NAHMIAS, Justice.
Appellees, defendants Carlester Jackson and Warren Woodley Smith, allegedly conspired with Jerold Daniels to rob a drug dealer at gunpoint. The victim, however, also turned out to be armed, and he shot and killed Daniels in self-defense. A Cobb County grand jury [647]*647indicted Jackson and Smith on three counts of felony murder along with other offenses. The defendants moved to dismiss the felony murder counts pursuant to State v. Crane, 247 Ga. 779 (279 SE2d 695) (1981). The trial court granted the motion to dismiss, and the State now appeals, asking us to overrule Crane. After thorough review, we conclude that Crane should be overruled, and we therefore reverse. The causation issue presented should be decided by a properly instructed jury at trial, using the customary proximate cause standard.
This should be an easy case for a Georgia appellate court. The question presented is what the term “causes” means as used in the felony murder statute. See OCGA § 16-5-1 (c) (“A person also commits the offense of murder when, in the commission of a felony, he causes the death of another human being irrespective of malice.”). In cases both before and after Crane, this Court interpreted that very term to require “proximate causation.” In addition, there are dozens of other cases from this Court and the Court of Appeals, before and after Crane, that hold that the same term as used in other homicide statutes and in many other criminal and civil contexts means proximate cause.
This case is difficult only because of Crane. There, in a short opinion that did not mention any of Georgia’s extensive causation case law, the Court held that the word “causes” in the felony murder statute requires not proximate causation, but that the death be “caused directly” by one of the parties to the underlying felony. Id. at 779. Applying this new and more restrictive conception of causation, the Court concluded that a defendant cannot be found guilty of felony murder when the intended victim of the underlying felony kills the defendant’s accomplice, because that death is “caused directly” by the victim rather than the defendant. See id.
As shown below, the opinion in Crane was poorly reasoned, and perhaps because it is so incongruous with the rest of Georgia law, it has not been consistently applied by this Court or the Court of Appeals in the ensuing three decades. Its holding has not been applied uniformly in the specific context of felony murder, nor has its reasoning been followed in construing the same causation language in other homicide statutes. The relevant facts of this case, however, are almost identical to Crane’s, and so today we must either reaffirm Crane or reject it. After careful consideration, we have concluded that Crane must be overruled. Stare decisis is an important doctrine, but it is not a straightjacket. Crane’s age and statutory nature are outweighed by the other factors undermining its precedential authority, and it is important that the Court refute its reasoning to ensure that the case can no longer be cited in efforts to pollute other streams of our law.
[648]*648
The Factual and Procedural Background of This Case
1. The parties stipulated, for purposes of the motion to dismiss, that Jackson, Smith, and Daniels conspired to commit an armed robbery of someone who the defendants believed was a drug dealer. Daniels approached the intended victim armed with a handgun, with Jackson nearby and Smith waiting in the getaway vehicle. The victim, who was also armed, exchanged gunfire with Daniels, and he ultimately shot and killed Daniels in self-defense. Jackson and Smith were later arrested. The indictment charged the defendants with, among others offenses, felony murder. Tracking the statutory language, Count 1 alleged that both Jackson and Smith “did cause the death of Jerold Daniels, a human being, . . . while in the commission of a felony, to wit: Aggravated Assault.” The indictment charged Smith with two more counts, alleging that he caused Daniels’s death while in the commission of the felony of possession of a firearm by a convicted felon.
The defendants moved to dismiss the felony murder charges. They argued that because the victim fired the shot that killed their co-conspirator, they did not directly cause Daniels’s death. The trial court, bound by this Court’s decision in Crane, granted the motion to dismiss. The State filed this direct appeal under OCGA § 5-7-1 (a) (1), asking us to overrule Crane.
“Cause” in Georgia’s Homicide Statutes Means Proximate Cause
2. The felony murder statute provides that “[a] person also commits the offense of murder when, in the commission of a felony, he causes the death of another human being irrespective of malice.” OCGA § 16-5-1 (c) (emphasis supplied). As in Crane, the question in this case is whether a defendant who commits a felony whose intended victim kills a co-conspirator “causes” that death. The answer should be straightforward. Georgia is a proximate cause state. When another meaning is not indicated by specific definition or context, the term “cause” is customarily interpreted in almost all legal contexts to mean “proximate cause” — “[t]hat which, in a natural and continuous sequence, unbroken by any efficient intervening cause, produces injury, and without which the result would not have occurred.” Black’s Law Dictionary 1103 (5th ed. 1979).
