(concurring and dissenting).
I concur with the conclusion of the majority that Stephen Ario is not entitled to common law official immunity on the claim that he operated without lights, but I would go further, and thus respectfully dissent with respect to the conclusion that St. Louis County is entitled to either statutory immunity or vicarious official immunity on the claim that Ario was otherwise negligent in grading against traffic. I would conclude that statutory immunity is not available to the county because Ario’s decisions (1) not to deadhead and (2) to grade against traffic without giving any warning, were operational, not discretionary. I would also conclude that, under these facts, the county’s vicarious official immunity should be no greater than its statutory immunity. Because there is sufficient evidence to create genuine issues of material fact that both of Ario’s decisions were negligently made and were the proxi*509mate cause of the death of Joshua Schroeder, I would reverse the summary judgment in favor of the county and remand for trial.
I begin, as our summary judgment standard requires, by viewing the evidence in the light most favorable to the Schroeders. This means that we should assume as true the following facts.
Joshua Schroeder was an innocent traveler on Highway 29, proceeding within the proper driving lane. The day was at dusk, but had already been so overcast that one could not actually see the sun set. Highway 29 was also banked with trees, limiting visibility. Joshua proceeded with the reasonable assumption that his driving lane was clear of obstructions or other traps because no signs warned of work ahead. Unknown to Joshua, the county grader was actually occupying his entire lane, some mile or two ahead, and the operator was negligently grading in the wrong direction at about 3 miles an hour, without any headlights on. As Joshua approached, the operator stopped the grader on the wrong side of the road, in Joshua’s driving lane. Joshua ran directly into the grader without warning. No horn was sounded. The presence of the grader on the wrong side of the road was the proximate cause of the crash and of Joshua’s death.
On its face, this fact situation suggests an injustice, a wrong for which the law should provide a remedy. As we have observed:
One of the paramount interests of the members of an organized and civilized society is that they be afforded protection against harm to their persons, properties, and characters. The logical extension of that interest is that, if harm is wrongfully inflicted upon an individual in such a society, he should have an opportunity to obtain a reasonable and adequate remedy against the wrongdoer, either to undo the harm inflicted or to provide compensation therefor. If the state is properly to serve the public interest, it must strive, through its laws, to achieve the goals of protecting the people and of providing them with adequate remedies for injuries wrongfully inflicted upon them. So long as the state fails to do so, it will be functioning in conflict with the public interest and the public good.
Nieting v. Blondell, 306 Minn. 122, 235 N.W.2d 597, 602-03 (1975).
If the grader operator had been working for a private party, he and his employer would surely be liable. In fact, if the private employer had made a policy decision to permit the creation of this dangerous condition, by weighing the cost of remedying the danger against the risk to human life, we would consider that policy reprehensible — a callous disregard for the safety of others that would warrant the imposition of punitive damages. The question then is whether comparable conduct by the government should escape liability, precisely because it involved a policy made by weighing the cost of remedying a dangerous condition against the risk to human life. I would first conclude that, under an appropriately narrow construction of our existing tort immunity law, the county should not escape liability. ' But if our existing tort immunity law does not yield this result, then I would conclude that it is time to seriously reexamine that law.
A. Statutory Immunity: the statutory exception to governmental tort liability for a “discretionary function” must be narrowly construed.
1. Discretionary function immunity is an exception to the general rule of governmental liability.
Although our decisions under the State Tort Claims Act (Minn.Stat. ch. 3) and the *510Municipal Tort Claims Act (Minn.Stat. ch. 466) frequently refer to “statutory immunity,” this is perhaps a misnomer because the primary effect of those acts is to establish governmental liability for the torts of its employees and any immunity that is extended by those acts is an exception to that liability. Minn.Stat. §§ 3.736 and 466.02 (2004). Both acts create an exception to liability for the “discretionary” acts of employees, generally referred to as the “discretionary function exception.” Minn. Stat. §§ 3.736, subd. 3(b), and 466.03, subd. 6 (2004); Holmquist v. State, 425 N.W.2d 230, 232 (Minn.1988).
Because of the risk that the discretionary function exception could swallow the general rule of tort liability, we have consistently interpreted that exception narrowly. Holmquist, 425 N.W.2d at 232 (citing Cairl v. State, 323 N.W.2d 20, 23 (Minn.1982), and Nusbaum v. County of Blue Earth, 422 N.W.2d 713 (Minn.1988)). Moreover, I would suggest that the discretionary function exception should be viewed even more narrowly in a case where, as here, the governmental entity has created a dangerous condition on its roads that gives rise to a duty to warn. A brief historical review of governmental tort immunity, specifically in the context of the government’s responsibility to care for the safety of travelers on its roads, will reinforce this-view.
