MEMORANDUM OPINION OF THE COURT
The issue in this workers’ compensation case involves the calculation of an employee's average weekly wage. Claimant was awarded compensation benefits upon a 50% occupational disability finding as a result of his contraction of coal workers’ pneumo-coniosis. The Administrative Law Judge (AU) calculated claimant’s average weekly wage pursuant to KRS 342.140(l)(d). The Workers’ Compensation Board (Board) reversed the AU on the wage calculation and held that KRS 342.140(l)(e) provided the proper method of calculation rather than subsection (l)(d). The Court of Appeals affirmed the decision of the Board.
KRS 342.140(l)(d) and (l)(e) govern the calculation of an employee’s average weekly wage when the employee’s wages are fixed by the day, hour, or by the output. The claimant here was paid on an hourly rate. If the employee has been employed for more than 13 weeks, the average weekly wage is calculated pursuant to (l)(d) by totalling the wages earned in either the first, second, third, or fourth quarter (13 week period), whichever quarter is more favorable to the employee, of the year immediately preceding the injury. That total *484is then divided by 13, and this sum is the average weekly wage.
If the employee has been employed for less than 13 weeks immediately preceding the injury, the average weekly wage is calculated pursuant to (l)(e). That subsection provides that the average weekly wage shall be computed to determine what would have been earned during the 13-week period immediately preceding the injury had the employee been employed by the employer the full 13 weeks and had worked, when work was available to other employees, in a similar occupation.
The following schedule illustrates claimant’s entire work history with C & D Bulldozing:
[[Image here]]
15 8/9/85 21 $147.00
14 8/16/85 39 $273.00
13 8/23/85 none $ 0.00
12 8/30/85 none $ 0.00
11 9/06/85 none $ 0.00
10 9/13/85 none $ 0.00
9 9/20/85 none $ 0.00
8 9/27/85 32 $224.00
7 10/04/85 32 $224.00
6 10/11/85 32 $224.00
5 10/18/85 39 1/2 $276.50
4 10/25/85 38 $266.00
3 11/01/85 none $ 0.00
2 11/08/85 33 $231.00
1 11/15/85 30 $210.00
The AU made one conclusive statement on this issue: “I find from the testimony that the Plaintiff was earning $127.35 per week while employed with C & D Bulldozing company by reviewing the highest quarter of payments for the last thirteen weeks.” Presumably, the AU believed that claimant had been in the employ of C & D for more than 13 weeks, and therefore calculated the average weekly wage per (l)(d) even though claimant had worked either 9 of 15 weeks, or 7 of 13 weeks. The AU’s calculation added the wages earned during the 13-week period preceding the injury and divided by 13 to arrive at an average weekly wage of $127.35.
We agree with the Board that (l)(d) was not the proper method of calculation and that (l)(e) should have been applied. However, the actual calculation by the Board per (l)(e) we believe was erroneous, and coincidentally in this case, the proper average weekly wage per (l)(e) does not differ from the AU’s result per (l)(d).
C & D Bulldozing Company has maintained throughout this appeal that since the time period between the original date of hiring and the date of injury exceeded 13 weeks, KRS 342.140(l)(d) was properly relied upon by the AU in calculating claimant’s average weekly wage. The Board reversed and held that KRS 342.140(l)(e) governed the wage calculation because, while claimant was first employed August 9, 1985, that employment was terminated August 23,1985. Claimant was then reemployed on September 27, 1985, through November 15, 1985.
This conclusion is supported by the testimony of the owner of C & D Bulldozing who stated that claimant was not employed during the weeks he received no wages. Moreover, the claimant explained that he worked for C & D when work was available and that for those weeks when he received no wages, he had been laid off due to the unavailability of work.
On appeal, C & D argues that the Board exceeded the proper scope of review *485and invaded the province of the fact finder by substituting its judgment for that of the AU and in conducting a de novo finding of fact. The AU did not state what testimony he relied upon in reaching a conclusion, nor was there included within the opinion and award any discussion of the evidence relative to his final determination. The AU made a conclusion regarding the applicable statutory calculation, without stating any factual findings. Furthermore, this conclusion was not consistent with claimant’s work record without some evidence to the contrary. The Board, of necessity, considered the testimony, which in fact did not support the AU’s conclusion, and from which it drew the only possible inference.
The Board also noted that there was no indication that during the 15-week period the claimant had any rights which he could assert against the employer, or had any connection, however tenuous, with the employer; that there was no indication that C & D had provided any benefits to its employees which continued for claimant when he received no wages, nor was there any indication that claimant had any priority with respect to reemployment; and that there was no indication that C & D continued any health and accident insurance benefits for claimant during the time he was off from work or that they were obligated to do so. C & D argues that because the Board noted a lack of evidence in support of C & D’s position, the Board, therefore, made its own findings of fact. C & D maintains that the absence of proof should not work to claimant’s favor. We believe, however, that the Board was recognizing that what little proof existed was contrary to the AU’s conclusion and that there was no other proof to support the AU’s conclusion. The Board’s opinion was well-reasoned and within the proper bounds of review.
C & D’s second argument is that the Board erred in establishing the criteria for determining whether the fact-finder should utilize KRS 342.140(l)(d) or (l)(e). C & D maintains that “all the fact-finder has to do is determine when the employee began his employment with the employer. If there are 13 weeks between that date and the date of the injury; then (l)(d) applies. If this period is less than 13 weeks, then (l)(e) applies.” C & D then complains that the Board created a variety of factors to consider in place of the simple statutory mandate, assuming the role of the General Assembly.
C & D has misstated the statutory scheme which does not turn upon the time period from an original date of hire to the date of injury, but turns upon the actual period of employment. The statutory language is clear that one should consider how long the employee “had been in the employ of the employer.” While this process would be very simple in the case of continuous employment, where the work is sporadic, a determination must be made on a case-by-case basis.
The work schedule submitted by the employer in this case clearly shows that claimant worked for C & D for at least nine weeks over a fifteen-week period. What remained unclear was the relationship between the parties during the other six weeks. The testimony of C & D’s owner and the claimant was that claimant was not an employee of C & D during those six weeks. The Board’s recitation of other factors, such as whether the claimant retained any rights he could assert against the employer during the dead periods, or whether any benefits were provided the employee by the employer during those periods, or whether the employee had priority with respect to reemployment, were simply examples of evidence having the potential to rebut the initial showing that claimant was in the employ of C & D for less than thirteen weeks.
Claimant was undoubtedly in the employ of C & D for at least nine weeks. The Board Opinion recognized that there was no evidence that claimant was in the employ of C & D for a longer period, and therefore, the AU’s conclusion was erroneous.
C & D also argues that even if KRS 342.140(l)(e) provides the proper method of calculation, the Board miscalculated claimant’s average weekly wage. The Board added claimant’s wages over the 15-week period totaling $2,075.00 and then divided this sum by 9 (the actual number of weeks *486worked) to arrive at an average weekly wage of $230.61. KRS 342.140(l)(e) provides:
342.140 Employe’s average weekly wage, how computed.
The average weekly wage of the injured employe at the time of the injury or last injurious exposure shall be determined as follows:
(1) If at the time of the ... last date of injurious exposure preceding death or disability from an occupational disease:
[[Image here]]
(e) The employe had been in the employ of the employer less than thirteen (13) calendar weeks immediately preceding the injury, his average weekly wage shall be computed under paragraph (d), taking the wages (not including overtime or premium pay) for such purpose to be the amount he would have earned had he been so employed by the employer the full thirteen (13) calendar weeks immediately preceding the injury and had worked, when work was available to other employes in a similar occupation.
Initially, we hold that the Board erred by considering weeks of employment outside the 13-week period immediately preceding the last injurious exposure. Calculations under either KRS 342.140(l)(d) or (l)(e) are made utilizing a 13-week period and the Board borrowed two weeks of employment from weeks 14 and 15 to reach a total of 9. The proper calculation would be based upon the wages earned for 7 weeks during the 13-week period preceding the injurious exposure under (l)(e).
Subsection (l)(e) is designed to arrive at an average weekly wage based upon an employee’s earning capacity, since under that subsection, the average weekly wage is determined not upon actual wages earned, but upon what would have been earned had the employee been employed by the employer for the full 13 calendar weeks immediately preceding the injury and had worked when work was available to other employees in a similar occupation. {See, Lexington Mining Company v. Richardson, 286 Ky. 418, 150 S.W.2d 889 (1941) for a discussion of the earning capacity approach.)
However, in this case there was no evidence that such work was available. Claimant testified that during the dead periods he simply did not work because work was unavailable. Consistently intermittent employment was the nature of the job, and claimant explained that it often took his employer several days to find other jobs.
The statutory scheme under (l)(e) was designed to compute an average weekly wage reflective of a realistic earning capacity, and therefore, (l)(e) includes the consideration of a normal 13-week period of hire so that an employee’s compensation will reflect his future loss of earnings in his regular employment. In claimant’s case, the evidence revealed that 7 weeks of work operating a bulldozer for C & D out of a 13-week period constituted a normal period of service. It is unfortunate that there is not a provision which is more narrowly tailored to accommodate consistently intermittent employment that is still not seasonal employment. However, the compensation scheme is based upon a determination of average weekly wages, and we must apply the statute as best we can to varying circumstances.
Claimant’s average weekly wage is $127.35 as calculated by dividing by 13 the total he earned during the 7 weeks he worked throughout the 13 calendar weeks immediately preceding his last injurious exposure. We believe this does in fact represent an average weekly wage based upon his future capacity to earn income under the normal and customary practices of hire in his employment with C & D in that irregular working time was a normal part of the job according to claimant’s testimony.
Therefore, the decision of the Court of Appeals is affirmed as to the application of KRS 342.140(l)(e) and is reversed as to the calculation pursuant to that subsection.
All concur.