Thus, this Court has explained that proximate cause is the standard for criminal cases in general. See, e.g., Skaggs v. State, 278 Ga. 19, 19-20 (596 SE2d 159) (2004) (In a criminal case, proximate cause exists when the accused’s “ ‘act or omission played a substantial part in bringing about or actually causing the (victim’s) injury or damage and . . . the injury or damage was either a direct [649]*649result or a reasonably probable consequence of the act or omission.’ ” (citations omitted)). We have also said that proximate cause is the standard for homicide cases in general. See, e.g., James v. State, 250 Ga. 655, 655 (300 SE2d 492) (1983) (“In Wilson v. State, 190 Ga. 824, 829 (10 SE2d 861) (1940), we set out the following test for determining causation in homicide cases: ‘Where one inflicts an unlawful injury, such injury is to be accounted as the efficient, proximate cause of death, whenever it shall be made to appear, either that (1) the injury itself constituted the sole proximate cause of the death; or that (2) the injury directly and materially contributed to the happening of a subsequent accruing immediate cause of the death; or that (3) the injury materially accelerated the death, although proximately occasioned by a pre-existing cause.’ ”).
Consistent with this general rule, we have held in many cases and for many decades that proximate causation is the standard for murder cases prosecuted under the murder statute, now codified as OCGA § 16-5-1. Thus, we have long held, in numerous cases, that proximate causation is the test for malice murder, a crime defined using the identical “he . . . causes” phrasing. See OCGA § 16-5-1 (a) (“A person commits the offense of murder when he unlawfully and with malice aforethought, either express or implied, causes the death of another human being.”).1 Finally, with respect to the statutory text at issue in this case, and in full accord with the general rule for criminal and homicide cases and with our construction of the identical language in subsection (a) of the same statute, we have repeatedly held, before and after Crane, that the phrase “he causes” in OCGA § 16-5-1 (c) establishes proximate causation as the standard for liability in felony murder cases.2
[650]*650Indeed, in virtually all of Georgia’s many homicide and feticide statutes, including the frequently charged voluntary and involuntary manslaughter and vehicular homicide statutes, the General Assembly has employed the same or very similar causation phrasing.3 And [651]*651to the extent those statutes have been interpreted by Georgia’s appellate courts, once again the term “cause” has been regularly construed as requiring proximate causation.4
[652]*652As an original matter, therefore, we would decide this case simply by applying the customary legal meaning of “cause,” which is supported by the ample precedent interpreting the felony murder provision at issue, its identical sister provision in the murder statute, and identical or substantially similar provisions in many other homicide statutes. We would hold that the phrase “he causes” as used in OCGA § 16-5-1 (c) requires the State to prove that the defendant’s conduct in the commission of the underlying felony proximately caused the death of another person. In the context of this case, proximate causation would exist if (to use “the rule” for felony murder that the Court stated a year after deciding Crane) the felony the defendants committed “directly and materially contributed to the happening of a subsequent accruing immediate cause of the death,” Durden, 250 Ga. at 329, or if (to use language from a case decided 16 years before Crane) “ ‘the homicide [was] committed within the res gestae of the felony’ . . . and is one of the incidental, probable consequences of the execution of the design to commit the robbery,” Jones, 220 Ga. at 902 (citations omitted).
Whether the evidence in this case would establish such proximate causation beyond a reasonable doubt is a harder question, in part because the stipulated facts we have before us are summary and the issue of proximate causation is so fact-intensive. That is why proximate cause determinations are generally left to the jury at trial. See McGrath, 277 Ga. App. at 829 (“What constitutes proximate cause is ‘undeniably a jury question and is always to be determined on the facts of each case upon mixed considerations of logic, common sense, justice, policy, and precedent.’ ” (citation omitted)).