2. Common law immunity did not apply to the failure of the government to warn of a dangerous condition it created on its roads.
Under common law governmental tort immunity, prior to its abolition, the paradigm was reversed- — immunity was the general rule and liability was the exception. Nusbaum, 422 N.W.2d at 718 (stating that “the general rule was immunity with limited exceptions of liability”); Anderson v. City of Minneapolis, 296 N.W.2d 383, 385-86 (Minn.1980). But, even under that regimen, our cases recognized that a county or municipality, having responsibility for a road, had a common law duty to maintain that road in a reasonably safe condition and could be held liable, as an exception to immunity, for its failure to do so upon notice of a dangerous condition. See, e.g., Anderson, 296 N.W.2d at 385-86 (describing the exception to common law governmental tort immunity for negligence in the maintenance and repair of roads under the government’s control).2
*511That duty included the exercise of reasonable care to warn travelers of a dangerous condition. See, e.g., Mix v. City of Minneapolis, 219 Minn. 389, 395, 18 N.W.2d 130, 134 (1945) (stating that “if, by reason of peculiar facts or circumstances, a pitfall, trap, or snare dangerous to a traveler proceeding with reasonable care is created in respect to a street, a municipality owes a duty to exercise reasonable care to warn or otherwise protect such a traveler for resulting danger.”) Although the duty to repair or warn generally required proof that the municipality had actual knowledge or constructive notice of a dangerous condition, it was uniformly recognized that such proof was satisfied where an act of the municipality’s employee created the dangerous condition. See, e.g., Larson v. Township of New Haven, Olmsted County, 282 Minn. 447, 454, 165 N.W.2d 543, 547 (1969) (stating that the general rule, that the municipality’s duty to warn arises only if the municipality has actual knowledge or constructive notice of dangerous conditions, is inapplicable if the “defect * * * is one which [the municipality] created.”).
Thus, even where immunity was the rule and liability the exception, this court ultimately held a governmental body liable for the failure to warn or remedy a dangerous condition that its employees created on one of its roads. Intuitively, one would expect that the liability of a governmental body that existed under common law governmental immunity would not be reduced by the abolition of that immunity and the substitution of statutory governmental tort liability. This was precisely the reaction of the court of appeals when it first addressed the discretionary function exception to state and municipal tort liability in the context of dangerous conditions on a road.
3. Does the statutory discretionary Junction exception to liability override a municipality’s duty to remedy or warn of dangerous conditions on the road?
In the first cases that followed the statutory establishment of governmental tort liability, the court of appeals concluded that the duty of the state or local government to remedy or warn of dangerous conditions on its roads could not be eliminated by the statutory discretionary function exception. Thus, in Holmquist v. State, 409 N.W.2d 243, 247-48 (Minn.App. 1987), rev’d, 425 N.W.2d 230, 232 (Minn. 1988), the court of appeals determined that the abrupt narrowing of the width of the shoulder on a road constituted a pitfall, trap, or snare, giving rise to a duty to warn, and that the discretionary function exception under the State Tort Claims Act was not applicable because the state had created this dangerous condition. Likewise in Nusbaum v. County of Blue Earth, the court of appeals determined that the discretionary function exception to state liability was not applicable where the state created a dangerous condition on its roads that gave rise to a duty to warn.3 411 N.W.2d 917, 922-23 (Minn.App.1987), affd on other grounds, 422 N.W.2d 713, 717 (Minn.1988). If those cases were applicable here, we would hold that the decision *512to grade against traffic created a dangerous condition on the county’s roads that gave rise to a duty to warn, making the discretionary function exception inapplicable.
The rationale of those court of appeals decisions was the same as the public policy considerations that led us to abolish governmental immunity in the first place. See, e.g., Nieting, 235 N.W.2d at 601 (stating that immunity was an exception to “the fundamental concept of tort law that liability follows tortiuous conduct and that individuals and corporations are responsible for the acts of their employees acting in the course of their employment.”); Spanel v. Mounds View Sch. Dist. No. 621, 264 Minn. 279, 285, 118 N.W.2d 795, 799 (1962) (stating that immunity was unfair to injured individuals in the era of expanding governmental services and is neither just nor reasonable). A governmental body that is charged with responsibility for the safety of the roads should obviously be discouraged from creating dangerous conditions on those roads. Further, a governmental body whose conduct creates a common law duty to warn its citizens about a dangerous condition should not be allowed to unilaterally immunize itself from liability by simply engaging in the process of weighing the risk of harm to the public, against some competing policy concerns of the government, especially purely economic concerns. And, because the government’s decision to forgo certain expenditures will provide benefits to all taxpayers, it is not fair to require a single innocent victim to bear all of the cost of the adverse consequences of that decision.
