BRUNETTI, Circuit Judge:
In this ease we are asked to decide whether the enactment of the Tenant Relocation Assistance Ordinance by the City of Seattle effected a taking of property in violation of the Fifth and Fourteenth Amendments.1
[804]*804I.
A.
During the 1980s, Seattle area low-income tenants were hurt by a booming redevelopment market that led to a sharp increase in rental prices and a corresponding decrease in the number of rental units affordable to low-income tenants. In 1990, as part of the Growth Management Act, the Washington State Legislature adopted legislation enabling municipal governments to enact relocation assistance provisions. RCW 59.18.440. In relevant part, the enabling legislation authorizes certain local jurisdictions:
to require, after reasonable notice to the public and a public hearing, property owners to provide their portion of reasonable relocation assistance to low-income tenants upon the demolition, substantial rehabilitation whether due to code enforcement or any other reason, or change of use of residential property, or upon the removal of use restrictions in an assisted-housing development....
RCW 59.18.440(1)
Through this legislation, the state hoped to encourage economic opportunity for all Washington citizens and to promote the availability of affordable housing as well as to preserve existing housing stocks. Washington Laws, 1990, 1st Ex.Sess., Ch. 17, § 2(4) and (5).
In July 1990, the City of Seattle took advantage of the State’s enabling legislation by promulgating the Tenant Relocation Assistance Ordinance (“TRAO”), which requires landlords to pay cash relocation assistance to low-income2 tenants they intend to displace by redeveloping their property. Specifically, the TRAO provides that:
Low-income tenants who are displaced by demolition, change of use, substantial rehabilitation, or removal of use restrictions and who comply with the requirements of [the TRAO], shall be paid a relocation assistance payment in the amount of two thousand dollars ($2,000.00)....
SMC 22.210.130(A).
Under the TRAO, the owner of the dwelling unit:
is responsible for payment of one-half (1/2) of the total amount of relocation assistance due to eligible tenants pursuant to [the TRAO]. The City is responsible for payment of the remaining one-half (1/2) of the relocation assistance.
SMC 22.210.110(A).
Within five days of receiving notice of tenant eligibility, the owner of a dwelling must provide the Director of the Department of Construction and Land Use with the owner’s portion of the relocation assistance to be paid to eligible tenants. SMC 22.210.110(B). After eligible tenants are paid the relocation assistance, any money remaining of the owner’s deposit is returned “thirty (30) days after final unappealed decisions regarding eligibility of all tenants of the affected units,.... ” SMC 22.210.130(F).
On June 7, 1990, the Seattle City Council held a public hearing on the TRAO. Approximately thirty citizens testified in favor of the ordinance, while no one testified against it. The parties have stipulated to the following summary of testimony given at the June 7, 1990 public hearing:
One of the purposes of the hearing was to receive testimony regarding the relocation expenses that displaced tenants might reasonably incur. At the June 7, 1990 public hearing, Karen White, an employee of the Department of Construction and Land Use (“DCLU”), the agency designated to administer the TRAO, testified regarding the results of an informal study she had conducted regarding the average costs of relocating for displaced tenants, using the cost categories provided in RCW 59.18.440. Ms. White called three or four local moving companies to determine the average cost of moving a household from a one-bedroom apartment to some other location in Seattle. She referred to a [805]*805monthly report published by a local realty company, Cain and Scott, to learn the average rent in Seattle, to use in determining first and last months’ rent costs. She also contacted a local real estate analyst who had recently completed a study of the typical amount required for damage and security deposits by local landlords. She also contacted the local utility companies to determine typical connection fees and deposits ____
Stip. Fact 4.
Karen White testified that the various moving costs, on average, totalled $2,191.00. Stip. Fact 4. That amount was based upon: $291 for physical moving costs, $1,000 for first and last month’s rent, $200 for damage deposit, $100 for utility connection fees and deposits, and $600 for one year’s increased rent of $50 per month. Stip. Fact 4.
B.
Owners of rental units subject to the TRAO3 brought suit in state court challenging the constitutionality of the TRAO and its enabling legislation. They asserted facial and as-applied takings challenges, and facial and as-applied substantive due process challenges. The City removed the case to federal district court. At the outset of discovery, the City served each plaintiff with a set of interrogatories and requests for production of documents. The City asked plaintiffs to produce information concerning the value of their properties and the projected impact the TRAO would have on the value of each property. Plaintiffs refused to respond to the requests, objecting that the documents were not relevant and that the requests were overly burdensome. In response to the City’s motion to compel discovery, plaintiffs brazenly confronted the district court by arguing: Plaintiffs admit they have not attempted to literally comply with [Requests for Production] 1 and 2, or any other discovery request. Rather, plaintiffs view discovery as a process to provide relevant information, not complying with the letter of discovery requests. If the information sought in the discovery requests can be conveyed without answering completely, plaintiffs will do so.
On November 21, 1994, the district court granted the City’s motion to compel discovery, ordering plaintiffs to produce the discovery forthwith. After plaintiffs ignored the court’s order, the City filed a motion for sanctions, seeking dismissal of plaintiffs’ as-applied constitutional claims. Hoping to avoid producing the requested discovery, plaintiffs agreed to dismiss their as-applied claims. Plaintiffs’ hopes that the discovery requests would finally cease were defeated when the court found that the City’s discovery requests were also relevant to plaintiffs’ facial takings claim. Thus, even though the as-applied claims were dismissed, the court ordered plaintiffs to comply with the City's discovery requests. When plaintiffs again flaunted the court’s order, the court imposed sanctions by prohibiting plaintiffs from introducing at trial or for any other purpose evidence “of the loss of value, the effect on the market value, or the economic impact of the TRAO.” The court acknowledged the potentially fatal consequences of its order. “[T]he inability to present evidence of impact on market value or economic impact may greatly hinder or make impossible plaintiffs’ ability to maintain a takings claim.”
Following this episode, the parties filed a “Stipulations of Fact,” concerning the City’s reasons for promulgating the TRAO, and the [806]*806effect the TRAO would have on -plaintiffs. As noted above, the parties stipulated to the testimony presented at the June 7, 1990 City Council meeting. In addition, the parties stipulated to other facts relevant to the constitutional claims raised on appeal, including:
6. The TRAO does not compel any of the plaintiffs to submit to the physical occupation of any of their property.
7. The TRAO does not deny any of the plaintiffs of all economically viable use of their property.
9. One of the purposes of the TRAO is to protect and financially assist residential tenants, especially low-income tenants, who were being displaced by demolition, change of use, or substantial rehabilitation of their rental units.
10. Protecting arid assisting residential tenants, especially low-income tenants, who are being displaced by private development is a legitimate public purpose.
11. One of the ways in which the TRAO protects and assists low-income tenants subject to displacement ... is to pay each of them $2,000 for the purpose of relocation assistance....
16. The TRAO relocation assistance payment may be effective in reducing the costs to owners of evictions and repairs.
17. The TRAO did not prevent the named plaintiffs from developing their property.
After the close of discovery, the City moved for summary judgment, requesting that the court determine the facial constitutionality of the State and municipal laws.4 The landowners cross-moved for summary judgment, requesting that the court declare the TRAO unconstitutional. The court granted the City’s motion for summary judgment. Rejecting the landowners’ facial takings claim, the court found that the TRAO is reasonably related to a legitimate state interest. Garneau, 897 F.Supp. at 1325 (citing Commercial Builders of Northern California v. City of Sacramento, 941 F.2d 872 (9th Cir.1991)). The court employed a nearly identical analysis in rejecting the landowners’ facial substantive due process claim. Id. at 1323-24 (citing Kawaoka v. City of Arroyo Grande, 17 F.3d 1227, 1234 (9th Cir.1994)). The landowners appeal the court’s denial of their facial takings claim.5 They also appeal the court’s discovery orders compelling them to produce financial documents and subsequently sanctioning them for failure to produce the documents.
II.
