(concurring in the judgment, with whom Ireland, C.J., joins). I agree with the court that where, as here, an underage social host knowingly allows underage guests to use her home to drink alcohol, but does not herself supply the alcohol, we should decline to impose social host liability on the underage host. I differ with the court in that I would limit our holding to an underage host, and would wait until we are presented with a case where a social host who has reached legal drinking age knowingly allows underage guests to use his or her home to drink alcohol before we decide whether to extend our holding to all social hosts.
I believe that such restraint is the more prudent course here for two related reasons. First, it is not difficult to imagine egregious circumstances where an adult of legal drinking age encourages underage guests to “bring your own beer or booze” to get drunk at his or her house, one of whom later kills or cripples someone while driving home, that might cause us to *543look differently at whether we should impose liability on such an irresponsible social host. It is worth recalling the tragic event that spurred the Legislature to adopt the amendments to G. L. c. 138, § 34 (St. 2000, c. 175), that made it a crime “to knowingly or intentionally supply, give, or provide to or allow a person under 21 years of age ... to possess alcoholic beverages on premises or property owned or controlled by the person charged.” In June, 1996, an eighteen year old teenager who had recently graduated from preparatory school attended a party hosted by the father of a fellow student, where teenagers drank from unsupervised beer kegs and adults played drinking games with the teenagers. The teenager died when he was trying to drive home in his automobile with a blood alcohol level of .19 per cent, more than twice the legal limit for an adult, and crashed into a telephone pole. The following year, the adult host of the party was acquitted of the misdemeanor charge of providing alcohol to a minor, which suggests that there was at least a reasonable doubt whether the adult host had supplied the alcohol for the party.
Under the court’s decision today, if similar tragic facts again arose, a social host who has reached the legal drinking age would not be liable in tort if he or she did not provide the alcohol for the party, even if the underage guest had killed a pedestrian, passenger, or another motorist while driving home. I am not yet convinced that this is the correct result under the common law where the social host is not himself or herself underage. Therefore, I think it wiser to limit our holding to underage social hosts, and wait for a case where the defendant social host is legally allowed to drink alcohol before extending our holding beyond the facts presented here.
Second, I am not yet persuaded that the policy reasons articulated by the court for denying social host liability apply with equal force where the social host is twenty-one years of age or older rather than underage. While I join the court in concluding that the underage defendant in this case is not liable in tort as a social host for knowingly allowing the underage drunk driver to drink alcohol in her home, I do so because of the “practical difficulties” identified by the court in Ulwick v. DeChristopher, 411 Mass. 401, 406 (1991) (Ulwick): “Hosts in *544these circumstances might be left with little alternative than to resort to physical force in order to discourage further drinking or to try to eject the guest, a solution that in many cases will aggravate the situation and put the drunk driver where he should not be — behind the wheel of a car.” In Ulwick, as here, an underage person was allowed to drink at the home of an underage host when the host’s parents were away. Id. at 402-403. I share the Ulwick court’s concern about these “practical difficulties” where an underage host, without the assistance of his or her parents, attempts to remove an underage guest who brought alcohol to a party in the host’s home. But I am not yet convinced that these practical difficulties are the same if the underage host’s parents are present or if the host himself or herself is twenty-one years of age or older. And I am not yet convinced that these practical difficulties are the same now as they were when the Ulwick case was decided because, nine years after the Ulwick decision, the Legislature specifically made the conduct at issue a crime.1
Nor am I persuaded by some of the court’s reasons for its broad holding. The court declares that “we are reluctant to impose a duty of care in the absence of ‘clear existing social values and customs’ supporting such a step,” ante at 537, quoting Remy v. MacDonald, 440 Mass. 675, 678 (2004), and claims there is no “ ‘community consensus’ regarding the proposed expansion of social host liability.” Ante at 537, quoting Schofield v. Merrill, 386 Mass. 244, 248 (1982). As best I can tell, there exists a “clear existing social value[]” that parents not allow the underage guests of their children to drink alcoholic beverages at *545their home. I have no clue whether there is a “clear existing social value[]” or a “community consensus” that these adults not be liable for the injuries caused when an underage guest they had permitted to drink at their home kills or maims someone while driving home. But I am convinced that the court cannot establish the absence of a “clear existing social value[]” or a “community consensus” based on “the Legislature’s subsequent refusals” to enact a statute imposing social host liability where the social host violates G. L. c. 138, § 34, but does not supply the alcohol. Ante at 537.