Moreover, if a governmental entity is immune, and therefore has no risk of loss for negligent acts, then it has nothing to balance against forgoing the expenditure. Carried to the extreme, immunity could encourage the government to abdicate many of its most important responsibilities by making policy decisions to save the expenses necessary to execute them. Could a county, for example, allow travelers to continue to use an unsafe bridge, without warning, because it weighed the safety of the travelers against budget constraints that made it financially difficult to make the bridge safe? One would hope not, but the extension of the discretionary function exception to the deliberate abdication of governmental responsibilities, purely for cost-saving reasons, could produce precisely that extreme result.
In fact, it is anomalous that a governmental body may obtain immunity from liability under the discretionary function exception by engaging in the very conduct that would increase the culpability of a private party. For example, products liability cases routinely hold that a product manufacturer who knows of a dangerous condition but defers correction by engaging in a cost-benefit analysis, balancing human lives against corporate profits, has demonstrated such “callous indifference to public safety” as to be subject to punitive damages. See, e.g., Grimshaw v. Ford Motor Co., 119 Cal.App.3d 757, 174 Cal. Rptr. 348, 384 (1981) (“There was evidence that Ford could have corrected the hazardous design defects at minimal cost but decided to defer correction of the shortcomings by engaging in a cost-benefit analysis balancing human lives and limbs against corporate profits. Ford’s institutional mentality was shown to be one of callous indifference to public safety.”); Hodder v. Goodyear Tire & Rubber Co., 426 N.W.2d 826, 835-36 (Minn.1988) (concluding that Goodyear’s inadequate distribution of warnings about the danger of exploding rims, based on a corporate policy to restrict advertising dollars for projects that are not promoting product sales, was “willful indifference to the safety of others.”); Gryc v. Dayton-Hudson Corp., *513297 N.W.2d 727 (Minn.1980) (upholding a punitive damages award against a manufacturer who continued, for profit reasons, to supply flammable nightwear when nonflammable material was available). We should expect our government to pay an even higher regard for the safety of its citizens. Yet the discretionary function exception, broadly applied, would make the government immune from liability precisely because it has engaged in “a cost-benefit analysis balancing human lives and limbs against [budget constraints].”4
For all of these reasons, if the issue were being presented to us for the first time, I would be inclined to agree with the early court of appeals decisions and-hold that the discretionary function exception to liability is not applicable to loss caused by dangerous conditions on the road that were created by the government itself and for which no warning was given. But the question appears to have been decided differently by our precedent, albeit without explicit discussion of these policy concerns. And the Schroeders have not asked us to overrule those cases. Although I believe we should be open to reexamine the question in a future case that presents it more fully and squarely, I would apply our past precedent to this case.
We decided Holmquist and Nusbaum in reverse order from that of the court of appeals. In Nusbaum, we did not reach the question of whether the discretionary function exception to state liability was inapplicable where the state created dangerous conditions because we determined that the state’s decision on road signs was not a discretionary function and thus the exception did not apply in the first place. Nusbaum, 422 N.W.2d at 717. But in Holmquist, we reached the question and ultimately rejected the court of appeals’ analysis, stating:
To say that the discretionary function exception is inapplicable to any decision concerning a highway condition created by the State without regard to the nature of the decision is inconsistent with the language and structure of the State Tort Claims Act. The question is not whether the State’s conduct resulted in a condition posing an unreasonable risk of harm; it is whether the conduct consisted of planning or policymaking decisions (protected) or operational level decisions (unprotected). To hold otherwise would effectively nullify the discretionary function exception because a claim of loss caused by the performance or failure to perform a discretionary duty frequently rests on an allegation that the State created a hazardous condition. * * * The fundamental inquiry must be whether the challenged decision involves a policymaking decision *514entrusted to the political branches of government and is, therefore, to be protected from judicial second-guessing.
Holmquist, 425 N.W.2d at 232.
