OPINION OF THE COURT
This workers’ compensation appeal concerns the average weekly wage of an individual who was employed for fewer than 13 weeks when injured and also concerns the July 14, 2000, amendments to KRS 342.730(l)(e)l and 2. Affirming decisions by an Administrative Law Judge (ALJ) and the Workers’ Compensation Board (Board), the Court of Appeals determined that the claimant’s wages from a previous 13-week period in a substantially similar employment were reasonably considered under KRS 342.140(l)(e). It also determined that the claimant’s award was properly enhanced under KRS 342.730(l)(c)l although he admitted that he presently earned more than when he was injured. This appeal by the employer followed.
During most of the claimant’s work life, he was a carpenter. Although he worked as a self-employed general contractor from the late 1970’s through the late 1990’s, he later began to work for others on various short-term jobs, sometimes as an uninsured independent contractor and sometimes as an employee. On December 1, 2000, he fractured his hip in a fall from the roof of a house that he was helping to frame, and he underwent right hip replacement surgery the following day. At that point, he had been working for Elwood Fawbush as a framing carpenter for only four weeks.
The claimant testified that he had been out of work for several weeks when his roommate told him that Fawbush was looking for help. Fawbush, a home builder, hired the claimant in mid-October, 2000, to help him frame a house. They also worked on three other houses, doing trim and repair work, and he indicated at the hearing that Fawbush spoke of building an addition as well as another home in the future. The claimant testified that he was the only full-time employee and that there would be an occasional helper for a day or two at a time. The work was fairly steady, but they did not work every day because Fawbush sometimes had other things to do. Agreeing with the defense, he indicated that work weeks of 30,19½, 35 ½, and 32 hours were “about right.” Faw-bush paid him $15.00 per hour, and he earned a total of $1,755.00 over a four-week period from November 6 through December 1, 2000. The claimant testified that Fawbush had expressed an interest in hiring him again after he recovered from his injury.
When deposed, the claimant submitted evidence of his self-employment earnings in June and November, 2000, but admitted that he was not covered by workers’ compensation insurance at the time. He also introduced documentary evidence of his earnings from KC Construction, which he described as “a very small company.” The claimant testified that he worked for KC from June, 2000, through September 29, 2000, framing houses for $15.00 per hour and working about forty hours per week. He stated that he worked as an employee and was covered by workers’ compensation insurance when doing so.
*8Responding to the ALJ’s request for information concerning the earnings of a similar employee, counsel for Fawbush alleged that there was no similar employee during the 13-week period before the injury but offered no testimony to that effect. Arguing that the claimant was not an employee and was not covered by insurance while working for KC, Fawbush maintained that wages from KC should not be considered when calculating his average weekly wage. Hale v. Bell Aluminum, Ky., 986 S.W.2d 152 (1998). Fawbush’s position was that the $1,755.00 the claimant earned during the four-week employment should be divided by 13 weeks, resulting in an average weekly wage of $135.00.
At the hearing, the claimant testified that he had returned to work for KC as a construction supervisor and earned $833.00 per week, an amount that exceeded his average weekly wage at the time of the injury. He pointed out, however, that he was not claiming to be totally disabled and that although the work was less strenuous than framing carpentry, it exceeded his medical restrictions. He explained that after his temporary total disability payments ceased, he had no income and relied on the generosity of friends at first. Eventually, he was forced to work because he needed an income. Although his present employer let him rest when needed and had a helper do the more strenuous work, he could only do the job when he took more than the prescribed amount of narcotic pain medication. Assured that the claimant was not alleging total disability, Fawbush agreed that his impairment was 8% and that he lacked the physical capacity to return to work as a framing carpenter. When the claim was submitted for a decision, the questions at issue concerned the proper method for calculating his average weekly wage under KRS 342.140(l)(e) and the proper method for calculating his partial disability benefit under KRS 342.730(l)(b) and (c).
