixed costs when a firm continues to operate is to punish the firm for mitigating its damages, as is clearly required under Louisiana law. See La.C.C. art. 2002 ("An obligee must make reasonable efforts to mitigate the damages caused by the obligor's failure to perform.").
As noted above, the trial court in this case stated in its Reasons for Judgment that it was deducting fixed expenses from the plaintiff's recovery, relying on the Second Circuit's decision in Graham v. Edwards, 614 So.2d 811 (La.App. 2d Cir.), writ denied, 619 So.2d 547 (La.1993). In that case, plaintiff, a self-employed barber, brought suit after he was injured in an automobile accident. After the trial court awarded plaintiff $1,534.76 in lost wages,[1] defendants appealed, arguing that the amount of lost wages should have been decreased by the rent plaintiff was required to pay for his space in the barber shop. According to the appellate court, "[t]he proper measure for damages for lost profits in the case of a sole proprietor is net loss, i.e., gross income less all business expenses incident to the operation of the business." Id. at 819. Applying that principle to the facts at hand, the court stated:
In calculating the lost wages award, the trial court may well have considered that, in view of [plaintiff's] marked decline in gross sales for the months in question, the evidence did not clearly show that he continued to pay rent to his mother[ who owned the barber shop], thus warranting further deductions from gross sales. Viewing the evidence in the light most favorable to the plaintiff and in view of the discretion afforded the trier of fact, we find no error in the trial court's award of damages for loss of past wages in the amount of $1,534.79.
Id. at 820. Graham thus demonstrates the principle discussed above that when a business continues to operate following a delict, the court should not deduct fixed expenses from the award for lost profits.
In support of its argument that fixed costs should not have been deducted by the trial court in this case, plaintiffs cite a number of cases, including Peacock's, Rosenblath v. Louisiana Bank & Trust Co., 432 So.2d 285 (La.App. 2d Cir.1983), and Fidelity & Guaranty Insurance Co. v. Central Plumbing & Heating, 339 So.2d 904 (La.App. 3d Cir. 1976). We find the rule of law announced in those cases and discussed above to be controlling, to-wit: When a business suffers a loss of damages through the fault of another but continues in its day-to-day operations, an award of damages for loss of profits should not deduct fixed expenses. Rather, the plaintiff should be made whole by an award of net profits (that is, gross sales minus cost of goods sold minus variable costs) including fixed costs. See Peacock's, 510 So.2d at 407; Rosenblath, 432 So.2d at 289 n. 3 ("Since plaintiff's normal overhead expenses continued... loss of gross profits is the proper award."); see also Fidelity & Guar. Ins. Co., 339 So.2d at 911. This principle guided the Second Circuit in Graham and we can discern no reason to depart from this sensible rule. Therefore, we find that the trial court erred in concluding that plaintiff's recovery should have been reduced by deducting fixed expenses.
However, despite stating that it would deduct fixed expenses from the plaintiff's recovery in this case, our review of the trial court's Judgment, Reasons for Judgment, and the expert reports received in evidence in this case convinces us that the trial court did not, in fact, deduct fixed expenses from plaintiff's recovery. The trial court awarded $27,157 in damages, the exact figure offered by defendants' expert economist, Dr. Wood. His report arrived at this figure as follows:
Lost Revenues: $512,924
Less: Cost of Goods Sold ($372,998)
Less: Variable Costs ($112,769)[2]
Lost Profit: $ 27,157
Nowhere does this computation reveal the deduction of fixed costs from the plaintiff's award of lost profits. As such, there is no evidence that plaintiff was deprived of recovery for its fixed costs. Accordingly, we find no merit in this assignment of error.
Regression Analysis
In its second assignment of error, plaintiff asserts that the trial court erred in relying on the regression analysis[3] offered by defendant's expert, Dr. Wood, to calculate plaintiff's variable expenses. Plaintiff contends that the calculations offered by its own expert, Mr. Berger, should have been accepted by the trial court. In order to resolve this dispute, it is necessary to understand the methodology employed by each expert.
Berger, who at the time of trial had been plaintiff's accountant for two years, testified that in preparing his reports, he reviewed the records of plaintiff's business, interviewed members of management, observed the actual operations of plaintiff's business, and relied on his best professional judgment. Based on the foregoing, he determined that only eight of plaintiff's 41 categories of expenses had any variable component. Those categories were advertising, drayage/freight, customer parking, non-officer salaries, casual labor, commissions, photography/printing, and postage. Based on this information, Berger determined that 8.8% of plaintiff's total costs were variable.
Berger's methodology was attested to by Dr. Kenneth Boudreaux, who was tendered by plaintiff and accepted by the court as an expert economist in the calculation of a business interruption claim. Boudreaux testified that he regularly conferred with Berger while Berger was making his calculations. Boudreaux further opined that Berger's methodology was "perfectly reasonable" and that there was "no good alternative to that given."
Defendant's expert, Dr. Wood, offered testimony on two points. First, he verified the accuracy of and critiqued Berger's methodology. Second, he prepared an independent analysis of plaintiff's variable costs. On the first point, Wood found a number of errors in Berger's computations using Berger's own methodology. According to Wood, if the court accepted Berger's methodology in theory, the 8.8% variable cost figure should be adjusted to 9.15% to reflect these mathematical errors.
On the second point, Wood's analysis indicated that 12 categories of expenses in addition to the eight identified by Berger had some variable component. These categories were travel, professional fees, equipment rental, car and truck, furniture repair, insurance, miscellaneous, office expense, officers' salaries, outside service, repairs and maintenance, taxes, telephone, and utilities. Based on this information, Wood concluded that 22% of plaintiff's costs were variable.
