Liskowitz v. Astrue

559 F.3d 736 (2008)

Lora LISKOWITZ, Plaintiff-Appellant,
v.
Michael J. ASTRUE, Commissioner of Social Security, Defendant-Appellee.

No. 08-1576.

United States Court of Appeals, Seventh Circuit.

Argued October 22, 2008. Decided March 24, 2009.

Barry A. Schultz, Attorney (argued), Law Offices of Barry A. Schultz, P.C., Evanston, IL, for Plaintiff-Appellant.

Suzanne E. Lohmeyer Duman, Attorney (argued), Social Security Administration, Office of the General Counsel, Region V, Chicago, IL, for Defendant-Appellee.

Before CUDAHY, MANION, and WILLIAMS, Circuit Judges.

CUDAHY, Circuit Judge.

Lora Liskowitz applied for disability insurance benefits more than eight years ago. An administrative law judge initially denied her claim, was reversed by the district court and subsequently denied her claim following a second administrative hearing. The district court affirmed the ALJ's second decision. Although this second decision is less clear than it might have been, the decision was supported by substantial evidence. We affirm.

I.

Lora Liskowitz was born with a congenitally deformed left hip. As a consequence of her hip condition, she underwent multiple surgical procedures as a child, including a procedure when she was twelve to equalize the length of her legs by removing "growth material" from her right tibia and femur plates. She reports that she has experienced pain in her knee, hips and back ever since. In spite of the pain, she was able to work for fifteen years in her parents' waterbed factory, splitting her time between the upholstery shop and the factory office, where she performed basic clerical tasks. However, she stopped working in 1998 because, by her own account, her pain grew progressively worse, and she became incapable of performing even sedentary clerical tasks.

Liskowitz testified that since she stopped working, she has been bedridden for all but a few hours each day and that she can remain seated without discomfort only for ten minutes at a time. Her assessment of her own condition is at least partly corroborated by her rheumatologist, Doctor Joseph Bretza. Liskowitz began seeing Bretza in 2003, after she initially tested positive for rheumatoid arthritis. In 2004, Bretza completed a questionnaire in which he indicated that Liskowitz can remain seated only for an hour at a time, that she can use her hands only for twenty percent of an eight-hour work-day and that she has suffered from these limitations since 1998.

The record shows, however, that prior to 2003, Liskowitz was more functionally capable than she now admits. For instance, in 2000 her examining physician noted, contrary to Bretza's retrospective assessment, that Liskowitz had no significant upper extremity limitations. Between 2001 and 2003, multiple healthcare providers noted that she was fully able to perform household and child care duties, and was not otherwise limited in activities of daily living. And in 2003, a healthcare provider — apparently a nurse practitioner — noted that Liskowitz had lost weight since giving birth to her third child, and that her exercise regime included "some walking."

Liskowitz's own statements also belie her claim that she has been incapable of sedentary activity since 1998. In 2000, she told her doctors at the Milwaukee Medical Clinic that she does "a lot of squatting, kneeling and lifting off the floor of her young children," who were five and seven months at the time.^[1] She expanded on this claim at her initial administrative hearing in 2001, where she admitted that she alone was responsible for the care of her children, and testified that she walked or drove her eldest daughter to school, changed diapers, made lunch and occasionally dinner, did some vacuuming and dusting, washed dishes, did laundry and shopped for groceries. She also testified that she was able to control her pain by taking Celebrex and Vicodin. (As she stated at the time, "[t]he Celebrex is awesome.")

Two state agency physicians concluded that Liskowitz was capable of standing for two hours and sitting for six hours in an eight-hour workday. The ALJ concurred, finding that she was capable of sedentary work and that her testimony to the contrary was not credible. Liskowitz appealed the ALJ's decision, and in 2004 the district court remanded the case for a new administrative hearing based on problems with the testimony of an expert witness. However, the district court affirmed the ALJ's findings regarding Liskowitz's credibility. (Liskowitz does not challenge, or indeed even mention, this aspect of the district court's 2004 decision on appeal.)

Following the remand, Liskowitz appears to have changed her theory of the case. In the second hearing, Liskowitz emphasized swelling and pain in her hands, which she did not even mention in the first hearing, as evidence of her disability. Based in part on her previous factual findings, the ALJ refused to fully credit either Liskowitz's own testimony or the corroborating testimony of her rheumatologist. Instead, the ALJ found that Liskowitz remained capable of sedentary work through December 2002, when her insured status expired. The ALJ credited the testimony from an expert witness who identified 4,000 unskilled jobs in the Milwaukee area that a person with Liskowitz's background and limitations would have been capable of performing. The Appeals Council denied review.