Thereafter, in Steinke v. City of And-over, we considered the comparable discretionary function exception contained in the Municipal Tort Claims Act, Minn.Stat. § 466.03, subd. 6. 525 N.W.2d 173, 175 (Minn.1994). The court of appeals had held that the discretionary function exception was not applicable to decisions that involve warning the public of known hazards. Steinke v. City of Andover, C3-93-865, 1993 WL 491248 at ⅜1 (Minn.App. Nov.30, 1993). We determined that “[wjhether a governmental agency was warning the public of known hazards is not relevant in determining whether the conduct involved discretionary decision making.” Steinke, 525 N.W.2d at 175.
A Given this precedent, what should be the proper interpretation of the words “discretionary function’’?
As discussed above, we have been clear that the discretionary function exception must be narrowly construed. This means, in part, that the exception should apply only when necessary to achieve separation of powers goals. Holmquist, 425 N.W.2d at 231, 233. But this broad, largely subjective formulation is difficult to apply consistently. Our decisions have attempted to develop more objective tests. In Nusb-aum, we mentioned several distinctions that could be drawn: between operations and planning; between implementation and policy; between scientific or professional decisions and policymaking decisions; and between challenges to government conduct pursuant to a policy and challenges to the policy itself. Nusbaum, 422 N.W.2d at 720-21. But we acknowledged that these distinctions were not always helpful, and that the ultimate test was whether “the challenged governmental action involves a balancing of policy with political, economic and social considerations.” Id. at 720. Similarly, in Gleason v. Metro Council Transit Operations, we said:
We have interpreted the discretionary function exception narrowly — protecting only those activities which require the balancing of policy objectives, “involving social, political, or economic considerations, rather than merely professional or scientific judgments.” [Zank v. Larson, 552 N.W.2d 719, 721 (Minn.1996)] (quoting Steinke v. City of Andover, 525 N.W.2d 173, 175 (Minn.1994)).
582 N.W.2d 216, 219 ,n. 5 (Minn.1998).
In Nusbaum, we determined that the decisions to place a sign ending a reduced speed zone before an unmarked curve and to provide no specific warning were not mandated by any policy considerations. 422 N.W.2d at 723. We said that the policy that required engineers to determine the appropriate speed on the roads by consideration of enumerated factors did not mandate the particular location of the sign, which was a matter of professional judgment. Id. As to the failure to warn, we said that the policy only determined that engineers may recommend warnings, not that they must, leaving room for professional judgment. Id. Accordingly, we held that the conduct of the engineers was not within the discretionary function exception. Id. In Holmquist, we again determined that the policy concerning warning signs left room for judgment in implementation, and that the state had failed to show that such judgment itself involved the balancing of public policy considerations. 425 N.W.2d at 234.
From these cases, we should conclude that, when applying a narrow interpretation of the discretionary function exception, the burden is on the governmental *515body to show that the policymaking decision mandates the conduct that is the basis for the plaintiffs claim. If, instead, the policy merely permits that conduct, and leaves it to the employee to determine, by use of operating and not discretionary judgment, whether to engage in the conduct, then the discretionary function exception has not been shown.
5. What was the county’s policy on grading against traffic, and what decisions were left to the operational judgment of Ario?
With these considerations in mind, I would conclude that the “policy” of the county was to “permit” grader operators to grade against traffic, but only under certain circumstances and with the use of certain precautions. Because the policy did not mandate that Ario grade against traffic at the time that this accident occurred, or prohibit Ario from warning that he was doing so, but left those choices to Ario’s operational judgment, I would conclude that Ario’s decisions were not discretionary and did not qualify for statutory immunity under the discretionary function exception.
Although it is clear that various employees of the county considered the question of whether it was permissible to grade gravel roads against traffic, it is less clear precisely what decisions were made on that issue. The majority concludes that a policy decision was made to “permit grade operators to choose when to grade against traffic.” I would agree with the use of the word “permit,” because grading against traffic was clearly not mandated, and I would add that the evidence demonstrates that even such permission was conditional. More specifically,- Ario’s choices not to deadhead, which would have allowed him to grade with traffic, and to grade against traffic without any warning to oncoming travelers, were operational choices that did not balance public policy concerns but basically focused on operating convenience— the extra 4-6 minutes that either measure might add to the completion of the job.
Because the county’s policy was not formal or written, we must look to several sources to determine what the scope of that policy was. The 1985 memorandum of Joe Varda, relied upon by the county as containing at least a partial statement of the policy, describes grading against traffic in permissive terms (“you may”) and as being subject to these conditions:5
(1) If you are operating your equipment with routine care and the normally accepted safety precautions are taken (beacon, working lights, flags, etc.) * * ⅜.
(2) [I]f there are cases where sight distance is short either vertically or horizontally, either an attempt to improve the condition should be made or special precautions should be taken.