Furthermore, although Wood opined that plaintiff had suffered no loss in sales that was attributable to the flood, he did conclude that based on Berger's computation of lost revenue, plaintiff's lost profits amounted to $27,157, the exact amount awarded by the trial court. Significantly, in offering this assessment of lost profits, Wood stated: "Therefore, I conclude that, if Mr. Berger's estimate of the lost revenues is correct (and I disagree that it is), then the total amount of lost profits to Goldberg Auction Galleries is $27,157.00, using the methodology recommended by Mr. Berger."
Our review of the record reveals no error by the trial court in relying on Wood's analysis. Wood's report, including the scatter diagrams contained therein, convincingly demonstrates that each of the 12 categories he identified varied to some degree with sales. As such, these expense categories constitute variable or mixed costs and thus should have been taken into account in computing plaintiff's lost profits. The reliability of Wood's approach is confirmed the by the high r2 value of his analysis (.73).[4], which, according to Wood, constitutes the most important confidence measure and thus shows his analysis to be "a good measure of the reliability of the results." Although the standard error of the coefficient was somewhat high (.07),[5] Wood indicated that this was due at least in part to the nature of the monthly expense data with which he was provided.
Ultimately, this is a question of discretion vested in the trial court in assessing the credibility of expert testimony and its probative value. Although there was sufficient evidence in the record to support either of the expert's positions on the question of variable expenses, the trial court is vested with the discretion to determine which expert's testimony was more persuasive and helpful. Cf. La.C.C. art. 1999 ("When damages are insusceptible of precise measurement, much discretion shall be left to the court for the reasonable assessment of those damages."). As the Second Circuit recently reiterated in Sledge v. Continental Casualty Co., 25,770 (La.App. 2d Cir. 6/24/94), 639 So.2d 805, 815, "the effect and weight to be given expert testimony depends upon the facts underlying the opinion and, also, rests within the broad discretion of the trier of fact." Thus, we find no merit in this assignment of error.
Period to Measure Lost Profits
In its third assignment of error, plaintiff complains that the trial court only awarded lost profits for the three-month period following the flood (June to August, 1991). Plaintiff avers that the proper time period for which damages should have been awarded is six months (June to November, 1991).
We find no merit in this argument. The only testimony in the record that lends credence to this argument is the unsubstantiated testimony by David Goldberg, president of the corporate plaintiff, that the flood affected the business for six months. Both of plaintiff's experts merely assumed that this assertion was correct. No independent evidence corroborated this claim.
The trial court, by adopting Dr. Wood's approach to damages, inferentially found that three months was the appropriate time frame in which to measure lost profits. In light of our review of the record, we are not prepared to say that this finding of fact is clearly wrong or manifestly erroneous. Thus, we find no basis on which to overturn this ruling by the trial court. Plaintiff's third assignment of error is without merit.
European Buying Trip
By its final assignment of error, plaintiff complains that the trial court should have awarded it damages for lost profits resulting from diminished capital and increased expenses associated with a trip by David Goldberg to Europe in August 1991 to purchase inventory for the business. Plaintiff alleges that as a result of the flood, it had reduced capital on hand and thus could not qualify for a loan from a bank to finance its purchasing trip. Although plaintiff did secure alternative financing from a European lender, it claims that the lender imposed charges above those that plaintiff normally would have paid and thus it should be reimbursed for these expenses.
We conclude based on our review of the record that the trial court did not err in refusing to award this item of damages. There was testimony in the record that plaintiff applied for a loan from the Whitney prior to the flood but was turned down due to an excessive debt to equity ratio. An additional application to finance the European trip submitted after the flood was declined for the same reasons. In fact, in this second application, plaintiff declared it was seeking a loan because it had spent approximately $80,000 moving the business to its new location and that it needed "capital for breathing room." There simply was no credible evidence linking the reduced profits from the European buying trip with the flood of June 1991. Accordingly, the trial court did not err in refusing to award this item of damages.
CONCLUSION
Based on the foregoing, we affirm the trial court's judgment.
NOTES
[1] Under the particular facts of Graham, lost wages were the equivalent of lost profits for a business. Graham was a self-employed barber operating as a sole proprietorship. Thus, whatever amounts he lost as wages were the same as the profits lost by his business enterprise. Accordingly, the principles relied on by the Graham court in reaching its decision are applicable, although not controlling, to our decision in this case.
[2] As discussed infra, we find that the calculation of variable costs by Dr. Wood is not clearly erroneous and thus is appropriate for use in computing plaintiff's lost profits.
[3] Stated quite broadly, regression analysis is used to ascertain a mathematical relationship between two (or more) variables. If a sufficient correlation between variables is found to exist, expectations of future changes in one variable (the independent variable) about which one is presumably relatively confident is used to predict changes in the other (dependent) variable. Regression analysis, therefore, is most useful when the variable one desires to estimate is difficult to estimate directly, but is well correlated with a variable with a more easily predicted value, which can then be used to project future values of the dependent variable.
Cede & Co. v. Technicolor, Inc., Civ.A. No. 7129, 1990 WL 161084, at *11 (Del.Ch. Oct. 19, 1990). According to Dr. Wood, regression analysis "is considered to be by the literature in cost accounting, Financial Analysis and Managerial Accounting to be the best method to discover the variable and fixed components of costs."
[4] This measure, r2, is a correlation coefficient that describes the strength of a linear relationship between two measures. Donald J. Koosis, Statistics 191 (3d ed. 1985). When r2 is between + 1.0 and 0, this indicates that "[h]igh scores on one measure tend to go with high scores on the other but the relationship is not perfect." Id.
[5] Another correlation coefficient measuring the reliability of a linear relationship.