Plaintiffs argue that the “unconstitutional exactions” cases provide the appropriate standard for reviewing the TRAO. See Dolan v. City of Tigard 512 U.S. 374, 114 S.Ct. 2309, 129 L.Ed.2d 304 (1994); Nollan v. California Coastal Comm’n, 483 U.S. 825, 107 S.Ct. 3141, 97 L.Ed.2d 677 (1987). They argue that we must invalidate the TRAO if we determine that the $1,000 per low-income tenant exaction is not roughly proportional to the harm caused by their proposed development. For reasons discussed in the next section, we conclude that these cases provide no support for plaintiffs’ claim. Instead, we find the more traditional regulatory takings analysis of Agins v. City of Tiburon, 447 U.S. 255, 100 S.Ct. 2138, 65 L.Ed.2d 106 (1980) applicable. Applying Agins, we conclude that plaintiffs have failed to provide evidence relevant to their claim, and we affirm the district court’s summary judgment in favor of the City.
Government regulation of private property violates the Takings Clause6 if it fails to [807]*807substantially advance legitimate state interests. Lucas v. South Carolina Coastal Council, 505 U.S. 1003, 1016, 112 S.Ct. 2886, 120 L.Ed.2d 798 (1992); Agins, 447 U.S. at 260, 100 S.Ct. 2138. Governmental regulation categorically violates the Takings Clause if it results in the physical invasion of property, or if it denies the owner all economically viable use of his property. Lucas, 505 U.S. at 1015, 112 S.Ct. 2886. Generally in these cases, no matter how minute the physical invasion, and no matter how important the public purpose behind the governmental action, just compensation will be required. Id. at 1015-16, 112 S.Ct. 2886.
By contrast, in non-categorical regulatory takings cases, the court must engage in an ad hoc, factual inquiry to determine whether the government regulation goes too far. Id. at 1015, 112 S.Ct. 2886; see also Pennsylvania Coal Co. v. Mahon, 260 U.S. 393, 415, 43 S.Ct. 158, 67 L.Ed. 322 (1922) (“The general rule at least is, that while property may be regulated to a certain extent, if regulation goes too far it will be recognized as a taking.”). Although this inquiry may focus on several factors, the Court has repeatedly recognized two as “keenly relevant.” Id. at 1019 n. 8, 43 S.Ct. 158. First, is “the economic impact of the regulation on the claimant”; second, is “the extent to which the regulation has interfered with distinct investment-backed expectations.” Id. (citing Penn Central Transportation Co. v. New York City, 438 U.S. 104, 124, 98 S.Ct. 2646, 57 L.Ed.2d 631 (1978)); see also Hodel v. Irving, 481 U.S. 704, 715, 107 S.Ct. 2076, 95 L.Ed.2d 668 (1987) (finding that Native Americans did not have reasonable investment-backed expectations in devise and descent of property); MacDonald, Sommer & Frates v. County of Yolo, 477 U.S. 340, 349, 106 S.Ct. 2561, 91 L.Ed.2d 285 (1986) (factors to be considered are economic impact of regulation, its interference with reasonable investment-backed expectations, and the character of the governmental actions); Williamson County Regional Planning Comm’n v. Hamilton Bank of Johnson City, 473 U.S. 172, 191, 105 S.Ct. 3108, 87 L.Ed.2d 126 (1985) (same);' Ruckelshaus v. Monsanto Co., 467 U.S. 986, 1005, 104 S.Ct. 2862, 81 L.Ed.2d 815 (1984) (same); Kaiser Aetna v. United States, 444 U.S. 164, 175, 100 S.Ct. 383, 62 L.Ed.2d 332 (1979) (same).
In facial takings claims, the inquiry is further limited to whether “mere enactment” of the regulation has gone too far. See Suitum v. Tahoe Regional Planning Agency, - U.S. -, -n. 10, 117 S.Ct. 1659, 1666 n. 10, 137 L.Ed.2d 980 (1997). Indeed, the Court recognized that facial takings challenges “face an uphill battle since it is difficult to demonstrate that mere enactment of a piece of legislation deprived the owner of economically viable use of his property.” Id. (internal citations and punctuation omitted). In a facial challenge, a court will look only to the regulation’s “general scope and dominant features ... leaving other [specific] provisions to be dealt with as cases arise directly involving them.” Village of Euclid v. Ambler Realty Co., 272 U.S. 365, 397, 47 S.Ct. 114, 71 L.Ed. 303 (1926).
