POSNER, Circuit Judge.
While driving to work early one morning Robert Farringer, a rural mailman, struck a car driven by the plaintiff’s decedent, Glenn Konradi, killing him. The suit is against the United States under the Federal Tort Claims Act, 28 U.S.C. §§ 1346(b), 2671 et seq., with a pendent-party claim under state law against Farringer. The basis of both claims is that Farringer’s negligence in failing to yield the right of way to Konradi at an intersection was the cause of the accident. The district judge dismissed the suit on the government’s motion for summary judgment. He ruled that the accident had not occurred within the scope of Farringer’s employment by the Postal Service, which let off the Service; he then relinquished jurisdiction over the pendent party claim.
The parties agree that the question whether the accident occurred within the scope of Farringer’s employment is gov[1209]*1209erned by Indiana law, 28 U.S.C. § 1346(b); Williams v. United States, 350 U.S. 857, 76 S.Ct. 100, 100 L.Ed. 761 (1955) (per curiam), that under Indiana law it is a question of fact, Gibbs v. Miller, 152 Ind. App. 326, 329, 283 N.E.2d 592, 594 (1972), and therefore that the judge was right to dismiss the case on summary judgment only if no reasonable jury, presented with the evidence that was before the judge when he ruled, could have answered the question in the plaintiffs favor. One could quarrel with “therefore,” since whether a question is one of fact or of law has been held to fall on the procedure side of the substance/procedure divide that Erie R. Co. v. Tompkins, 304 U.S. 64, 58 S.Ct. 817, 82 L.Ed. 1188 (1938), established. Byrd v. Blue Ridge Rural Elec. Coop., Inc., 356 U.S. 525, 78 S.Ct. 893, 2 L.Ed.2d 953 (1958); Nunez v. Superior Oil Co., 572 F.2d 1119, 1125 (5th Cir.1978); Deland v. Old Republic Life Ins. Co., 758 F.2d 1331, 1335 (9th Cir.1985). Although the present case is not a diversity case, Erie was an interpretation of the Rules of Decision Act, 28 U.S.C. § 1652, and its principles apply to any case in federal court in which state law supplies the rule of decision, Morgan v. South Bend Community School Corp., 797 F.2d 471, 474 (7th Cir.1986); Hernas v. City of Hickory Hills, 507 F.Supp. 103, 105 (N.D.Ill.1981); Wright, Miller & Cooper, Federal Practice and Procedure § 4515 (1982), as it does here by virtue of 28 U.S.C. § 1346(b). Still, circumstances alter cases — or at least may. Byrd and the cases following it rely heavily on the Seventh Amendment, which has no application to the Federal Tort Claims Act, for the proposition that federal law determines when a question is factual, and therefore a jury issue, in a case tried in a federal court; the case for applying federal law to the law-fact issue in this case is therefore weakened. But so is the case for applying state law to the issue. Congress has given the federal courts exclusive jurisdiction over tort claims against the federal government, incorporating local law for the convenience of the federal government rather than to vindicate state policies— though on the other hand the states do have an interest in conduct of federal employees that injures the state’s citizens.
All this is as academic as it is interesting. No party argues in this case that federal law rather than state law should determine whether scope of employment is to be treated as a legal or as a factual question— perhaps believing, plausibly enough, that the question would be decided the same way under either law. Without further ado, therefore, we can turn to the merits.
The general rule is that an employee is not within the scope of his employment when commuting to or from his job. As the Supreme Court of Indiana put it the last time it addressed the issue, more than three decades ago, “an employee on his way to work is normally not in the employment of the corporation.” Biel, Inc. v. Kirsch, 240 Ind. 69, 73, 161 N.E.2d 617, 618 (1959) (per curiam). The rub is “normally,” and though omitted in the statement of the rule in Pursley v. Ford Motor Co., 462 N.E.2d 247, 249 (Ind.App.1984), this weasel word is definitely required for the sake of accuracy. In State v. Gibbs, 166 Ind.App. 387, 336 N.E.2d 703 (1975), the employer furnished the employee with a car for use on the job but also allowed him to take it home at night. The accident occurred while he was driving home, and the employer was held liable. In Gibbs v. Miller, supra, the employer was held liable for an accident that occurred when its traveling salesman, who used his own car to make his rounds, was driving home for lunch from an appointment with a customer; he had other appointments scheduled for that afternoon. On the other hand, in City of Elkhart v. Jackson, 104 Ind.App. 136, 10 N.E.2d 418 (1937), which also involved an employee driving the company car at lunch time — this time he was returning to work after lunch when the accident occurred— the accident was held to be outside the scope of employment. Biel, Inc. v. Kirsch, supra, was another company-car case, and again the accident (which occurred while the employee was driving the car to work one morning) was held to be outside the scope of employment — but the employee happened also to be the employer’s owner, [1210]*1210and, it seems, was using the car for her personal convenience rather than on company business. In City of Crawfordsville v. Michael, 479 N.E.2d 102 (Ind.App.1985), the employee was using the company car (actually truck) for personal business on his day off when the accident occurred; he too was held not to have been acting within the scope of his employment.
It is impossible to find the pattern in this carpet without a conception of what the law is trying to accomplish by making an employer liable for the torts of his employees committed within the scope of their employment and by excluding commuting from that scope — “normally.” The Indiana decisions are few and not articulate on these issues, and although there are plenty of cases in other states, they use a similar approach and are similarly reticent about the considerations that animate their decisions. Annot., Employer’s Liability for Negligence of Employee in Driving His Own Car, 52 A.L.R.2d 287, 303, 311 (1957); Annot.,
Free access — add to your briefcase to read the full text and ask questions with AI
POSNER, Circuit Judge.
While driving to work early one morning Robert Farringer, a rural mailman, struck a car driven by the plaintiff’s decedent, Glenn Konradi, killing him. The suit is against the United States under the Federal Tort Claims Act, 28 U.S.C. §§ 1346(b), 2671 et seq., with a pendent-party claim under state law against Farringer. The basis of both claims is that Farringer’s negligence in failing to yield the right of way to Konradi at an intersection was the cause of the accident. The district judge dismissed the suit on the government’s motion for summary judgment. He ruled that the accident had not occurred within the scope of Farringer’s employment by the Postal Service, which let off the Service; he then relinquished jurisdiction over the pendent party claim.
The parties agree that the question whether the accident occurred within the scope of Farringer’s employment is gov[1209]*1209erned by Indiana law, 28 U.S.C. § 1346(b); Williams v. United States, 350 U.S. 857, 76 S.Ct. 100, 100 L.Ed. 761 (1955) (per curiam), that under Indiana law it is a question of fact, Gibbs v. Miller, 152 Ind. App. 326, 329, 283 N.E.2d 592, 594 (1972), and therefore that the judge was right to dismiss the case on summary judgment only if no reasonable jury, presented with the evidence that was before the judge when he ruled, could have answered the question in the plaintiffs favor. One could quarrel with “therefore,” since whether a question is one of fact or of law has been held to fall on the procedure side of the substance/procedure divide that Erie R. Co. v. Tompkins, 304 U.S. 64, 58 S.Ct. 817, 82 L.Ed. 1188 (1938), established. Byrd v. Blue Ridge Rural Elec. Coop., Inc., 356 U.S. 525, 78 S.Ct. 893, 2 L.Ed.2d 953 (1958); Nunez v. Superior Oil Co., 572 F.2d 1119, 1125 (5th Cir.1978); Deland v. Old Republic Life Ins. Co., 758 F.2d 1331, 1335 (9th Cir.1985). Although the present case is not a diversity case, Erie was an interpretation of the Rules of Decision Act, 28 U.S.C. § 1652, and its principles apply to any case in federal court in which state law supplies the rule of decision, Morgan v. South Bend Community School Corp., 797 F.2d 471, 474 (7th Cir.1986); Hernas v. City of Hickory Hills, 507 F.Supp. 103, 105 (N.D.Ill.1981); Wright, Miller & Cooper, Federal Practice and Procedure § 4515 (1982), as it does here by virtue of 28 U.S.C. § 1346(b). Still, circumstances alter cases — or at least may. Byrd and the cases following it rely heavily on the Seventh Amendment, which has no application to the Federal Tort Claims Act, for the proposition that federal law determines when a question is factual, and therefore a jury issue, in a case tried in a federal court; the case for applying federal law to the law-fact issue in this case is therefore weakened. But so is the case for applying state law to the issue. Congress has given the federal courts exclusive jurisdiction over tort claims against the federal government, incorporating local law for the convenience of the federal government rather than to vindicate state policies— though on the other hand the states do have an interest in conduct of federal employees that injures the state’s citizens.
All this is as academic as it is interesting. No party argues in this case that federal law rather than state law should determine whether scope of employment is to be treated as a legal or as a factual question— perhaps believing, plausibly enough, that the question would be decided the same way under either law. Without further ado, therefore, we can turn to the merits.
The general rule is that an employee is not within the scope of his employment when commuting to or from his job. As the Supreme Court of Indiana put it the last time it addressed the issue, more than three decades ago, “an employee on his way to work is normally not in the employment of the corporation.” Biel, Inc. v. Kirsch, 240 Ind. 69, 73, 161 N.E.2d 617, 618 (1959) (per curiam). The rub is “normally,” and though omitted in the statement of the rule in Pursley v. Ford Motor Co., 462 N.E.2d 247, 249 (Ind.App.1984), this weasel word is definitely required for the sake of accuracy. In State v. Gibbs, 166 Ind.App. 387, 336 N.E.2d 703 (1975), the employer furnished the employee with a car for use on the job but also allowed him to take it home at night. The accident occurred while he was driving home, and the employer was held liable. In Gibbs v. Miller, supra, the employer was held liable for an accident that occurred when its traveling salesman, who used his own car to make his rounds, was driving home for lunch from an appointment with a customer; he had other appointments scheduled for that afternoon. On the other hand, in City of Elkhart v. Jackson, 104 Ind.App. 136, 10 N.E.2d 418 (1937), which also involved an employee driving the company car at lunch time — this time he was returning to work after lunch when the accident occurred— the accident was held to be outside the scope of employment. Biel, Inc. v. Kirsch, supra, was another company-car case, and again the accident (which occurred while the employee was driving the car to work one morning) was held to be outside the scope of employment — but the employee happened also to be the employer’s owner, [1210]*1210and, it seems, was using the car for her personal convenience rather than on company business. In City of Crawfordsville v. Michael, 479 N.E.2d 102 (Ind.App.1985), the employee was using the company car (actually truck) for personal business on his day off when the accident occurred; he too was held not to have been acting within the scope of his employment.
It is impossible to find the pattern in this carpet without a conception of what the law is trying to accomplish by making an employer liable for the torts of his employees committed within the scope of their employment and by excluding commuting from that scope — “normally.” The Indiana decisions are few and not articulate on these issues, and although there are plenty of cases in other states, they use a similar approach and are similarly reticent about the considerations that animate their decisions. Annot., Employer’s Liability for Negligence of Employee in Driving His Own Car, 52 A.L.R.2d 287, 303, 311 (1957); Annot., Employer’s Liability for Employee’s Negligence in Operating Employer’s Car in Going to or from Work or Meals, 52 A.L.R.2d 350, 354, 362-63 (1957). There is however a rich scholarly literature on vicarious liability, specifically of employers, from which clues can be gleaned. Sykes, The Boundaries of Vicarious Liability: An Economic Analysis of the Scope of Employment Rule and Related Legal Doctrines, 101 Harv.L.Rev. 563 (1988), is particularly helpful; we have relied on it previously, in an opinion by Judge Manion, Wilson v. Chicago, Milwaukee, St. Paul & Pac. R.R., 841 F.2d 1347, 1352, 1356 n. 2 (7th Cir.1988), to help decide a scope of employment issue.
Often an employer can reduce the number of accidents caused by his employees not by being more careful — he may already be using as much care in hiring, supervising, monitoring, etc. his employees as can reasonably be demanded — but by altering the nature or extent of his operations: in a word by altering not his care but his activity. This possibility is a consideration in deciding whether to impose strict liability generally. Anderson v. Marathon Petroleum Co., 801 F.2d 936, 939 (7th Cir.1986); Bethlehem Steel Corp. v. EPA, 782 F.2d 645, 652 (7th Cir.1986); Shavell, Strict Liability versus Negligence, 9 J. Legal Stud. 1 (1980). The liability of an employer for torts committed by its employees — without any fault on his part — when they are acting within the scope of their employment, the liability that the law calls “respondeat superior,” is a form of strict liability. It neither requires the plaintiff to prove fault on the part of the employer nor allows the employer to exonerate himself by proving his freedom from fault. The focus shifts from changes in care to changes in activity. For example, instead of dispatching its salesmen in cars from a central location, causing them to drive a lot and thus increasing the number of traffic accidents, a firm could open branch offices closer to its customers and have the salesmen work out of those offices. The amount of driving would be less (an activity change) and with it the number of accidents. Firms will consider these tradeoffs if they are liable for the torts of their employees committed within the scope of their employment, even if the employer was not negligent in hiring or training or monitoring or supervising or deciding not to fire the employee who committed the tort. This liability also discourages employers from hiring judgment-proof employees, which they might otherwise have an incentive to do because a judgment-proof employee, by definition, does not have to be compensated (in the form of a higher wage) for running the risk of being sued for a tort that he commits on his employer’s behalf. He runs no such risk; he is not worth suing.
If it is true that one objective of the doctrine of respondeat superior is to give employers an incentive to consider changes in the nature or level of their activities, then “scope of employment” can be functionally defined by reference to the likelihood that liability would induce beneficial changes in activity. It becomes apparent for example that the employer should not be made liable for a tort committed by the employee in the employee’s home, for there is no plausible alteration in the activity of [1211]*1211the employer that would substantially reduce the likelihood of such a tort. This overstates the case a bit; one can imagine a plaintiffs arguing that if the employer had not made the employee work so hard the employee would have been more alert and therefore more careful and the accident would not have occurred. But the law has to draw some lines for ease of administration, and a rough-and-ready one is between accidents on the job and accidents off the job — including accidents while commuting — in recognition of the fact that the employer’s ability to prevent accidents by employees is normally much less when the employees are not at work. Indiana recognizes, however, that the line is indeed a rough one, and it allows juries to cross it when particular circumstances make the line inapt to the purpose that it seeks to implement. Whether it is wise to give juries such discretion is not our business, at least given the parties’ agreement that the state rule empowering the jury to decide whether the accident was within the scope of the employment governs this case.
The Postal Service, Parringer’s employer, requires its rural postal carriers to furnish their own vehicle (Parringer’s was a pick-up truck) in making their rounds. Postal Operations Manual § 634.21 (1985). The alternative would be for the Service to buy or lease mail trucks for these carriers to use. A possible consequence of the choice it has made is to increase the amount of driving over what it would be if the Service furnished the vehicles. No family with one car (and precious few with two) would want to leave its car at work and thereby have to find an alternative method of commuting. The Postal Service’s rule pretty much guarantees that its mailmen will drive to and from work, and by doing this it increases the amount of driving compared to a system in which, since the mailman does not need to have his own car at work, he can take a train or bus or join a car pool. One cost of more driving is more accidents, and this cost can be made a cost to the Postal Service, and thus influence its choice between furnishing its mailmen with vehicles and requiring them to furnish their own, if the scope of employment is defined for purposes of tort law as including commuting in all cases in which the employee is required to furnish a vehicle for use at work. The argument for liability might actually seem stronger than if the employer had furnished the vehicle. But it must be borne in mind that the question of the employer’s liability in cases involving a company-owned vehicle arises only when the vehicle is being used outside of work time; and here it can be argued that a person furnished with a company car is apt to drive more, and more carelessly, than if he were using his own ear.
All this is highly speculative. The Postal Service’s rule is limited to rural deliverymen, and neither public transportation nor car pooling is common in rural America. Especially since any expansion in tort liability is bound to be a source of litigation costs and judicial burdens, we could not be sufficiently confident concerning the effects of liability to be justified in laying down a general rule that employers who require their employees to use their own vehicle on the job, or permit them to use a company vehicle off the job, are liable for the employees’ accidents while commuting in that vehicle; nor would that be a plausible extrapolation from the Indiana cases. But additional evidence in this case points to employer liability. According to testimony that for purposes of this appeal (only) we must take to be true, Farringer’s postmaster required the postal carriers to take the most direct route in driving to and from work, and hence not to divagate for personal business. Nor was the carrier to stop for such business, or give anyone a ride. And he was to fasten his seatbelt (this was before Indiana passed a seatbelt law). The record does not reveal the reasons for these requirements. They may just reflect the Postal Service’s fear of being held liable for commuting accidents and its concomitant desire to minimize the length of the trip and number of persons in the .employee’s car in order to reduce the likelihood of accidents. If this is right the plaintiff can do nothing with the requirements, because a person’s fear of being held liable is not a reason for the law’s [1212]*1212holding the person liable. An employer should not be held liable, and therefore penalized, for taking steps rationally designed to minimize its liability by increasing the safety of its operations.
Another possibility, however, is that the Postal Service was trying to minimize time lost by its employees from work and its workers’ compensation costs, for the government interprets the federal employees’ compensation law to include the commuting accidents of postal workers. U.S. Dept. of Labor, Wage and Labor Standards Administration, Bureau of Employees’ Compensation, FECA Memorandum No. 104 (Oct. 24, 1969), interpreting 5 U.S.C. § 8102(a); but see Avasthi v. United States, 608 F.2d 1059 (5th Cir.1979). This is not to suggest that the scope of liability for workers’ compensation purposes is identical to that for liability to third-party victims of those workers’ torts; in fact it is broader. Calabresi, Some Thoughts on Risk Distribution and the Law of Torts, 70 Yale L.J. 499, 544-45 (1961). The point is different. If the Postal Service insists for time’s sake that the carrier always travel to and from work by the most direct route, which may not be the safest route, the Service should be liable for the accidents that result from this directive; it has made them more likely.
The rules of commuting that the postmaster has imposed upon his carriers may also or instead reflect a belief that the work of a rural deliveryman begins when he gets into his car in the morning and ends when he gets out of it in the evening. For during all that time he has control over an essential instrumentality of postal service — the delivery van — albeit supplied by the deliverer. This underscores the earlier point that the Postal Service has made a choice between buying its own fleet of vans and pressing its employees’ vans into service, and may bring the case within the orbit of State v. Gibbs. It is as if the Postal Service had decided to store its vans at night in its employees’ garages. But against this is the fact that the mileage allowance which the Service gives its employees when they use their own vehicle on the job does not include the use of the vehicle in commuting. Moreover, there is no general rule making the employer liable for a commuting accident merely because he supplied the vehicle. Biel no doubt is a special case, and so is Michael, where the employer-owned car was being used purely for personal business, since it was the employee’s day off; and in State v. Gibbs the employer was held liable. On the other hand Jackson seems indistinguishable from State v. Gibbs, yet was decided in favor of the employer. The analysis sketched above suggests that there is no magic to the employer’s supplying the car; the functional argument (promotion of safety) for employer liability is as strong if he makes the employee bring his own car. Gibbs v. Miller imposed liability in such a case. Of course the two Gibbs cases are factually different from our case — every case is factually different from every other case — but that is no warrant for refusing to follow them in this case unless the factual differences between them and this case are connected with a difference in principle.
After and because of the accident, the Postal Service fired Farringer. This may have been because it feared that he might have a similar accident, for which the Service would incontestably be liable, while on the job; another possibility however is that the Service considered the tort he did commit to have occurred on the job. But this consideration seems merely to duplicate the one discussed in the preceding paragraph, and it is therefore entitled to no weight.
Not only may the imposition of liability on the Postal Service be consistent with most of the Indiana cases (indeed all but Jackson); it is consistent with all three of the formulas that courts in Indiana and elsewhere intone when they are trying to generalize about scope of employment. Reed v. House of Decor, Inc., 468 So.2d 1159, 1161 (La.1985). By driving to and from work Farringer conferred a benefit on his employer because he was bringing an essential instrumentality of the employer’s business. (True, the employer would not have cared if Farringer had left his truck in the post office parking lot and [1213]*1213thumbed a ride to work, but few employees would thus forgo all personal use of their vehicle.) The employer exerted substantial control over the employee’s commuting, as shown by the regulations discussed earlier. And finally the employee while commuting was in the service of the employer because he was keeping and maintaining the instrumentality.
These “tests” should not be thought conclusive. Tests divorced from purposes tend not to be useful, let alone conclusive, and the linkage between the tests and what the discussion in this opinion conjectures is the underlying purpose of the scope of employment concept is obscure. The law has drawn a line between at work and at home but treats commuting as an intermediate zone that can be placed within or outside the scope of employment depending on circumstances, though the presumption is in favor of outside. The purpose of a doctrine determines what circumstances are relevant. The purpose of this doctrine may be to induce the employer to consider activity changes that might reduce the number of accidents. One possible change might be to substitute a fleet of postal vans for the employees’ personal vehicles driven to and from work daily perhaps over substantial distances.
In attempting to predict how the Indiana courts would have decided this case had it been brought in such a court (which it could not have been, of course, because the federal government has not consented to be sued in state courts), this court necessarily is speculating. There is no reason to apologize for this acknowledgment. The decision of a federal court in a diversity case, or in any other case in which state law supplies the rule of decision, is an exercise in predicting how the highest court of the state would decide the case if it were presented to it. When the relevant decisions of the state’s courts do not articulate the grounds that animate their results, and the issue is not suitable for certification to the state supreme court (perhaps because as here it is highly fact-specific), the federal court has no choice but to speculate as to what the true grounds might be or to supply grounds that it thinks might recommend themselves to the state’s courts in the future. It is in that spirit that this opinion has sought to bring modern scholarship to bear on the vexing issue of scope of employment in commuting cases.
The unavoidably speculative character of the analysis furnishes an additional reason for believing that the district court acted prematurely in granting summary judgment. The more nebulous or unsettled the legal standard, the more difficult it should be to exclude contested facts from consideration on the ground that they are immaterial. If the Indiana rule excluding commuting from the scope of employment were strict, the Postal Service would be right to argue that it is immaterial why it fired Farringer. But as it is not strict, the question may be material. Perhaps the Postal Service fired Farringer because it considers commuting to be part of working and because it has a policy of firing people who kill tortiously in the course of their employment with the Service. The scope of federal employees’ compensation may be irrelevant to scope of employment for liability purposes, but then again it may not be. Perhaps as the facts are developed it will become evident that Farringer was not acting within the scope of his employment, but on the basis of the record compiled thus far it cannot be said that no reasonable jury could find that he was.
It should go without saying that the recitation of facts in this opinion is tentative; the facts may appear quite different after further proceedings. In particular the nature of and authority for the various rules to which the Knightstown Post Office where Farringer worked, as distinct from the Postal Service itself, subjected Farringer are wholly unclear.
One final point. The plaintiff joined Farringer as a pendent party defendant, and the district court dismissed Farringer from the suit without prejudice after deciding to dismiss the main claim, that is, the claim against the United States. The court did this on the familiar ground that, with immaterial exceptions, when the main claim [1214]*1214drops out before trial the court should relinquish jurisdiction over all pendent claims. United Mine Workers v. