delivered the opinion of the Court,
in which PHILLIPS, Chief Justice, and GONZALEZ, HIGHTOWER, HECHT, ENOCH and SPECTOR, Justices, join.Respondent’s Motion for Rehearing is overruled. We withdraw our prior opinion and substitute the following in its place.
This case requires us to clarify the standards governing the imposition of punitive damages in the context of bad faith insurance litigation. The parties have asked us to address three issues. First, in a bad faith case, how should Texas courts apply the definition of gross negligence from Burk Royalty Co. v. Walls, 616 S.W.2d 911 (Tex.1981), to determine whether punitive damages are appropriate? Second, what constitutes legally sufficient evidence of gross negligence to support an award of punitive damages? Third, what limits do the Due Process clause of the Fourteenth Amendment to the United States Constitution and the Due Course clause of the Texas Constitution, Tex. Const. art. 1, § 19, place on punitive damages?1 We hold *13that Juan Moriel did not present legally sufficient evidence of gross negligence. Therefore, Moriel is not entitled to punitive damages. It necessarily follows that the constitutional issues — whether the size of the punitive damages award or the procedures the trial court followed violated Transportation’s due process rights — are questions that must await another day. City of San Antonio v. Schautteet, 706 S.W.2d 103, 106 (Tex.1986) (per curiam) (explaining that constitutional challenges should not be addressed when a ease may be decided on nonconstitutional grounds). Because the court has not previously addressed punitive damages in the bad faith context, and because this opinion represents a substantial clarification of the gross negligence standard that will apply in all cases, we remand this case for a new trial in the interest of justice.
I
On March 15, 1986, Juan Moriel, an employee of Cashway Building Materials in El Paso, was injured when a stack of counter-tops fell on him. Moriel suffered three broken ribs, a broken wrist, and a fractured pelvis. As a result, he was hospitalized for twelve days. His hospitalization costs were paid by Cashway’s workers’ compensation carrier, Transportation Insurance Company.
A few days after leaving Providence hospital, Moriel experienced periodic loss of movement in one leg. He returned to Providence to undergo tests for possible nerve damage, but the record does not reveal the results of those tests. Six weeks after the accident, Moriel attempted to resume sexual relations with his wife but discovered he was impotent. He had never before experienced this problem.
Moriel’s orthopedist, Dr. Toni Ghiselli, referred him to Dr. Abel Garduño, a urologist. Dr. Garduño ordered tests for Moriel at Pathlab in El Paso. The Pathlab tests, however, revealed no physical cause for his complaint. Dr. Garduño prescribed hormones, but these were no help. Dr. Garduño then referred Moriel to Dr. Gonzalo Diaz at the Sun Towers Sleep Disorder Center in El Paso for further testing. An equipment failure rendered the Sun Towers tests inconclusive and precluded further testing for “months.” Therefore, Dr. Diaz recommended that Moriel undergo testing at the Baylor College of Medicine Sleep Disorders and Research Center in Houston.
On August 9, 1986, Moriel asked Less Huss, the adjustor handling his workers’ compensation claim, to authorize payment for the Baylor tests in Houston. Huss was an employee of Crawford & Company, an independent adjusting company that handled claims for Transportation. According to Moriel’s testimony, Huss required Moriel to obtain an authorization letter from Dr. Ghi-selli, which Moriel did within two or three days. Huss then instructed Moriel that he would need a letter from his urologist, Dr. Garduño. Moriel complied with this request within three days. According to Moriel, Huss then instructed him to obtain yet another letter from Dr. Diaz. Moriel again complied. Huss then indicated that he could not personally authorize the Baylor tests, but that he needed approval from his superiors at Transportation’s Dallas offices.
On September 10, 1986, Transportation told Moriel’s attorney that it would cover the tests but not Moriel’s travel expenses to Houston. Moriel’s attorney accepted the proposal the same day. Moriel testified that the month that had elapsed between his request and Transportation’s approval forced him to reschedule the tests. After a ten-day *14delay, Moriel underwent testing in Houston on September 25-27. The results of the testing at Baylor indicated that Moriel’s impotence problem was at least partially physical. Although no specific treatment for the physical problem was indicated, the Baylor report recommended that Moriel obtain counseling for emotional problems, for which Dr. Garduño referred Moriel to a psychiatrist, Dr. Oscar Perez. Moriel testified that he personally delivered a copy of the Baylor report to Huss, and obtained Huss’s authorization for Perez’s treatment.
Dr. Perez treated Moriel until April 1987. At trial, Dr. Perez testified that Moriel’s impotence had both physical and mental components, and that Moriel was able to overcome the mental component through therapy and resume sexual relations with his wife.
The $3,155.00 bill for Moriel’s Baylor tests was presented to Transportation on November 4,1986. Although Transportation authorized the testing in advance, it delayed payment of this bill for more than two years. Transportation claimed that it initially delayed payment because no medical report accompanied the bill, but Moriel testified that he personally delivered the report to Huss shortly after the tests were completed. Even after the date Transportation conceded receiving the report, though, it continued to deny payment on the ground that Moriel’s impotence was unrelated to his on-the-job injury.
Transportation also received bills for Dr. Perez’s services totalling $2,075.00, but delayed paying them for more than a year on the ground that it had never received his medical report. Yet Perez testified that he had sent a detailed report promptly at the conclusion of Moriel’s treatment.
Transportation also delayed paying a $382.25 bill from Providence Hospital for follow-up outpatient tests. Huss’s correspondence indicates he sent the bill to Transportation along with the other Providence bills, which totalled nearly $7000.00, and that Transportation paid the other Providence bills. Transportation also paid the bill for follow-up testing in September 1987, after Providence filed a collection action against Moriel.
Finally, Transportation failed to pay the $238.20 Pathlab bill before the commencement of Moriel’s lawsuit. However, the evidence reflects that Pathlab mailed the bill to the wrong address and that Transportation did not receive it before the suit was filed.
While he was being tested and treated, Moriel filed a workers’ compensation claim against Transportation, securing a $30,022.77 award from the Industrial Accident Board (IAB) on July 17, 1987. After Transportation appealed that award to district court, Moriel counterclaimed for additional compensation, unpaid medical bills, and bad faith claims practices. In July 1988, Moriel and Transportation settled the workers’ compensation claim, leaving the bad faith claim extant.
At the trial of the bad faith claim, the jury found that Transportation delayed paying the medical bills without a reasonable basis, that it knew or should have known that it had no reasonable basis to delay payment, and that it acted “with heedless and reckless disregard” for Moriel’s rights. The jury awarded Moriel $1000.00 in actual damages, excluding mental anguish, $100,000.00 in mental anguish damages, and $1,000,000.00 in punitive damages. Moriel also obtained findings that Transportation had engaged in unfair, deceptive or misleading acts or practices prohibited by statute that caused him actual damages, including mental anguish. The trial court entered judgment on the bad faith findings, and overruled Transportation’s motions for judgment notwithstanding the verdict, new trial, remittitur, and to disregard the jury findings.
The court of appeals affirmed, 814 S.W.2d 144, with one justice dissenting, 814 S.W.2d at 151 (Koehler, J., dissenting).
II
We must initially determine whether the settlement of Moriel’s workers’ compensation claim precludes his recovery of punitive damages. The agreed Partial Judgment setting forth the parties’ workers’ compensation settlement recited that “it [appears] to the Court that the extent of the injury and liability for compensation or medical expenses are *15uncertain....” Transportation argues that this judgment recital precludes the later finding of conscious indifference on its part, and thus eliminates Monel’s right to pursue a claim for punitive damages. In support of this position, Transportation cites Izaguirre v. Texas Employers’ Ins. Ass’n, 749 S.W.2d 550 (Tex.App.—Corpus Christi 1988, writ denied), and Price v. Texas Employers’ Ins. Ass’n, 782 S.W.2d 938 (Tex.App.—Tyler 1989, no writ), which held that similar recitals agreed to by the claimant in settling a workers’ compensation claim constituted judicial admissions estopping the claimant from later asserting a bad-faith claim. See also National Union Fire Ins. Co. v. Dominguez, 793 S.W.2d 66 (Tex.App.—El Paso 1990), rev’d on other grounds, 873 S.W.2d 373 (Tex.1994).2
While we do not disagree with these cases, they are not dispositive of Monel’s claims under the facts of this case. Moriel’s workers’ compensation counterclaim included a demand for additional lost wages as well as unpaid medical expenses. The Partial Judgment states only that “the extent of the injury and liability for compensation or medical expenses are uncertain....” (emphasis added). This recital indicates, at most, that Transportation’s total liability was uncertain; it does not constitute an admission by Moriel or a finding by the court that Transportation’s liability for each and every medical bill, or even any of the medical bills, was uncertain. Indeed, the settlement required Transportation to pay the disputed medical bills in full. Furthermore, Moriel’s bad faith contentions are primarily related to Transportation’s bad faith delay in paying Moriel’s medical bills. The recital in the Partial Judgment is therefore not inconsistent with Mor-iel’s claim that Transportation had a clear and certain duty to promptly pay at least some of the medical expenses, particularly those which Transportation authorized in advance.
Moreover, unlike the cases Transportation relies on, the Partial Judgment which Moriel approved specifically reserved his bad-faith claim as follows:
It is further ORDERED that this partial judgment does not release and discharge Plaintiff TRANSPORTATION INSURANCE COMPANY and any and all of its agents or representatives of and from any and all liability of any character arising out of claims made pursuant to allegations of bad faith, violations to the Texas Insurance Code, and Deceptive Trade Practices Act and Breach of Fiduciary Duty, as well as the manner in which TRANSPORTATION INSURANCE COMPANY failed to pay or delayed payment of the medical bills, as more fully alleged in Defendant JUAN CARLOS MORIEL’S Fifth Amended Counterclaim, Paragraphs XV through XXTV and that these claims are not subject to this partial judgment and remain pending in the above styled and numbered cause. It is further ORDERED that this partial judgment does not release and discharge Plaintiff TRANSPORTATION INSURANCE COMPANY and any and all of its agents or’ representatives of and from any and all liability of any character as alleged in Defendant JUAN CARLOS MORIEL’S Fifth Amended Counterclaim, Paragraph XXV as it pertains to those allegations of bad faith, violations of the Texas Insurance Code, Deceptive Trade Practices Act and Breach of Fiduciary Duty, as well as the manner in which TRANSPORTATION INSURANCE COMPANY either failed to pay or delayed payment of the medical bills.
Transportation concedes this point, stating in its brief that “[bjecause ... Moriel preserved his right to sue Transportation for bad faith, [Transportation] does not argue that the agreement and judgment preclude such a cause of action.” Transportation thus asserts its estoppel argument only as to the conscious indifference finding underlying the punitive damage award. But it appears that Moriel likewise preserved the punitive damages claim. The reservation in the Partial Judgment extends to “any and all liability of any character arising out of claims made *16pursuant to allegations of bad faith ... as well as the manner in which [Transportation] failed to pay or delayed payment of the medical bills....”3 This specific reservation of Monel’s bad-faith and punitive damages claim, combined with the fact that Moriel’s bad faith complaints primarily arise out of Transportation’s delay in paying the medical bills in dispute rather than the extent of his injuries, controls over the more general recital that Transportation’s liability was “uncertain.” See O’Connor v. O’Connor, 694 S.W.2d 162, 155 (Tex.App.— San Antonio 1985, writ ref'd n.r.e.).
Accordingly we hold, under the facts of this case, that the recital in the Partial Judgment concerning Transportation’s uncertain liability does not preclude Moriel’s claim for punitive damages.
Ill
A. The Exceptional Nature of Punitive Damages
The typical remedy in a civil case is an award of money damages sufficient to compensate the injured plaintiff. Cavnar v. Quality Control Parking, Inc., 696 S.W.2d 549, 552 (Tex.1985); Reaugh v. McCollum Exploration Co., 139 Tex. 485, 163 S.W.2d 620, 621 (1942). Compensatory damages are intended to make the plaintiff “whole” for any losses resulting from the defendant’s interference with the plaintiffs rights.4 Cavnar, 696 S.W.2d at 552; W. Page Keeton, et al., PROSSER AND KEETON ON THE LAW OF Torts §§ 1, 4 (5th ed. 1984) [hereinafter Prosser & Keeton].
Punitive damages have an altogether different purpose. Cavnar, 696 S.W.2d at 555; Smith v. Sherwood, 2 Tex. 460, 463-64 (1847). “The idea of punishment, or of discouraging other offenses usually does not enter into tort law_ In one rather anomalous respect, [punitive damages], the ideas underlying the criminal law have invaded the field of torts.” Prosser & Keeton § 2. More than 100 years ago, in one of its first opinions addressing punitive damages, this court recognized the close connection between punitive damages and the criminal law:
Such indifference is morally criminal, and if it leads to actual injury may well be regarded as criminal in law. A mere act of omission or non-feasance, to be punishable by exemplary damages, should reach the border-line of a quasi-criminal act of commission or malfeasance.
Southern Cotton Press & Mfg. Co. v. Bradley, 52 Tex. 587, 600-601 (1880) (citations omitted).
Punitive (or exemplary) damages are levied against a defendant to punish the defendant for outrageous, malicious, or otherwise morally culpable conduct. Id.; Tex.Civ. Prac. & Rem.Code Ann. § 41.001(3) (Vernon Supp.1994) (defining “exemplary damages” as “any damages awarded as an example to others, as a penalty, or by way of punishment”). The legal justification for punitive damages is similar to that for criminal punishment,5 and like criminal punishment, puni*17tive damages require appropriate substantive and procedural safeguards to minimize the risk of unjust punishment.
Although punitive damages are levied for the public purpose of punishment and deterrence,6 the proceeds become a private windfall.7 See 26 U.S.C. §§ 104,104(a)(2) (excluding personal injury compensatory damages but not punitive damages from income for tax purposes); Commissioner v. Miller, 914 F.2d 586, 591 (4th Cir.1990) (explaining that punitive damages, unlike personal injury damages, are not excludable from income because “[s]uch damages are a windfall ... over and above any award of compensatory damages”). In contrast, criminal fines are paid to a governmental entity and used for a public benefit. Our duty in civil cases, then, like the duty of criminal courts, is to ensure that defendants who deserve to be punished in fact receive an appropriate level of punishment, while at the same time preventing punishment that is excessive or otherwise erroneous.
B. Bad Faith Insurance Disputes
Defining the substantive standard for punitive damages in the context of a bad faith insurance dispute requires an examination of the relationship of a claim for punitive damages to the underlying claims for breach of contract and bad faith. Our law recognizes a three-tier framework for measuring damages in an insurance coverage dispute, and each level is associated with distinctly different policies, substantive definitions, and measures of proof. A bad faith case can potentially result in three types of damages: (1) benefit of the bargain damages for an accompanying breach of contract claim, (2) compensatory damages for the tort of bad faith, and (3) punitive damages for intentional, malicious, fraudulent, or grossly negligent conduct. It is important to preserve distinct legal boundaries between the three bases of recovery to prevent arbitrariness and confusion at the critical thresholds. See, e.g., Lyons v. Millers Casualty Ins. Co., 866 S.W.2d 597, 600 (Tex.1993) (“This focus on evidence and its relation to the elements of bad faith is necessary to maintain a distinction between a contract claim on the policy, and a claim of bad faith delay or denial.”).
As we said in Viles v. Security Nat'l Ins. Co., 788 S.W.2d 566, 567 (Tex.1990), “[A] breach of the duty of good faith and fair dealing will give rise to a cause of action in tort that is separate from any cause of action for breach of the underlying insurance contract.” The contract aspect of a coverage dispute concerns either the factual basis for the claim, the proper legal interpretation of the policy, or both. If the loss is covered, then the insurer is obligated to pay the claim according to the terms of the insurance contract. See Crisp v. Security Nat’l Ins. Co., 369 S.W.2d 326, 328-29 (Tex.1963); Klein v. Century Lloyds, 154 Tex. 160, 275 S.W.2d 95 (1955). An insurer’s nonpayment of a covered claim ordinarily is a breach of contract, and does not alone entitle a plaintiff to mental anguish, Dean v. Dean, 837 F.2d 1267 (5th Cir.1988) (citing St. Elizabeth Hosp. v. Garrard, 730 S.W.2d 649 (Tex.1987), overruled on other grounds, Boyles v. Kerr, 855 S.W.2d 593 (Tex.1993)), or exemplary damages. Jim Walter Homes, Inc. v. Reed, 711 S.W.2d 617 (Tex.1986).
The threshold of bad faith is reached when a breach of contract is accompanied by an independent tort. Evidence that merely shows a bona fide dispute about the insurer’s liability on the contract does not rise to the level of bad faith. National Union Fire Ins. Co. v. Dominguez, 873 S.W.2d 373, 376-77 (Tex.1994); National Union Fire Ins. Co. v. Hudson Energy Co., 780 S.W.2d 417, 426 (Tex.App.—Texarkana 1989), aff'd, *18811 S.W.2d 552 (Tex.1991).8 Nor is bad faith established if the evidence shows the insurer was merely incorrect about the factual basis for its denial of the claim, or about the proper construction of the policy. Lyons, 866 S.W.2d at 601 (“[T]he issue in bad faith focuses not on whether the claim was valid, but on the reasonableness of the insurer’s conduct in rejecting the claim.”). A simple disagreement among experts about whether the cause of the loss is one covered by the policy will not support a judgment for bad faith. Id. To the contrary, an insured claiming bad faith must prove that the insurer had no reasonable basis for denying or delaying payment of the claim, and that it knew or should have known that fact. Arnold v. National County Mut. Fire Ins. Co., 725 S.W.2d 165, 167 (Tex.1987); Aranda v. Insurance Co. of N. Am., 748 S.W.2d 210, 213 (Tex.1988).
The final critical threshold is that separating the tort of bad faith from conduct subject to punishment. Even if the insurer has “no reasonable basis” to deny or delay payment of the claim, the plaintiff may not recover punitive damages on that basis alone. Arnold, 725 S.W.2d at 168 (“the same principles allowing recovery of [punitive] damages in other tort actions” apply with respect to bad faith) (citing Ware v. Paxton, 359 S.W.2d 897, 899 (Tex.1962) (“The fact that an act is [tortious] is not of itself ground for an award of exemplary or punitive damages.”) (quoting Jones v. Ross, 141 Tex. 415, 173 S.W.2d 1022, 1024 (1943) (quoting 25 C.J.S. Damages § 123 at 726 (1966)))). The bad faith of the insurer justifies an award of compensatory damages and nothing more. Only when accompanied by malicious, intentional, fraudulent, or grossly negligent conduct does bad faith justify punitive damages. Id.
It is as important to maintain the distinction between punishment and compensation in the context of bad faith as it is in the remainder of tort law. The reason the law of torts recognizes compensation, rather than punishment, as its paramount objective is that civil punishment can result in overdeter-rence and overcompensation. Every tort involves conduct that the law considers wrong, but punitive damages are proper only in the most exceptional cases. Unless bad faith is accompanied by aggravated conduct by the insurer, then compensatory damages alone are the proper remedy.
The particular difficulty at this threshold is to articulate the distinction between simple and aggravated bad faith in a manner that judges and juries may accurately and fairly apply. In a helpful effort to draw this distinction the Arizona Supreme Court has observed:
Such damages are recoverable in bad faith tort actions when, and only when, the facts establish that the defendant’s conduct was aggravated, outrageous, malicious or fraudulent. Indifference to facts or failure to investigate are sufficient to establish the tort of bad faith but may not rise to the level required by the punitive damage rule. The difference is no doubt harder to articulate in legalistic terms than it is to differentiate on the facts. To obtain tort damages, for instance, the plaintiff must prove only that the defendant ... [acted without] the reasonable basis required for denying the claim. To obtain punitive damages, plaintiff must also show that the evil hand that unjustifiably damaged the objectives ... [of] the insurance contract was guided by an evil mind which either consciously sought to damage the insured or acted intentionally, knowing that its conduct was *19likely to cause unjustified, significant damage to the insured.
Rawlings v. Apodaca, 151 Ariz. 149, 726 P.2d 565, 578 (1986) (emphasis in original) (citations omitted). Likewise, the Alabama Supreme Court in Continental Assur. Co. v. Kountz, 461 So.2d 802 (Ala.1984), developed a two-part standard for punitive damages in a bad faith insurance dispute. The Kountz court held that before punishment is justified, an insurer (1) must have intentionally breached the duty of good faith and fair dealing, and (2) must additionally have engaged in conduct that involved “malice, willfulness or wanton and reckless disregard of the rights of others.” Id. at 808-09.
