OPINION
{1} This case requires us to examine the scope and manner of exercise of a court’s inherent power to sanction litigants for spo*710bating evidence before initiating litigation. Plaintiffs appeal the dismissal of their suit, which they filed against Defendants for damages caused by a fire at a restaurant in Carlsbad, New Mexico. Before filing suit, Plaintiffs allowed a fire-suppression system to be destroyed during renovation of the restaurant. The fire-suppression system allegedly failed, abowing the fire to become more destructive than it should have been. All Defendants maintain they needed access to it to prepare their respective defenses. After learning of the loss of the fire-suppression system, Defendants all moved either for summary judgment or dismissal. The district court granted each of the motions because Plaintiffs had permitted the destruction of evidence that should have been preserved. We reverse and remand for reconsideration in bght of this opinion.
FACTS
{2} The underlying facts are largely undisputed. In the late afternoon of February 1, 1995, employees on duty at Restaurant Management Company’s (RMC’s) Carlsbad, New Mexico, Long John Silver’s restaurant heard popping noises, smelled burning electrical insulation, and saw smoke in the area of one of the restaurant’s fryers. On closer inspection, the employees noticed the fryer’s heating cobs glowing bright red. They turned the fryer off, but the ob in it caught fire nonetheless. The employees attempted unsuccessfully to control the fire; it soon spread to other parts of the restaurant, causing considerable damage.
{3} Prior to the fire, RMC hired Hobart Corporation (Hobart) to examine the fryer because it had not been heating properly. The Hobart technicians who examined the fryer recommended that RMC have new heating elements installed. RMC agreed and purchased the elements from Hobart. Two days before the fire occurred, RMC hired Vance Galloway, doing business as Quality Heating, Air Conditioning and Refrigeration of Carlsbad (Quality), to install the new heating elements. Also prior to the fire, Fire Safety Industries, Inc. (Fire Safety), had for some time periodically inspected and maintained the fire-suppression system in the restaurant. Kidde-Fenwal, Inc. (Kidde-Fenwal), manufactured the fire-suppression system that allegedly failed to activate during the fire and that Plaintiffs allowed to be destroyed.
{4} RMC had fire insurance on the building with St. Paul Fire and Marine Insurance Company (St.Paul). On February 6, 1995, several days after the fire, Raymond Marion, an independent adjuster working for St. Paul, examined the restaurant. Based on his investigation he concluded that St. Paul might have subrogation claims against Hobart, Quabty, and Kidde-Fenwal. Two days later, on February 8, 1995, St. Paul claims adjuster Juan Azcarate hired Richard Skinner, an independent fire investigator, to examine the restaurant and attempt to determine the cause of the fire. Skinner concluded that a problem in the fryer caused it to overheat, which in turn caused the oil to catch fire. He recommended that St. Paul remove and save the fryer. He also concluded that the fire-suppression system located in the hood above the fryer had failed to activate during the fire, and he recommended that St. Paul hire an engineer to examine both the fryer and the fire-suppression system.
{5} St. Paul hired engineer Lewis Poe to investigate the fire further. Poe concluded that “[t]he probable cause of the fire was an electrical short in the fryer which over heated [sic] the fat to its ignition point.” Poe also found that the fire-suppression system “failed to operate in a timely manner,” suggesting that it was designed improperly. Poe submitted his report to St. Paul on May 15, 1995, but by that time the fire-suppression system had already been destroyed during the razing of the restaurant preparatory to reconstruction.
{6} Plaintiffs filed their suit on October 21, 1996. It included claims for breach of implied warranty and negligence against all four Defendants, and for strict liability against Defendants Kidde-Fenwal and Hobart. Upon learning of the destruction of the fire-suppression system, Defendants each moved either for summary judgment or dismissal of the complaint. The district court granted each of the motions. The district court did not enter any findings of fact or *711conclusions of law in support of its decision. It did send the parties a letter ruling in which it stated: “In my opinion the defense Motions To Dismiss and for Summary Judgment must be granted because of the destruction of necessary evidence which should have been preserved.”
