Davis Memorial Goodwill Industries (Goodwill) petitions for review of a decision and order of the National Labor Relations Board (Board) finding that Goodwill violated section 8(a)(1) and (5) of the National Labor Relations Act (Act), 29 U.S.C. § 158(a)(1), (5), by refusing to bargain with the union that the Board had certified as the exclusive representative of a unit of Goodwill's employees. Goodwill contends that it was under no duty to bargain because the unit in question was comprised of workers who did not qualify as "employees" under the Act either because they were in a primarily rehabilitative relationship with Goodwill or because they were temporary. We grant Goodwill's petition for review and deny the Board's cross-petition for enforcement.

Goodwill is a nonprofit District of Columbia corporation that administers a work training program for handicapped individuals. Participants in the program receive preliminary training from Goodwill's Rehabilitation Division on matters such as punctuality, personal hygiene, simple legal problems, housing and personal finance. Participants are then referred to Goodwill's Contracts Division where they receive on the job training by working on government contracts awarded to Goodwill pursuant to the Javits-Wagner-O'Day Act (JWOD Act), 41 U.S.C. §§ 46-48.1 Since 1982 a number of handicapped individuals in the Goodwill work training program have performed custodial work under a series of JWOD Act contracts awarded to Goodwill by the Bureau of Engraving and Printing (BEP). Goodwill attempted to meet the obligations on its BEP contract with a workforce comprised entirely of handicapped workers but when this was not possible, Goodwill supplemented its workforce with non-handicapped workers.2

In August 1994 the International Association of Machinists and Aerospace Workers, AFL-CIO (Union) sought to represent a bargaining unit of handicapped and non-handicapped individuals performing custodial work under Goodwill's contract with BEP. Goodwill opposed the inclusion of the handicapped workers in the bargaining unit on the ground that as participants in a rehabilitative work training program they were not statutory "employees" under the Act.3 Goodwill also contended that the non-handicapped workers in the proposed bargaining unit were not statutory "employees" because their positions were temporary. The Board rejected Goodwill's position and directed that a representation election be held. Davis Mem'l Goodwill Indus., 318 N.L.R.B. 1044, 1995 WL 538981 (1995). The Union prevailed in the election by a vote of 44 to 11. Goodwill, however, refused to bargain with the Union. The Board found that Goodwill's refusal to bargain violated section 8(a)(1) and (5) of the Act, from which decision Goodwill filed the petition for review now before the court.4 Davis Mem'l Goodwill Indus., 320 N.L.R.B. No. 151, 1996 WL 188997 (Apr. 17, 1996).II.

In Goodwill Industries of Southern California, 231 N.L.R.B. 536, 537-38, 1977 WL 8988 (1977), the Board acknowledged that collective bargaining in the context of a rehabilitative work training program may not always effectuate the purposes of the Act. Collectively bargained terms of employment that would represent obvious gains for employees in another setting can work to the detriment of participants in a rehabilitative work training program. For example, collective bargaining might secure higher wages. Higher wages, however, can force the employer to employ more productive workers who often have less to gain from rehabilitative training. See id. at 537. Recognizing that in the rehabilitation setting the employer may very well safeguard employee interests more effectively than a union, the Board has adopted a case by case approach in which it asks whether a program has such rehabilitative elements that the participants do not qualify as statutory employees under the Act. The Board has summarized its approach as follows:

When the [employment] relationship is guided to a great extent by business considerations and may be characterized as a typically industrial relationship, statutory employee status has been found. When the relationship is primarily rehabilitative and working conditions are not typical of private sector working conditions, however, the Board has indicated it will not find statutory employee status.

Goodwill Indus. of Denver, 304 N.L.R.B. 764, 765, 1991 WL 525201 (1991).

In this case the Board concluded that Goodwill's handicapped workers at BEP were in a "typically industrial" rather than a "primarily rehabilitative" relationship on the basis of four factors. Specifically, the Board found that the workers experienced long terms of employment, were disciplined in the same manner as non-handicapped individuals, were subject to productivity standards and received only limited counseling services. The Board's factual findings will be affirmed if they are supported by substantial evidence in the record as a whole. See Gold Coast Restaurant, 995 F.2d at 263. Moreover, we give substantial deference to the Board's inferences drawn from the facts. See Peoples Gas Sys., Inc. v. NLRB, 629 F.2d 35, 42 (D.C.Cir.1980). We will not, however, "merely rubber stamp NLRB decisions." Avecor, Inc. v. NLRB, 931 F.2d 924, 928 (D.C.Cir.1991). Using these standards, we conclude the Board's findings are not supported by substantial evidence and we therefore grant the petition for review.

Because the ultimate purpose of a rehabilitative job training program is to place participants in jobs elsewhere, the Board considers long-term employment in a work training program itself as detracting from the rehabilitative character of the program. See, e.g., Lighthouse for the Blind of Houston, 244 N.L.R.B. 1144, 1147, 1979 WL 9621 (1979). Here, the Board found evidence of a typically industrial relationship because the handicapped workers at BEP "experience long periods of employment" with Goodwill. JA 312. We find that there is absolutely no evidence in the record to support this finding.

Handicapped workers at BEP were assigned positions either as janitors or as elevator operators. The record shows that in filling the elevator positions, which were less demanding, Goodwill gave preference to individuals who were physically unable to work as janitors. These individuals had little prospect of placement outside Goodwill because they were physically unable to perform janitorial work in the private sector and elevator operator positions, which they could perform, were scarce. JA 141. Based on the unlikelihood of placement and the fact that 27 of the 35 handicapped workers at BEP were elevator operators, the Board concluded "[t]his ... strongly suggests that the vast majority of the Employer's handicapped workers are retained for long periods of employment." JA 315. The Board's reasoning is flawed. Although Goodwill stated a preference for placing individuals unable to perform janitorial work in elevator operator positions, this tells us nothing about how many of the 27 elevator operators were actually incapable of performing janitorial work (and thus unlikely to be placed in the private sector). To answer that question, it is necessary to examine the record for evidence of the physical condition of the elevator operators rather than reasoning, as the Board did, solely from Goodwill's placement policy. Although the record indicates that one elevator operator suffered from a fused back and therefore supports the Board's conclusion of long-term employment in that instance, it does not manifest how many (if any) other elevator operators were similarly disabled.

The Board's conclusion is further undermined by the fact that Goodwill employs five full time placement counselors that were available to place handicapped workers at BEP in private sector employment. Based on its finding that the majority of handicapped workers at BEP was incapable of performing private sector work, the Board concluded that the counselors were "of limited service to the handicapped workers at the BEP location." JA 315. As discussed above, the Board's reasoning with respect to the number of handicapped workers capable of private sector employment was flawed because unsupported by the record. The Board's failure to give sufficient weight to the role of the five placement counselors was likewise flawed. Cf. NLRB v. Lighthouse for the Blind of Houston, 696 F.2d 399, 403 (5th Cir.1983) (absence of formal placement program is evidence of typically industrial relationship).

The Board also supported its typically industrial relationship conclusion with the finding that "handicapped workers are subject to the same disciplinary system as non- handicapped workers." JA 315. There is no dispute that handicapped workers and non-handicapped workers received the same employee handbook in which Goodwill stated work rules and personnel policies. However, as the Board itself has emphasized, the relevant question in determining whether a disciplinary system evidences a typically industrial or a primarily rehabilitative relationship is not what rules handicapped workers must follow but the penalties they face for breaking those rules. See Goodwill Indus. of Tidewater, 304 N.L.R.B. 767, 768, 1991 WL 187497 (1991).

The record does not support the Board's finding that handicapped and non-handicapped employees were subject to the same disciplinary sanctions. Handicapped workers with disciplinary problems could not be discharged by the Contracts Division but instead were referred to the Rehabilitation Division for counseling and further training. They then either went back to BEP or were placed in some other position within Goodwill. By contrast, non-handicapped workers who violated workplace rules could be disciplined (and ultimately terminated) by the Contracts Division itself.

Although handicapped workers at BEP were subject to productivity standards, we disagree with the Board's conclusion that the productivity standards evidenced a typically industrial relationship. Productivity standards were important in three respects. First, handicapped workers were paid a wage keyed to their productivity. For example, a handicapped worker who performed at 77% productivity made 77% of the Department of Labor mandated wage. JA 49-50. Second, handicapped workers had to be capable of working at 75% productivity before Goodwill referred them to the Contracts Division, JA 70, and, once referred, those workers who failed to maintain a level of 75% productivity were sent back to the Rehabilitation Division for further training. JA 71. Finally, once a handicapped worker surpassed 80% productivity, Goodwill attempted to place that worker in private sector employment through its Placement Department. JA 104. Merit pay increases have been considered evidence of a typically industrial setting. See Lighthouse for the Blind of Houston, 696 F.2d at 406. Here the variable wage structure may have encouraged handicapped workers to increase their productivity but any resemblance to a typically industrial relationship is outweighed by the fact that handicapped workers who exceeded 80% productivity were referred for placement in the private sector--a practice antithetical to what one would expect to find in a typically industrial relationship. The typical employer does not use a variable wage rate to increase productivity in order to find a different employer for the most productive of its workers. Moreover, the purpose of a rehabilitative job training program is to get participants to the level where they are productive enough to find employment in the private sector. To the extent the variable wage rate did increase productivity, this result is consistent with a primarily rehabilitative relationship. Also, the 80% referral standard was consistent with a primarily rehabilitative relationship because it allowed Goodwill to assess when a handicapped worker had reached a sufficient level of productivity to make private sector placement feasible.

With respect to the 75% productivity standard for obtaining and keeping a job at BEP, the Board suggests that the Contracts Division acted like a typically industrial employer who maintains productivity minima. The Board, however, relied on an artificial distinction between Goodwill's Rehabilitation Division and Contracts Division. Although the bargaining unit in this case was limited to Goodwill workers who, through Goodwill's Contracts Division, worked at BEP, the employer in this case is Goodwill--not its Contracts Division. In determining whether Goodwill's 75% productivity standard supports the Board's finding of a typically industrial relationship, therefore, we must look beyond how Goodwill treated its workers while they were in the Contracts Division. Here Goodwill's Rehabilitation Division first raises the employees to the 75% standard and then offers additional training if the employees should fall below the standard while in its Contracts Division. The situation, therefore, is different from that of a private sector employer who hires workers already able, it is anticipated, to satisfy productivity standards and then discharges those who fail to meet expectations. In this respect the Board failed to adequately distinguish Goodwill Industries of Denver, 304 N.L.R.B. 764, and Goodwill Industries of Tidewater, 304 N.L.R.B. 767, two cases in which the Board found handicapped workers to be in a primarily rehabilitative relationship. In both cases the Board relied on the fact that the handicapped workers were allowed to work at their own pace and were not subject to discipline. See Denver, 304 N.L.R.B. at 765; Tidewater, 304 N.L.R.B. at 768. Similarly, Goodwill's handicapped workers were allowed to work at their own pace, so long as they did not fall below 75% of the norm. Moreover, as noted, handicapped workers who did fall below that level were not disciplined but received further training.

Not surprisingly, an important ingredient of primarily rehabilitative employment relationships is the availability of counseling. See Goodwill Indus. of Denver, 304 N.L.R.B. at 765; Goodwill Indus. of Tidewater, 304 N.L.R.B. at 768. The Board found that Goodwill "provided only limited counseling services." JA 312. Again the record provides scant evidence to support the Board's finding.

A Goodwill counselor visited the BEP site weekly and counseling was available any time off-site at the handicapped worker's request. Furthermore, the Board failed to give weight to the fact that Goodwill counseled and trained workers extensively through its Rehabilitation Division before placing them at BEP. As discussed above in the context of productivity standards, we see no reason why it should matter whether the counseling was provided under the auspices of the Rehabilitation Division or the Contracts Division. Because the employer in this case is Goodwill, not its Contracts Division, the question we must ask is whether Goodwill provided counseling services. The record does indeed show that Goodwill counselors worked with handicapped individuals for periods ranging from several weeks to several years before referring them to the Contracts Division for on the job training. JA 133-34.

Temporary workers do not qualify as statutory employees. See Pen Mar Packaging Corp., 261 N.L.R.B. 874, 1982 WL 24502 (1982). An employee is considered temporary if "the prospect of termination was sufficiently finite on the eligibility date to dispel reasonable contemplation of continued employment beyond the term for which the employee was hired." St. Thomas--St. John Cable TV, 309 N.L.R.B. 712, 713, 1992 WL 363320 (1992).

We find that the Board erred in concluding that the non-handicapped employees working at BEP were not temporary. All non-handicapped workers were required to sign a form containing an express statement that they were being hired for temporary positions:

It is fully understood that if I'm non-handicapped and hired for one of Goodwill's janitorial contract sites, I'm strictly in a ninety-day temporary position, unless otherwise stated by my supervisor or granted an extension, and that I can be moved or terminated without notice if my slot in [sic] needed for a handicapped individual.

JA 256. The Board contends that this statement is inconclusive in light of a provision in Goodwill's employee handbook that states "staff and service employees are reviewed once a year." JA 229. The annual evaluation provision was applicable to Goodwill's non-handicapped workers because they fell within the definition of "service employees." Thus the provision appears to conflict with their acknowledged status as temporary employees hired for only ninety days. The "staff and service employees" language in the annual evaluation provision covers groups of Goodwill employees broader than the non-handicapped workers performing custodial work at BEP. For example, the provision covers employees such as counselors and administrative personnel who have long term positions with Goodwill. Considering the provision's broad sweep, we can fairly conclude that the non-handicapped workers were nonetheless temporary without rendering the evaluation provision meaningless. Finally, we note that the record is devoid of empirical evidence showing that the non-handicapped workers' time on the payroll in fact exceeded the ninety-day term stated in the form they signed. We therefore conclude that they were temporary workers and thus did not qualify as statutory employees.

* * *

The record in this case shows that Goodwill disciplined its handicapped workers in a manner significantly different from its discipline of non-handicapped workers, allowed its handicapped workers to work at their own pace, provided counseling and training and had as a goal finding permanent employment for its handicapped workers in the private sector. These are the same factors the Board has previously used to declare handicapped workers to be in a primarily rehabilitative relationship. See, e.g., Goodwill Indus. of Denver, 304 N.L.R.B. 764, 765; Goodwill Indus. of Tidewater, 304 N.L.R.B. 767, 768-69. The Board has had its chance to develop a record that would distinguish the instant case from its decisions in Denver and Tidewater and has failed. Because the Board has pointed to nothing to convince us that further development of the record could serve to distinguish Denver and Tidewater, we decline to remand the case to the Board for further findings. For the foregoing reasons, we grant the petition for review and deny the cross-petition for enforcement.

So ordered.