The question presented by the parties on this appeal is whether a third party injured by an intoxicated person has a common-law cause of action against a tavernkeeper for the negligent sale of intoxicating beverages to a patron the tavern-keeper knew or should have known was intoxicated and who the tavernkeeper knew or should have known would be driving.

*486However, in its decision the trial court recognized that the public policy decision involved was broader than the liability of a licensed tavern operator and in so doing stated:

“. . . the time has come to object and lay to rest those social considerations which for many years past have exculpated negligent dispensers of liquor, whether they be commercial or social, . . .”

This court is again being asked to abrogate the common-law rule that dispensers of intoxicating beverages have no liability to third parties injured by the intoxicated patron. In 1970, this court refused to change the rule of nonliability. Garcia v. Hargrove, 46 Wis.2d 724, 176 N.W.2d 566 (1970). This decision was reaffirmed in Garcia v. Hargrove, 52 Wis.2d 289, 190 N.W.2d 181 (1971). The basis of the common-law rule is the theory that the proximate cause of the injury is the act of the purchaser in drinking the liquor, and not the act of the vendor in selling it. In Garcia, the majority based its retention of the common-law rule of nonliability on considerations of public policy. The court did not rely on the rationale that the drinking and not the sale was the proximate cause of the injury or that resolution of the issue had been preempted by the legislature. The court stated:

“Therefore, in the final analysis, the outcome of this case is not determined by application of ‘proximate cause’ or whether the legislature has or has not acted. The controlling consideration is one of public policy, and it is the determination of this court that reasons of public policy dictate liability not be extended as proposed by plaintiff.” Garcia v. Hargrove, supra, 46 Wis.2d at 732, 733.

The trial court here is creating a cause of action unknown at common law by rejecting public policy considerations vitally important to such a decision. When *487the ramifications of this action are considered grave questions arise regarding its wisdom.

Admittedly the specific factual context presented in this case lends itself to a superficial application of the principles of negligence. Certainly in view of the current practice of defining proximate cause by using the substantial factor test, lack of proximate cause as a basis for the common-law rule has been seriously eroded and can no longer be relied upon as a bar to the action. However, the considerations that were formerly incorporated into the concept of proximate cause remain in the law as public policy factors imposed by the court to cut off liability. Hass v. Chicago & North Western Ry. Co., 48 Wis.2d 321, 326, 179 N.W.2d 885 (1970). The label may have changed but the reason for the common-law rule remains. The responsibility for the act remains with the person whose voluntary intoxication resulted in the injury. The fact that today there may be more intoxicated persons behind the wheels of more autos is not a sufficient reason to change the rule. In those states with dram shop acts and those recognizing a common-law cause of action liability is not limited to the injuries negligently caused by intoxicated drivers. The action can also cover intentional assaults by intoxicated persons and may even cover injuries inflicted by the drinker on himself. Vance v. United States, 355 Fed. Supp. 756 (D. Alas. 1973) ; Hull v. Rund, 150 Colo. 425, 374 Pac.2d 351 (1962) ; Ramsey v. Anctil, 106 N.H. 375, 211 Atl.2d 900 (1965) ; Schelin v. Goldberg, 188 Pa. Super. 341, 146 Atl.2d 648 (1958). The problems posed by the presence of a seriously intoxicated person in a public place are no different now than they were in the past. The only difference is in numbers — now there are more people and perhaps more people driving in an intoxicated condition.

*488A court should alter a common-law rule where the reason for the rule no longer exists. Shier v. Freedman, 58 Wis.2d 269, 206 N.W.2d 166, 208 N.W.2d 328 (1973) ; Dippel v. Sciano, 37 Wis.2d 443, 155 N.W.2d 55 (1967) ; Moran v. Quality Aluminum Casting Co., 34 Wis.2d 542, 150 N.W.2d 137 (1967) ; Widell v. Holy Trinity Catholic Church, 19 Wis.2d 648, 121 N.W.2d 649 (1963) ; Holytz v. Milwaukee, 17 Wis.2d 26, 115 N.W.2d 618 (1962) ; Kojis v. Doctors Hospital, 12 Wis.2d 367, 107 N.W.2d 131 (1961). The public policy factors which militated against this cause of action in the past are as persuasive today.

