The husband appeals a final judgment of dissolution of marriage, raising several points regarding the financial aspects of the judgment. In particular, he contends that the trial court erred in the valuation of his sole proprietorship C.P.A. practice by including an amount for goodwill and further erred in its equitable distribution of the assets of the marriage. He also claims that the court erred in its order requiring him to pay all medical expenses for the children of the mar*1215riage -without limitation and to pay the contributions to the prepaid college tuition plan established by the parties for their children. We reverse as to the valuation of the business, the equitable distribution of the marital assets, and the contribution to the medical expenses of the children, but affirm as to the contributions to the prepaid college tuition plan.
One issue of major dispute at trial was the valuation of goodwill in the husband’s C.P.A. practice. In Thompson v. Thompson, 576 So.2d 267 (Fla.1991), the supreme court established the method for valuing goodwill attributable to professional practices for the purpose of making equitable distribution of marital assets. It first cited with approval the Supreme Court of Missouri’s definition of goodwill in Hanson v. Hanson, 738 S.W.2d 429 (Mo.1987), which held that goodwill within a professional practice constituted the “value of the practice which exceeds its tangible assets and which is the tendency of elients/patients to return to and recommend the practice irrespective of the reputation of the individual practitioner.” Thompson, 576 So.2d at 269. The court then noted that personal goodwill represented no more than probable future earning capacity which was not a proper consideration in asset distribution on dissolution, although it would be relevant in determining alimony. Id. at 270 (citing Taylor v. Taylor, 222 Neb. 721, 386 N.W.2d 851 (1986)). Personal goodwill should not be included in the value of the professional practice for purposes of equitable distribution. The court then directed that the fair market value approach should be used as the exclusive method of measuring goodwill of a professional practice, which it described as what a willing buyer would pay and a willing seller would accept, neither acting under duress for a sale of the business. “The excess over assets would represent goodwill_ Actual comparable sales are not required, so long as a reliable and reasonable basis exists for an expert to form an opinion.” Thompson, 576 So.2d at 270.
Applying Thompson, the court in Young v. Young, 600 So.2d 1140 (Fla. 5th DCA 1992), noted that a determination of goodwill valuation involves a two step process: first, there must be proof of the existence of goodwill separate and apart from reputation, and second, there must be proof of its value. In that case, there was proof of neither element. Later, in Weinstock v. Weinstock, 634 So.2d 775 (Fla. 5th DCA 1994), the court held that comparable sales where the selling professional remained with the buyer for a period of time could not be used to establish evidence of goodwill, because Thompson required that, to be a marital asset, goodwill must exist separate and apart from the reputation or continued presence of the marital litigant. Id. at 778. In other words, to be comparable, the sale must be one which eliminates any further personal influence which the seller might have over the business. A similar conclusion was reached in the pre-Thompson case of Spillert v. Spillert, 564 So.2d 1146 (Fla. 1st DCA 1990). The Spillert court criticized the capitalization of future income method for valuing a sole medical practice partly on the ground of the assumptions it made that a noncompete agreement would exist on any sale and that the seller would stay with the business for a period of time.
In line with these principles, we examine the facts of this case. The husband’s practice, Walton & Company, was a successful one in Boca Raton, and he employed two other C.P.A.’s and two support personnel in his office. While the other C.P.A.S and support staff had contact with the clientele, the majority of conferences with clients were conducted by the husband himself. There was no evidence that anyone besides the husband brought in clients to the company.
Some years prior to the divorce the husband had submitted a loan application which valued the business at $300,000, but the husband explained that this valuation was not made with the principles of Thompson in mind. Instead he valued it as though it represented a bundle of contracts, including a covenant not to compete, interest free loans, a contract for introduction services as well as for consultation during the transition period, and a retention contract.
The husband’s expert testified that the fair market value of the company was $41,733, *1216employing the liquidation valuation method. In essence, this was the value of the assets of the corporation, the expert failing to find any professional goodwill attributable to the business other than the personal reputation and efforts of the husband.
The wife’s expert, who completed his valuation of the business the night before his testimony, determined that institutional goodwill made up 15% of the total goodwill of the business. He used two methods for calculating the value of goodwill. The first was the excess earnings approach. According to the ledger sheets submitted into evidence, he first obtained an average net practice income of $162,391, from which he deducted an average compensation of $91,155. We are not told where this “average” compensation comes from, although it clearly was not the husband’s average compensation rate.1 The difference between those figures was the “excess earnings before cap rate”. We are not told what a “cap rate” is but the ledger shows that it is 25%. Total goodwill was then calculated at $284,944, which was the excess earnings divided by the cap rate. He then assigned institutional goodwill of 15% of the total, or $42,742. There was no testimony or documents to explain how the 15% was arrived at, and it seems that it was an arbitrary figure picked by the wife’s expert. He gave no market data to support it. Moreover, his use of this figure could not be based on his experience in valuing C.P.A. practices, as this was the first one he had ever valued.
His second method of valuation was what he called the “market approach.” That method began with the “average fee income” of $420,000, to which was applied a “market %” of 75% to arrive at a “market goodwill” of $315,000. Applying the 15% figure for institutional goodwill to this figure gave an institutional goodwill value of $47,250.
