Plaintiff and defendant were married on 1 June 1975, separated on 10 October 1991, and divorced on 15 December 1992. Prior to receiving evidence at an equitable distribution hearing, the parties made numerous stipulations regarding the identity, classification, valuation and distribution of a substantial amount of property. On 24 October 1994, an equitable distribution trial was held with regard to the property issues on which the parties were unable to agree. This appeal pertains specifically to the methodology employed by the trial court in determining the value of two marital assets, a ten-acre tract of land and plaintiffs medical practice.
Plaintiff is a cardiologist trained in the highly specialized field of interventional cardiology, a sub-specialty of invasive cardiology. Invasive cardiologists perform diagnostic procedures, such as coronary catheterization, to determine whether a patient has heart disease or blockages in the arteries. Interventional cardiologists are trained to perform the same diagnostic procedures as invasive cardiologists, and in addition perform therapeutic treatments designed to remove blockages from the arteries.
Plaintiff was employed by Quadrangle Medical Specialists, P.A. from 1987 until he resigned in January 1989 and established his own cardiology practice known as Eastern Cardiology. Plaintiff selected a site located in a medical park on Stantonsburg Road in Pitt County (hereinafter Stantonsburg property) on which to build his medical offices.
In January 1990, plaintiff purchased the Stantonsburg property from Park West Properties for $389,000.00. Plaintiffs deed included a provision requiring the grantor to construct an access road from a public highway at the grantee’s request. The cost of the road construction would be shared by the grantor and grantee, with the grantee’s portion not to exceed $25,000.00.
Pursuant to an equitable distribution order entered 4 October 1995 nunc pro tunc 15 April 1995, the trial judge made numerous findings of fact valuing the marital assets of the parties. We turn first to *90the contested finding regarding the valuation of the Stantonsburg property.
[1] The trial court found as a fact that the value of the Stantonsburg property was $300,000.00. In addition the court made the following finding to which plaintiff assigns error:
24. (i) The deed of conveyance to plaintiff required the grantees to build an access road from Stantonsburg Road back to Plaintiffs ten-acre tract. The Court finds from the evidence presented that the cost to build this road, which would have to be built in order to get to plaintiffs land, was estimated to be no less than $75,000.00. The deed required the grantee, plaintiff, to pay up to, but no more than, $25,000.00 toward the road construction costs. The Court finds as a fact that the value of plaintiffs land is increased over the $30,000.00-per-acre value because the deed to plaintiff required the grantor to spend a sum (which the Court finds would be not less than $75,000.00 for the road), which is greater than the maximum ($25,000.00) plaintiff has to spend to build the road. The value of plaintiffs land is therefore increased because of the obligation of the grantor to build this road for plaintiff, and the increased value is $50,000.
(i) [sic] Therefore, the gross fair market value of the 10 acres was $350,000.
Plaintiff contends that because the access road was never actually constructed, the trial court improperly based its valuation of the Stantonsburg property upon a fact not in existence at the date of separation.
N.C. Gen. Stat. § 50-21(b) (1995) provides that “[f]or purposes of equitable distribution, marital property shall be valued as of the date of the separation of the parties.” In making a determination as to net market value of a marital asset, the trial court is required to only consider evidence of the value of the property as of the date of separation. Christensen v. Christensen, 101 N.C. App. 47, 55, 398 S.E.2d 634, 639 (1990). As of 10 October 1991, the date of separation, the access road had not been constructed. The evidence of the value the grantor’s promise to build the road may have added to the land was mere speculation and improperly considered by the trial court. However, defendant contends that the trial court properly considered evidence of the promise to build a road because the obligation was a fact in existence as of the date of separation.
*91Prior to ordering an equitable distribution of marital property, the trial judge is required to calculate the net fair market value of the property. Beightol v. Beightol, 90 N.C. App. 58, 63, 367 S.E.2d 347, 350, disc. review denied, 323 N.C. 171, 373 S.E.2d 104 (1988). “Fair market value is defined as the price which a willing buyer would pay to purchase the asset on the open market from a willing seller, with neither party being under any compulsion to complete the transaction.” Brett R. Turner, Equitable Distribution of Property § 7.03, at 505 (2d. ed. 1994). The trial court calculates the net fair market value, by reducing the fair market value of the property by the value of any debts that are attached to the asset. Id. at 505.
