BRONSON, P.J.

Petitioner appeals as of right from the Tax Tribunal's determination of the value of its four parcels of real property located in Leoni Township, County of Jackson. The parcels of land are the site of a heavy industrial plant owned by petitioner. The four parcels were valued as a single entity and will be treated as such for purposes of this appeal. This appeal concerns the assessments of the subject realty and industrial plant for the years of 1978 and 1979. The Tax Tribunal found the true cash value of the property to be $8,487,766 for 1978 and $8,934,490 for 1979.

Petitioner's appraiser offered only a market approach to valuation of the subject property. Under petitioner's market analysis, the true cash value for each year was determined to be $5,000,000. The Tax Tribunal rejected petitioner's market analysis, finding that the properties used by petitioner's appraiser were not really comparable to the subject property. In reaching its valuation conclusions, the tribunal agreed with respondent's appraiser that a replacement cost minus depreciation analysis was most likely to result in an accurate true cash value determination for the property. This holding by the tribunal is the subject of petitioner's first assignment of error.

The Tax Tribunal found "that the best and most appropriate method of appraising an owner-occupied, on-going use industrial facility wherein the present use represents the highest and best use of the property FOR ASSESSMENT PURPOSES is to calculate its replacement (as distinct from reproduction) cost, allow for physical, functional and economic obsolescence (or depreciation) and add the land value to the appraisal". Petitioner objects to this "value in use" approach as having no relationship to the usual selling price, which is the statutory touchstone for determining the cash value of a property. MCL 211.27; MSA 7.27. Petitioner contends that the Tax Tribunal's decision improperly substitutes the hypothetical standard of "value in use" for the correct standard of "true cash value" as represented by "fair market value". As such, petitioner's claim is best characterized as an assertion that the tribunal adopted a wrong appraisal principle. This presents a legal issue which we are constitutionally empowered to review. Const 1963, art 6, § 28, Northwood Apartments v Royal Oak, 98 Mich. App. 721, 724; 296 N.W.2d 639 (1980).

There are three accepted methods for determining true cash value: market analysis using the selling prices of comparable properties, cost analysis using reproduction costs less depreciation, and income analysis using the capitalization of income approach. Wolverine Tower Associates v Ann Arbor, 96 Mich. App. 780, 781; 293 N.W.2d 669 (1980). It is the duty of the Tax Tribunal to select the method which is the most accurate after considering all the facts before it. Ramblewood Associates v City of Wyoming, 82 Mich. App. 342, 345-347; 266 N.W.2d 817 (1978).

Petitioner relies on Safran Printing Co v Detroit, 88 Mich. App. 376; 276 N.W.2d 602 (1979), lv den 411 Mich. 880 (1981), for its contention that a market approach should have been used instead of the cost analysis adopted by the Tax Tribunal. However, Safran was decided on its particular facts and does not control the instant case. In Safran, the Tax Tribunal insisted on reaching a valuation determination on a "value in use" basis where both the petitioner's and respondent's appraisers agreed that the subject plant was obsolete, inefficient, and could not be sold as a printing plant, its current use. Moreover, the tribunal specifically held that a true cash value based on an analysis of usual selling price was a mere guide, and not the statutorily imposed standard for determining the fair market value of a property. The Safran holding is easily summarized as follows: a particular property cannot be valued by reference to its current use where all evidence shows that, due to the property's inefficiency and obsolescence, no buyer would consider purchasing the property with the purpose of utilizing it in conformity with its present use.

In this case, we are confronted with a factual scenario quite different than that posed by Safran. Unlike the situation in Safran, all the appraisers in this case agreed that the subject property's current use is also its highest and best use. Indeed, petitioner's appraiser's market analysis report includes the following statement:

"The subject property was originally designed as a manufacturing plant and has been used for this purpose continuously since its conception. Although it has several obsolete design features, it is still modern enough to be considered for continued use for an industrial purpose. Moreover, it is currently occupied and used as an industrial plant and its owner-occupant has expressed no desire to abandon the property even though recent adjustments have been made in employee levels and product lines. Based upon the consistency of use exhibited by the above factors, the subject's highest and best use was estimated to be consistent with its current use as a manufacturing plant."

Contrary to petitioner's apparent contention, the Court in Safran did not hold that a cost analysis based on value in use could never be used to determine usual selling price. The Safran Court specifically noted that "existing use may be indicative of the use to which a potential buyer would put the property and is, therefore, relevant to the fair market value of the property". Id., 382. See also First Federal Savings Loan Ass'n of Flint v City of Flint, 104 Mich. App. 609, 617-618; 305 N.W.2d 553 (1981), which analyzes Safran and concludes, on grounds very similar to those we adopt, that Safran is not controlling.

We agree with petitioner, however, that the following excerpt from the Tax Tribunal's opinion flies in the face of Safran and is a misstatement of an applicable legal principle:

"Above all else, it [cost analysis based upon highest and best use] will get away from this fictional nonsense that an existing functioning industrial manufacturing or processing facility is so `outmoded' that its current use should be ignored and it should be valued on the speculative basis of what it would be worth if its current use were abandoned and if it were put up for sale on the `market' and ended up in a secondary use (which for some mysterious reason seldom rises above a nominal `warehouse usage', or at least so it seems in most of such similar appeals of similar facilities that are heard by this tribunal.) Indeed, rather than engage in the fiction of ignoring the current usage by the current owner altogether, as this sort of speculation requires, then why not approach the problem in a more reasonable manner by means of a `value in use' appraisal as though the property were to be continued in its present use or alternatively if sold that it might well be sold as a part of a going concern, where its tailor-made construction and fixtures will continue to fulfill their present use and efficiencies."