There is not yet any uniformity among states concerning the proper test for punitive damages in a bad faith case. See John C. MoCaRthy, Recovery of Damages for Bad Faith § 1.60 (1990). However, a review of relevant cases in a number of jurisdictions reveals two core requirements: (1) the insurance carrier must act on the basis of an aggravated mental state, such as intent, recklessness, or gross negligence, that involves either deliberate and purposeful injury to the insured or actual subjective awareness that serious injury is highly probable, and (2) this intended or probable injury must be independent and qualitatively different from the breach of contract and the compensable harms associated with it. See Kountz, 461 So.2d 802; Rawlings, 726 P.2d 565; Linthicum v. Nationwide Life Ins. Co., 150 Ariz. 326, 723 P.2d 675 (1986) (en banc); Borland v. Safeco Ins. Co., 147 Ariz. 195, 709 P.2d 552 (Ct.App.1985); Weisman v. Blue Shield of Cal., 163 Cal.App.3d 61, 209 Cal.Rptr. 169, 173-74 (1984); Newton v. Standard Fire Ins. Co., 291 N.C. 105, 229 S.E.2d 297, 302 (1976); Shimola v. Nationwide Ins. Co., 25 Ohio St.3d 84, 495 N.E.2d 391, 393 (1986) (per curiam); Anderson v. Continental Ins. Co., 85 Wis.2d 675, 271 N.W.2d 368, 378-79 (1978); Guarantee Abstract & Title Co. v. Interstate Fire & Casualty Co., 232 Kan. 76, 652 P.2d 665, 668 (1982). We believe the general thrust of these cases is essentially correct.9 However, no jurisdiction has a gross negligence standard or bad faith jurisprudence identical to those of Texas. Moreover, none of these jurisdictions has addressed the critical distinctions in a manner we find completely satisfactory. Therefore, the duty falls on this court to add our refinements to these earlier analyses in a manner that meshes properly with the rest of Texas law.
IV
Three steps are required to analyze whether the evidence in this case is legally sufficient to support punitive damages. The first is to examine our traditional definition of gross negligence and identify its basic elements. The second is to determine how these elements should be applied in the context of a bad faith insurance coverage dispute. Only then can we reach step three and determine meaningfully whether legally sufficient evidence supports the verdict.
A.
1. The Definition of Gross Negligence
In Burk Royalty, 616 S.W.2d at 920, we analyzed the history óf gross negligence evidentiary review in Texas at length and determined that “all roads” lead to the following definition of gross negligence:
Gross negligence, to be the ground for exemplary damages, should be that entire want of care which would raise the belief that the act or omission complained of was the result of a conscious indifference to the right or welfare of the person or persons to be affected by it.
This definition is taken verbatim from Missouri Pac. Ry. v. Shuford, 72 Tex. 165, 171, 10 S.W. 408, 411 (1888), and is little changed since this court first defined gross negligence in Southern Cotton Press, 52 Tex. 587 (1880) *20(“that entire want of care which would raise a presumption of conscious indifference to consequences”). In 1987, however, the legislature modified the common law definition:
“Gross negligence” means more than momentary thoughtlessness, inadvertence or error of judgment It means such an entire want of care as to establish that the act or omission was the result of actual conscious indifference to the rights, safety, or welfare of the person affected.10
Tex.Civ.Prac. & Rem.Code Ann. § 41.001(5) (Vernon Supp.1994) (emphasis added). The statutory definition of gross negligence, when compared with the common law definition, emphasizes that the evidence must “establish” the defendant’s actual conscious indifference, rather than raise the mere belief that conscious indifference might be attributable to a hypothetical reasonable defendant.11 John T. Montford & Will G. Barber, 1987 Texas Tort Reform: The Quest for a Fairer and More Predictable Texas Civil Justice System, 25 Hous.L.Rev. 245, 323-24 (1988). Because the Tort Reform Act codified the common law definition and made no change affecting the basic elements of gross negligence, Act of June 3, 1987, 70th Leg., 1st C.S., ch. 2, § 2.12, 1987 Tex.Gen.Laws 44 (codified as Tex.Civ.Prac. & Rem.Code Ann. § 41.001(5)), the language defining gross negligence to be applied in a particular case should not turn on the applicability of the Act.
Our cases before Burk Royalty tended to focus only on one portion of the Shuford definition, “entire want of care,” and reasoned that any defendant who exercised “some care” could not be grossly negligent. Burk Royalty, 616 S.W.2d at 918 (citing Sheffield Div., Armco Steel Corp. v. Jones, 376 S.W.2d 825 (Tex.1964)). Burk Royalty cautioned, however, that “entire want of care” must be “understood in the context of the whole [definition],” and correctly instructed reviewing courts to look for evidence of the defendant’s subjective mental state rather than the defendant’s exercise of care. Id. at 922 (“What lifts ordinary negligence into gross negligence is the mental attitude of the defendant”).
Although Burk Royalty exhorted a reviewing court to look to “all of the surrounding facts,” our usual process of “no evidence” review constrains us to look only to the evidence that favors the verdict. Garza v. Alviar, 395 S.W.2d 821, 824 (Tex.1965), cited in Burk Royalty, 616 S.W.2d at 922. Applying traditional legal sufficiency review since Burk Royalty, we have consistently disre*21garded facts that tend to contradict the verdict. General Chem. Corp. v. De La Lastra, 852 S.W.2d 916, 921 (Tex.1993), cert. dismissed, — U.S. -, 114 S.Ct. 490, 126 L.Ed.2d 440 (1993); Williams v. Steves Indus., Inc., 699 S.W.2d 570, 574 (Tex.1985); International Armament Corp. v. King, 686 S.W.2d 595, 597 (Tex.1985); Trenholm v. Ratcliff, 646 S.W.2d 927, 931 (Tex.1983); Neely v. Community Properties, Inc., 639 S.W.2d 452, 454 (Tex.1982).
Unfortunately, other aspects of Burk Royalty caused the courts to have difficulty applying the appropriate standard of review in gross negligence cases. First, Burk Royalty admonished courts to concentrate on the defendant’s mental state, Burk Royalty, 616 S.W.2d at 922, but offered no other ground to distinguish gross from ordinary negligence.12 Second, Burk Royalty correctly permitted an inference of subjective awareness from surrounding circumstances. Id. As a result, our usual “no evidence” review process, when applied to the Burk Royalty analysis of gross negligence eroded the distinction between gross and ordinary negligence. If evidence of carelessness could support a finding of negligence, then the evidence viewed in its most favorable light would often support a circumstantial inference of subjective awareness as well. Id. Because subjective awareness was the only ground the court identified to separate gross from ordinary negligence, id., the Burk Royalty analysis readily permitted an inference of gross negligence from evidence of “some” carelessness. See, e.g., General Chem. Corp., 852 S.W.2d at 921 (holding chemical manufacturer grossly negligent for failure to include word “death” in extensive warning label based on evidence of manufacturer’s awareness of incident involving misuse of unlabelled product 15 years earlier); but see Wal-Mart Stores, Inc. v. Alexander, 868 S.W.2d 322 (Tex.1993) (expressly distinguishing between legally sufficient evidence of negligence and legally sufficient evidence of gross negligence). In this way, Burk Royalty established a “some carelessness” standard of legal sufficiency review that was as favorable to punitive damage awards as the earlier- “some care” test had been hostile to them.
2. Toward a Functional Interpretation
“Some carelessness” review invariably confuses two very different things: (1) some evidence of carelessness, and (2) legally sufficient evidence that the defendant was consciously indifferent to the rights, safety, or welfare of others. The fundamental defect in the “some carelessness” approach is that it fails to consider the definition of gross negligence as a whole. The entire definition of “gross negligence” is “such an entire want of care as to establish that the act or omission was the result of actual conscious indifference to the rights, safety, or welfare of the person affected.” Tbx.Civ.PRAC. & Rem.Code Ann. § 41.001(5) (Vernon Supp.1994) (emphasis added).
Gross negligence thus involves two components: (1) the defendant’s act or omission, and (2) the defendant’s mental state. As defined, the act or omission element must involve behavior that endangers the rights, safety, or welfare of the person affected.13 Tex.Civ.Prac. & Rem.Code Ann. § 41.001(5) (Vernon Supp.1994). Gross negligence, then, differs from ordinary negligence with respect to both elements — the defendant must be “consciously indifferent” and his or her conduct must “create an extreme degree of risk.” Williams, 699 S.W.2d at 573; see also Wal-Mart, 868 S.W.2d at 326 (“We reaffirm that a gross negligence finding may be upheld on appeal only if there is [legally sufficient] evidence that a) the defendant’s conduct created an extreme risk of harm, and b) the defendant was aware of the extreme risk.”).
As we have recently reiterated, the test for gross negligence “contains both an *22objective and a subjective component.” Wal-Mart, 868 S.W.2d at 326. Subjectively, the defendant must have actual awareness of the extreme risk created by his or her conduct. Id. Objectively, the defendant’s conduct must involve an “extreme degree of risk,” a threshold significantly higher than the objective “reasonable person” test for negligence. Id. Extreme risk is a function of both the magnitude and the probability of the anticipated injury to the plaintiff. As we said in Wal-Mart, the “extreme risk” prong is not satisfied by a remote possibility of injury or even a high probability of minor harm, but rather “the likelihood of serious injury” to the plaintiff. Id. at 327. (emphasis omitted).
An act or omission that is merely thoughtless, careless, or not inordinately risky cannot be grossly negligent. Only if the defendant’s act or omission is unjustifiable14 and likely to cause serious harm can it be grossly negligent.
It should not be surprising, perhaps, that many efforts to analyze gross negligence findings confuse the defendant’s mental state with the nature of the defendant’s act or omission. The mental state involves awareness of the risk, and thus the two are interrelated. As Prosser and Keeton point out, most jurisdictions distinguishing gross from ordinary negligence have focused on either the difference in the defendant’s mental state or the difference in riskiness of the defendant’s act, but the two definitions have tended to merge and “take on the same meaning, of an aggravated form of negligence....” PROSSER & Keeton § 34 at 214 (5th ed. 1984), cited in Williams, 699 S.W.2d at 572. But this is not so in Texas. Ours is “a hybrid definition, distinctive to this state ... com-bin[ing] both of the traditional tests for gross negligence.” Williams, 699 S.W.2d at 572-73.
Williams also suggested that gross negligence could be proved if “under the surrounding circumstances a reasonable person would have realized that his conduct created an extreme degree of risk to the safety of others.” Id. at 573. This language is potentially misleading, however. It is true that the relative riskiness of the defendant’s action can be determined by an objective “extreme risk” test, but before a gross negligence finding can be sustained the evidence must show both that the act was likely to result in serious harm and that the defendant was consciously indifferent to the risk of harm. Wal-Mart, 868 S.W.2d at 326. The requirement of conscious indifference, which our law requires, is superfluous unless it requires proof that the defendant had actual subjective knowledge of an extreme risk of serious harm. Williams, therefore, should not be read to import the objective “reasonable person” standard for ordinary negligence into the distinctly different process of determining subjective mental state in gross negligence cases.