The record firmly supports the district court’s grant of summary judgment in favor of the City on plaintiffs’ facial takings claim. Plaintiffs bear the burden of proving their facial takings claim at trial. See Suitum, 117 S.Ct. at 1666 n. 10; Lucas, 505 U.S. at 1016 n. 6, 112 S.Ct. 2886. Summary judgment is therefore appropriate if the City has pointed out that there is an absence of evidence to support plaintiffs’ claim. See Celotex Corp. v. Catrett, 477 U.S. 317, 325, 106 S.Ct. 2548, 91 L.Ed.2d 265 (1986).
Plaintiffs have stipulated that the TRAO neither physically invades their property, nor denies them all economically viable use of their property. Accordingly, we must engage in the ad hoc, factual inquiry appropriate in non-categorical regulatory takings cases. See Lucas, 505 U.S. at 1015, 112 S.Ct. 2886. As noted above, plaintiffs must show that the TRAO “goes too far” in regulating their property by introducing evidence of the economic impact of the enactment of the TRAO on their property. Thus, plaintiffs must show that the value of their property diminished as a consequence of the TRAO. Further, plaintiffs must show that the diminution in value is so severe that the TRAO [808]*808has essentially appropriated their property for public use.
Not only have plaintiffs failed to show the type of “extreme circumstances” necessary to sustain a regulatory takings claim, see United States v. Riverside Bayview Homes, Inc., 474 U.S. 121, 126, 106 S.Ct. 455, 88 L.Ed.2d 419 (1985), plaintiffs have refused to produce any evidence of economic impact. Plaintiffs refused to produce evidence regarding the value of their property before enactment of the TRAO. They have refused to produce evidence regarding the value of their property after enactment of the TRAO. They have refused to produce evidence regarding the anticipated economic consequences of the TRAO. Plaintiffs have not generally alleged that the TRAO makes it' commercially impracticable for them to continue operating their apartment buildings. See Keystone Bituminous Coal Ass’n v. DeBenedictis, 480 U.S. 470, 495-96, 107 S.Ct. 1232, 94 L.Ed.2d 472 (1987). Indeed, not a single member of the plaintiff class has pointed to a single apartment building that can no longer be operated for profit. Id. at 496, 107 S.Ct. 1232. Furthermore, as a consequence for their refusals, the district court prohibited plaintiffs from introducing any of this evidence “at trial or ... for any purpose whatsoever.”
There is very little in the record in this case from which we may determine the economic impact of the TRAO on plaintiffs’ property. What little evidence there is relates only to an as-applied, not a facial claim. For example, on the value-loss side of the ledger, plaintiffs have asserted that, collectively, members of the plaintiff class have been required to pay almost $80,000 in relocation assistance under the TRAO. Although they have provided no documentation in support of these claims, plaintiffs assert that the Garneaus paid $6,000, Klepinger paid $2,000, Fedan paid $1,000, Ju paid $8,187.96, and Triad paid $59,850. In their complaint, plaintiffs also agree that the City refunded $11,000 to Triad. Thus, in order to get their development permits, plaintiff class members suffered approximately $70,000 in out of pocket expenses.
On the other side of the ledger, plaintiffs have stipulated that the TRAO “may be effective in reducing the costs to owners of evictions and repairs.” Stip. Fact 16. This is so because the TRAO includes a valuable quid pro quo. - In exchange for paying money to the displaced tenants, developers are protected from liability and litigation costs they might incur under Seattle’s Just Cause Eviction Ordinance. See SMC 22.206.160(C). Prior to enactment of the TRAO, Seattle tenants could only be evicted for “just or good cause.” This protection is not waivable by the landlord or the tenant. By seeking a permit to develop their property under the TRAO, landlords can avoid litigation over the displacement of low-income tenants. Although plaintiffs have stipulated that the TRAO may benefit them by reducing the cost of evictions, they offer no evidence indicating the amount of that benefit. None of the evidence on either side of the ledger goes very far in determining the economic impact the TRAO’s enactment had on plaintiffs’ property.