The defendants here planned an armed robbery of someone they believed to be a drug dealer, who also turned out to be armed, an occurrence not unusual among drug dealers. When their co-conspirator Daniels approached the victim with a handgun to execute the robbery, the victim defended himself and killed Daniels. Perhaps more detailed evidence would show that, despite the dan[653]*653gerous and violent nature of armed robbery and drug dealing, circumstances existed that made the fatal result of the defendants’ felonious conduct improbable in this case, or made the drug dealer victim’s actions an “efficient intervening cause.” On the limited record before us, however, a jury could rationally conclude that the defendants’ felonies played a “substantial part in bringing about” their accomplice’s death when they confronted at gunpoint a drug dealer, whose deadly response could be viewed as a “reasonably probable consequence” of their acts. Skaggs, 278 Ga. at 19-20 (citations and punctuation omitted). Thus, as an original matter, we would have little hesitation reversing the trial court’s order and remanding the case for trial and decision by a jury properly charged on causation using language adapted from our proximate cause homicide cases.
State v. Crane
3. This is not, however, an original matter. The same legal issue was presented, in much the same factual scenario, nearly 30 years ago in Crane. In that case, Crane and three confederates were burglarizing a home when the homeowner shot and killed one of them in defense of himself and his property. See 247 Ga. at 779. The Court recognized that the case turned on whether the term “he causes,” as used in the felony murder statute, can extend to the death of an accomplice killed by the intended victim. Id. In its one-and-a-half page opinion, however, the Crane Court did not consider the customary legal meaning of “cause” or look to our then-existing case law interpreting that term as used in the felony murder statute, the malice murder statute, and homicide and other criminal statutes in general. Instead, the Court baldly asserted that it was faced with the choice between limiting felony murder to deaths “caused directly by one of the parties to the underlying felony” or construing the statute “to include also those deaths indirectly caused by one of the parties.” Id. (footnote omitted; emphasis supplied). Reflecting on the only two interpretations of “he causes” that it considered, the Court stated that “[w]e would, if allowed a choice, favor the construction which would criminalize the conduct involved in the present case.” Id. at 780. Because a criminal statute was being interpreted, however, the Court concluded that “we are constrained by principle to rule in behalf of the accuseds.” Id.
We agree that the rule of lenity would require the Court to adopt the interpretation that favored the accuseds if, after applying all other tools of statutory construction, the Court determined that “directly causes” and “indirectly causes” were the only possible [654]*654meanings of the word “causes” in OCGA § 16-5-1 (c) and that equally strong reasoning supported either interpretation, leaving the statute ambiguous. See Banta v. State, 281 Ga. 615, 617-618 (642 SE2d 51) (2007) (“ ‘The rule of lenity . . . applies only when, after consulting traditional canons of statutory construction, we are left with an ambiguous statute.’ ” (quoting United States v. Shabani, 513 U. S. 10, 17 (115 SC 382, 130 LE2d 225) (1994)). But the Crane Court did not apply the traditional canons of statutory construction before jumping to that conclusion, and the binary reading of the causation element proposed by the Crane Court finds no foundation in our legal tradition or our case law, none of which the Court mentioned.5 Indeed, other than Crane and cases discussing Crane, we have found not a single instance in our extensive causation case law where the Court has suggested that the word “causes” can mean only “directly causes” or “indirectly causes.”
To the contrary, we have consistently employed the more nuanced concept of proximate causation, which does not track the binary, and often unhelpful, direct-indirect dichotomy of Crane. Proximate causation imposes liability for the reasonably foreseeable results of criminal (or, in the civil context, tortious) conduct if there is no sufficient, independent, and unforeseen intervening cause. That definition would include, at least in some factual scenarios, a deadly response against one of the perpetrators by the intended victim of a dangerous felony like burglary or armed robbery.