Relying on the claimant’s deposition testimony and exhibit, the ALJ determined that the claimant was an employee when he worked for KC and, therefore, was a covered employee when he did so. After pointing out that the claimant did not work in an inconsistent pattern for Fawbush, the ALJ determined that the method Faw-bush proposed for calculating his average weekly wage was inappropriate. The ALJ noted that during the 18-week period immediately preceding the injury, the claimant worked two weeks for KC, earning a total of $1,200.00, and also worked four weeks for Fawbush, earning a total of $1,755.00. Whereas, during the previous 13-week period, from June 9 to September 1, 2000, the claimant had worked every week for KC and earned a total of $7,605.00. Noting that the goal of KRS 342.140(l)(e) was to obtain a realistic estimate of what the worker would expect to earn in a normal period of employment, the ALJ concluded that it would both permit and require a consideration of the claimant’s work for KC. See Huff v. Smith Trucking Co., Ky., 6 S.W.3d 819 (1999). Furthermore, relying on the most favorable 13-week period, during which time the claimant worked entirely for KC, the ALJ determined that his average weekly wage was $585.00.
Turning to the benefit calculation, the ALJ noted that the claimant had returned to work at a weekly wage that exceeded his wage at the time of the injury and also that he did not retain the physical capacity to return to framing carpentry. Furthermore, the ALJ observed that as amended effective July 14, 2000, KRS 342.730(l)(c)l and 2 are separated by the word “or,” indicating that they are mutually exclusive and that only one may be applied. Faced with having to determine which provision *9to apply, the ALJ concluded that because the claimant had an impairment from his injury that permanently altered his ability to labor and earn money and because paragraph (c)l appeared first in the statute, it should be applied. "Whereupon, the employer appealed on both issues.
AVERAGE WEEKLY WAGE
Fawbush points out that although the average weekly wage calculation is made under KRS 342.140(l)(d), the plain language of subsection (l)(e) limits the calculation to what the claimant would have earned while working for the same employer during the 13 weeks immediately preceding the injury. It maintains, therefore, that the claimant’s earnings with KC from June 9 to September 1, 2000, should not have been considered for two reasons: 1.) they were from a different employment, and 2.) they were not from the 13 weeks immediately preceding the injury. Noting that the claimant worked for a total of 6 weeks during the 13-week period from September 1 to December 1, 2000, Faw-bush asserts that there was no evidence in the record to support a finding that he would have worked full-time for the entire period. Therefore, even if the claimant’s earnings with KC were viewed as being some indication of what he would have earned if he had worked for Fawbush during the entire 13-week period that immediately preceded his injury, only the KC earnings from that period were relevant.
As applied to these facts, KRS 342.140 directs the average weekly wage to be determined as follows:
(1) If at the time of injury ...
[[Image here]]
(d) The wages were fixed by the day, hour, or by the output of the employee, the average weekly wage shall be the wage most favorable to the employee computed by dividing by thirteen (13) the wages (not including overtime or premium pay) of said employee earned in the employ of the employer in the first, second, third, or fourth period of thirteen (13) consecutive calendar weeks in the fifty-two (52) weeks immediately preceding the injury.
(e) The employee 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 that 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 employees in a similar occupation.
We are not persuaded that Hale v. Bell Aluminum, supra, governs these facts except to the extent that the ALJ was correct in refusing to consider the claimant’s concurrent self-employment income on the ground that he did not carry workers’ compensation insurance on himself. See Holman Enterprise Tobacco Warehouse v. Carter, Ky., 536 S.W.2d 461 (1976); Wright v. Fardo, Ky.App., 587 S.W.2d 269 (1979). In Hale v. Bell Aluminum, supra, the worker installed aluminum siding on a self-employed basis and also worked as an insured subcontractor for Bell Auminum, but the longstanding pattern of employment with Bell was sporadic. When appealing to this Court, the worker indicated that KRS 342.140(l)(e) provided the most appropriate method for calculating his average weekly wage Id. at 154. We concluded, therefore, that only the wages that he earned from Bell in the 13-week period immediately preceding the injury should be considered. See C & D Bulldozing v. Brock, Ky., 820 S.W.2d 482, 486 (1991).
*10In Huff v. Smith Trucking, supra, the timber cutting business in which the injured worker was employed had existed for fewer than 13 weeks. Furthermore, it ceased operating immediately after the injury, a situation that the ALJ attributed primarily to the injury. Thus, no evidence was available concerning the earnings of an individual who performed similar work for the same employer. Addressing the average weekly wage calculation, we pointed out that the goal of KRS 342.140(l)(e) is to obtain a realistic estimate of what the individual would have expected to earn in a normal period of employment by considering the unique facts and circumstances of the case. We determined, therefore, that it was appropriate to consider evidence that timber cutting work was available in the area, that such work paid $75.00 per day, and that during the relevant period, the weather permitted the work to be performed about 50% of the time.