In other words, the permission did not relieve the operator of the need to use due care in determining what precautions should be taken if he was to grade against traffic.
The affidavit of David Skelton confirms the county’s view that the Varda memorandum does state the county’s “policy,” presumably including the conditions stated by Varda. Skelton’s affidavit also confirms that the “policy” was permissive, not mandatory (“St. Louis County has had a *516policy for a number of years wherein we allow our grader operators to grade against traffic on gravel roads.” (Emphasis added.)). In addition, Skelton’s affidavit suggests that the “policy” includes these further conditions:
(1) “sign usage where practical”; and
(2) “deadheading where practical.”
Finally, Skelton’s affidavit refers to the county’s reliance on a “booklet entitled, ‘Blading Aggregate Surfaces’ from the National Association of County Engineers Training Guide Series in conjunction with the U.S. Department of Transportation Federal Highway Administration.” Although this booklet advises grading operators to “[periodically blade surface of the road against the flow of traffic to eliminate drifting of aggregate onto ends of bridges, culverts, intersections, and railroad crossings[,]” it also makes these safety recommendations:
• Turn on grader headlights when blad-ing against traffic.
[[Image here]]
• Use signing and proper flaggers where needed to warn traffic of work in progress or as warning if grader is left unattended.
From these descriptions of the unwritten “policy” of the county, I would conclude that the county’s grading protocol permitted operators to grade against traffic only when deadheading was not practical and the operators were able to do so with appropriate precautions. This protocol left it to the operators to decide when deadheading was practical and when warning signs were practical. The policy did not require Ario to grade against traffic on the third or fourth pass and did not prohibit Ario from either deadheading after the second pass or placing warning signs after the first, if he was going to grade against traffic.
On this record, there is evidence to show that deadheading was practical. According to Ario, he intended to make four passes on this 2-mile segment of County State-Aid Highway 29. At a normal grading speed of 3 miles per hour, the four passes would have taken a total of 2 hours and 40 minutes. Ario’s first two passes were on the right side of the road, with traffic. The choice to deadhead before passes three and four would have added only 4-6 minutes to the job, given the ability of the grader to deadhead at 23-30 miles per hour.6 Ario acknowledged that the only advantage to grading against traffic that he considered was the time saved (4-6 minutes) by not driving to the other end. He did not raise any operational objections to deadheading.
Although the county may have been concerned that routine deadheading, by all operators on all jobs, would significantly add to the overall expense of grading and might require a reduction in grading practices, that concern is not presented by these facts. In this circumstance, deadheading was surely practical and would not have added any significant cost. Likewise, the delay to place a warning sign at the north end of the west lane, after Ario completed his first pass to that end, would have been insignificant. In fact, to the extent that Ario considered the time cost of deadheading or placing a warning sign, it was so de minimis as to not amount to a public policy consideration.7
*517Further, there is evidence from which one could conclude that the decision to not deadhead, but to grade against traffic without a warning sign, was particularly dangerous in this case because of the time of day and the worsening visibility conditions. Ario testified that when the accident occurred, it was dusk on a day that was already so cloudy that one could not see the sunset. He also said that the road was mostly banked with trees. Because I conclude that the county is not entitled to discretionary function immunity, I would hpld that this evidence is sufficient to create genuine issues of material fact on the question of whether Ario’s decision not to deadhead, but to grade against traffic without warning, was negligent.
B. yicarious Official Immunity: public policy considerations suggest that governmental vicarious liability should be broader than official liability.
The majority essentially concludes that the vicarious official immunity of the county is coextensive with the official immunity of Ario and that the county is thus immune from the claim that Ario was negligent in grading against traffic (but presumably not from the claim that Ario negligently operated the grader at dusk without lights). Because I would conclude that the county is not entitled to statutory immunity, as discussed above, I would also conclude that, at least under these facts, the official immunity of Ario should not be extended to the county where to do so would create vicarious official immunity that would eviscerate the county’s statutory liability. In so doing, I would find this to be one of the circumstances, as recognized in our prior cases, where the governmental entity cannot vicariously benefit from its employee’s personal immunity. Pletan v. Gaines, 494 N.W.2d 38, 42 (Minn. 1992) (quoting Holmquist, 425 N.W.2d at 238 n. 1, “[N]ot infrequently a governmental entity is required to compensate for the harm done by a public official even though the official is not held personally liable.”).
1. The origins and scope of our common law of official immunity are unclear.
Curiously, the common law doctrine of official immunity seems to have evolved wholly independent of that of governmental immunity. The criticisms that were made of common law governmental immunity — the irrelevance of its historical rationale tied to the British monarchy, the development of the American tort concept that liability should flow from fault, and the recognition of the social policy that the government is better able to spread the loss — apply equally to official immunity. But our cases abolishing sovereign immunity make no mention of official immunity. See, e.g., Spanel, 264 Minn. 279, 118 N.W.2d 795; Nieting, 306 Minn. 122, 235 N.W.2d 597.
Also, when the legislature enacted the State Tort Claims and Municipal Tort Claims Acts, it did not address the question whether the newly created governmental liability was meant to replace or otherwise affect employee liability. Specifically, section 466.02 addresses the tort liability of municipalities (including the vicarious liability of a municipality for the torts of its employees), but does not mention the limits of employee immunity.
But the Municipal Tort Claims Act did adopt four provisions that essentially eliminated the policy concerns that underlie *518official immunity: (1) it established liability limits that also apply to direct claims against public officers and employees (Minn.Stat. § 466.04, subd. la (2004)); (2) it required that a notice of claim be given within 180 days after loss or within 1 year for wrongful death (Minn.Stat. § 466.05 (2004)); (3) it authorized the municipality to procure insurance against liability, including that for liability of “its officers, employees, and agents” (Minn.Stat. § 466.06 (2004)); and (4) it required the municipality to defend and indemnify its “officers and employees,” subject to certain qualifications (Minn.Stat. § 466.07, subd. 1 (2004)).
Yet, despite these protections for government officials, the official immunity doctrine appears to have thrived after 1963. It has been applied to acts that involve operational discretion so long as they are not purely ministerial and so long as the official is not guilty of a willful or malicious wrong. See, e.g., Elwood v. County of Rice, 423 N.W.2d 671, 677-79 (Minn.1988) (applying the official immunity doctrine to trespass and battery claims brought against peace officers well after the Municipal Torts Claim Act was enacted).
The policy considerations that underlie official immunity have been seriously questioned. In Dan B. Dobbs, The Law of Torts, the author states:
The supposition is that immunity is required in at least some cases to assure that the ardor of public officials for performing their tasks will not be dampened. Courts usually assume that official ardor is desirable. They also assume it can be dampened unduly. The first point is a question of values; the second is a question of data. Neither point is demonstrated in most of the case discussions. The Supreme Court has also suggested that unfounded lawsuits entail social costs, including expenses of litigation and diversion of official energies. That argument, however, seems wide of the mark, since the result of immunity is to avoid the trial that could tell us whether the suit was unfounded or not.
Section 273, p. 733 (2000) (footnotes omitted). Yet we have continued to ground official immunity on the fear that “personal liability * * ⅜ might deter independent action and impair effective performance of [a public official’s] duties.” Terwilliger, 561 N.W.2d at 913. And, in fact, we have suggested that the immunity offered to public officials is even broader than that offered to the governmental entity. Holmquist, 425 N.W.2d at 233 n. 1; Pie-tan, 494 N.W.2d at 41 (“Official immunity involves the kind of discretion which is exercised on an operational rather than a policy making level * * *.”). See also Michael K. Jordan, Finding a Useful Path Through the Immunity Thicket, 61-Oct Bench & Bar of Minnesota 24, 28-29 (“[CJommon law official immunity does provide an extra layer of protection for officials who must exercise their discretion in situations where, out of necessity, the application of policy to facts occurs in a myriad of ways and cannot be fully explicated by the legislative or executive branches.”).
In so doing, we have not addressed the effect of the statutory protections available to public officials. Further, our cases on common law official immunity have not addressed the question of whether there should be an exception, comparable to that available under common law governmental immunity, for dangerous conditions in the road created by the acts of a public employee. It is difficult to imagine a rationale that would support such an exception to governmental immunity but not to official immunity.
*519The discussion thus far would support the suggestion that we should abolish the common law doctrine of official immunity, as we did with common law governmental immunity, or at least recognize an exception to common law official immunity for the creation of a dangerous condition on the roads. Although I would be open to consider that suggestion, the Schroeders do not present it here and our repeated, though unquestioned, application of official immunity after Ehvood may make abolition impractical in this case. See, e.g., Terwilliger, 561 N.W.2d at 913; Wiederholt v. City of Minneapolis, 581 N.W.2d 312, 315 (Minn.1998); Gleason v. Metro. Council Transit Operations, 582 N.W.2d 216, 220 (Minn.1998). But, even if we do not abolish official immunity or create an exception to it, this discussion is immediately relevant to the question of whether we should extend official immunity vicariously to the county under these facts.