Plaintiffs have not met their burden of providing evidence that the enactment of the TRAO effected a taking or harmed them at all. We find the absence of any evidence of the economic impact of the TRAO dispositive. On the evidence presented to the district court, no rational jury could have found in favor of plaintiffs on their facial takings claim. Accordingly, we affirm the judgment of the district court.
III.
In rejecting plaintiffs’ facial takings claim, we conclude that the Supreme Court’s “unconstitutional exactions” cases provide no support for plaintiffs’ claim. See Dolan v. City of Tigard, 512 U.S. 374, 114 S.Ct. 2309, 129 L.Ed.2d 304 (1994); Nollan v. California Coastal Comm’n, 483 U.S. 825, 107 S.Ct. 3141, 97 L.Ed.2d 677 (1987). Plaintiffs argue that evidence of the TRAO’s economic impact is irrelevant under Nollan and Dolan. They argue that where the government demands an exaction in exchange for a permit, all that is relevant is whether the exaction is “roughly proportional” to the harms caused by the proposed development: We first review the [809]*809Court’s analytical framework in the unconstitutional exactions cases, and then conclude that the analysis does not support plaintiffs’ facial takings claim in this case.
In Nollan and Dolan, the government used its regulatory power over land use to force an exaction from a permit applicant. In Nollan, the California Coastal Commission forced a couple seeking a building permit to give the public an easement over their beachfront property. 483 U.S. at 831, 107 S.Ct. 3141. In Dolan, the City of Tigard forced a store owner to dedicate a strip of his land for public use. 512 U.S. at 384, 114 S.Ct. 2309. In both cases the Supreme Court found the government’s use of its police power illegitimate. Stripped of its legitimacy, government imposition of the exactions were found to constitute takings in violation of the Fifth and Fourteenth Amendments.
Read together, Nollan and Dolan establish a three-part test. First the court asks whether government imposition of the exaction would constitute a taking. Second is the “essential nexus” test, which asks whether the government has a legitimate purpose in demanding the exaction. Third is the “rough proportionality” test, which asks whether the exaction demanded is roughly proportional to the government’s legitimate interests.
1.
The first inquiry ignores the government’s land use power, and asks only whether government imposition of the exaction would be a taking. The exaction is the concession sought by the government, or the condition upon which granting the permit depends. The government may seek land, money, or other concessions in return for the permit. Governments use exactions to mitigate the harms associated with the proposed development. In Nollan, the Court addressed this first inquiry by explaining:
Had California simply required the Nol-lans to make an easement across their beachfront available to the public on a permanent basis in order to increase public access to the beach, rather than conditioning their permit to rebuild their house on their agreeing to do so, we have no doubt there would have been a taking.
483 U.S. at 831, 107 S.Ct. 3141. The Nollans wanted permission to rebuild their bungalow, and the government demanded a public easement across their beach in return. This first inquiry focuses only upon the government’s demand for a public easement. The question thus becomes, “Does the government’s demand for a public easement effect a ‘taking’?” Because the answer was “yes,” the Nollan Court proceeded to the second inquiry. If the answer had been “no,” the inquiry would end.
The Court began with this step in Dolan as well. 512 U.S. at 384, 114 S.Ct. 2309. Dolan wanted to expand her electric supply store located on Main Street, while the government wanted to turn a strip of her land into a greenway. As it had done in Nollan, the Court in Dolan explained that, “[w]ithout question, had the city simply required petitioner to dedicate a strip of land along Fanno Creek for public use, rather than conditioning the grant of her permit to redevelop her property on such a dedication, a taking would have occurred.” Id.