{7} Subsequent to the filing of the appeal, Kidde-Fenwal settled its dispute with Plaintiffs, and it has been dismissed as a party.
STANDARD OF REVIEW
{8} The parties agree that the motions forming the basis of this appeal were directed to the court’s inherent power. In addition, to the extent the court explained its ruling, we can infer that it intended to exercise its inherent power. We therefore analyze the district court’s ruling in the context of case law discussing the inherent power of courts. Cf. Baliotis v. McNeil, 870 F.Supp. 1285, 1289 (M.D.Pa.1994) (invoking court’s own inherent power because, “[although raised in the context of a summary judgment motion, the arguments of [the defendants] address[ed the e]ourt’s discretion to impose appropriate sanctions for spoliation of evidence”). We review the district court’s exercise of its inherent power to dismiss Plaintiffs claims for abuse of discretion. See Chambers v. NASCO, Inc., 501 U.S. 32, 55, 111 S.Ct. 2123, 115 L.Ed.2d 27 (1991); Gonzales v. Surgidev Corp., 120 N.M. 151, 157, 899 P.2d 594, 600 (1995). Our review is deferential. See Marrocco v. General Motors Corp., 966 F.2d 220, 223 (7th Cir.1992) (‘We cannot understate the difficulty of the task litigants face when challenging a district court’s choice of sanctions.”).
DISCUSSION
{9} The California Court of Appeal has defined spoliation broadly as “the destruction or significant alteration of evidence, or the failure to preserve property for another’s use as evidence, in pending or future litigation.” Willard v. Caterpillar, Inc., 40 Cal.App.4th 892, 48 Cal.Rptr.2d 607, 616 (1995), overruled in part on other grounds by Cedars-Sinai Med. Ctr. v. Superior Court, 18 Cal.4th 1, 74 Cal.Rptr.2d 248, 954 P.2d 511, 521 n. 4 (1998). “Destruction of potentially relevant evidence obviously occurs along a continuum of fault — ranging from innocence through the degrees of negligence to intentionality. The resulting penalties vary correspondingly.” Welsh v. United States, 844 F.2d 1239, 1246 (6th Cir.1988).
{10} The vehicles for imposing penalties likewise vary. For example, a majority of states, including New Mexico, see NMSA 1978, § 30-22-5 (1963), have laws criminalizing the destruction of evidence. See Scott S. Katz & Anne Marie Muscaro, Spoilage of Evidence—Crimes, Sanctions, Inferences, and Torts, 29 Tort & Ins. L.J. 51, 53 & n. 17 (1993). In addition, a handful of jurisdictions, also including New Mexico, have provided for civil redress against third parties by adopting the tort of intentional spoliation of evidence. See Coleman v. Eddy Potash, Inc., 120 N.M. 645, 647-49, 905 P.2d 185, 187-89 (1995); see also Iain D. Johnston, Federal Courts’ Authority to Impose Sanctions for Prelitigation or Preorder Spoliation of Evidence, 156 F.R.D. 313, 314 n. 8 (1994).1 Given the parties’ agreement that the district court was exercising its inherent power, criminal and tort-based responses to destruction of evidence are not directly relevant to our inquiry. Except as necessary to respond to the parties’ arguments, we limit our discussion to the proper exercise of inherent powers in this context.