Accordingly, although I concur that, under our present caselaw, Ario is entitled to official immunity on the claims that he negligently graded against traffic (because the decisions to not deadhead and to not place warning signs were within the operational discretion left open by the county’s grading policy), I disagree with the majority’s conclusion that the county is entitled to vicarious official immunity on those claims.
2. Under our facts, vicarious official immunity should be no greater than the county’s statutory immunity.
Although we have said that a governmental employer may be vicariously immune where its employee has official immunity (e.g., Terwilliger, 561 N.W.2d at 913), we have never gone so far as to say that the two must be equated. In fact, in Pletan we suggested that they need not necessarily be equal. 494 N.W.2d at 42. We did acknowledge that some of our decisions extended the benefit of employee immunity to the employer; we noted that they did so “without any explanation.” Id. We also observed that dicta in Holmquist suggested that we “might under some circumstances find an exception to vicarious official immunity.” Id.
Both Holmquist and Pletan cited a comprehensive law review article on this subject. George A. Bermann, Integrating Governmental and Officer Tort Liability, 77 Colum. L.Rev. 1175, 1186-87 (1977). Bermann surveyed the caselaw and concluded that the rationale of those cases recognizing that the government may be liable under some circumstances even though its officers are not, “rests on firmer ground.” He stated:
Even if we may assume that the prospect of governmental liability in damages will trouble any official who is sensitive to the disfavor of his superiors, it should not dampen his zeal nearly as much as the prospect of personal liability. Unless the government’s exposure to liability can genuinely be expected to impair seriously the official’s performance of duty, the government should not enjoy immunity from liability simply because the official is immune.
More important, as a policy matter, any rigid equation of governmental with officer immunity assumes a happy congruence between the compensatory purposes of tort law, on the one hand, and its deterrent and retributive purposes, on the other, that simply does not exist. Situations frequently arise in which it is appropriate to require the government to compensate for harm done by a public official, even though it is inappropriate to hold the official personally liable. For example, in the well-known case of Miller v. Horton, a state health officer, acting under a statute requiring him to *520destroy horses infected with glanders, ordered plaintiffs horse put to death in the reasonable though mistaken belief that it had the disease. Imposing liability on the government rather than the innocent officer in that case would have provided the victims a remedy, while distributing the loss over the entire community in whose interest the program was presumably initiated. This would have been a fair and sensible result. Moreover, even if an official has acted culpably, placing the full monetary burden on his shoulders may be out of proportion to his fault. Consider the case of a municipal power plant operator whose slight delay in responding to danger signals paves the way for a blackout with untold financial consequences for the entire community.
Id.
Recently, we reviewed the dicta in Holmquist and the discussion in Pletan in the context of a case where the employee’s official immunity was grounded in his adherence to a policy adopted by his employer. In Anderson v. Anoka Hennepin School District 11, we said:
Since Pletan, this court has further refined its vicarious official immunity standard. The court applies vicarious official immunity when failure to grant it would focus “stifling attention” on an official’s performance “to the serious detriment of that performance.”
678 N.W.2d at 664 (quoting Olson v. Ramsey County, 509 N.W.2d 368, 372 (Minn. 1993)). But in Anderson, the question of vicarious official immunity was virtually the same as that of statutory immunity because the challenged acts of the employee were in compliance with the employer’s policy. Thus the challenge to those acts was necessarily a challenge to the policy. We said that vicarious official immunity was necessary because liability would discourage the employer from developing policies and, instead, place “stifling attention” on the work of the employee. Anderson, 678 N.W.2d at 665 (quoting Olson, 509 N.W.2d at 372).
Anderson is distinguishable because the policy directed the precise actions complained of. Here, the policy left the precise actions to the operating judgment of Ario. For that reason, the claim here does not require a challenge to the county’s grading policy, and thus allowing the claim to proceed would not discourage future policy development.
There are several policy considerations that support the conclusion that the county should not be immune on the claims for which Ario is immune. First, the primary source for determining the liability of the county should be the Municipal Tort Claims Act. That act represents the solution achieved by the joint efforts of this court, which abolished common law governmental tort immunity, and the legislature, which enacted governmental tort liability. All of those efforts would have been for naught if they could simply be eclipsed by common law vicarious official immunity, which would replace the statutory discretion function exception with far broader common law immunity for all but ministerial acts. Second, because our common law official immunity doctrine is not firmly grounded in reasoned analysis, it should not be extended to governmental employers where there is no compelling policy reason to do so. Here, the policy concerns that underlie official immunity do not compel vicarious official immunity. As Bermann noted, the inhibiting effect of an official’s fear of government liability is not nearly as severe as the official’s fear of personal liability. And the county’s incentive to make policy will not be lessened because its policy is not challenged. Third, by requiring indemnification, the *521legislature has determined that the risk of liability should be shifted from the employee to the governmental body. Moreover, the legislature has limited the government’s liability by setting caps and authorizing insurance. Finally, this result would not put “stifling attention” on the work of grader operators, who make operating decisions, not policy, and who are protected by liability caps, insurance, and indemnification.