The Court had no trouble in either case finding that government imposition of the exaction would amount to a taking. In both cases the government demanded permanent physical occupation of some portion of the applicant’s land. Courts have generally found that where government action leads to the physical invasion of private property, it constitutes a per se taking. See Loretto v. Teleprompter Manhattan CATV Corp., 458 U.S. 419, 433, 102 S.Ct. 3164, 73 L.Ed.2d 868 (1982); Levald v. City of Palm Desert, 998 F.2d 680, 684 (9th Cir.1993). Similarly, if the government action denies the owner of all economically viable use of his property, it is also a per se taking. See Lucas, 505 U.S. at 1015, 112 S.Ct. 2886. By contrast, in non-categorical takings cases, courts must undertake “complex factual assessments of the purposes and economic effects of government actions.” Yee v. City of Escondido, 503 U.S. 519, 523, 112 S.Ct. 1522, 118 L.Ed.2d 153 (1992). Because of the difference in the Court’s approach, “much turns on the classi[810]*810fication of the government’s action.” Levald, Inc. v. City of Palm Desert, 998 F.2d 680, 685 (9th Cir.1993).
2.
The second and third inquiries seek to determine whether the government may shield itself from a takings claim through the use of its police powers. In Nollan, the Court framed the second question by asking: “Given, then, that requiring uncompensated conveyance of the easement outright would violate the Fourteenth Amendment, the question becomes whether requiring it to be conveyed as a condition for issuing a land-use permit alters the outcome.” 483 U.S. at 834, 107 S.Ct. 3141. This “essential nexus test” compares the government’s purpose in seeking the concession with its legitimate land use interests. If the government’s purposes are not connected, then the government’s demand for the exaction is not a legitimate exercise of its police power, but “an out-and-out plan of extortion.” Id. at 837, 107 S.Ct. 3141.
Where the government may legitimately use its power to block development, however, “it would be strange to conclude that providing the owner an alternative to that prohibition which accomplishes the same purpose is not a taking.” Id. at 836-37, 107 S.Ct. 3141. In Nollan, the government’s stated purpose for demanding the beachfront easement was to protect views of the beach from the street. The Court held that giving the public the right to walk along the Nollan’s beachfront property was completely unrelated to the public views of the beach from the street. Since the government’s demand was unrelated to its legitimate interest, the Court found that there was no essential nexus.
The Court reached the opposite conclusion in Dolan. The City of Tigard’s interest in the dedicated strip of land was to provide more porous surface in the city to help flood control. The Court found that the exaction was directly related to the government’s concern because the strip of land could be used by the City for flood control purposes. Thus, in Dolan, the government passed the essential nexus test.
3.
The third inquiry is simply a refinement of the second. In Nollan the Court explained that the exaction must be related to the burdens imposed by the development. Left open, however, was “how close a ‘fit’ between the condition and the burden is required.” Id. at 838, 107 S.Ct. 3141. The Court answered this question by announcing the “rough proportionality” test. Dolan, 512 U.S. at 391, 114 S.Ct. 2309. In determining this federal constitutional standard, the Court looked to the experience of the state courts.
We think the ‘reasonable relationship’ test adopted by a majority of the state courts is closer to the federal constitutional norm than either of those previously discussed. But we do not adopt it as such, partly because the term ‘reasonable relationship’ seems confusingly similar to the term ‘rational basis’ which describes the minimal level of scrutiny under the Equal Protection Clause of the Fourteenth Amendment. No precise mathematical calculation is required, but the city must make some sort of individualized determination that the required dedication is related both in nature and extent to the impact of the proposed development.
Id.
The Court found that the City of Tigard’s exaction was not roughly proportional to its flood control interests because the City could have accomplished the same goal with a less invasive measure. For example, the City could have required Dolan to maintain part of her property as a greenway, but kept it private for her customers. “The difference to [Dolan], of course, is the loss of her ability to exclude others. As we have noted, this right to exclude others is ‘one of the most essential sticks in the bundle of rights that are commonly characterized as property.’” Id. at 393, 114 S.Ct. 2309 (citing Kaiser Aetna, 444 U.S. at 176, 100 S.Ct. 383). The rough proportionality test ensures that the “price” of the government permit is not significantly higher than the social harm caused by the proposed development.