Despite this incorrect statement of the law, we need not reverse. The quoted excerpt was mere dictum in this case since, as noted previously, all the appraisers agreed the subject property could still be used for manufacturing purposes. Had this case involved facts like those in Safran, reversal would have been required, and, had this case involved a factual dispute over whether the subject property was too outmoded to be sold for use consistent with its present use, the matter would have had to be remanded for specific factual findings on this disputed issue.

The problem with valuing large industrial plants is a problem with the statutory standard itself. The reality is that these types of industrial plants are rarely bought and sold, so that a determination of "usual selling price" constitutes a metaphysical exercise which puts the Tax Tribunal in the position of having to resolve a question somewhat akin to how many angels can dance on the head of a pin. Petitioner may well be correct in its assertion that there is no market for its industrial plant at its current use. However, as we construe MCL 211.27; MSA 7.27, to the extent that an industrial plant is not so obsolete that, if a potential buyer did exist who was searching for an industrial property to perform the functions currently performed in the subject plant, said buyer would consider purchasing the subject property, the usual selling price can be based upon value in use. To apply MCL 211.27; MSA 7.27, a hypothetical buyer must be posited, although, in actuality, such a buyer may not exist. To construe MCL 211.27; MSA 7.27 as requiring the taxing unit to prove an actual market for a property's existing use would lead to absurd undervaluations. Large industrial plants are constructed to order, in accordance with the exact specifications of the purchasing user. Such plants are not constructed like small commercial buildings or residential structures with only a mere hope or expectation on the builder's part that the plant will be sold. When a large corporate entity such as Ford or General Motors builds a factory, it is probable that absolutely no market exists for the resale of that factory consistent with its current use. It is ludicrous to conclude, however, that such a brand new, modern, industrial facility is worth significantly less than represented by its replacement cost premised on value in use because, in actuality, such industrial facilities are rarely bought and sold. Thus, we hold that, to the extent a large industrial facility is suited for its current use and would be considered for purchase by a hypothetical buyer who wanted to own an industrial facility which could operate in accordance with the subject property's capabilities, said facility must be valued as if there were such a potential buyer, even if, in fact, no such buyer (and therefore no such market) actually exists. This is in accordance with First Federal Savings, supra, 619-620, wherein the Court states:

"In summary, we find that unwavering adherence to the `cost on the open market' approach would consistently undervalue those buildings which are especially suited to a particular use, are not obsolete, and are being used for the particular use for which the building was designed or altered. Because the subject property in the instant case is such a building, the Tax Tribunal did not commit an error of law or adopt a wrong principle in sustaining the assessment as made by respondent city."

As was the case in First Federal Savings, the subject property here remains suited to its particular use and is not obsolete.

Petitioner's second claim of error is that the Tax Tribunal used a depreciation factor unsupported by the evidence and, consequently, its cost analysis resulted in a true cash value for the subject property far in excess of that which was advocated by the respondent's appraiser, namely, $7,932,452 for 1978 and $7,843,800 for 1979. The tribunal rejected respondent's depreciation factor of 52%, on the basis that it was overgenerous and in excess of that allowed industrial facilities of comparable age in the Marshall-Swift Valuation Manual. The tribunal allowed a depreciation factor of only 40%. In its opinion, it found that respondent's appraiser had not been in the subject plant for five years and was not in a position to make an informed judgment as to the amount of depreciation to be allowed.

We find nothing per se objectionable in the Tax Tribunal's determination that a particular property has a value in excess of that advanced by either party. It is the tribunal's duty to ascertain the true cash value of a property when a dispute over its valuation comes before it. If, due to methodological flaws in the parties' analyses, neither side's valuation figure is accurate, the tribunal should be free to reject both. Nonetheless, on the facts of this case, we believe a remand for further proceedings is in order.

In the instant case, the Tax Tribunal reached its depreciation rate by adopting a figure from the Marshall-Swift Valuation Manual. However, this manual had not been introduced into evidence. The only expert testimony concerning the appropriate depreciation allowance was from respondent's appraiser, and his testimony supported a higher depreciation factor than that allowed by the tribunal. It is one thing for the tribunal to reject the respondent's proposed depreciation figure because its assessor lacked up-to-date knowledge of the subject facility. It is quite another matter to conclude that, because there exists reason to believe that respondent's depreciation figure is inaccurate, the tribunal should simply substitute some other figure equally lacking record support.

We recognize that the petitioner has the burden of proving true cash value, MCL 205.737(3); MSA 7.650(37)(3), and that, here, petitioner presented no evidence of an appropriate depreciation rate, relying solely on a market analysis to the valuation problem. In our opinion, however, all this constitutes is a concession that, if the respondent's cost analysis be adopted by the tribunal, petitioner does not claim error in the depreciation factor respondent's appraiser has found to be applicable.

In our opinion, in the relatively rare case in which the tribunal largely accepts the valuation approach advanced by the taxing unit, but also concludes that some portion of the unit's approach is not supported by competent evidence, resulting in a belief by the tribunal that the property should probably be valued in excess of the taxing unit's claim, the parties must be allowed to present further evidence on that portion of the case. To hold to the contrary denies the adversely affected party an opportunity to know and possibly rebut the evidence which the tribunal will rely on in reaching its decision. Compare, CAF Investment Co v State Tax Comm, 392 Mich. 442, 456-457; 221 N.W.2d 588 (1974).

Affirmed in part and remanded to the Tax Tribunal to allow the parties to present additional evidence on the issue of the proper depreciation factor to be applied to the subject property. We do not retain jurisdiction.