More recently, in Affordable Aluminum, Inc. v. Coulter, Ky., 77 S.W.3d 587 (2002), the worker earned an average of $801.35 per week during the three weeks that he was employed by Affordable Aluminum (Affordable) as an insured subcontractor. Affordable offered no evidence concerning the regularity with which it had work for a similar employee in the 13 weeks preceding the injury or concerning what such an individual would have earned. Whereas, the worker’s unrebutted testimony indicated that a similar worker in that area would have expected to earn between $800 to $1000 per week, permitting an inference that the claimant would have been able to earn a similar amount had he been employed by Affordable. We concluded, therefore, that it was not unreasonable for the ALJ to arrive at an average weekly wage of $800.31.
The claimant had the burden to prove every element of his claim, including his average weekly wage. Calculating the average weekly wage is straightforward for workers who are paid by the week, month, or year. KRS 342.140(l)(a) — (c). When workers are paid by the day, hour, or output and, therefore, may not receive an income that is spread evenly throughout the year, KRS 342.140(l)(d) directs the ALJ to calculate the average weekly wage from the most favorable 13-week period in the year before the injury. In contrast, when an injury is sustained in an employment of less than 13 weeks’ duration, KRS 342.140(l)(e) bases the average weekly wage on what the injured worker would have been able to earn in the same employment had it existed for a full 13 weeks before the injury. It estimates the injured worker’s weekly earnings by averaging the earnings of an individual who worked for the same employer, performing similar work, during the 13 weeks immediately preceding the injury. On the present facts, the relevant period extended from September 1 to December 1, 2000.
Although using the 13-week averaging method of subsection (l)(d), KRS 342.140(l)(e) clearly focuses on the 13-week period immediately preceding the injury and on the employment in which the injury occurred, including the availability of work for those performing a similar job. Fawbush alleged that there was no similar employee during the relevant period but offered no evidence concerning what, if any, work the company had available during that time. The available evidence indicated that during the 13-week period immediately preceding the injury, the claimant worked for KC for the first two weeks, was unemployed for the following seven weeks, and then worked for Faw-bush for four weeks before being injured. It is apparent, therefore, that although he did not work every week during the period, there was no pattern of consistently *11intermittent employment with Fawbush. Although there was evidence concerning the claimant’s earnings during the four weeks .that he worked for Fawbush, no direct evidence was offered concerning the amount of work that Fawbush had available during the preceding nine weeks or what an individual such as the claimant would have been likely to earn. The ALJ’s dilemma, therefore, was how to reach a realistic estimate of that amount.
The claimant’s actual wages demonstrated his wage earning capacity and the availability of work with Fawbush during the final four weeks before his injury. Furthermore, to the extent that direct evidence was not available, Huff v. Smith Trucking and Affordable Aluminum v. Coulter would permit the ALJ to estimate what a framing carpenter who worked for Fawbush was likely to have earned by considering the availability of substantially similar work in the area during the relevant 13-week period. For that reason, the extent to which substantially similar work was available in the area during the 13 weeks immediately preceding the injury may be viewed as being some evidence of the likely availability of work with Faw-bush at that time. That is not to say, however, that the ALJ was authorized to determine the claimant’s average weekly wage with Fawbush by relying solely on his KC earnings from a previous 13-week period, entirely disregarding the probable availability of work with Fawbush during the 13 weeks that immediately preceded his injury.
Although the claimant stated that he was hired to work 40 hours per week, he was paid $15.00 per hour for the hours that he worked. The evidence established that he worked an average of 2954 hours per week with Fawbush, and there was no evidence that the average would have been any greater had he been employed by Fawbush for the entire 13-week period. Thus, there was no substantial evidence to support an inference that he would have worked sufficient hours to earn an average of $585.00 per week. It is apparent, therefore, that the finding concerning the claimant’s average weekly wage was erroneous as a matter of law and that the claim must be remanded for further consideration.