Six policy reasons are recognized by this court for cutting off liability:

. . (1) The injury is too remote from the negligence; or (2) the injury is too wholly out of proportion to the culpability of the negligent tort-feasor; or (3) in retrospect it appears too highly extraordinary that the negligence should have brought about the harm; or (4) because allowance of recovery would place too unreasonable a burden on the negligent tort-feasor; or (5) because allowance of' recovery would be too likely to open the way for fraudulent claims; or (6) allowance of recovery would enter a field that has no sensible or just stopping point. . . .” Coffey v. Milwaukee, 74 Wis.2d 526, 541, 247 N.W.2d 132 (1976).

The possibility that an accident will occur if an intoxicated person continues to drink may be foreseeable, but the act of serving the liquor is nonetheless a remote cause of any resulting harm. As we said in Garcia v. Hargrove, 46 Wis.2d 724, 736, 176 N.W.2d 566 (1970) :

“. . . Human beings, drunk or sober, are responsible for their own torts. Allowing plaintiff’s cause of action against the defendant would only erode that responsibility. . . .”

*489The cause of an auto accident involving an intoxicated driver is the driver’s inability to control his auto and this inability is the direct result of his voluntarily becoming intoxicated. Assigning the blame to the one who provided that means for that intoxication carries the chain a link too far.

The argument that the intoxicated driver is not absolved of liability but merely shares it with the dispenser is difficult to comprehend. Given a situation in which the voluntarily intoxicated driver is solely responsible for the accident, the task of assigning some percent of the negligence to the dispenser appears to be impossible. A mere five percent causal negligence assessed to the dispenser shifts part of the burden from the driver and says that to the extent of that five percent he is not responsible. Formulating a cause instruction to explain this situation to a jury could prove impossible. The dispenser and the driver are not joint tort-feasors. Their negligence does not combine to produce the result. Nonetheless, one must assume that if this action were recognized the driver would be entitled to recover contribution from the dispenser. If the practice of some states is followed the driver would be able to recover for his own injuries. Any shifting of the blame from the shoulders of the intoxicated driver is contrary to sound public policy. Further, because this action would be resorted to most often where the intoxicated driver is financially irresponsible, in many cases the dispenser would bear the whole of the judgment. In such a situation the dispenser who fails to notice that the drink his patron is being served will be the “one too many” for that driver will certainly be bearing a burden wholly out of proportion to his culpability.

Yet another problem with this created cause of action provides ample reason to adhere to our former position on this issue. It is one thing to think in terms of con*490sidering the application of such a law to the individually licensed country tavern. However, an entirely different situation is presented when considering its application to licensed hotels in a metropolitan center. On a given occasion such a licensee accommodates several hundred people at perhaps seven or eight special events, each with special arrangements for dispensing beverages and a corresponding increase in the number of bartenders and waitresses and waiters.

As the trial court recognizes, there can be no realistic, sensible common-law distinction when describing one who gives, sells or serves intoxicating beverages to one who is visibly intoxicated. If it is negligent when done by a licensed tavernkeeper, it is negligent when done by a bartender, a waiter or waitress, a social host or fellow-guest or companion. Such liability must be expanded to include others participating in serving or giving or supplying a beverage to the person with “one too many drinks.” It must necessarily include the social situation as well as the commercial. We know of no rule of law which would support the proposition that such a cause of action can be created only against licensed tavernkeepers while others participating in or performing the identical act under different situations can be excluded. Some such situations that come to mind are county fairs, American Legion Fourth of July picnics, and wedding receptions whether held in a private club or at a home. Such a cause of action places an unreasonable burden on potential tort-feasors. The determination of when a person is visibly intoxicated is highly discretionary. A jury of 12 will most likely have 12 different opinions on this question. Likewise, each bartender, waiter and host will have equally different perceptions. To expect, for example, a hotel to be able to monitor the condition and behavior of guests who are free to circulate from social event to restaurant to bar *491or to the room of a hotel guest or to his own room is to impose an unreasonable burden. At many social events it is possible to obtain a drink without even confronting a hotel employee. The possibilities for fraud are evident.

A change in the law which has the power to so deeply affect social and business relationships should only be made after a thorough analysis of all the relevant considerations. The respondents cite a number of statutes reflecting that the legislature, acting under its police power, has established penalties for certain acts committed by holders of licenses and operators of liquor dispensing establishments. If it is somehow intended that this liability should be restricted to licensed dispensers, then reference to these statutes supports the proposition that regulation of the industry is left to the wisdom of the legislature. The type of analysis required is best conducted by the legislature using all of the methods it has available to it to invite public participation.

The problem presented by this issue is not one of adequate remedies for an injured plaintiff. There seems to be little reason to provide the person injured by an intoxicated driver with a deeper pocket than the person injured by the stupid driver, the careless driver or the drugged driver. The problem here is one of controlling drinking and driving. This is a legislative problem within the police power to regulate liquor. Holding dispensers liable for injuries caused by intoxicated drivers is not likely to reduce such injuries. Realistically the most likely result would instead be an increase in tavern-keepers’ insurance premiums. For the court to create such a cause of action would be an attempt at legislating which would no doubt prove to have missed its mark.

The respondents point to a trend in other jurisdictions to recognize this common-law cause of action. In fact no firm trend is evident on this issue. For each state *492that has recognized a common-law cause of action there is one that has refused to do so. In several states the court and the legislature are both attempting to deal with the question. California is a notable example. California is a jurisdiction in which the court recognized a common-law cause of action against both commercial vendors and social hosts. The legislative response to those holdings is contained in two new statutes which prohibit such liability. Cal. Civil Code, sec. 1714 (1978), Cal. Bus. & Prof. Code, sec. 25602 (1978). Both the Iowa and Minnesota legislatures moved to nullify those courts’ holdings that social hosts could also be held liable. Iowa Code, sec. 123.92 (1979) ; Minn. Stats., sec. 340.95 (1978). Eight courts have held that no common-law cause of action exists and - that the formulation of a remedy belongs with the legislature. Profitt v. Canez, 118 Ariz. 235, 575 Pac.2d 1261 (1977) ; Carr v. Turner, 238 Ark. 889, 385 S.W.2d 656 (1965) ; Meade v. Freeman, 93 Idaho 389, 462 Pac.2d 54 (1969) ; State v. Hatfield, 197 Md. 249, 78 Atl.2d 754 (1951) ; Runge v. Watts, 589 Pac.2d 145 (Mont. 1979) ; Holmes v. Circo, 196 Neb. 496, 244 N.W.2d 65 (1976); Hamm v. Carson City Nugget, Inc., 85 Nev. 99, 450 Pac.2d 358 (1969) ; Marchiondo v. Roper, 90 N.M. 367, 563 Pac.2d 1160 (1977). Six others have concluded that no common-law cause of action exists apart from the provisions of their dram shop acts. Phillips v. Derrick, 36 Ala. App. 244, 54 So.2d 320 (1951) ; Nelson v. Steffens, 170 Conn. 356, 365 Atl.2d 1174 (1976) ; Keaton v. Kroger Co., 143 Ga. App. 23, 237 S.E.2d 443 (1977) ; Edgar v. Kajet, 375 N.Y.S.2d 548, 84 Misc.2d 100 (1975) ; Griffin v. Sebek, 245 N.W.2d 481 (S.D. 1976) ; Parsons v. Jow, 480 Pac.2d 396 (Wyo. 1971). Of the states recognizing a common-law cause of action, some, such as Illinois and Minnesota, have recognized such an action for those situations not *493covered by their dram shop act, others have recognized a cause of action in the absence of a dram shop act and the majority recognize a cause of action based on negligence per se for a violation of a criminal liquor control statute. These differing positions illustrate the difficulty with a judicial resolution of the problem and reinforce our belief that the issue is one best dealt with by the legislature. This type of action, if it is necessary at all, needs to be carefully developed with standards and definitions. Leaving it to a piecemeal judicial construction will result in a lengthy period of confusion and uncertainty as litigants attempt to apply the principle enunciated here — that it is negligent to provide liquor to a visibly intoxicated person.