*23Recognizing the practical difficulty of producing direct evidence of conscious indifference short of the defendant’s admission, Williams quite reasonably stated that “the plaintiff need not prove the defendant’s subjective state of mind by direct evidence,” and authorized proof of this element by circumstantial evidence. Williams, 699 S.W.2d at 673 (“reaffirm[ing] [the] holding in Burk Royalty,” 616 S.W.2d at 922). We hereby reaffirm our holding that the defendant’s subjective mental state can be proven by direct or circumstantial evidence.
Determining whether an act or omission involves extreme risk or peril requires an examination of the events and circumstances from the viewpoint of the defendant at the time the events occurred, without viewing the matter in hindsight. In every negligence or gross negligence case, some injury has allegedly occurred. However, the magnitude of the injury may be entirely disproportionate to the riskiness of the behavior. For example, inadvertently dropping a wooden board into the metal hold of a ship may constitute negligence, but cannot be gross negligence. This is so even though the board, upon landing, triggers a Rube Goldberg chain reaction, eventually causing the whole ship to explode.15 See In re Polemis, [1921] 3 K.B. 560. If somebody has suffered grave injury, it may nevertheless be the case that the behavior which caused it, viewed prospectively and without the benefit of hindsight, created no great danger. In such a case, punitive damages are not appropriate. In summary, the definition of gross negligence includes two elements:
(1) viewed objectively from the standpoint of the actor, the act or omission must involve an extreme degree of risk, considering the probability and magnitude of the potential harm to others, and (2) the actor must have actual, subjective awareness of the risk involved, but nevertheless proceed in conscious indifference to the rights, safety, or welfare of others.
B. Application to Bad Faith Insurance Disputes
An insurer’s delay or refusal to pay an insured’s claim, even when the insurer has no reasonable basis for doing so, is not gross negligence. It is bad faith. An insurer is hable for punitive damages only if the bad faith tort was accompanied by gross negligence.16 Aranda, 748 S.W.2d at 215; Arnold, 725 S.W.2d at 168 (“the same principles allowing [the] recovery of [punitive] damages in other tort actions” apply with respect to bad faith) (citing Ware, 359 S.W.2d 897); cf. Lyons, 866 S.W.2d at 600. Although the multiple grounds of recovery in bad faith lawsuits might seem to complicate the gross negligence issue, the proper analysis is relatively simple — gross negligence in the context of insurance is no different from gross negligence in any other context.
In addition to conscious indifference, an insured who alleges gross negligence must prove that the insurer committed an act that was likely to cause serious injury. In essence, the issue in determining whether bad faith involved an independent likelihood of “serious injury” is whether the insurer engaged in the sort of outrageous behavior that the law seeks to punish. Ware, 359 S.W.2d at 899 (“The fact that an act is [tortious] is not of itself ground for an award of exemplary or punitive damages.”).
It may be easier to describe “serious injury” in a bad faith case by first saying what it cannot be. It cannot be the injury associated with the breach of contract. Other jurisdictions considering this issue uniformly require some serious injury that is *24independent and qualitatively different from the breach of the insurance contract. Linthicum, 723 P.2d 675; Borland, 709 P.2d 552; Weismon, 209 Cal.Rptr. at 173-74; Newton, 229 S.E.2d at 302; Shimola, 495 N.E.2d at 393; Anderson, 271 N.W.2d 368; Guarantee Abstract, 652 P.2d at 668. Nor do mere inconvenience, annoyance, or delay satisfy this requirement. Some inconvenience accompanies the breach of any contract, and is not ordinarily considered punishable, even if the breach is intentional. Jim Walter Homes, 711 S.W.2d at 618. Although the bad faith breach of an insurance contract, unlike the breach of most contracts, may justify an award of consequential damages for mental anguish,17 Arnold, 725 S.W.2d at 168, the existence of mental anguish damages alone does not ordinarily warrant legal punishment.
Some consequential injuries may satisfy the “independent and qualitatively different” test for punitive damages. For example, if an insured requires medical treatment, yet her insurer denies coverage in bad faith with full knowledge of the danger to the insured, then the insured will probably suffer serious injury not ordinarily associated with mere breach of contract. Also, some forms of insurance, such as disability insurance, may be such that insurers frequently can expect that bad faith denial of coverage will result in such an extraordinary degree of hardship and oppression that punishment would be warranted. ,See Delgado v. Heritage Life Ins. Co., 157 Cal.App.3d 262, 203 Cal.Rptr. 672, 680-81 (1984).
Because the fact patterns arising in insurance coverage disputes vary so widely, we must rely on the evolution of the common law to increase the precision of this distinction over time. In general, though, an insurance carrier’s refusal to pay a claim cannot justify punishment unless the insurer was actually aware that its action would probably result in extraordinary harm not ordinarily associated with breach of contract or bad faith denial of a claim — such as death, grievous physical injury, or financial ruin. See, e.g., Silberg v. California Life Ins. Co., 11 Cal.3d 452, 113 Cal.Rptr. 711, 521 P.2d 1103 (1974) (holding carrier subject to punitive damages when it knew that its bad faith denial of health insurance claim for over two years caused great harm to the plaintiff, including denial of surgical treatment, inability to afford pain medication, failure of his small business, two nervous breakdowns, and repossession of plaintiffs wheelchair). The harm must be independent and qualitatively different from the sort of injuries that typically result from bad faith or breach of contract. Jim Walter Homes, 711 S.W.2d at 618; Ware, 359 S.W.2d at 899.
y
To sustain the punitive damages awarded to Moriel in this case, the record must contain legally sufficient evidence of Transportation’s gross negligence. Otherwise, we are obligated to reverse the judgment of the court below.
However, we are not simply directed to determine whether evidence exists that has some remote relation to the verdict. We must also determine whether the evidence is legally sufficient. Robert W. Calvert, “No Evidence” and “Insufficient Evidence” Points of Error, 38 Tex.L.Rev. 361 (1960); William Powers, Jr. & Jack Ratliff, Another Look at “No Evidence” and “Insufficient Evidence,” 69 Tex.L.Rev. 515 (1991). If the evidence is legally sufficient, we may not reverse the trial court’s judgment; if it is not legally sufficient, we are compelled to do so.
“The evidence presented, viewed in the light most favorable to the prevailing party, must be such as to permit the logical inference [that the jury must reach].” Lyons, 866 S.W.2d at 600; Dominguez, 873 S.W.2d at 376. There must necessarily be a logical connection, direct or inferential, between the evidence offered and the fact to be proved. Lyons, 866 S.W.2d at 600 (citing Pittman v. Baladez, 158 Tex. 372, 312 S.W.2d 210, 216 (1958)). However, we must also bear in mind the difference between materiality of the evidence and the issue of evidentiary sufficiency. Simply because a piece or pieces of *25evidence are material in the sense that they make a “fact that is of consequence to the determination of the action more ... or less probable,” Tex.R.Civ.Evid. 401, does not render the evidence legally sufficient. As Professor McCormick succinctly put it, “a brick is not a wall.” Chaeles T. McCormick, HANDBOOK OF THE LAW OF EVIDENCE § 152 (West ed. 1954); see also Fed.R.Evid. 401 advisory committee’s note; Richard D. Friedman, The Elements of Evidence 59-60 (West ed. 1991).
In evaluating legal sufficiency, we are required to determine whether the proffered evidence as a whole rises to a level that would enable reasonable and fair-minded people to differ in their conclusions. William Powers, Jr. & Jack Ratliff, Another Look at “No Evidence’’ and “Insufficient Evidence,” 69 Tex.L.Rev. 515, 522, 523 (1991) (“[T]he court must be persuaded that reasonable minds could not differ on the matter.... Ultimately, the test for “conclusive evidence” ... is similar to the test for “no evidence” ... the court asks whether reasonable minds could differ about the fact determination to be made by the jury.”); Calvert, supra at 363-65 (“The rule as generally stated is that if reasonable minds cannot differ from the conclusion that the evidence lacks probative force it will be held to be the legal equivalent of no evidence.”); see also Litton Indus. Prods., Inc. v. Gammage, 668 S.W.2d 319, 324 (Tex.1984); Woods v. Townsend, 144 Tex. 594, 192 S.W.2d 884 (1946); Joske v. Irvine, 91 Tex. 574, 44 S.W. 1059 (1898); McDonough v. Zamora, 338 S.W.2d 507, 515-16 (Tex.Civ.App.—San Antonio 1960, writ ref'd n.r.e.).
As we have previously stated, there is evidence that Transportation agreed in advance to pay for Moriel’s testing but not his travel, that its process of deciding whether to pay caused a ten-day delay in testing, that it delayed for two years to pay the testing expenses and for over a year to pay subsequent treatment expenses. Transportation eventually paid only after the threat of collection efforts.18 Transportation says it disputed coverage of the treatment expenses because it believed that Moriel’s impotence was not work-related, but Transportation’s reasons for delay are irrelevant because it concedes for purposes of this appeal that legally sufficient evidence supports the jury’s conclusion that it acted without any reasonable basis and is thus liable for bad faith.19
The jury found that Transportation acted in “heedless and reckless disregard” of Mor-iel’s rights. The jury’s other findings, (1) that Transportation had no reasonable basis to delay payment, and (2) that Transportation “knew or should have known” it had no reasonable basis, only support the unchallenged award of bad faith damages. “Heedless and reckless disregard,” as defined in the jury instruction, is gross negligence. See Rowan v. Allen, 134 Tex. 215, 134 S.W.2d 1022, 1024 (1940).
Under a gross negligence theory, Transportation cannot be liable for punitive damages unless it was “consciously indifferent.” The question is not whether Transportation was unreasonable or whether it “should have known” it was acting in bad faith, but whether Transportation was actually aware of an extreme risk — some genuine and unjustifiable likelihood of serious harm to Moriel that was independent and qualita*26tively different from the inconvenience of an unreasonable delay in payment.
The evidence presented does not support this finding. Viewed in its most favorable light, Monel’s evidence does not support either (1) the inference that Transportation had any subjective awareness that Moriel would probably suffer serious injury because of Transportation’s delay, or (2) the inference that Transportation’s actions created any risk of serious harm to Moriel.