The Inconsistent Application of Crane’s Holding
4. No later cases have bolstered Crane’s reasoning, nor do the dissents today make any effort to do so. Indeed, neither this Court nor the Court of Appeals has consistently applied Crane’s holding that the words “he causes” in the felony murder statute “require the death to be caused directly by one of the parties to the underlying felony.” 247 Ga. at 779 (footnote omitted). In nearly three decades, the Court has applied Crane wholeheartedly on just two occasions. [655]*655The first came a year after Crane, when the Court reversed a felony murder conviction where a police officer killed a bystander during a shootout with the defendant. See Hill v. State, 250 Ga. 277, 278-280 (295 SE2d 518) (1982).6 The second time was in Hyman v. State, 272 Ga. 492 (531 SE2d 708) (2000). Police came to Hyman’s home looking for a murder suspect, and he falsely told them that the suspect was not there. When the police were allowed to search the house, the suspect shot and killed one of the officers. See id. at 493. Hyman was charged with murder while in the commission of the felony of making a false statement, but the Court held that the “direct cause” of the officer’s death was the suspect, with whom Hyman was not acting in concert, and so under Crane his felony murder conviction was reversed. See 272 Ga. at 493. It is possible that the same result would have been reached under the proximate cause test, consideration of which the Hyman Court pretermitted. See id.
In another case, however, the Court upheld the defendant’s felony murder conviction based upon the death of a bystander killed by someone who was engaging in a gunfight with the defendant. See [656]*656Smith v. State, 267 Ga. 372, 375-376 (477 SE2d 827) (1996). To reach that result, the Court had to redefine the Crane test as whether the death of the bystander was “directly caused” by “a willing participant” (rather than a co-party) in the gunfight. 267 Ga. at 375. The Court struggled to distinguish Crane and Hill as cases in which “the homicides were not committed by either the defendant or someone acting in concert with him.” 267 Ga. at 376. The shooter in Smith, however, was plainly “one of the parties to the [defendant’s] underlying felony,” Crane, 247 Ga. at 779 (footnote omitted), and it is questionable whether someone charged with committing an aggravated assault against the defendant by shooting at him, see Smith, 267 Ga. at 372, n. 1, can really be said to be “acting in concert with him,” id. at 376.
In other cases since Crane, we have upheld felony murder convictions where the death could hardly be said to have been “caused directly” by the defendant’s acts. See McCoy v. State, 262 Ga. 699, 700 (425 SE2d 646) (1993) (upholding felony murder conviction by finding that the death of a firefighter who fell into a well behind a burning house and died of asphyxiation was “directly attributable” to the defendant’s felonious conduct in setting fire to the house); Durden, 250 Ga. at 329 (affirming felony murder conviction where a storeowner responding to a burglary died of a heart attack after exchanging shots with the defendant). In several other felony murder cases, we have simply ignored Crane and applied the proximate cause test. See, e.g., the post-1981 cases cited in footnote 2 above.
Moreover, if Crane's reasoning is solid and its holding deserving of precedential value, as Justice Thompson’s dissent suggests, see Dissenting Op. at 663, then the term “causes” and the identical or substantially similar causation language used in Georgia’s other homicide statutes should also be susceptible to the same “directly causes” versus “indirectly causes” ambiguity posited in Crane. And because all those statutes are also penal, the rule of lenity should require that the “directly causes” interpretation be applied in those contexts as well. But that has not happened. To the contrary, this Court and the Court of Appeals have continued to apply the traditional proximate cause standard in those situations. See, e.g., the post-1981 cases cited in footnotes 1 and 4 above.
Crane has caused the most tension in vehicular homicide cases, which, like felony murder cases, sometimes involve deaths that are “directly” caused by innocent third parties acting as a result of the defendant’s precipitating criminal acts. Thus, in Hill, this Court held that, under Crane, a defendant did not “cause” the death of another person and so was not guilty of felony murder when a police officer at whom the defendant was shooting shot back and killed an [657]*657innocent bystander. See 250 Ga. at 280. Yet the Court of Appeals, in a case involving almost the same causation language and a similar fact pattern, held that a defendant was guilty of vehicular homicide when a police officer from whom he was illegally fleeing bumped his truck in an effort to stop it (much like an officer returning fire to stop an ongoing felony) and caused the truck to crash, killing an innocent bystander (a baby riding in the truck). See Pitts, 253 Ga. App. at 374. The Pitts court reached this conclusion by simply ignoring Crane and applying the usual proximate cause test. See id. at 374-375.