II.

The ALJ denied benefits initially in 2001, and again in 2005. Only the 2005 decision is at issue here. Because the Appeals Council declined to review the ALJ's second ruling, this ruling constitutes the Agency's final decision. Getch v. Astrue, 539 F.3d 473, 480 (7th Cir.2008). We review this decision directly without giving deference to the district court's decision. Elder v. Astrue, 529 F.3d 408, 413 (7th Cir.2008). We will uphold the ALJ's decision if it is supported by "substantial evidence," see 42 U.S.C. § 405(g), which means "such relevant evidence as a reasonable mind might accept as adequate to support a conclusion." Richardson v. Perales, 402 U.S. 389, 401, 91 S. Ct. 1420, 28 L. Ed. 2d 842 (1971).

To qualify for disability benefits, a claimant must be "disabled," 42 U.S.C. § 423(a)(1)(E), which the Social Security Act defines as an "inability to engage in any substantial gainful activity by reason of any medically determinable physical or mental impairment which can be expected to result in death or which has lasted or can be expected to last for a continuous period of not less than 12 months." Id. at § 423(d)(1)(A). Further, a claimant must show that the disability arose while he or she was insured for benefits. See 42 U.S.C. §§ 423(a)(1)(A), (c)(1); Sienkiewicz v. Barnhart, 409 F.3d 798, 802 (7th Cir. 2005) (per curiam).^[2]

Social Security regulations prescribe a five-step test for determining whether a claimant is disabled within the meaning of the Act. See 20 C.F.R. §§ 404.1520, 416.920. Only the fifth step — which requires the ALJ to determine the claimant's residual functional capacity and to ascertain whether there are a significant number of jobs that the claimant could perform — is at issue here.

A.

In considering whether a claimant is capable of work, an ALJ must determine the claimant's "residual functional capacity," which is the work he or she can still do despite her physical and mental limitations. 20 C.F.R. § 404.1545(a); Hickman v. Apfel, 187 F.3d 683, 688-89 (7th Cir. 1999). In the present case, because the ALJ found that Liskowitz was capable of sedentary work in April 2001 and Liskowitz has not challenged this finding on this appeal, our review is limited. The only question that remains is whether there was substantial evidence for the ALJ's determination that Liskowitz's condition did not become dramatically worse in the roughly one and one-half years between the ALJ's first decision and expiration of Liskowitz's insured status.

Liskowitz argues that it was error for the ALJ to refuse to credit her own testimony regarding her functional limitations. We are not persuaded. First, and most obviously, the ALJ had found Liskowitz to be a non-credible witness in the first hearing, and was entitled to rely on this finding in the second hearing. See Berger v. Astrue, 516 F.3d 539, 546 (7th Cir.2008) (an ALJ is entitled to view the testimony of an applicant who has been deceptive with skepticism). This is not to say that having found Liskowitz to have exaggerated her symptoms once, the ALJ could automatically disregard her testimony in the second hearing. But this is not what happened here. Instead, the ALJ found that Liskowitz's testimony at the second hearing was inconsistent both with the ALJ's prior findings and with Liskowitz's own testimony during the first hearing. Having already found the witness to have exaggerated her symptoms, and having reasonably observed that the witness's account of those symptoms materially changed from the first hearing to the second, there was nothing improper about the ALJ's adherence to her previous credibility determination.

But even if it were somehow improper for the ALJ to adhere to her previous credibility assessment, this would still not justify a second remand. A second problem with Liskowitz's argument is that her testimony at the second hearing did not speak to the one narrow point regarding her functional limitations that was still at issue in the second hearing. Again, the ALJ had previously found that Liskowitz was capable of sedentary work prior to 2001. Liskowitz appealed this finding to the district court and lost. She has not pursued her challenge to the ALJ's 2001 findings in this appeal. Thus, to be entitled to disability benefits, Liskowitz would have had to show that her condition became dramatically worse between 2001, when the ALJ rendered her initial decision, and 2002, when Liskowitz's insured status expired. However, Liskowitz had practically nothing to say during the second hearing about how her condition had changed. On the contrary, although she stated that she has "gradually declined" over the years, she also testified that her functional limitations have remained the same since she initially applied for benefits.^[3] Thus, even if the ALJ were somehow required to credit Liskowitz's testimony, this testimony did not speak to a rapid deterioration in her condition between 2001 and 2002.