Finally, one last problem is essential to a full consideration of the issue. This court has recognized that the availability of insurance to protect against a particular liability is a factor that should be considered in abrogating an immunity. Goller v. White, 20 Wis.2d 402, 412, 122 N.W.2d 198 (1963) ; Kojis v. Doctors Hospital, supra. Although not mentioned by the parties, it appears that liability insurance is available for this type of action. However, this insurance is generally purchased separately and one can easily conclude that the majority of Wisconsin tavernkeepers have not purchased insurance which they have not needed. Also, at least some existing insurance contracts contain specific exclusions for such coverage. For these reasons any change in the law should at least be prospective. This court has said that a decision should be made prospective wherever a compelling judicial reason exists. Such a reason is found where the parties have relied upon the existing law and will be adversely affected by a retroactive effect. Antoniewicz v. Reszczynski, 70 Wis.2d 836, 236 N.W.2d 1 (1975) ; Shier v. Freedman, supra; *494Dupuis v. General Casualty Co., 36 Wis.2d 42, 152 N.W.2d 884 (1967) ; Goller v. White, supra; Widell v. Holy Trinity Catholic Church, supra; Holytz v. Milwaukee, supra; and Kojis v. Doctors Hospital, supra. Again, this suggests the need for legislative, not judicial, resolution of the issue.

For the reasons stated above we reverse the order of the trial court and remand with directions to grant the motion of the appellant for summary judgment dismissing the complaint against him on its merits.

By the Court. — Order reversed and cause remanded for further proceedings consistent with this opinion.

(dissenting). I dissent for the reasons stated in the dissenting opinion of Mr. Justice Hallows in Garcia v. Hargrove, 46 Wis.2d 724, 176 N.W.2d 566 (1970).

I am authorized to state that Justice Abrahamson joins in this dissent.

(dissenting). I dissent and would affirm the trial court in this case.

In the last twenty years, several states have recognized common law causes of action against vendors of intoxicating beverages.1 Rappaport v. Nichols, 31 N. J. 188, 156 *495A.2d 1, 75 A.L.R.2d 821 (1959), was a landmark case in this area. In that case, the court noted that New Jersey had had a Civil Damage Law (or Dram Shop Act) during Prohibition which had imposed strict liability for compensatory and punitive damages upon unlawful sellers of alcoholic beverages. The act was repealed in 1934.

Applying normal principles of tort liability, the court said:

“The repealer left unimpaired the fundamental negligence principles which admittedly prevail in New Jersey and upon which the plaintiff grounds his common law claim. . . . Negligence is tested by whether the reasonably prudent person at the time and place should recognize and foresee an unreasonable risk or likelihood of harm or danger to others. . . . And correspondingly, the standard of care is the conduct of the reasonable person of ordinary prudence, under the circumstances

“The negligence may consist in the creation of a situation which involves unreasonable risk because of the expectable action of another. . . . Where a tavern keeper sells alcoholic beverages to a person who is visibly intoxicated or to a person he knows or should know from the circumstances to be a minor, he ought to recognize and foresee the unreasonable risk of harm to others through action of the intoxicated person or the minor.

“The defendants contend that, assuming their conduct was unlawful and negligent as charged in the complaint it was nevertheless not the proximate cause of the injuries suffered. But a tortfeasor is generally held answerable for the injuries which result in the ordinary course of events from his negligence and it is generally sufficient if his negligent conduct was a substantial *496factor in bringing about the injuries.” Rappaport v. Nichols, supra, 156 A.2d at 8, 9.

Three members of the court in Garcia v. Hargrove, 46 Wis.2d 724, 176 N.W.2d 566 (1970), took a similar approach in the dissent authored by the late Chief Justice Hallows:

“The time has arrived when this court should again exercise its inherent power as the guardian of the common law and hold upon general principles of common-law negligence a person, who, when he knows or should have known a person is intoxicated, sells or gives intoxicating liquor to such a person, is guilty of a negligent act; and if such negligence is a substantial factor in causing harm to a third person, he should be liable with the drunken person under our comparative-negligence doctrine. Conceded, the common law in this state for almost one hundred years has been to the contrary. Dillon v. Linder, 36 Wis. 344 (1874) ; Demge v. Feierstein, 222 Wis. 199, 268 N.W. 210; (1936); Seibel v. Leach, 233 Wis. 66, 288 N.W. 774 (1939) ; Farmers Mut. Automobile Ins. Co. v. Gast, 17 Wis.2d 344, 117 N.W.2d 347 (1962), but the basis upon which these cases were decided is sadly eroded by the shift from commingling alcohol and horses to commingling alcohol and horsepower.” Garcia, supra, 46 Wis.2d at 737.