KRS 342.730(l)(c)
KRS 342.730(l)(b) sets forth a method for calculating the income benefit for permanent partial disability under which the benefit is the product of the worker’s average weekly wage, AMA impairment, and a statutory factor. As amended effective July 14, 2000, KRS 342.730(l)(c) provides, in pertinent part, as follows:
1. If, due to an injury, an employee does not retain the physical capacity to return to the type of work that the employee performed at the time of injury, the benefit for permanent partial disability shall be multiplied by three (3) times the amount otherwise determined under paragraph (b) of this subsection, but this provision shall not be construed so as to extend the duration of payments; or
2. If an employee returns to work at a weekly wage equal to or greater than the average weekly wage at the túne of injury, the weekly benefit for permanent partial disability shall be determined under paragraph (b) of this subsection for each week during which that employment is sustained. During any period of cessation of that employment, temporary or permanent, for any reason, with or without cause, payment of weekly benefits for permanent partial disability during the period of cessation shall be two (2) times the amount otherwise payable under paragraph (b) of this subsection. This provision shall not be con*12strued so as to extend the duration of payments.
KRS 342.730(l)(c)3 recognizes that “limited education and advancing age impact an employee’s post-injury earning capacity” and, in certain instances, enhances the multiplier to be applied under paragraph (c)l. Finally, KRS 342.730(l)(c)4 provides that, notwithstanding KRS 342.125, a claim may be reopened at any time during the period of partial disability to conform the award to KRS 342.730(l)(c)2.
Since December 12, 1996, KRS 342.730 has limited an ALJ’s discretion in determining the extent of permanent partial disability. The formula for calculating income benefits that was enacted at that time was weighted to favor more severely impaired workers who were more likely to have a greater occupational disability. Other considerations were the worker’s physical capacity to return to the pre-injury employment and post-injury earnings. Those who were the most severely impaired were entitled to benefits for a longer period of time. See Adkins v. R & S Body Co., Ky., 58 S.W.3d 428 (2001).
As amended in 2000, the formula for calculating a permanent partial disability benefit was further refined. The statutory factors in subsection (l)(b) were decreased. In subsection (l)(c), paragraphs (c)l and 2 were amended, and the word “or” was inserted between them. Furthermore, paragraph (c)3, which contains additional multipliers based upon age and education, was added. The Board’s opinion referred to decisions wherein it determined that the pre-2000 versions of KRS 342.730(l)(c)l and 2 could be applied concurrently where appropriate. Thus, the legislature presumably knew of those decisions when it inserted the word “or” at the end of paragraph (c)l and, by doing so, evinced an intent for only one of the provisions be applied to a particular claim. See Whitley County Board of Education v. Meadors, Ky., 444 S.W.2d 890 (1969).
Although the employer maintains that paragraph (c)2 modifies the application of paragraph (c)l and, therefore, takes precedence, we note that the legislature did not preface paragraph (c)2 with the word “however” or otherwise indicate that one provision takes precedence over the other. We conclude, therefore, that an ALJ is authorized to determine which provision is more appropriate on the facts. If the evidence indicates that a worker is unlikely to be able to continue earning a wage that equals or exceeds the wage at the time of injury for the indefinite future, the application of paragraph (c)l is appropriate.
Here, the ALJ based the decision to apply paragraph (c)l upon a finding of a permanent alteration in the claimant’s ability to earn money due to his injury. The claimant’s lack of the physical capacity to return to the type of work that he performed for Fawbush was undisputed. Furthermore, although he was able to earn more money than at the time of his injury, his unrebutted testimony indicated that the post-injury work was done out of necessity, was outside his medical restrictions, and was possible only when he took more narcotic pain medication than prescribed. It is apparent, therefore, that he was not likely to be able to maintain the employment indefinitely. Under those circumstances, we are convinced that the decision to apply paragraph (c)l was reasonable.
The decision of the Court of Appeals is affirmed in part and reversed in part, and the claim is remanded to the ALJ for further consideration of the claimant’s average weekly wage.
LAMBERT, C.J., and JOHNSTONE, KELLER, STUMBO, and WINTERSHEIMER, JJ., concur. *13COOPER, J., concurs in part and dissents in part by separate opinion in which GRAVES, J., joins.