Liskowitz's treatment records, including her rheumatologist Dr. Bretza's opinion, also evidence no dramatic deterioration of her condition. Even without more, this undermines Liskowitz's somewhat more plausible argument that the ALJ impermissibly "played doctor" by refusing to credit the opinion of her treating physician. "[A]n ALJ cannot play the role of doctor and interpret medical evidence." Murphy v. Astrue, 496 F.3d 630, 634 (7th Cir.2007). Along the same lines, "an ALJ cannot disregard medical evidence simply because it is at odds with the ALJ's own unqualified opinion." Id. Here, there is at least a strong argument to be made that the ALJ improperly substituted her own, non-professional opinion for that of Liskowitz's treating physician. What the ALJ said was:

There is no basis for relating back [Dr. Bretza's] findings from 2003 to the period from the alleged onset date in 1998 to December 2002, the date last insured. Dr. Bretz's [sic.] assessment conflicts with the claimant's lack of treatment for a condition which could be rheumatoid arthritis in the year 2002. It also conflicts with claimant's infrequent treatment for hip dysplasia during the period from 1998 to 2002. It is noted that Dr. Bretz [sic] found in March 2003 that the claimant had a full range of motion in all joints with no synovitis.

We are troubled by this aspect of the ALJ's decision. It is quintessentially a matter for medical judgment whether disabling rheumatoid arthritis is consistent with "a full range of motion" or "joint synovitis." Perhaps the ALJ is right that disabling rheumatoid arthritis would result in significant joint swelling. See Stedman's Medical Dictionary 1773 (27th ed.2000) (defining "synovitis" as inflamation, especially that of a joint). But we do not know this; and the ALJ does not know either.

Standing alone, the ALJ's remarks give us pause. Were this the ALJ's first decision on the matter, there would be a strong case for reversal. However, when we consider the above-quoted remarks together with the ALJ's findings following the first administrative hearing, we conclude that there was substantial evidence for the ALJ's decision. Again, the ALJ made these remarks only after she had already found that Liskowitz was capable of certain forms of sedentary work. Liskowitz appealed this aspect of the ALJ's first decision, and lost.^[4] Because the factual findings that furnished the basis for the ALJ's first decision to deny benefits were affirmed on appeal, the ALJ was entitled to rely on these findings in the second hearing.

Indeed, even if the ALJ had elected to reconsider the issue of Liskowitz's functional limitations prior to 2001, Bretza's retrospective functional assessment does not strike us as particularly strong evidence, especially in the light of the contrary assessments of Liskowitz's contemporaneous examining physicians. "A retrospective diagnosis may be considered only if it is corroborated by evidence contemporaneous with the eligible period." Estok v. Apfel, 152 F.3d 636, 640 (7th Cir.1998); see also Evangelista v. Sec. of Health & Human Servs., 826 F.2d 136, 140 (1st Cir.1987); Adams v. Chater, 93 F.3d 712, 714 (10th Cir.1996); Perez v. Chater, 77 F.3d 41, 48 (2d Cir.1996); Jones v. Chater, 65 F.3d 102, 103-04 (8th Cir.1995). In the present case, Bretza's retrospective opinion was that since 1998, Liskowitz has been able to use her hands for one-fifth of an eight-hour workday. This opinion was inconsistent, not only with the opinions of the two state agency doctors, but also with the opinion of Liskowitz's examining physician, who noted in 2000 that Liskowitz suffered from no significant upper extremity limitations.

Faced with competing opinions, the ALJ had to decide which opinion to credit. We cannot say, as a matter of law, that the ALJ made the wrong choice here. See Donahue v. Barnhart, 279 F.3d 441, 444 (7th Cir.2002) ("[T]he resolution of competing arguments based on the record is for the ALJ, not the court."). There may be situations where it would be appropriate for an ALJ to repudiate previous factual findings in the light of new medical evidence. However, it was not error for the ALJ to refuse to credit a treating physician's opinion four years after the fact where this opinion was inconsistent with contemporaneous medical evidence that the ALJ had previously and properly credited.