I see no reason not to apply normal principles of tort liability instead of making a special exception for the vendors of alcohol. As the court stated in A. E. Investment Corp. v. Link Builders, Inc., 62 Wis.2d 479, 483, 484, 214 N.W.2d 764 (1974) :

“. . . . The duty of any person is the obligation of due care to refrain from any act which will cause foreseeable harm to others even though the nature of that harm and the identity of the harmed person or harmed interest is unknown at the time of the act. . . .

“A defendant’s duty is established when it can be said that it was foreseeable that his act or omission to act may cause harm to someone. A party is negligent *497when he commits an act when some harm to someone is foreseeable. Once negligence is established, the defendant is liable for unforeseeable consequences as well as foreseeable ones. In addition, he is liable to unforeseeable plaintiffs.”

• To call remote the link between a fatal automobile accident and the serving of alcohol to an intoxicated person is to ignore reality. The New York court in Berkeley v. Park, supra, 262 N.Y.S.2d at 293 commented that the Rwppaport and Waynick cases:

“. . . rejected as simply unreal the distinction that the selling of alcohol is only a remote cause of resulting intoxication while the consumption is the proximate cause. On the question of foreseeability, the cases, citing reports of the National Safety Council, modern conditions, drunkenness and driving, and the almost universal use of automobiles, stated that the consequences of serving an intoxicated person who drives an automobile more alcohol, were easily foreseeable to the reasonable person.”

Wisconsin Accident Facts 1977, published by the Traffic Accident Data Section of the Division of Motor Vehicles, shows that 945 persons were killed in Wisconsin motor vehicle traffic accidents in 1977. Of the 514 drivers who were killed in accidents, 422 were tested for blood alcohol content. 221 of those drivers were legally intoxicated with blood alcohol levels of .10 or above.

I conclude that the vendor in this case was guilty of a negligent act if he sold intoxicating beverages to a person who he knew or should have known was intoxicated and who he knew or should have known would be driving in an intoxicated condition. If such negligence was a substantial factor in causing harm to a third person; the commercial vendor here should be liable with the intoxicated person under the comparative negligence rule.

*498The defendant points to the first Garcia case as the “law” that should be followed and the majority of this court agrees. But the Garcia I opinion was by a four to three vote. It fails to take into account the fact that we bear a tremendous social cost from the death and maiming that takes place on our highways because of drivers whose senses are dulled by alcohol.

The tragedy of this family, where the wife and mother was killed returning from work is another example of the failure of a policy of placing sole responsibility on the consumer of alcohol. Until those who furnish alcohol to one who has obviously had more than he can handle are made to share responsibility, the scene will be repeated again and again and the list of innocent victims will continue to grow. Following repeal of prohibition, the legislature wisely made it a crime to serve one who is intoxicated or bordering on intoxication.2 When a *499purveyor of alcohol flouts that law he. should share with the intemperate consumer the blame for the often tragic results. If money could be lost as well as made from serving those already intoxicated it would undoubtedly cut down the number roving the highways unfit to drive because of over indulgence in alcohol.

I would hold that violation of the statutes making it a crime to serve an intoxicated person is negligence per se, and gives rise to a civil cause of action. Recently, the' New York Supreme Court Appellate Division held that an airline’s breach of a Federal Aviation Administration regulation that forbids service of alcoholic beverages to intoxicated aircraft passengers gives rise to a cause of action against the airline for an assault committed by an intoxicated passenger. Manfredonia v. American Airlines, Inc., 47 U.S.L.W. 2745 (May 29, 1979). If a cause of action may be grounded upon a regulation, certainly breach of the criminal statutes is a proper basis for a common law cause of action.

I do not accept the view that the legislature had preempted the field. In Garcia I, this court made clear that its decision was not motivated by considerations of legislative pre-emption. Rather, the Garcia decision was based strictly on the court’s view of public policy.

We should not hesitate to depart from stare decisis when the facts warrant a change in the law.

The facts in the case before us present an appropriate framework to correct the mistake of Garcia 1. I would do so.