In this case, the trial court first determined the fair market value of the Stantonsburg property to be $300,000.00. The court then added $75,000.00 representing the cost of the road construction and then subtracted the $25,000.00 debt plaintiff would incur as a result of the improvement; calculating a net added value of $50,000.00.
The case sub judice is analogous to equitable distribution actions in which some trial courts have erroneously reduced the value of an asset by the cost of projected expenses associated with a possible future sale of the asset. The majority of jurisdictions hold that when determining the net fair market value of an asset, “the court should deduct only debts which are reasonably certain to exist in the near future.” Id. at 506. Accordingly, most jurisdictions hold that costs associated with hypothetical sales of assets should not be subtracted from the fair market value of the property. See e.g. McDaniel v. McDaniel, 829 P.2d 303 (Alaska 1992); Taber v. Taber, 626 So. 2d 1089 (Fla. Dist. Ct. App. 1993); In re Benkendorf, 252 Ill. App. 3d 429, 624 N.E.2d 1241 (1993); Goodwin v. Goodwin, 640 So. 2d 173 (Fla. Dist. Ct. App. 1994). When an actual sale of an asset is not imminent, “expenses of sale are hypothetical liabilities which may well never be incurred.” Turner, supra § 7.03 n.73 (Supp. 1996).
Moreover, the expenses of a future sale of an asset are uncertain in both occurrence and amount. Id. For example, the property owner may die and thus never sell the asset. Id. In any event, even if the sale does take place in the future, unless the sale is imminent, there is no reasonable basis upon which to predict the amount of expenses related to the sale. Id.
The majority rule precluding the deduction of expenses associated with future sales is equally applicable to the facts of this case. Plaintiff presented evidence showing that on several occasions he *92made oral and written demands to Park West requesting construction of the access road. Yet, as of the date of the equitable distribution hearing, a full three years after the date of separation, plaintiff was unaware of any plans to begin road construction on the Stantonsburg property. In fact, plaintiff ultimately decided to select another site on which to locate his medical practice due to the failure of Park West to provide the promised road.
In making a determination as to the fair market value of the Stantonsburg property, the trial court must ascertain the price a willing buyer would pay to purchase the land on the open market from a willing seller as of the date of the parties’ separation. The value, if any, of the obligation to build an access road on the property is intrinsic to the fair market price and should have been included in the trial court’s fair market valuation of the real property. The trial court’s findings of fact numbered 24(i) and (i) [sic] are vacated and the case is remanded for a determination of the fair market value of the Stantonsburg property on the date of separation, to include the value, if any, of the obligation to construct a road to the property.
[2] Next, we examine plaintiff’s assignment of error regarding the trial court’s valuation of plaintiff’s medical practice, particularly the “goodwill”, component of the practice. Goodwill, the most difficult element of a professional practice to value, is “commonly defined as the expectation of continued public patronage.” Poore v. Poore, 75 N.C. App. 414, 420, 331 S.E.2d 266 271, disc. review denied, 314 N.C. 543, 335 S.E.2d 316 (1985). “It is an intangible asset which defies precise definition and valuation.” Id. “There is no set rule for determining the value of the goodwill of a professional practice; rather, each case must be determined in light of its own particular facts.” Id. at 421, 331 S.E.2d at 271 (citations omitted). If it appears that the trial court, based on competent evidence and a sound valuation method, reasonably approximated the goodwill value of plaintiff’s medical practice, that valuation will not be disturbed on appeal. Id. at 422, 331 S.E.2d at 272.
The record shows that the trial court carefully considered the evidence presented by three different experts and determined that “capitalization of excess earnings” was the appropriate method for determining the fair market value of plaintiff’s medical practice. This Court has described the capital excess earnings method as a proper and legitimate means of measuring the present value of goodwill. Id. at 421, 331 S.E.2d at 271. Under this approach, the trial court first *93determines the difference between plaintiff’s actual earnings and the earnings of the “average” similarly situated physician. Turner, supra § 7.07, at 533; see also Poore, 75 N.C. App. at 421-422, 331 S.E.2d at 271-72. The difference between the compared earnings is then multiplied by a number (the factor) between one and five to yield the final value. Turner, supra § 7.07, at 533. The accuracy of this approach depends significantly upon the accuracy of the “average” statistics used in the comparison. Id. at 535.