Similarly, in Ponder v. State, 274 Ga. App. 93 (616 SE2d 857) (2005), the defendant, who was under the influence and recklessly fleeing the police, caused a pursuing police car to veer into oncoming traffic, where the police car collided with a Buick, killing the officer. See id. at 94-96. Like the homeowner who fired the fatal shot in Crane, the “direct cause” of the officer’s death was the driver of the Buick. But the Court of Appeals, again without mention of Crane, upheld the conviction because the evidence supported the jury’s finding that the defendant’s criminal conduct was the proximate cause of the officer’s death. See 274 Ga. App. at 95-96.
In McGrath v. State, 277 Ga. App. 825 (627 SE2d 866) (2006), the chain of causation was even more indirect. McGrath, who was driving recklessly and under the influence on 1-85, crashed into a car driven by Kar. Both vehicles were wrecked, and McGrath and Kar were injured. Burroughs-Brown, a nurse, saw the wreck and stopped to assist. Another car driven by Ramirez, who could not see Burroughs-Brown until it was too late due to poor visibility, hit her. She was pinned briefly between Kar’s and Ramirez’s cars, but then she fell onto the highway, where two other vehicles ran over her. See id. at 826-827. Citing Crane, McGrath argued that he did not directly cause Burroughs-Brown’s death, and faithful application of Crane’s reasoning would indeed have required reversal. But the Court of Appeals again upheld the conviction under the proximate cause test. See McGrath, 277 Ga. App. at 828-830. In a footnote, the court distinguished Crane on the ground that it “involved the felony murder statute, which was subject to two interpretations” and asserted that “[s]uch is not the case here, since the vehicular homicide statute has been consistently interpreted and applied.” Id. at 830, n. 4. The Court of Appeals distinguished Crane similarly in an earlier vehicular homicide case. See Johnson, 170 Ga. App. at 434 {“Crane is clearly inapposite to the instant case where there is no evidence of indirect causation and which involves construction of an entirely different statute.”).
Vehicular homicide and felony murder may be defined in “entirely different” statutes, in terms of their Code sections, but the relevant causation language is indistinguishable, compare OCGA § [658]*65840-6-393 (a) {“Any person who, without malice aforethought, causes the death of another person through the violation of [various code sections] commits the offense of homicide by vehicle in the first degree.. ..” (emphasis supplied)), with OCGA § 16-5-1 (c) (“A person also commits the offense of murder when, in the commission of a felony, he causes the death of another human being irrespective of malice. ...” (emphasis supplied)). If Crane is good law, then this Court’s construction of the causation language in OCGA § 16-5-1 (c) should be binding on the Court of Appeals when it interprets the virtually identical causation language in the vehicular homicide statute. See Ga. Const, of 1983, Art. VI, Sec. VI, Par. VI (“The decisions of the Supreme Court shall bind all other courts as precedents.”). Crane is, however, no longer good law.
Stare Decisis Considerations
5. Stare decisis is an important principle that promotes the rule of law, particularly in the context of statutory interpretation, where our incorrect decisions are more easily corrected by the democratic process. See Smith v. Salon Baptiste, 287 Ga. 23, 30 (694 SE2d 83) (2010) (Nahmias, J., concurring specially). However, stare decisis is not an “ ‘inexorable command,’ nor ‘a mechanical formula of adherence to the latest decision.’ . . . Stare decisis is instead a ‘principle of policy.’ ” Citizens United v. Fed. Election Commn., 558 U. S. __ (130 SC 876, 920, 175 LE2d 753) (2010) (Roberts, C. J., concurring) (citations omitted). In considering whether to reexamine a prior erroneous holding, we must balance the importance of having the question decided against the importance of having it decided right. Id. In doing so, we consider factors such as the age of the precedent, the reliance interests at stake, the workability of the decision, and, most importantly, the soundness of its reasoning. See Montejo v. Louisiana, 556 U. S. __ (129 SC 2079, 2088-2089, 173 LE2d 955) (2009).