This leaves Dr. Bretza's retrospective opinion on rather shaky footing. Like Liskowitz herself, Bretza did not say that Liskowitz's functional limitations became more pronounced between 2001 and 2002. (Indeed, even this claim would not have been supported by the record. During this period, Liskowitz reported to her doctors that she was experiencing decreased pain through her knee and hips, and that she was capable of performing all household and child care duties.) Because Bretza had nothing to say on the one novel question concerning Liskowitz's functional limitations that was properly before the ALJ, the ALJ's findings concerning Liskowitz's residual functional capacity were supported by substantial evidence.

B.

We are also unpersuaded by Liskowitz's claim that the Commissioner did not properly show that there were a significant number of jobs that she was capable of performing. An individual is disabled only if he or she "cannot, considering his age, education, and work experience, engage in any other kind of substantial gainful work which exists in the national economy ... in significant numbers either in the region where such individual lives or in several regions of the country." 42 U.S.C. § 423(d)(2)(A).^[5] The Commissioner bears the burden of showing that there are a significant number of jobs that the claimant is capable of performing. See 20 C.F.R. § 404.1560(c)(2); Britton v. Astrue, 521 F.3d 799, 803 (7th Cir.2008) (per curiam). The Commissioner typically uses a vocational expert ("VE") to assess whether there are a significant number of jobs in the national economy that the claimant can do. Lee v. Sullivan, 988 F.2d 789, 793 (7th Cir.1993).^[6]

In the present case, the VE testified that there were approximately 4,000 jobs in the Milwaukee area that a person with Liskowitz's functional limitations would be capable of performing. Liskowitz does not argue that 4,000 jobs is insignificant; nor would such an argument be plausible. As few as 174 jobs has been held to be significant, see Allen v. Bowen, 816 F.2d 600, 602 (11th Cir.1987), and it appears to be well-established that 1,000 jobs is a significant number. See Lee, 988 F.2d at 794; see also Hall v. Bowen, 837 F.2d 272, 275 (6th Cir.1988) (1,350 jobs); Barker v. Sec. of Health & Human Servs., 882 F.2d 1474, 1479 (9th Cir.1989) (1,266 jobs); Trimiar v. Sullivan, 966 F.2d 1326, 1330-32 (10th Cir.1992) (850-1,000 jobs); Jenkins v. Bowen, 861 F.2d 1083, 1087 (8th Cir.1988) (500 jobs).

Although the VE indisputably identified a significant number of jobs, Liskowitz argues that it was error for the ALJ to credit the VE's testimony for two reasons: first, the VE was not able to testify as to the reliability of the data she used to reach her conclusions; second, the VE was unable to identify the number of part-time jobs that were included in her data set.

As to this first argument, it is not entirely true that the VE failed to vindicate the reliability of the data on which she relied. The VE initially admitted that she could not assess the degree of accuracy of the data sources on which she was relying. However, on follow-up questioning, she added that these sources were "widely recognized as acceptable sources in the vocational rehabilitation area." Perhaps ideally the VE would have been able to say a bit more, but this does not go without saying. The witness was testifying as a vocational expert, not as a census taker or statistician. Indeed, even if the VE had happened to know something about the statistical basis for her testimony, she arguably still would not be in a position to fully vindicate her conclusions. After all, statisticians use arithmetic operations, but few probably have studied the foundations of arithmetic in set theory. Is the statistician's use of arithmetic therefore unjustified? Clearly not. In administrative proceedings, no less than in ordinary life, "explanations come to an end somewhere." LUDWIG WITTGENSTEIN, § 1 PHILOSOPHICAL INVESTIGATIONS (G.E.M. Anscombe trans., 1968).

In addition to testifying that her sources were widely recognized as acceptable, the VE actually identified her sources. Two of these sources were published by the United States Department of Labor and the Wisconsin Department of Workforce Development. These are government sources of which the ALJ was required to take administrative notice. See 20 C.F.R. § 404.1566(d). Although the record does not include information concerning the nature of the VE's third source, the Occupational Employment Quarterly (OEQ), we note that this does indeed seem to be a source on which VEs customarily rely. See, e.g., Britton, 521 F.3d at 802. Liskowitz argues for the first time in her reply brief that the VE should not have relied on the OEQ because it was published by a private company.^[7] But she forfeited this argument by failing to object to the VE's testimony during the hearing. See Barrett v. Barnhart, 355 F.3d 1065, 1067 (7th Cir.2004); Donahue, 279 F.3d at 447 ("Raising a discrepancy only after the hearing ... is too late."). The VE had the OEQ with her while she testified. Had Liskowitz actually objected to the VE's testimony, the VE could have said more about the kind of information the OEQ contains. At the very least, the VE could have identified which of her conclusions were based on the OEQ. As it stands, however, the VE's testimony was both unobjected to and uncontradicted. Thus, the ALJ was entitled to credit this testimony.