[811]*811In footnote eight of its Dolan opinion, the Court added one more twist to this test. See id. at 391 n. 8, 114 S.Ct. 2309. The Court recognized that “in most generally applicable zoning regulations, the burden properly rests on the party challenging the regulation to prove that it constitutes an arbitrary regulation of property rights.” Id. The Court explained that the burden shifts to the government, however, where the government makes an “adjudicative decision to condition [an] application for a building permit on an individual parcel.” Id. The rationale for this burden shifting appears to rest on the Court’s concern that where the government demands individual parcels of land through adjudicative, rather than legislative* decision making, there is a heightened risk of extortionate behavior by the government See Nollan, 483 U.S. at 837, 107 S.Ct. 3141; see also Ehrlich v. City of Culver City, 12 Cal.4th 854, 50 Cal.Rptr.2d 242, 257, 911 P.2d 429, 444, cert. denied, - U.S. -, 117 S.Ct. 299, 136 L.Ed.2d 218 (1996); id., 50 Cal.Rptr.2d at 272, 911 P.2d at 272 at 459 (Mosk, J., concurring)-(noting that at ad hoc decision making-aimed at individuals increases risk of extortionate government behavior).
The scope of Dolan’s rough proportionality test in takings cases is in considerable doubt. For example, the Supreme Court has left unsettled the-question whether Dolan’s rough proportionality test applies to legislative, as opposed to administrative exactions.7 Lower courts and commentators have also struggled to determine whether the test applies to fee exactions, as opposed to physical exactions.8 In this case, however, we find Dolan inapplicable for two different reasons. First, we conclude that Dolan applies only to as-applied takings challenges, not to facial takings challenges. Second, Dolan does not address when a taking has occurred, instead, it addresses only how close a fit the exaction (which would otherwise constitute a taking) must have to the harms caused by development. Thus, for the important question in this ease whether the TRAO effects a, taking, Dolan simply. does not address the issue.
The Dolan analysis cannot be applied in facial takings claims. “A facial challenge involves a claim that-the mere enactment of a statute constitutes a taking, while an as-applied challenge involves- a claim that the particular impact of a government action on a specific piece of property requires the payment of just compensation.” Carson Harbor Village, Ltd. v. City of Carson, 37 F.3d 468, 473-74 (9th Cir.1994). In Dolan, the Court analyzed whether the government’s exaction was too costly. However, in a facial claim, the court does not analyze the exaction at all. Rather, the court looks to whether the enactment of the ordinance, under which exactions will be imposed, effects a taking. An example of a successful facial claim is Lucas v. South Carolina Coastal Council, 505 U.S. 1003, 112 S.Ct. 2886, 120 L.Ed.2d 798 (1992), where government regulation prohibited all beneficial use of Lucas’ beachfront property, for which he -had paid $975,000. The Court found that mere enactment of this use restriction rendered Lucas’ property “valueless.”
If this were an as-applied challenge, we would determine the TRAO’s effect on each parcel of land. For example in this ease, Triad claims to have paid $50,000 in relocation assistance under the TRAO. At the other end* Fedan paid only $1,000 for a permit. Each as-applied - regulatory ' takings claim must be evaluated independently to determine whether the total exaction is roughly proportional to the harm caused by each development. Because in a facial claim we do not analyze the exactions, Dolan’s test for when the exaction costs too much does not apply.
[812]*8122.
A second reason why Nollan and Dolan provide no support for plaintiffs’ takings claim, is that they do not address when a taking occurs. The first step in the unconstitutional exactions cases is to determine whether government imposition of the exaction would be a taking. Because Nollan and Dolan both involved physical invasions of private property, the Court found the exac-tions were per se takings. Assuming these cases apply outside the context of physical invasions, a plaintiff still must show in the first step that government imposition of the exaction would constitute a taking. If plaintiffs’ claim were an as-applied challenge, each plaintiff would be required to show that the exaction applied to them constituted a taking of their property. If the court determined that imposition of a $1,000 per tenant fee constituted a taking, it would then ask whether requiring the fee as a condition for issuing a land-use permit alters the outcome. See Nollan, 483 U.S. at 834, 107 S.Ct. 3141. Neither Nollan nor Dolan provide a court with any guidance to determine whether imposition of a $1,000 per tenant fee constitutes a taking. It is this first step in the analysis that plaintiffs have entirely ignored in litigating this case.