When the interests of the county, as thus analyzed, are weighed against the interests of an innocent victim to obtain compensation, I would conclude that vicarious official immunity should not be available.
. Prior to Anderson, this exception to governmental tort immunity was applied to municipalities but not to the state or to towns and counties that were regarded as “involuntary corporations” organized as political subdivisions of the state. See, e.g., Altnow v. Town of Sibley, 30 Minn. 186, 190, 14 N.W. 877, 877 (1883) (stating that "whatever may be the reasons assigned, and whether they are consistent, or, in all instances, sensible or not, the distinction between [municipalities on the one hand and towns and counties on the other] is clearly and firmly established.”), overruled, in part, by, Spanel v. Mounds View Sch. Dist., 264 Minn. 279, 118 N.W.2d 795 (1962); Tholkes v. Decock, 125 Minn. 507, 147 N.W. 648 (1914) (towns); Hitchcock v. County of Sherburne, 227 Minn. 132, 135, 34 N.W.2d 342, 344 (1948) (towns and counties); Gaare v. Bd. of County Comm’rs of Clay County, 90 Minn. 530, 531-32, 97 N.W. 422, 423 (1903) (counties); Stevens v. Lycan Co., 259 Minn. 106, 108, 105 N.W.2d 889, 891 (1960) (noting that "from early times” we held municipalities liable but, "[pjeculiarly, we have followed the opposite rule with respect to liability of towns and counties holding that neither the town or county or its officers are liable except for affirmative misconduct.” (footnotes omitted)). In Anderson, we eliminated the distinction between municipalities and the state and its subdivisions by holding "that the exceptions to common law governmental tort immunity [for municipalities] apply to other governmental entities, including the state, as well as to municipalities.” Anderson, 296 N.W.2d at 387. We said:
If the reason for municipal liability was that municipalities have control over city *511streets and the power to control defects, the same rationale would apply equally to the state with regard to state-controlled roads.
Id.
. Although these cases arose under the State Tort Claims Act, Minn.Stat. § 3.736, subd. 3(b), we have said that the discretionary function exception for liability of municipalities (including counties) is identical. Terwilliger v. Hennepin County, 561 N.W.2d 909, 912 (Minn. 1997) (stating that in reviewing the exception under section 466.03, subdivision 6, the court will be guided by the decisions under section 3.376, subdivision 3(b)).
. Of course, the analogy to private parties is not perfect. Perhaps government should be given more flexibility in allocating limited resources because it is choosing one public need over another, not preferring a private need over a public one. But this understates the public needs that can also be served by corporate profits. They may provide economic stability to a region, for the benefit of the government and its citizens; they often provide income to thousands or even millions of shareholders, including pension funds; they provide employment to citizens and tax revenues to local, state, and federal governments; and they may provide valuable goods and services. Further, governmental bodies are already given benefits not available to private parties — liability limits, notice of claim requirements, liability insurance, and employee indemnification. Although the comparison may not be fully appropriate, one would hope that the principles of law by which we encourage good conduct and discourage bad can be applied in a similar fashion to government as to private parties. In fact, one could suggest that government should provide leadership by being a role model of the type of conduct it would hope to see from private parties.
. The focus of Varda’s memorandum appears to be primarily on the question whether operators could expect to be ticketed, and only secondly on safety considerations. Thus, it does not truly reflect a policymaking decision, achieved by balancing competing political, social and economic concerns.
. Although there is evidence that graders could reach 30 miles per hour, Ario testified that the grader he was using could only reach a maximum speed of 23 miles per hour.
. Although we have not considered the issue, I agree with the court of appeals’ determination in Christensen v. Mower County, that consideration of de minimis cost factors is not a meaningful exercise of discretion. 587 *517N.W.2d 305, 308 (Minn.App.1998). See also Nguyen v. Nguyen, 565 N.W.2d 721, 724 (Minn.App.1997) (noting that "such relatively inexpensive remedies as guardrails or warning signs * * * may not require a policy decision regarding the allocation of resources.”).