The only harm that Transportation’s delay caused Moriel was the anxiety of knowing that his bills were not paid until Providence Hospital filed a collection action in violation of the Workers’ Compensation Act.20 While Moriel’s anxiety and embarrassment were genuine, this injury does not rise to the level of serious harm sufficient to justify punitive damages.21 Cf. Fidelity & Guar. Ins. Underwriters v. Saenz, 865 S.W.2d 108, 113 (Tex.App. —Corpus Christi 1993, writ granted); Cronin v. Bacon, 837 S.W.2d 265, 269 (Tex.App.—Fort Worth 1992, writ denied) (defining “mental anguish”).
Although the jury found that Moriel suffered substantial mental anguish, this consequential injury does not support the punitive damage award under the circumstances of this case. We cannot rule out the possibility that evidence of aggravating circumstances might, in a different case, justify punitive damages for an insurer’s conscious indifference to the probability of mental anguish. However, even assuming that sufficient evidence of aggravating circumstances were in this record, Moriel has still failed to introduce any evidence whatsoever that Transportation had subjective awareness. Because Moriel alleged no other form or risk of consequential injury, and introduced no evidence of the same, Transportation’s “no evidence” point must be sustained.
Ordinarily, reversal on this point of error would result in judgment in Transportation’s favor on the issue of punitive damages. However, the trial was conducted at a time when no opinion of this court specifically addressed the standards governing the imposition of punitive damages in bad faith lawsuits. Moreover, today’s decision represents a substantial clarification of the Texas gross negligence standard and “no evidence” review of gross negligence findings. We have broad discretion to remand for new trial in the interest of justice. Tex.R.App.P. 180; Exxon Corp. v. Tidwell, 867 S.W.2d 19, 23 (Tex.1993); Boyles, 855 S.W.2d at 603; Murray v. San Jacinto Agency, Inc., 800 S.W.2d 826, 830 (Tex.1990); Scott v. Liebman, 404 S.W.2d 288, 294 (Tex.1966). We therefore reverse the judgment of the court of appeals and remand this case for a new trial. We are convinced that a retrial of the case, applying the standards announced today, is warranted in the interest of justice.
VI
Procedural Standards
Inasmuch as we are remanding this case for retrial, we consider it advisable, in the exercise of our common law duties, to articulate procedural standards for the trial court. The standards we announce apply to all punitive damage cases tried in the future. In *27Pacific Mut. Life Ins. Co. v. Haslip, 499 U.S. 1, 18, 111 S.Ct. 1032, 1043, 113 L.Ed.2d 1 (1991), the Supreme Court expressed concern that punitive damages might “run wild,” recognizing that unlimited discretion on the part of a jury or reviewing court “may invite extreme results that jar one’s constitutional sensibilities.” The Court did not, however, set forth specific due process guidelines, but merely held that the Alabama system under scrutiny in Haslip passed constitutional muster. The Court focused on the jury instruction, which provided some limit on the jury’s discretion, and the process of post-verdict review by Alabama trial and appellate courts.
The jury instruction in Haslip explained that the purpose of punitive damages is to “punish the defendant” and “for the added purpose of protecting the public by [deterring] the defendant and others from doing such wrong in the future.” Id. at 19, 111 S.Ct. at 1044. The instruction further explained that punitive damages are not compulsory, and that in awarding damages a jury “must take into consideration the character and the degree of the wrong as shown by the evidence and necessity of preventing similar wrong.” Id. The Court held that the jury’s discretion under this instruction, although significant, was not unlimited, as it “was confined to deterrence and retribution, the state policy concerns sought to be advanced.” Id. The Court concluded that this instruction “reasonably accommodated [the defendant’s] interest in rational decisionmaking and Alabama’s interest in meaningful individualized assessment of appropriate deterrence and retribution.”22 Id. at 20, 111 S.Ct. at 1044.
The Court next examined post-verdict review of punitive damages by Alabama trial courts. These courts must scrutinize each award, considering such factors as “the culpability of the defendant’s conduct,” the “desirability of discouraging others from similar conduct,” the “impact upon the parties,” and “other factors, such as the impact on innocent third parties.” Id. (quoting Hammond v. City of Gadsden, 493 So.2d 1374, 1379 (Ala.1986)). Alabama trial courts must also state on the record their reasons for interfering with the verdict or refusing to do so. Id. The Court held that these procedures provide “meaningful and adequate review by the trial court whenever a jury has fixed the punitive damages.” Id.
The Court finally focused on appellate review of Alabama punitive damage awards, which “provides an additional check on the jury’s or trial court’s discretion.” Id. 499 U.S. at 20-22, 111 S.Ct. at 1045. The Alabama Supreme Court first undertakes a comparative analysis, and then applies “detailed substantive standards” to evaluate the award,23 making sure that the award “does not exceed an amount that will accomplish society’s goals of punishment and deterrence.” Id. at 21, 111 S.Ct. at 1045 (citing Green Oil Co. v. Hornsby, 539 So.2d 218, 222 *28(Ala.1989); Wilson v. Dukona Corp., 547 So.2d 70, 73 (Ala.1989)). “This appellate review makes certain that the punitive damages are reasonable in their amount and rational in light of their purpose to punish what has occurred and to deter its repetition.”24 Id.
In conclusion, the Court held that the Alabama procedures did not violate due process, as their application “imposes a sufficiently definite and meaningful constraint on the discretion of Alabama factfinders in awarding punitive damages. The Alabama Supreme Court’s postverdict review ensures that punitive damages awards are not grossly out of proportion to the severity of the offense and have some understandable relationship to compensatory damages.” Id. at 22, 111 S.Ct. at 1045.
The Supreme Court again examined the constitutional issues surrounding punitive damages in TXO Prod. Corp. v. Alliance Resources Corp., — U.S. -, 113 S.Ct. 2711, 125 L.Ed.2d 366 (1993). That case involved a West Virginia slander of title action in which the jury awarded compensatory damages of $19,000 and punitive damages of $10 million. The Court upheld the punitive award against a procedural and substantive due process attack, although there was no majority opinion.
The West Virginia procedures in TXO differed from those previously examined in Haslip in that the trial court did not articulate its reasons for upholding the large punitive damages verdict. The Court nevertheless concluded that this was not a constitutional defect, id. — U.S. at-, 113 S.Ct. at 2724 (plurality opinion of Stevens, J., joined by Rehnquist, C.J., and Blackmun & Kennedy, JJ.); id. — U.S. at-, 113 S.Ct. at 2726 (Scalia, J., concurring, joined by Thomas, J.), but the plurality emphasized that the trial judge did “give counsel an adequate hearing on TXO’s postverdict motions.” Id. — U.S. at -, 113 S.Ct. at 2742.
Many aspects of the procedures employed up to now by Texas courts do not compare favorably to those examined in Haslip and TXO. First, although a defendant may obtain trial court review of a punitive damage award by filing a motion for new trial, Tex. R.Crv.P. 320, there has been no requirement that trial courts scrutinize each award, setting forth reasons on the record for refusing to disturb a jury verdict. Indeed, trial courts have not even been required to rule on motions for new trial as the passage of time may serve to overrule a new trial motion by operation of law. Tex.R.Civ.P. 329b(c). Thus, the judge who heard first-hand the evidence in a case, and who is uniquely positioned to determine whether the jury’s award was based on passion or prejudice rather than reason, is not required to perform a meaningful review of the award.
Second, although this Court has set forth factors to guide Texas courts of appeals in evaluating punitive damages awards, see Alamo Nat’l Bank v. Kraus, 616 S.W.2d 908, 910 (Tex.1981), a court of appeals may overturn the jury’s verdict and order the case retried only if the verdict “is so against the great weight and preponderance of the evidence as to be manifestly unjust.” Pool v. Ford Motor Co., 715 S.W.2d 629, 635 (Tex.1986).25 *29Moreover, this court, unlike its counterparts in Alabama, West Virginia, and most other states, is precluded from reviewing the evidence supporting a punitive damages award for factual sufficiency. Tex. Const. art. V, § 6; Tex.Gov’t Code § 22.225(a); Kraus, 616 S.W.2d at 910. Although due process does not require two levels of appellate review, this aspect of Texas procedure means that the evidence supporting a punitive award is scrutinized less closely on appeal to the highest court here than in some other states.
The broad jury discretion that is the hallmark of the common law punitive damages system must be complemented by procedural safeguards that will ensure against excessive or otherwise inappropriate awards.26 We conclude that the Texas procedures for assessing and reviewing punitive damage awards, both at the trial and appellate levels, in some respects may not adequately ensure that such awards “are not grossly out of proportion to the severity of the offense and have some understandable relationship to compensatory damages.” Haslip, 499 U.S. at 22, 111 S.Ct. at 1045. Although we do not reach the question whether the procedures previously employed violated either the state or federal27 constitutions, pursuant to our authority under the common law, we announce two changes in the procedures by which punitive damages may be sought. In our view, these modifications will not only enhance the overall fairness and predictability of punitive damage awards, but they should also ameliorate some of the potential arbitrariness that inheres in our current system.
Although Transportation argues for procedural changes only on constitutional grounds, various amici curiae alternatively argue for common-law reform. Justice Kennedy suggested this approach in Haslip:
We do not have the authority, as do judges in some of the States, to alter the rules of the common law respecting the proper standard for awarding punitive damages and the respective roles of the jury and the court in making that determination. Were we sitting as state-court judges, the size and recurring unpredictability of punitive damages awards might be a convincing argument to reconsider those rules or to urge a reexamination by the legislative authority.
499 U.S. at 42, 111 S.Ct. at 1056 (Kennedy, J., concurring). Other state supreme courts have recently modified their punitive damage procedures without holding those procedures unconstitutional. See Gamble v. Stevenson, 305 S.C. 104, 406 S.E.2d 350 (1991); Hodges v. S.C. Toof & Co., 833 S.W.2d 896 (Term.1992); Crookston v. Fire Ins. Exch., 817 P.2d 789 (Utah 1991).
A. Bifurcation
We held in Lunsford v. Morris, 746 S.W.2d 471 (Tex.1988), that evidence of a defendant’s net worth is relevant in determining the proper amount of punitive damages, and therefore may be subject to pretrial discovery. This decision aligned Texas with the overwhelming majority of other jurisdictions on this issue. See Lunsford, 746 S.W.2d at 472 n. 2. As we noted in Luns-ford, the amount of punitive damages necessary to punish and deter wrongful conduct depends on the financial strength of the defendant. “That which could be an enormous penalty to one may be but a mere annoyance to another.” Id. at 472.
*30However, evidence of a defendant’s net worth, which is generally relevant only to the amount of punitive damages, by highlighting the relative wealth of a defendant, has a very real potential for prejudicing the jury’s determination of other disputed issues in a tort case. We therefore conclude that a trial court, if presented with a timely motion, should bifurcate the determination of the amount of punitive damages from the remaining issues. See Wal-Mart, 868 S.W.2d at 829-82 (Gonzalez, J., concurring). Under this approach, the jury first hears evidence relevant to liability for actual damages, the amount of actual damages, and liability for punitive damages (e.g., gross negligence), and then returns findings on these issues. If the jury answers the punitive damage liability question in the plaintiff’s favor, the same jury is then presented evidence relevant only to the amount of punitive damages, and determines the proper amount of punitive damages, considering the totality of the evidence presented at both phases of the trial.
At least thirteen states now require bifurcation of trials in which punitive damages are sought.28 Ten of these, California, Georgia, Kansas, Missouri, Montana, Nevada, Ohio, Tennessee, Utah, and Wyoming, generally follow the procedure outlined above, in which the amount of punitive damages is bifurcated from the remaining issues. The other states require bifurcation of the entire punitive damage claim, including liability and amount. We believe the former approach is preferable, as some of the evidence relevant to punitive damage liability, such as evidence of gross negligence, will also be relevant to liability for actual damages. Bifurcating only the amount of punitive damages therefore eliminates the most serious risk of prejudice, while minimizing the confusion and inefficiency that can result from a bifurcated trial. See Lunsford, 746 S.W.2d at 477 (Phillips, C.J., dissenting on motion for reh’g).29
The issue in this Court is not whether bifurcation of punitive damage claims is constitutionally required, but whether our system of imposing punitive damages, on the whole, provides adequate procedural safeguards to protect against awards that are grossly excessive. Concluding that the current procedures are not adequate, we hereby adopt the requirement of bifurcated trials in punitive damage cases.
B. Court of Appeals Review of the Evidence
A court of appeals may vacate a damage award or suggest a remittitur only if the award is “so factually insufficient or so against the great weight and preponderance of the evidence as to be manifestly unjust.” Pope v. Moore, 711 S.W.2d 622, 624 (Tex.1986); see Pool, 716 S.W.2d at 635; In Re King’s Estate, 150 Tex. 662, 244 S.W.2d 660, 661 (1951). A court of appeals is also governed by this same standard when reviewing a trial court’s suggestion of remittitur. Larson v. Cactus Util. Co., 730 S.W.2d 640 (Tex.1987). While we do not alter this level of *31appellate deference,30 we emphasize that courts of appeals must carefully scrutinize punitive awards to ensure that they are supported by the evidence. We have already held in Pool that courts of appeals, when reversing on insufficiency grounds, should detail the evidence in their opinions and explain why the jury’s finding is factually insufficient or is so against the great weight and preponderance of the evidence as to be manifestly unjust. 715 S.W.2d at 635. Due to the jury’s broad discretion in imposing punitive damages, we believe that a similar type of review is appropriate when a court of appeals is affirming such an award over a challenge that it is based on insufficient evidence or is against the great weight and preponderance of the evidence. This will ensure careful appellate review of the punitive award, and allow this court to determine whether the court of appeals correctly applied the Kraus factors. Thus, we hold that the court of appeals, when conducting a factual sufficiency review of a punitive damages award, must hereafter detail the relevant evidence in its opinion, explaining why that evidence either supports or does not support the punitive damages award in light of the Kraus factors.
C. Clear and Convincing Evidence
Clear and convincing evidence is “that measure or degree of proof which will produce in the mind of the trier of fact a firm belief or conviction as to the truth of the allegations sought to be established.” State v. Addington, 588 S.W.2d 569, 570 (Tex.1979); Tex.Fam.Code § 11.16(c) (Vernon 1986). Many states have raised the eviden-tiary burden for gross negligence and punitive damages from “a preponderance of the evidence” to “clear and convincing evidence.” See Linthicum, 723 P.2d at 681; Masaki v. General Motors Corp., 71 Haw. 1, 780 P.2d 566, 575-75 (1989); Tuttle v. Raymond, 494 A.2d 1353, 1363 (Me.1985); Owens-Illinois, Inc. v. Zenobia, 325 Md. 420, 601 A2d 633, 657 (1992); Hodges v. S.C. Toof & Co., 833 S.W.2d 896, 901 (Tenn.1992); Wangen v. Ford Motor Co., 97 Wis.2d 260, 294 N.W.2d 437, 443 (1980); Ala.Code § 6-11-20 (Supp.1993); Alaska Stat. § 09.17.020 (Supp.1991); Cal.Civ.Code § 3294(a) (West Supp.1994); GaCode Ann. § 51-12-5.1(b) (Supp.1991); Ind.Code Ann. § 34-4-34-2 (Burns Supp. 1993); Iowa Code Ann. § 668A.l(l)(a) (West 1987); Kan.StatAnn. § 60-3701(c) (Supp. 1992); Ky.Rev.StatAnn. § 411.184(2) (Mi-chie/Bobbs-Morril Supp.1993); Minn.Stat. Ann. § 549.20(l)(a) (West Supp.1994); Mont. Code Ann. § 27-1-221(5) (1991); Nev.Rev. StatAnn. § 42.005(1) (Supp.1993); N.D.Cent.Code § 32-03.2-11 (Supp.1993); Ohio Rev.Code Ann. § 2307.80(A) (Anderson 1991); OR.Rev.Stat. § 30.925 (1991); S.C.Code Ann. § 15-33-135 (Law Coop.Supp. 1991); Utah Code Ann. § 78-18-1 (Supp. 1992); see also Colo.Rev.Stat. § 13-25-127(2) (1989) (requiring proof beyond a reasonable doubt); FlaStatAnn. Ch. 768.-73(l)(b) (Harrison Supp.1992) (punitive damages in excess of cap allowed if claimant demonstrates to the court by clear and convincing evidence that award is not excessive); OklaStatAnn. tit. 23, § 9 (West 1987) (clear and convincing evidence required to award punitive damages in excess of cap); S.D.Codified Laws Ann. § 21-1-4.1 (1987) (discovery on the punitive damages claim and submission of the punitive damages claim to the jury cannot occur unless the court finds “after a hearing and based on clear and convincing evidence” that there is a reasonable basis to believe that the defendant engaged in willful, wanton, or malicious conduct).
This change has also been endorsed by commentators, see Dorsey D. Ellis, Punitive Damages, Due Process, and the Jury, 40 Ala.Rev. 975, 993-995 (1989); American Law Institute, Reporters’ Study, II Enterprise Responsibility for Personal Injury— Approaches to Legal and Institutional Change 248-249 (1991), and favorably noted by the Supreme Court. Haslip, 499 U.S. at 23 n. 11, 111 S.Ct. at 1046 n. 11 (while not *32constitutionally required, “[t]here is much to be said in favor of [the clear and convincing standard] or, even, ‘beyond a reasonable doubt,’ ... as in the criminal context”).
Notwithstanding this support in other jurisdictions,31 our Legislature expressly considered and rejected the clear and convincing evidence standard for punitive damages as part of its 1987 tort-reform legislation. See Montford & Barber, 25 Hous.L.Rev. at 333. As this higher burden is not constitutionally compelled, see Haslip, 499 U.S. at 23 n. 11, 111 S.Ct. 1046 n. 11, the usual preponderance standard would remain the law in cases governed by the Act. Inasmuch as application of different evidentiary standards, depending on whether a case is governed by tort-reform legislation would be unnecessarily confusing, we decline to adopt the clear and convincing standard pursuant to our common-law authority at this time.
D. Trial Court Articulation
The Supreme Court emphasized in Haslip that trial courts in Alabama are required to scrutinize punitive damage awards, reflecting on the record their reasons for upholding or refusing to uphold the award. 499 U.S. at 20, 111 S.Ct. at 104445. The Court concluded that this procedure ensured “meaningful and adequate review by the trial court whenever a jury has fixed the punitive damages.” Id. In Texas, trial courts have not been required to articulate their reasons for upholding punitive awards, or even to take affirmative steps to consider post trial objections to a punitive damage award.
As noted by the Utah Supreme Court in adopting this requirement:
The reason that any determination as to whether the jury exceeded its proper bounds is best made in the first instance by the trial court is that the trial judge is present during all aspects of the trial and listens to and views all witnesses. Therefore, he or she can best determine if the jury has acted with “passion or prejudice” and whether the award is too small or too large in light of the evidence.
Crookston v. Fire Ins. Exch., 817 P.2d 789, 804 (Utah 1991). Meaningful trial court review is especially important in Texas because of the deference which our courts of appeals must give to jury verdicts, see Pool, 715 S.W.2d at 635, and the absence of this court’s jurisdiction to review factual sufficiency points.
At least eight jurisdictions now expressly require the trial court to articulate its reasons for refusing to disturb a punitive damage award. See Hammond v. City of Gadsden, 493 So.2d 1374, 1379 (Ala.1986); O’Dell v. Basabe, 119 Idaho 796, 810 P.2d 1082, 1092 (1991); Medical Mut. Liab. Ins. Soc’y v. B. Dixon Evander & Assocs., 92 Md.App. 551, 609 A.2d 353, 368-69 (1992) (not per se required, but trial court’s failure to explain reasoning may constitute abuse of discretion); GN Danavox, Inc. v. Starkey Labs., Inc., 476 N.W.2d 172, 177 (Minn.App.1991, review denied), cert. denied, — U.S.-, 112 S.Ct. 2940, 119 L.Ed.2d 565 (1992); Gamble v. Stevenson, 305 S.C. 104, 406 S.E.2d 350, 354 (1991); Hodges v. S.C. Toof & Co., 833 S.W.2d 896, 902 (Tenn.1992); Crookston v. Fire Ins. Exch., 817 P.2d 789, 811 (Utah 1991) (requiring trial court articulation whenever ratio of punitive damages to actual damages exceeds 3:1); Games v. Fleming Landfill, Inc., 186 W.Va. 656, 413 S.E.2d 897, 910 (1991). Most of these jurisdictions have adopted this requirement after Haslip. Moreover, several federal appellate courts reviewing punitive damage awards since Haslip have remanded the cases to the trial court with instructions to articulate the reasons for upholding the award. See Cole v. Control Data Corp., 947 F.2d 313, 321 (8th Cir.1991); Union Nat’l Bank v. Mosbacher, 933 F.2d 1440, 1448 (8th Cir.1991), cert. denied, — U.S. -, 112 S.Ct. 870, 116 L.Ed.2d 775 (1992); American Employers Ins. Co. v. Southern Seeding Servs., Inc., 931 F.2d 1453, 1458 (11th Cir.1991); Robertson Oil Co. v. Phillips Petroleum Co., 930 F.2d 1342, 1347 (8th Cir.1991).
*33We are sensitive to the demands already placed on our trial courts, which are overworked and understaffed. More than most states, Texas permits local governments to decide what levels of staffing, housing and equipment are appropriate to operate the state courts. National CenteR FOR State COURTS, SURVEY (1985). While some counties make excellent provision for their courts, in many instances this “reliance on the counties to fund most of the courts is one source of chronic fiscal crisis.” AmeRican Bah Association Special Committee on Funding The COURT System, Funding The Justice System: A Call To Action 54 (1992) (analyzing Texas court funding). As a result, Texas trial courts are generally badly understaffed. See id. A recent survey indicates that about seventy percent of Texas district judges have no access to secretarial services, while even fewer have paid law clerks. Office Of COURT ADMINISTRATION OF TEXAS, DISTRICT COURT RESOURCES Survey (June 1990).
Although we have on occasion suggested that our trial courts make specific findings not required by statute or rule, see Chrysler v. Blackmon, 841 S.W.2d 844 (1992), we do not do so lightly. While such a practice would facilitate meaningful post-verdict review of punitive damage awards at both the trial and appellate levels, we will not require such findings as a prerequisite for a judgment including punitive damages. We do observe that such findings “would obviously be helpful,” Transamerican Natural Gas v. Powell, 811 S.W.2d 913, 919 n. 9 (Tex.1991), and urge that such findings be made to the extent practicable.
For the foregoing reasons, we reverse the judgment of the court of appeals and remand this case for a new trial in accordance with this opinion.
. The manner and extent to which due process requires procedural safeguards and substantive limits on punitive damages is a matter of continuing controversy throughout the United States. We note that the United States Supreme Court has granted certiorari on this issue frequently in recent years, but has given lower courts no bright-line guidance. See Browning-Ferris Indus., Inc. v. Kelco Disposal, Inc., 492 U.S. 257, 109 S.Ct. 2909, 106 L.Ed.2d 219 (1989); Pacific Mut. Life Ins. Co. v. Haslip, 499 U.S. 1, 111 S.Ct. 1032, 113 L.Ed.2d 1 (1991); TXO Prod. Corp. v. Alliance Resources Corp., — U.S. -, 113 S.Ct. 2711, 125 L.Ed.2d 366 (1993). The Court has recently granted certiora-ri again on this important issue. See Oberg v. *13Honda Motor Co., 316 Or. 263, 851 P.2d 1084 (1993), cert. granted, — U.S. -, 114 S.Ct. 751, 127 L.Ed.2d 69 (1994). State supreme courts and legislatures have adopted a variety of measures to regulate the frequency and magnitude of punitive damages awards, including absolute caps on punitive awards or maximum ratios with respect to actual damages, see, e.g., Tex.Civ.Prac. & Rem.Code § 41.007 (Vernon Supp.1994), bifurcated trials upon proper motion, see Wal-Mart Stores, Inc. v. Alexander, 868 S.W.2d 322, 329-32 (Tex.1993) (Gonzalez, J., concurring), or elevated burdens of proof such as “beyond a reasonable doubt,” or "clear and convincing evidence.” See Haslip, 499 U.S. at 23 n. 11, 111 S.Ct. at 1046 n. 11. Many states following Haslip have required detailed jury instructions explaining the factors governing the size of punitive damages awards. See generally Alamo Nat'l Bank v. Kraus, 616 S.W.2d 908, 910 (Tex.1981); Haslip, 499 U.S. at 28-30, 111 S.Ct. at 1048-50 (upholding award based on Alabama list of governing factors).
. Transportation also cites Torchia v. Aetna Casualty & Sur. Co., 804 S.W.2d 219 (Tex.App.—El Paso 1991, writ denied). In that case, however, the claimant’s bad-faith cause of action was expressly included in the parties’ settlement agreement.
. The reservation, which is quoted in full above, refers to specific paragraphs of Moriel’s Fifth Amended Counterclaim, which is not in the appellate record. Transportation does not contend, however, that Moriel's claim for punitive damages falls outside of the referenced paragraphs.
. "Contract liability is imposed by the law for ... a single limited interest, that of having contracts performed. Quasi-contractual liability is created for the prevention of unjust enrichment. The criminal law is concerned with the protection ... of the public at large_ There remains a body of law which is directed toward the compensation of individuals, rather than the public, for losses they have suffered within the scope of their legally recognized interests generally ... where the law considers that compensation is required. This is the law of torts- The purpose of the law of torts is to adjust these losses, and to afford compensation for injuries sustained as a result of the conduct of another." W. Page Keeton, et. al., Prosser and Keeton On The Law Of Torts § 1, at 5-6 (5th ed. 1984) (emphasis added).
.The traditional justifications for criminal punishment include (l) rehabilitation of the defendant, Wayne R. LaFave & Austin W. Scott, Jr, Substantive Criminal Law § 1.5(a)(3) (West ed. 1986), (2) disabling, id. at § 1.5(a)(2), or deterring, id. at § 1.5(a)(1), the defendant from committing further crime, (3) deterring persons other than the defendant from crime, id. at § 1.5(a)(4), (5), (4) desert or retribution — punishing the defendant because he or she deserves to be pun- • ished, id. at § 1.5(a)(6), and (5) vengeance — satisfying the urge of the community to return to the defendant the injury done. Oliver W. Holmes, Jr, The Common Law 40-41 (1881).
. "Unlike compensatory damages, which are purely civil in character, punitive damages are, by definition, punishment. They operate as ... fines levied ... to advance governmental objectives.” Haslip, 499 U.S. at 47, 111 S.Ct. at 1058-59 (1991) (O'Connor, J., dissenting).
. "It is generally agreed that punitive damages are a windfall to the plaintiff and not a matter of right....” Prosser and Keeton § 2 at 14; see also Haslip, 499 U.S. at 11, 111 S.Ct. at 1040 (describing punitive damages as a “windfall”) (quoting Newport v. Fact Concerts, Inc., 453 U.S. 247, 271, 101 S.Ct. 2748, 2762, 69 L.Ed.2d 616 (1981) (“The impact of such a windfall recovery is likely to be both unpredictable and, at times, substantial.”) (emphasis added)).
. It is not apparent why JUSTICE DOGGETT mischaracterizes Hudson Energy as a means of "ffee[ing] from responsibility” insurers that "willfully ... [delay] or den[y] coverage with no reasonable basis,” or as creating "a perverse incentive for insurers to manufacture ... coverage disputes” for the purpose of delay. 879 S.W.2d 10, 36. By definition, such insurers are guilty of bad faith. Aranda v. Insurance Co. of N. Am., 748 S.W.2d 210 (Tex.1988). What he calls "the bona fide dispute rule” is nothing more than a shorthand notation for the observation that the parties to an insurance contract will sometimes have a good faith disagreement about coverage. Under such circumstances, the parties may require a court to interpret policy language for them, or a jury to resolve factual disputes. Simply because the parties go to court does not raise an issue of the insurer's bad faith. We do agree with JUSTICE DOGGETT, however, when he says that "[C]laims for insurance contract coverage are distinct from those in tort for bad faith; resolution of one does not determine the other.” 879 S.W.2d at 40.
. We disagree with JUSTICE DOGGETT’S reading of the authorities cited. 879 S.W.2d at 40, n. 16, n. 17, n. 18. In determining whether contract damages alone will support punitive damages against a breaching insurer, each of these jurisdictions adopts a "separate injury” rationale as opposed to a "same injury” rule. Although not all of these jurisdictions recognize the tort of bad faith, each decision is relevant because each uses similar standards to govern punitive damages in this context.
. Consistent with our earlier observations about the quasi-criminal nature of punitive damages, it is noteworthy that the civil standard for gross negligence bears a remarkable resemblance to the statutory definition of criminal recklessness:
A person acts recklessly, or is reckless ... when he is aware of but consciously disregards a substantial and unjustifiable risk that the circumstances exist or the result will occur. The risk must be of such a nature and degree that its disregard constitutes a gross deviation from the standard of care that an ordinary person would exercise under all the circumstances as viewed from the actor’s standpoint.
Tex.Penal Code Ann. § 6.03(c) (Vernon 1974). This definition was adopted from the American Law Institute Model Penal Code. Under the Model Penal Code, both criminal negligence and criminal recklessness require extremely risky behavior:
A person acts with criminal negligence ... when he ought to be aware of a substantial and unjustifiable risk that the circumstances exist or the result will occur. The risk must be of such a nature and degree that the failure to perceive it constitutes a gross deviation from the standard of care that an ordinary person would exercise under all the circumstances as viewed from the actor’s standpoint.
TexPenal Code Ann. § 6.03(d) (Vernon 1974). Criminal negligence and recklessness differ from one another only in terms of mental state. A criminally negligent defendant “ought to be aware" of a "substantial and unjustifiable" risk, while a reckless defendant is subjectively aware of an identical risk but disregards it. Importantly, nobody is subject to criminal punishment if they are aware of a relatively minor risk or simply negligent.
It should not be surprising that the civil definition of gross negligence and the criminal definition of recklessness are virtually identical. Both serve the same purpose — identifying when it is appropriate to punish an individual for consciously disregarding an unjustifiable risk.
. This statutory change was prompted, in part, by a 1985 case that suggested that common law gross negligence may be established by reference to a purely objective standard — what "a reasonable person would have realized” trader the same or similar circumstances. See Williams v. Steves Indus. Inc., 699 S.W.2d 570, 573 (Tex.1985).
. Thus, it has been argued that under Burk Royalty, gross negligence is simply "a conscious decision to act negligently.” . This argument eliminates the distinction between negligence and gross negligence because a jury can always infer that a person knew what he or she was doing.
. Burk Royalty described gross negligence as involving "peril” to the injured party. Burk Royalty, 616 S.W.2d at 922.
. The justification issue is essential for both courts and juries in determining whether a defendant was grossly negligent when the defendant introduces evidence of mitigating circumstances. See, e.g., Burk Royalty, 616 S.W.2d at 926 (Greenhill, C.J., concurring):
For example: there is evidence that the defendant, in a non-defective car, continues to drive at 65 miles per hour into a small town. The defendant runs a red light and passes a car over the center line before there is an accident. That is "some evidence” of gross negligence. If the court only considers that evidence, it must affirm a gross negligence finding. Then the court finds that the defendant’s wife and daughters, or other persons, are bleeding to death in the back seat of defendant's car; and they will die if they do not receive immediate medical attention. Will the court then use the "traditional no evidence” test to evaluate gross negligence? There may be negligence, yes. But is there conscious indifference to human life?