IV.
We now turn to the district court’s discovery rulings. Specifically, plaintiffs challenge two of the district court’s orders. First, plaintiffs challenge the court’s order compelling them to produce evidence relating to the economic impact of the TRAO. Second, plaintiffs challenge the court’s dismissal of their as-applied takings claims as a sanction for their failure to produce discovery. We review the district court’s discovery rulings for an abuse of discretion. See Sopcak v. Northern Mountain Helicopter Serv., 52 F.3d 817, 818 (9th Cir.1995).
On appeal, plaintiffs argue that the district court abused its discretion because the information sought by the City was not relevant to any of its claims.9 Relevance for purposes of discovery is defined very broadly. See Hickman v. Taylor, 329 U.S. 495, 506-507, 67 S.Ct. 385, 91 L.Ed. 451 (1947) (“Information is relevant to the subject matter if it might reasonably assist a party in evaluating the case, preparing for trial or facilitating settlement.”). The district court determined that the propounded discovery was relevant to plaintiffs’ facial and as-applied takings and substantive due process claims. The relevance of the economic impact of a government regulation challenged under the Takings Clause has been recognized since Pennsylvania Coal Co. v. Mahon, 260 U.S. 393, 43 S.Ct. 158, 67 L.Ed. 322 (1922) (Holmes, J.). See supra Section II. The district court did not abuse its discretion by ordering plaintiffs to produce evidence regarding the economic impact of the TRAO.
Plaintiffs also appeal the district court’s dismissal of their as-applied claims as a sanction for them violation of the court’s order. The district court found that plaintiffs agreed to the dismissal of their as-applied claim as a sanction for their refusal to produce discovery. In a letter to the district court, plaintiffs’ counsel explained their unwillingness to comply with the court’s order:
Plaintiffs will not comply with the court’s order. Plaintiffs urge the court to enter an appealable order so that this dispute can be heard by the 9th Circuit.
Plaintiffs believe it would be futile to comply with the court’s order regarding discovery. As long as the court believes the disputed discovery sought by the City regarding the economic impact of the TRAO on these particular plaintiffs is relevant to the TRAO’s constitutionality, there is little chance the court will find it results in a taking or a violation of substantive due process.
The legal issues l'egarding what is relevant in determining the constitutionality of the TRAO will ultimately be decided by [813]*813the appellate courts. It is best to get to that forum sooner rather than later. Plaintiffs wish the court would rule in their favor on these issues. The next best alternative is an order dismissing plaintiffs’ case as a sanction, which plaintiffs can appeal to the 9th Circuit.
This letter shows that plaintiffs willfully refused to follow the district court’s order. It further shows that so long as the court continued to find the requested discovery relevant, plaintiffs preferred to have their as-applied claims dismissed. We conclude that the district court did not abuse its discretion in finding that plaintiffs agreed to dismissal of their as-applied claims as a sanction.
CONCLUSION
We affirm the district court’s grant of summary judgment thereby upholding the validity of the TRAO. We have been forced to uphold Seattle’s relocation assistance ordinance in large part because of the way plaintiffs have chosen to litigate this case. We do not uphold the ordinance because we find it a wise solution to a difficult problem. Instead, we uphold it because the economic impact of the ordinance is relevant to plaintiffs’ takings claim, and plaintiffs have steadfastly refused to produce any evidence of the TRAO’s impact. Plaintiffs have litigated this case with a belief that Dolan may be used to strike down any attempt by the City to make them pay for problems they do not believe they created. As we have explained, more is necessary.
AFFIRMED.