We have long recognized the need to be wary of any supposed inference based on legislative nonaction, especially where, as here, “the Legislature’s subsequent refusals” are nothing more than bills failing to emerge from the committee where they were filed. See, e.g., Simon v. State Examiners of Electricians, 395 Mass. 238, 247 (1985) (“postenactment history is not ordinarily considered as showing legislative disapproval of the rejected amendments”). In the 2003-2004 legislative session, one senator submitted a bill proposing that any person violating the criminal statute would be liable in tort where an underage driver under the influence of alcohol causes injury or death. The bill was referred to committee where it remained with an order for further study. See 2003 Senate Doc. No. 1100, 2003 Senate J. at 93A; 2004 Senate Doc. No. 2288. For the next .three sessions, the same senator reintroduced the identical bill and each bill again remained in committee without triggering debate or further legislative action. See 2005 Senate Doc. No. 1020; 2006 Senate Doc. No. 2633; 2007 Senate Doc. No. 968; 2007 Senate J. at 81 A; 2009 Senate Doc. No. 1775. There was no discussion or vote by the Legislature on the proposal.
The “fallacy” in the court’s reasoning is that “no one knows why the legislature did not pass the proposed measures. . . . The practicalities of the legislative process furnish many reasons for the lack of success of a measure other than legislative dislike for the principle involved in the legislation.” Franklin v. Albert, 381 Mass. 611, 615-616 (1980), quoting Berry v. Branner, 245 Or. 307, 311 (1966). Even if there were a policy reason why these bills did not emerge from committee, we cannot know whether the reason was a legislative judgment that these matters of civil liability should be left to the courts to decide *546under the common law. See Franklin v. Albert, supra at 616, quoting H. Hart & A. Sacks, The Legal Process: Basic Problems in the Making and Application of Law 1395-1396 (tent. ed. 1958) (listing reasons legislators may not support particular bill, including “[bjelief that the matter should be left to be handled by the normal processes of judicial development of decisional law . . .”). As this court has stated in the past, “we reject the suggestion that defeated legislative proposals have the power to disable us, in a proper case, from considering the questions presented by such proposals and from abandoning prior conclusions that now seem inappropriate.” Franklin v. Albert, supra at 617. See Superintendent of Schs. of Leominster v. Mayor of Leominster, 386 Mass. 114, 118 n.11 (1982) (“consideration and rejection of various proposed measures do not control our decision”).
The court also appears to fear what it characterizes as “[pjotentially vast consequences to liability insurance” if social hosts who violate G. L. c. 138, § 34, by allowing underage guests to consume alcohol in their home were subject to civil liability. Ante at 537 n.17. This fear is probably overstated because liability insurance policies often exclude criminal acts from coverage. See 7A G. Couch, Insurance § 103:40 (3d ed. 2005). But even if an expansion of social host liability were to result in an increase in liability insurance claims, that is not a sufficient reason to rule against such an expansion where the common-law rules of liability need to be changed. See, e.g., Papadopoulos v. Target Corp., 457 Mass. 368, 369 (2010) (abolishing distinction between natural and unnatural accumulations of snow and ice). Nor can I accept the court’s suggestion that the Legislature’s regulation of insurance coverage somehow affects whether we should revise a common-law duty of care, because any change in the common-law duty of care will likely have consequences for liability insurers and policy holders. If we were to conclude that we should defer to the Legislature regarding a common-law duty of care where it will affect insurance coverage, we would frustrate the necessary evolution of the common law of tort liability.
Finally, the court professes concern about “the difficulties,” which it contends are “manifold,” that judges and juries would *547face if the existence of a duty turned on control of the premises rather than on the supply of alcohol. Ante at 537. This “difficulty” is not so dire. General Laws c. 138, § 34, imposes criminal liability on any person who knowingly or intentionally allows an underage person to possess alcohol “on premises or property owned or controlled by the person charged.” If civil liability were to mirror criminal liability, courts would need to determine when a person “controlled” a premises, which courts routinely do in other matters of common law. See, e.g., Bishop v. TES Realty Trust, 459 Mass. 9, 19 (2011), citing Humphrey v. Byron, 447 Mass. 322, 328-329 (2006) (in determining tort liability, court must evaluate whether lessor of commercial premises retains control).
Therefore, I concur in the judgment, but would limit the holding to underage hosts, and wait for a case where the host was at least twenty-one years of age before establishing a rule that would protect all social hosts from civil liability for allowing their homes to be used by underage guests for “bring your own beer or booze” parties.
Justice Botsford, in her concurrence, correctly notes that the Legislature does not allow persons to drink alcohol until they are twenty-one years of age but allows persons to serve alcohol in licensed establishments at eighteen years of age. G. L. c. 138, § 34. She concludes from this that “the fact that the adult host cannot herself yet drink alcohol legally seems irrelevant” in determining the scope of social host liability. Ante at 539 n.1 (Botsford, J., concurring). Such a distinction in the common law would derive, not from logic, but from our experience that older hosts who allow underage guests to drink alcohol in their homes are deemed more culpable than underage hosts and are better able to enforce such a prohibition without force or violence. I note, as does the court, that three of the six States (Hawaii, Minnesota, and Texas) that by statute have imposed civil liability on social hosts who allow underage drinking in their homes limit liability to hosts who are twenty-one years of age or older. See ante at 538.