The answer, of course, is that we should apply our traditional "no evidence” test. By taking an extreme risk, this hypothetical defendant greatly increased the likelihood (1) that an accident would prevent his wife and daughters from ever obtaining medical care, and (2) that additional medical attention would be required for both himself and bystanders. Jurors could nevertheless decide that the risk was excusable under the circumstances, which is their prerogative.
Assuming that the jury decides this hypothetical defendant was grossly negligent, then this court could only reverse the verdict if the evidence in the record were legally insufficient to permit an inference that the defendant had actual subjective awareness of the risk he created. See William Powers, Jr. & Jack Ratliff, Another Look at “No Evidence" and "Insufficient Evidence," 69 TexJL.Rev. 515, 522-23 (1991).
. In contrast, an act or omission that creates a great peril but results in only a minor injury may be grossly negligent. If, for example, a person fires a gun randomly into a crowd of schoolchildren, but the shooting only results in the destruction of a pair of sunglasses, the defendant would still be grossly negligent. See TXO Prod. Corp. v. Alliance Resources Corp.,-U.S.-,-, 113 S.Ct. 2711, 2720-21, 125 L.Ed.2d 366, 380 (1993).
. A literally correct statement of the law is, "Only if gross negligence, intentional injury, fraud or malice accompany the bad faith tort are punitive damages proper." This case deals only with gross negligence, however. We therefore express no opinion on the circumstances in which an insurer’s intentional misconduct will justify punitive damages.
. Transportation has not challenged recovery of mental anguish damages for its bad faith or the legal sufficiency of Ae evidence concerning Mor-iel’s mental anguish.
. The Workers’ Compensation Act makes the employee’s obligation to pay medical bills secondary to that of the insurer. Smith v. Stephenson, 641 S.W.2d 900, 902 (Tex.1982). Moreover, it prohibits health care providers from trying to collect from the employee unless and until the IAB rules that the carrier is not responsible for payment. Id. at 903. Therefore, Transportation did not expose Moriel to any risk of legitimate collection efforts by his health care providers when it simply brought a dispute before the IAB. Transportation cannot be held legally accountable on the assumption that Moriel’s providers would try to collect in violation of the law.
. Less than two weeks before we issued this opinion, Transportation moved to amend its application for writ of error to challenge the legal sufficiency of the bad faith findings. This court may allow such an amendment “at any time when justice requires upon such reasonable terms as the court may prescribe." TexR.App.P. 131(h). Transportation could have, but did not, raise this point of error in its original application; we have overruled its motion to amend.
. Moriel testified that he was embarrassed because Dr. Perez had not been paid for the services he rendered, causing Moriel to terminate the therapy prematurely. Thus, he testified that Transportation's failure to pay the Perez bill, at least indirectly, hindered his psychiatric treatment. However, there is no evidence in the record that this was ever communicated to Transportation.
. JUSTICE DOGGETT mistakenly suggests that punitive damages are the only deterrent available at law, and that the absence of punitive damages should be taken as an indication that a tort is "not worth deterring." 879 S.W.2d at 37. We disagree. Not only does the prospect of being sued and having to defend oneself have a deterrent effect in most circumstances, but the threat of compensatory damages also serves as a deterrent, even for such "repeat-players” in litigation as insurance companies. In cases in which punitive damages are not appropriate, an insurer denying a valid claim can face a judgment that includes prejudgment interest and the claimant’s attorneys' fees, over and above the policy benefits. Moreover, if an insurer denies its insured's claim without any reasonable basis or knowingly engages in a statutorily prohibited practice, then it faces either extracontractual bad faith damages or statutory treble damages, respectively. Finally, insurers must ordinarily pay their own litigation costs even if they prevail on the coverage issue.
. The jury instruction used in this case is more comprehensive than that in Haslip, and thus would presumably pass scrutiny under the United States Constitution. It provided as follows:
The term "punitive damages" is an amount which you may, in your discretion, award as an example to others and as a penalty or by way of punishment, in addition to any amount you may have found as actual damages.
In determining that amount, you may consider—
1. the nature of the wrong,
2. the frequency of the wrongs committed,
3. the character of the conduct involved,
4. the degree of culpability of the wrongdoer,
5. the situation and sensibilities of the parties concerned,
6. the extent to which such conduct offends a public sense of justice and propriety, and
7. the size of the award needed to deter similar wrongs in the future.
. These standards include:
a) whether there is a reasonable relationship between the punitive damages award and the harm likely to result from the defendant’s conduct as well as the harm that actually has occurred; b) the degree of reprehensibility of the defendant's conduct, the duration of that conduct, the defendant’s awareness, any concealment, and the existence and frequency of similar past conduct; (c) the profitability to the defendant of the wrongful conduct and the desirability of removing that profit and of having the defendant also sustain a loss; (d) the "financial position” of the defendant; (e) all the costs of litigation; (f) the imposition of criminal sanctions on the defendant for its conduct, these to be taken in mitigation; and (g) the existence of other civil awards against the defendant for the same conduct, these also to be taken in mitigation.
499 U.S. at 21-22, 111 S.Ct. at 1045-46.
. Appellate review in Alabama determines whether the amount of the award is excessive under the circumstances of the case, but the Court viewed this as a procedural due process protection, as it provides a post-hoc check on the jury’s discretion. 499 U.S. at 22-23, 111 S.Ct. at 1045-46. Justice O’Connor questioned whether appellate review, even if it did ensure that the amount of the punitive award was not excessive, could cure a jury instruction she found to be unconstitutionally vague. Id. at 43-44, 111 S.Ct. at 1056-57 (O’Connor, J., dissenting).
. The burden in Texas for reversing a punitive damages verdict on appeal does not appear functionally different from that in Alabama and West Virginia. A Texas court of appeals may overturn the jury’s verdict and order a new trial if the verdict "is so against the great weight and preponderance of the evidence as to be manifestly unjust.” Pool, 715 S.W.2d at 635; In Re King’s Estate, 150 Tex. 662, 244 S.W.2d 660, 661 (1951). In Alabama, the "presumption of correctness” accorded to the jury's verdict may be overcome by “a clear showing that the amount of the verdict is excessive,” Wilson v. Dukona Corp., 547 So.2d 70, 72 (Ala.1989), while in West Virginia an award may be reversed on appeal if it is "so unreasonable as to demonstrate such passion and prejudice that a new trial is warranted.” TXO Prod. Corp. v. Alliance Resources Corp., 187 W.Va. 457, 419 S.E.2d 870, 890 (1992), aff'd,-U.S. -, 113 S.Ct. 2711, 125 L.Ed.2d 366 *29(1993). Although the words used are quite different, we are not convinced that there is a meaningful difference among these respective standards that would produce a different result.
. The only aspect of Texas procedure to clearly meet or exceed the minimal requirements has been the instructions given to juries.. A Texas jury, although possessing significant discretion in awarding punitive damages, receives more guidance than did the jury in Haslip. See supra note 21. Texas jurors are instructed on the factors set forth in Kraus, 616 S.W.2d at 910, and thus determine the amount of the award based on the same criteria that the trial and appellate courts use in determining whether the award is excessive.
. The United States Court of Appeals for the Fifth Circuit has considered the existing Texas procedures for imposing punitive damage awards in light of Haslip, and concluded that they do not violate federal due process. Glasscock v. Armstrong Cork. Co., 946 F.2d 1085 (5th Cir.1991), cert. denied, -U.S. -, 112 S.Ct. 1778, 118 L.Ed.2d 435 (1992).
. Cal.Cw.Code § 3295(d) (West Supp.1993); Ga. Code Ann. § 51-12-5.1(d) (Supp.1992); Kan.Stat. Ann. § 60-3701(a) (Supp.1993); Minn.StatAnn. § 549.20(4) (West Supp.1993); Mo.Ann.Stat. § 510.263 (Vernon Supp.1992); Mont.Code Ann. § 27-1-221(7) (1991); Nev.Rev.Stat. § 42.005(3) (1991); N.J.Stat.Ann. § 2A:58C-5(b) (West 1987); N.D.Cent.Code § 32-03.2-11(2) through (4) (1993); Ohio Rev.Code Ann. § 2315.21(C) (Baldwin 1993); Utah Code Ann. § 78-18-1(2) (1992); Campen v. Stone, 635 P.2d 1121, 1132 (Wyo. 1981); Hodges v. S.C. Toof & Co., 833 S.W.2d 896 (Tenn.1992).
. Despite the authority of trial courts to order separate trials under Tex.R.Civ.P. 174(b), we have previously held that liability and damages may not be bifurcated in a personal injury action. Iley v. Hughes, 158 Tex. 362, 311 S.W.2d 648, 651 (1958). Citing our “long standing policy and practice” against "piecemeal trials,” id. 311 S.W.2d at 651, the court said: “[T]he public interest, the interests of litigants and the administration of justice [are] better served by rules of trial which avoid a multiplicity of suits.” Id. Arguably, this holding would apply to punitive, as well as actual, liability and damages. But see Beverly Enterprises of Texas, Inc. v. Leath, 829 S.W.2d 382, 387 (Tex.App.—Waco 1992, no writ) (the court, without citing Iley v. Hughes, held that a trial court does have discretion under Rule 174(b) to bifurcate the amount of punitive damages from punitive liability); Miller v. O'Neill, 775 S.W.2d 56, 59 (Tex.App.—Houston [1st Dist.] 1989, orig. proceeding) (same). Although we remain resolute that piecemeal trials as a general rule should be avoided, given the importance of the considerations we have discussed, we conclude that punitive damage cases should be the exception to the rule.
. The Supreme Court in Haslip distinguished Alabama's appellate review, under which the court applies detailed substantive standards, from the Vermont scheme, which allows an award to be set aside only if it is “manifestly and grossly excessive.” 499 U.S. at 21 n. 10, 111 S.Ct. at 1045 n. 10. Although Texas also uses a "manifestly unjust” standard, our review is actually more aligned with Alabama’s in that the court of appeals makes this determination by applying the specific Kraus factors.
. Accord. Ford Motor Co. v. Durrill, 714 S.W.2d 329, 347 (Tex.App.—Corpus Christi 1986), judgm't vacated by agreement, 754 S.W.2d 646 (Tex.1988) (acknowledging, although ultimately rejecting, the strength of arguments in support of clear and convincing standard for punitive damages).