As demonstrated above, Crane’s reasoning is unsound and contrary to the body of our law. Crane’s holding may be workable in its specific context — the death of a co-party directly caused by the intended victim of the underlying felony. As just discussed, however, this Court and the Court of Appeals have been unable or unwilling to apply Crane’s reasoning to all felony murder cases, much less to the many other homicide statutes that use the same causation language. In addition, Crane affects no property or contract issues and establishes no substantive rights, so it creates no meaningful reliance interests. (To be sure, the potential conspiring felon who is well-read in the law might be slightly less deterred from committing a dangerous felony by the belief that if one of his co-conspirators is [659]*659killed by the intended victim or a police officer, he will not face a murder charge, but that is not the sort of reliance the law usually recognizes in the stare decisis analysis.)
That leaves, on the side of reaffirming Crane, only its age and its statutory nature. That is all Justice Thompson’s dissent relies upon. See Dissenting Op. at 663-665. Crane is indeed nearly three decades old, and in Crane and the only two subsequent cases in which we actually applied its holding, the Court expressly noted that the General Assembly could correct the result. See Crane, 247 Ga. at 780 (“The choice of whether or not the conduct in the present case should be violative of our criminal statutes lies with the General Assembly.”); Hyman, 272 Ga. at 493 (“If this result be viewed as a defect in our felony murder statute, the remedy lies with the legislature.” (quotingHill, 250 Ga. at 280)).7 “Without strong reason to set aside a long-standing interpretation,” Justice Thompson’s dissent says, “we will not do so in the face of legislative acquiescence.” Dissenting Op. at 664. But see Durrence v. State, 287 Ga. 213, 216, n. 5 (695 SE2d 227) (2010) (Thompson, J.) (unanimously overruling a 26-year-old statutory interpretation case in a footnote, briefly explaining why the precedent was decided incorrectly but not mentioning “legislative acquiescence”).
We have explained at length the strong reasons that exist to overrule Crane, which the dissents do not refute. Moreover, Crane’s odd reasoning and the inconsistent application of its holding by both appellate courts make resort to “legislative acquiescence” particularly dubious.8 In large part because our Court and the Court of [660]*660Appeals have not consistently applied Crane, it has not had the sort of obviously far-reaching effects that are likely to stimulate a legislative response. Moreover, prosecutors will only rarely go to the trouble of charging felony murder where Crane appears to apply, much less appealing the issue when the trial court follows our precedent (as the trial courts must). Consequently, most of Crane’s direct effect — the felony murder prosecutions that are never brought — goes unseen.
Furthermore, it is not clear how the General Assembly would go about correcting Crane. If the legislature revised the “he causes” language in OCGA § 16-5-1 (c) to say “he proximately causes,” without simultaneously revising all the other homicide statutes that use similar causation language (including the malice murder provision in subsection (a) of the same statute), the effort could backfire. We could expect to see appeals by defendants arguing that the legislature’s revision of one provision indicates that the language remaining in all the other provisions means something else — what we said such language meant in Crane, that is, “directly causes.” Nor do legislatures commonly undertake to enact the highly detailed amendment that would be required to respond very specifically to Crane — assuming that, in light of the inconsistent application of Crane, the General Assembly could even tell for sure what it needed to correct.
In light of these considerations, we do not believe “that we can properly place on the shoulders of [the General Assembly] the burden of the Court’s own error.” Girouard v. United States, 328 U. S. 61, 69 (66 SC 826, 90 LE 1084) (1946). “Certainly, stare decisis should not be applied to the extent that an error in the law is perpetuated,” Etkind v. Suarez, 271 Ga. 352, 357 (519 SE2d 210) (1999), and it would not foster the objectives of predictability, stability, and consistent development of legal principles to reaffirm a decision that branched away from the path of prior and subsequent causation law, has rarely been followed, and if truly followed would disrupt many areas of settled law.
Conclusion
6. For these reasons, we hereby overrule State v. Crane, 247 Ga. 779, and our subsequent cases relying upon Crane. We hold that the felony murder statute requires only that the defendant’s felonious conduct proximately cause the death of another person. We therefore reverse the order of the trial court and remand the case for the jury to decide the causation question at trial.
Judgment reversed and case remanded.
All the Justices concur, except Hunstein, C. J., and Benham and Thompson, JJ., who dissent.