{11} A remedy for the destruction of evidence may be available pursuant to the inherent power of the courts “to impose sanctions on both litigants and attorneys in order to regulate their docket[s], promote judicial efficiency, and deter frivolous claims.” Martinez v. Martinez, 1997-NMCA-096, ¶ 23, 123 N.M. 816, 945 P.2d 1034. The courts’ inherent power exists apart from established criminal and civil *712remedies. See State ex rel. State Highway & Transp. Dep’t v. Baca, 120 N.M. 1, 4, 5, 896 P.2d 1148, 1151, 1152 (1995) (stating that these “powers inhere in judicial authority and exist independent of statute” and are not displaced by statute or rule). The rationale underlying the existence of the inherent power of the courts is that “a court must be able to command the obedience of litigants and their attorneys if it is to perform its judicial functions.” Id. at 4, 896 P.2d at 1151. As the United States Supreme Court explained many years ago, there are “[cjertain implied powers” of the courts “which cannot be dispensed with in a Court, because they are necessary to the exercise of all others.” United States v. Hudson, 11 U.S. (7 Cranch) 32, 34, 3 L.Ed. 259 (1812).
{12} Recent cases have discussed the contours and limits of inherent judicial power. The power must be exercised cautiously and judiciously. As this Court has recently noted:
The power to control the cases on its docket is not the power to dismiss cases without cause, but the power to “‘supervise and control the movement of all cases on its docket from the time of filing through final disposition,’ and to apply sanctions when reasonable efforts to manage the court’s caseload have failed.”
State v. Ahasteen, 1998-NMCA-158, ¶ 28, 126 N.M. 238, 968 P.2d 328 (quoting State v. Ericksen, 94 N.M. 128, 131, 607 P.2d 666, 669 (Ct.App.1980)). This Court has also previously said that, under its inherent power, “the district court’s authority to impose sanctions does not extend to pre-litigation conduct.” Martinez, 123 N.M. 816, 945 P.2d 1034,1997-NMCA-096, ¶ 23 (citing Baca, 120 N.M. at 7, 896 P.2d at 1154). That is, the inherent power to sanction “does not extend to the conduct that gave rise to the underlying cause of action.” Baca, 120 N.M. at 7, 896 P.2d at 1154 (emphasis added). The reason for this is clear: If courts were permitted, of their own accord, to punish litigants for the conduct that gave rise to the underlying cause of action, litigants could thereby be made to answer twice for the same conduct. Under appropriate circumstances, however, a court may use its inherent power to sanction prelitigation conduct that does not give rise to the underlying cause of action. See Dillon v. Nissan Motor Co., 986 F.2d 263, 267 (8th Cir.1993).
{13} As we have already suggested, “inherent powers must be exercised with restraint and discretion.” Chambers, 501 U.S. at 44, 111 S.Ct. 2123. In some cases a court may be justified in using its inherent power to dismiss. See Ahasteen, 1998-NMCA-158, ¶ 28, 126 N.M. 238, 968 P.2d 328. But “dismissal is an extreme sanction to be used only in exceptional cases.” State v. Bartlett, 109 N.M. 679, 680, 789 P.2d 627, 628 (Ct.App.1990). Thus, we are persuaded that, in exercising their inherent power to address prelitigation spoliation of evidence, courts should consider several factors before imposing sanctions, in particular dismissal. Courts should consider
(1) the degree of fault of the party who altered or destroyed the evidence; (2) the degree of prejudice suffered by the opposing party; and (3) whether there is a lesser sanction that will avoid substantial unfairness to the opposing party and, where the offending party is seriously at fault, will serve to deter such conduct by others in the future.
Schmid v. Milwaukee Elec. Tool Corp., 13 F.3d 76, 79 (3d Cir.1994); cf. Enriquez v. Cochran, 1998-NMCA-157, ¶ 48, 126 N.M. 196, 967 P.2d 1136 (“In determining the nature of the sanctions to be imposed [for a violation of Rule 1-037], the trial court must balance the nature of the offense, the potential prejudice to the parties, the effectiveness of the sanction, and the imperative that the integrity of the court’s orders and the judicial process must be protected.”).