None of the accountants who testified provided any figures for comparable sales of similar businesses. The wife’s accountant was asked his opinion as to the value of the business if the husband sold his practice without a noncompete clause in the contract so that the clients could follow him to a new practice. The expert testified, “Most of the time if one is going to sell his practice there is going to be a non-compete agreement. Nobody is going to buy a practice and let that accountant go across the street and practice basically.” (emphasis supplied).
The trial court valued the business in accordance with the testimony of the wife’s expert, finding that the husband’s own valuations of the business gave evidence that the company “has a value in excess of its computers, adding machines, and office furniture.” While that was true, the court did not make the key distinction that only that part of the value independent of the husband’s continued presence in the business amounted to a marital asset.
The reliance on the goodwill valuation of the wife’s expert was error. First, there was no proof of the existence of goodwill separate from the husband’s reputation. The husband’s name was the only one “on the door,” and the other C.P.A. employees were there to assist in the work, not in garnering clients. The most telling evidence of a lack of any institutional goodwill was the wife’s expert’s testimony that no one would buy the practice without a noncompete clause. If the business only has value over and above its assets if the husband refrains from competing within the area that he has traditionally worked, then it is clear that the value is attributable to the personal reputation of the husband. Secondly, the valuation testimony of the expert was not supported by competent substantial evidence. There is no explanation whatsoever as to how he arrived at his various factors used to fix goodwill. The only reliance on established data which we can find in the record was his reliance on a survey of the Texas Society of Accountancy as to the purchase price of a practice. However, there was no evidence that this shows comparable sales for the state of Florida or the region in which the husband’s practice is located. Nor were there any figures given from the survey which would establish the percentage of institutional goodwill in a sole *1217proprietorship. In short, as in Young and Weinstock, we find no competent evidence from which the trial court could have determined the existence of goodwill separate from the reputation of the husband. Any testimony in that regard is sheer speculation. On remand, we direct the court to exclude any value of goodwill attributable to the business.
Because this case must be remanded to redistribute the assets of the marriage in light of our exclusion of the goodwill value of the husband’s business, we also address the method the trial court used to equitably distribute the marital assets. In the final judgment the trial court first awarded net marital assets to each party and then allocated the full liabilities of the marriage. Although the trial court believed that it had equitably divided the assets of the parties (indicating that there was only a $15,000 difference in asset distribution), when the gross value of each asset is assigned to each party in accordance with the court’s formulation and then the liabilities are assigned to each again according to the final judgment, the husband ends up with thirty percent of the net assets while the wife receives seventy percent. There are no reasons stated for such an inequitable distribution, and we conclude that the court did not intend this result but merely erred in the mathematical calculations made, as the judgment indicates that the court believed that the asset distribution favored the husband. On the other hand, if the court intended the result, then it is not supported with written findings justifying the disparity of distribution. § 61.075(3), Fla. Stat. (1993). On remand, we encourage the court to allocate the gross value of the assets to each party and the full amount of liabilities to each party and then net those figures out to determine the equitable distribution of the assets.
The final judgment ordered the husband to “continue to cover the children for all medical insurance purposes and he will be responsibilities [sic] for all premiums, deductibles, and uncovered medical insurance expenses of the minor children.” The husband argues, and the wife concedes, that the expenses should be limited to those that are “reasonable and necessary.” We agree. Pauley v. Pauley, 652 So.2d 488 (Fla. 4th DCA 1995). See Armstrong v. Armstrong, 623 So.2d 1216, 1219 (Fla. 4th DCA 1993); Watford v. Watford, 605 So.2d 1313 (Fla 4th DCA 1992). On remand the judgment should be corrected in this regard.
Finally, the husband challenges the portion of the final judgment ordering that he “shall continue to contribute to and be responsible to contribute to the pre-paid college tuition program for the three minor children.” The husband contends that he is under no legal obligation to pay the anticipated college expenses for the minor children, citing Leaird v. Leaird, 540 So.2d 243 (Fla. 4th DCA 1989), which held that there is no legal obligation to furnish a child with a college education.
There is scant evidence in the record concerning the obligation to the prepaid college tuition fund. The husband listed it as a liability on his financial affidavit. There was no testimony that we could find in the record referring to it. The motion for rehearing did not object to the imposition of this requirement in the final judgment.
Here the parties entered into a legal obligation to advance fund the college tuition of their children through the Florida Prepaid College Tuition Program years before the contemplation of dissolution. This is a program which allows the purchase of a contract to pay in advance the costs of postsecondary education at a guaranteed level for the duration of undergraduate enrollment. § 240.551(1), Fla.Stat. (1993). The intent of the Legislature in establishing this program was to foster timely financial planning for college expenses. Id. Each participant signs a contract to make periodic payments until the contract is paid in full. See § 240.551(6)(a). It is a benefit that the parties have conferred on their children by their wise financial planning. Where the parties themselves have agreed to undertake this obligation and signed a contract to that effect and already committed resources to its funding, we see no reason why this should not be treated as any other liability of the marriage and allocated according to the discretion of the judge.
*1218Because we have directed that the equitable distribution of the marital assets be reconsidered and have eliminated a portion of the value of the husband’s business from the calculation, this cause must be remanded for further proceedings. As this may change the court’s overall plan, we vacate the entire final judgment except for the dissolution of marriage itself, so that the court may refashion an appropriate financial plan for these parties. We affirm as to the remaining issues raised.
Reversed and remanded for further proceedings.
DELL, C.J., and HERSEY, J., concur.