Finally, we are not persuaded by Liskowitz's argument that the VE was required to identify the number of part-time jobs included in the approximately 4,000 jobs she claimed Liskowitz was capable of performing. Liskowitz's argument is based on Social Security Ruling 96-8p, which provides that

[o]rdinarily, RFC [residual functional capacity] is an assessment of an individual's ability to do sustained work-related physical and mental activities in a work setting on a regular and continuing basis. A "regular and continuing basis" means 8 hours a day, for 5 days a week, or an equivalent work schedule.

In the present case, the VE testified that she had "no way of knowing" if the jobs she had identified were full-time or part-time. According to Liskowitz, this admission renders the VE's testimony unreliable.

As a threshold matter, it is far from clear, and neither party addresses, the level of deference to which Ruling 96-8p is entitled. Social Security Rulings "represent precedent [sic] final ... interpretations that we have adopted." 20 C.F.R. § 402.35(b)(1). We generally defer to an agency's interpretations of the legal regime it is charged with administering. See Lauer v. Bowen, 818 F.2d 636, 639 (7th Cir.1987) (per curiam); United Fire Ins. Co. v. Commissioner of Internal Revenue, 768 F.2d 164, 169 (7th Cir.1985). However, we are not invariably bound by an agency's policy statements.^[8] Since neither party has briefed the issue of the appropriate level of deference to apply to Social Security Rulings, we assume without deciding that violations of Ruling 96-8p constitute reversable error. What follows?

On its face, Ruling 96-8p applies to the ALJ's functional capacity determination, not to the VE's testimony. Again, having determined that a claimant has a severe impairment, the ALJ must, inter alia, assess her residual functional capacity and then determine whether there are a significant number of jobs in the national economy that she can perform. 20 C.F.R. § 404.1520. Ruling 96-8p provides that "capacity" means sustainable capacity. To take an example: the fact that a person can run down the block does not mean that she has the functional capacity to be a professional runner.

Ruling 96-8p does not say, nor do we interpret it to imply, that a VE may permissibly testify only as to the availability of full-time jobs. On the contrary, to say that the ALJ may deny benefits only if she finds the claimant capable of some form of full-time work is quite different from saying that only full-time jobs can constitute significant work in the national economy. To return to our previous example, a person who is functionally capable of running professionally should not be deemed disabled simply because some of the jobs that are available for professional runners are part-time jobs.

Our conclusion is not at odds with the Eleventh Circuit's decision in Kelley v. Apfel, 185 F.3d 1211 (11th Cir.1999) (per curiam). In Kelley, on the Commissioner's motion, the Eleventh Circuit clarified its basis for affirming the Commissioner's denial of disability benefits, stating in dicta that "if the government is correct in its interpretation [expressed in Ruling 96-8p], a claimant could pass Step Five and be entitled to benefits even though capable of working on a part-time basis." Id. at 1214-15. However, the Eleventh Circuit did not say that a VE may testify only as to the existence of full-time jobs. Indeed, the court explicitly observed that at step one of the Commissioner's sequential analysis, "there is no per se rule that part-time work cannot constitute substantial gainful activity." Id. at 1214 (citing 20 C.F.R. § 404.1572(a)).

Liskowitz's interpretation of Ruling 96-8p also has significant practical problems. Once again, the VE testified that she had "no way of knowing" how many of the jobs that she had identified were part-time jobs. In the colloquy that immediately preceded this remark, the VE made it clear that the reason she had no way of knowing was that this information was not contained in the data sources on which she based her testimony. Indeed, Liskowitz's counsel conceded at oral argument that no government data source contains this information. Surely, this is a sign that Liskowitz expects too much.

We decline Liskowitz's invitation to impose impossible burdens on the VE. We hold instead that a VE may, consistent with Ruling 96-8p, testify as to the numbers of jobs that a claimant can perform without specifically identifying the percentage of those jobs that are part-time. The claimant, of course, may respond to the VE's testimony by offering evidence of her own that the jobs the VE identified do not constitute "substantial gainful work" within the meaning of Section 423(d)(2)(A). There may even be circumstances in which a claimant can accomplish this by showing that a substantial percentage of the jobs that the VE has identified are part-time jobs. However, Liskowitz made no effort to rebut the VE's testimony in this case. Where, as here, the VE identifies a significant number of jobs the claimant is capable of performing and this testimony is uncontradicted (and is otherwise proper), it is not error for the ALJ to rely on the VE's testimony.

III.

It would have been better if the ALJ gave a better-reasoned basis for rejecting Liskowitz's treating physician's opinion. Be that as it may, the ALJ's decision was supported by substantial evidence.

AFFIRMED.

NOTES

[1] In 2002, Liskowitz gave birth to a third child.

[2] A claimant who cannot establish that she was disabled while she was insured may still receive Supplemental Security Income benefits if she can established that she is disabled and has limited means. 42 U.S.C. §§ 1381a, 1382; Sienkiewicz, 409 F.3d at 802.

[3] Her testimony was as follows: "Question: Now, I had asked you a mess of questions about limitations and your condition. And I just wanted to see if I'm clear about this. Is this basically essentially the way your condition has been since June 1 of 1998? Answer: Yes."

[4] Liskowitz has not pursued any further challenge to the ALJ's first set of factual findings on this appeal.

[5] Section 423(d)(2)(A) overturned the Kerner doctrine. See Kerner v. Flemming, 283 F.2d 916 (2d Cir.1960) (Friendly, J.). See Lance Liebman, The Definition of Disability in Social Security and Supplemental Security Income: Drawing the Bounds of Social Welfare Estates, 89 Harv. L.Rev. 833, 853 (1976). Under Kerner, once a claimant showed that she was unfit for her former work, the burden shifted to the Secretary to prove that the claimant had a realistic possibility of obtaining work that was near her home. 283 F.2d at 921. In response to the Federal courts' widespread adoption of the Kerner doctrine, Congress amended the Social Security Act in 1967 to provide that an impairment could not be considered a total disability unless it rendered the claimant unable to perform any kind of substantial gainful work, "regardless of whether such work exists in the immediate area in which he lives, or whether a specific job vacancy exists for him, or whether he would be hired if he applied for work." Social Security Amendments of 1967, Pub.L. No. 90-248, § 158(d)(2)(A).

[6] In the past, we have expressed confusion over how a person becomes a "vocational expert." See Donahue, 279 F.3d at 446. We note, therefore, that Liskowitz's counsel stated in oral argument that vocational experts are typically job placement specialists. In the present case, the VE for the first hearing had worked as a consultant who recruited and evaluated job applicants for corporate clients. There is no record of the professional background of the VE for the second hearing.

[7] She cites to the publisher's website as authority for the proposition that the OEQ is not a government source. Information gleaned from a company website, of course, is not part of the record on appeal. Further, it is not clear that the website helps Liskowitz's cause, as it states that "[a]ll data provided by [the Publisher] is derived from government sources." See http://www.uspublishing.net/ references.html (visited 2/6/09).

[8] In Prince v. Sullivan, 933 F.2d 598 (7th Cir. 1991), we suggested that Social Security Rulings are entitled to Skidmore deference. Id. 602 (quoting Skidmore v. Swift & Co., 323 U.S. 134, 140, 65 S. Ct. 161, 89 L. Ed. 124 (1944)). This is partially true, but incomplete. Where a policy statement or ruling interprets an Agency's authorizing statute, this will typically be entitled to Skidmore deference. See Christensen v. Harris County, 529 U.S. 576, 587, 120 S. Ct. 1655, 146 L. Ed. 2d 621 (2000). On the other hand, where, as here, the Agency's policy statement interprets an Agency's own regulations, the appropriate deference regime is established by Seminole Rock. See Auer v. Robbins, 519 U.S. 452, 461, 117 S. Ct. 905, 137 L. Ed. 2d 79 (1997) (reaffirming Bowles v. Seminole Rock & Sand Co., 325 U.S. 410, 65 S. Ct. 1215, 89 L. Ed. 1700 (1945)).