Plaintiff contends that the trial judge committed reversible error by utilizing information regarding the average salaries of invasive cardiologists as published in a national survey entitled Physician Compensation and Production Survey: 1992 Report Based on 1991 Data (hereinafter “Physician Survey”) to make a determination as to the average salary of a similarly situated physician. Plaintiff contends that the trial court was required to utilize evidence presented as to the average salary of interventional cardiologists in Pitt County. We disagree.
The trial court relied on the testimony and evidence presented by Edward Strange, plaintiff’s expert in the field of evaluation of medical practices. Mr. Strange testified that in calculating the goodwill component of plaintiff’s medical practice, he relied on the Physician Survey because it was the only salary survey he could find that gathered information regarding the specialty of invasive cardiology. He also testified that it was a very good source because it included information on fringe benefits and compensation amounts “in terms of total production” as well as general compensation information.
To ascertain the salary of a physician similarly situated to plaintiff, Mr. Strange adopted the Physician Survey salary for an invasive cardiologist in the 90th percentile. Mr. Strange explained in his Valuation Report, that he used the 90th percentile column because he believed that the “conditions for earning ability in Greenville, North Carolina for invasive cardiologists would be represented in the upper strata.” During direct examination, he noted that the salary he selected was somewhat lower than the average earnings reported by Pitt County interventional cardiologists.
In finding the goodwill component of plaintiff’s medical practice, the trial judge rejected utilizing either the salary of the invasive cardiologist in the 90th percentile or the average salary of Pitt County interventional cardiologists. Rather, he determined the similarly situated physician salary by relying on the average salary of an invasive *94cardiologist in the 75th percentile as reported in the Physician Survey. In selecting the disputed salary, the trial judge explained his reasoning as follows:
[Finding #28(c)]... The Court finds from the evidence that a similarly situated invasive cardiologist would earn $418,000.00 per year. This salary figure is higher (in the 75th percentile) than the medial salary, because the plaintiff is a very hard-working individual, who strives to perfection, and who is ambitious in the best sense of the word. He is skilled in marketing his practice, and he is a “hard-driving” physician. The Court finds that Dr. Carlson also possesses an exceptionally higher level of skill than other invasive cardiologists. . . .
[T]he average income of interventional cardiologists practicing in Pitt County should not be considered because this group represents too small of a statistical sample, and all of the interven-tional cardiologists practicing in Pitt County may have practice good will. . .
[T]he Court does not find these physicians to be similarly situated, because a comparison of plaintiff to approximately six cardiologists in Pitt County provides too small a statistical basis or sample for this Court to find this comparison meaningful or reliable in finding the similarly situated physician. Additionally, Dr. Carlson actively markets his practice with advertising, clinics, and the like, and there is no evidence that the other Pitt County physicians market at all, much less to the degree of Dr. Carlson.
Plaintiff objects to the trial court’s finding that the number of interventional cardiologists in Pitt County was too small a sample to provide a useful comparison. He contends the trial judge was required to utilize the average salaries of local interventional cardiologists. We disagree.
Other courts have approved of the use of national statistics in determining the goodwill component of a business valuation “[w]here the number of businesses in the field is small and the market is essentially nationwide” or a “small local sample size prevents collection of reliable information.” Turner, supra § 7.07 at 536; see e.g. In re Bookout, 833 P.2d 800 (Colo. Ct. App. 1991) (using American Physical Therapists Association survey data); Clark v. Clark, 782 S.W.2d 56 (Ky. Ct. App. 1990) (using American Medical Association survey of obstetrical practices data). Given the facts and circumstances of this *95case, the trial judge did not abuse his discretion by utilizing national salary statistics to calculate the goodwill component of plaintiff’s medical practice.
The trial court reasonably approximated the goodwill value of plaintiffs medical practice on the basis of competent evidence and a sound valuation method. The court’s findings and conclusions with regard to the valuation of plaintiff’s professional practice will not be disturbed on appeal.
Vacated and remanded in part, affirmed in part.
Judges EAGLES and McGEE concur.