{14} As we have indicated, “[d]estruction of potentially relevant evidence ... occurs along a continuum of fault.” Welsh, 844 F.2d at 1246. The first prong of the analysis requires the court to determine where along the continuum the offending party’s conduct fell. We do not undertake to define here all possible levels of culpability that might support sanctions. We acknowledge, though, that “a finding of ‘bad faith’ or ‘evil motive’ is not a prerequisite to imposition of sanctions for destruction of evidence.” *713Baliotis, 870 F.Supp. at 1291; see also Gonzales, 120 N.M. at 158, 899 P.2d at 601. The point is simply that “[t]he motive for the destruction of evidence is ... relevant to determining what sanctions, if any, should be imposed. Schmid makes clear that it is ‘the degree of fault’ that is at issue.” Baliotis, 870 F.Supp. at 1291.
{15} The second prong requires the court to look closely at the relevance of the destroyed evidence to the various causes of action, and more specifically at the effect that the loss of the evidence might have on the non-spoliating party’s ability to prepare and present a case.
[T]he destroyed evidence must be relevant to the issue or matter for which the party seeks the [sanction]. For example, the spoliation of a machine may raise an adverse inference with respect to a claim that that particular machine was defective, but such an inference may not be drawn with respect to a claim based upon design defect when the destruction would not hinder the defense.
Beers v. Bayliner Marine Corp., 236 Conn. 769, 675 A.2d 829, 833 (1996); see also Schmid, 13 F.3d at 79. Thus in this case, the district court must evaluate the effect of the loss of the fire-suppression system on each Defendants’ ability to defend against Plaintiffs’ claims and to lay off responsibility on others potentially at fault for the loss. Again, we leave these determinations to the district court on remand.
{16} Favoring outright dismissal, Defendants Hobart and Quality argue specifically that the loss of the fire-suppression system hampers their ability to defend the negligence claims by diminishing their ability to establish relative fault. Although Defendants’ point is well-taken, it is precisely the reason the court has available to it, and should consider, a whole range of possible sanctions. Moreover, Defendants’ situation is not unlike that of a defendant who seeks to lay off liability on an “absent” defendant. Defendants can still argue Kidde-Fenwal’s liability to the jury in regard to the faulty fire-suppression system, see Wilson v. Gillis, 105 N.M. 259, 261, 731 P.2d 955, 957 (Ct.App.1986), and must demonstrate to the court why a sanction short of dismissal is inadequate.
{17} The third prong of the analysis requires the court to balance the degree of fault of the spoliator’s conduct against the magnitude of prejudice to the non-spoliating party, and to levy a sanction accordingly. The range of possible sanctions allows the court to try to offset, to the extent possible, whatever prejudice non-spoliating parties face, while at the same time to permit offending parties whose conduct does not rise to the level of bad faith to continue to pursue their claims, though under an appropriate handicap. In addition, it allows the court to levy severe sanctions “where the offending party is seriously at fault ... to deter such conduct by others in the future.” Schmid, 13 F.3d at 79.
{18} One of the sanctions a court might appropriately impose is a “spoliation inference”; that is, “[a]n instruction to the jury that it may consider that the lost evidence would be unfavorable to the [spoliating party].” Baliotis, 870 F.Supp. at 1292. If a spoliation inference is given to a jury, the jury should also “be instructed that it is not required to draw the inference that the destroyed evidence would be unfavorable but that it may do so” based on its view of the first two factors. Beers, 675 A.2d at 833.
{19} Alternatively, a court may exclude certain of the spoliator’s evidence. See Unigard, 982 F.2d at 368; Howell v. Maytag, 168 F.R.D. 502, 507-08 (M.D.Pa.1996) (mem.). For example, in this case it might be appropriate to preclude investigators Marion, Skinner, and Poe from testifying about the fire-suppression system. Under some circumstances the exclusion of evidence might mean that a plaintiff is unable to meet his burden of proof, thus making summary judgment or dismissal appropriate. See Unigard, 982 F.2d at 369; State Farm Fire & Cas. Co. v. Frigidaire, a Div. of Gen. Motors Corp., 146 F.R.D. 160, 163 (N.D.Ill.1992) (mem.).2
*714{20} Outright dismissal of a spoliator’s case might sometimes be appropriate. See Howell, 168 F.R.D. at 507. But again, “dismissal is an extreme sanction to be used only in exceptional cases.” Bartlett, 109 N.M. at 680, 789 P.2d at 628. Thus, although “a finding of ‘bad faith’ or ‘evil motive’ is not a prerequisite to imposition of sanctions for destruction of evidence,” Baliotis, 870 F.Supp. at 1291, in its absence the prejudice to the non-spoilating party would have to be severe to support dismissal.
{21} In addition to the foregoing, a court might also draw on Rule 1—037(B) for an appropriate sanction. We have already indicated that we do not think Rule 1-037 provides authority for sanctioning prelitigation spoliation, but we believe it is sufficiently analogous that a court exercising its inherent powers could reasonably fashion sanctions similar to those provided in Rule 1-037(B). See Dillon, 986 F.2d at 268-69; Capellupo, 126 F.R.D. at 551.
{22} Plaintiffs rely on Coleman to argue against the imposition of any sanction. They reason that under Coleman they cannot be deemed to be under any duty to preserve the fire-suppression system. See 120 N.M. at 651, 905 P.2d at 191. Given this lack of duty, they reason, they cannot be punished for their failure to preserve the evidence. Our response is two-fold. First, the intentional spoliation tort recognized in Coleman simply does not apply. Defendants are not seeking affirmative relief in damages as was the plaintiff in Coleman. See 120 N.M. at 647, 905 P.2d at 187. It would be inappropriate to import the requirements of the tort into this context. Second, our holding is not dependent on a concept of duty in the sense in which that term is used in analyzing potential tort liability. The concept of inherent power concerns itself with the proper functioning of the court system. It protects the integrity of the judicial process. Destruction of evidence has the potential to harm that integrity through the difficulties it creates for other parties. In short, we are responding to potential prejudice. We need not find breach of a formal duty to preserve evidence to address the prejudice caused by its loss.
Necessity of Findings and Conclusions
{23} We address one final issue briefly. Review of the merits of the sanction imposed has been made all but impossible by the lack of findings and conclusions setting forth the basis of the district court’s dismissal orders. The necessity of findings and conclusions is implicit in the inherent-power cases upon which we rely. See, e.g., Dillon, 986 F.2d at 269 (“The district court did not abuse its discretion in making the predicate findings to support the sanction of excluding ... testimony ____” (Emphasis added.)); Gonzales, 120 N.M. at 158, 899 P.2d 594; Martinez, 1997-NMCA-096, ¶ 24, 123 N.M. 816, 945 P.2d 1034. It is implicit in our discussion here as well, for reasons that are apparent. A reviewing court has no way of knowing, for example, whether the district court considered alternative sanctions without findings and conclusions to that effect. We therefore make the requirement of findings and conclusions explicit in cases in which a court draws on its inherent power to sanction a litigant for spoliating evidence. This requirement is no different from mandating findings and conclusions explaining judicial decisions in other contexts. See Allsup’s Convenience Stores, Inc. v. North River Ins. Co., 1999-NMSC-006, ¶ 17, 127 N.M. 1, 976 P.2d 1 (requiring court, when ordering remittitur, to “provide a clear articulation of how and why damages are excessive”); Enriquez, 1998-NMCA-157, ¶¶ 44-45, 126 N.M. 196, 967 P.2d 1136 (affirming Rule 1-037 sanctions where it was clear from the record that district court made requisite findings).
*715 CONCLUSION
{24} On remand, the district court is to assess the spoliation in light of the factors discussed in this opinion and then enter specific findings of fact and conclusions of law with respect to what it decides to do and why.
{25} IT IS SO ORDERED.
APODACA and BOSSON, JJ., concur.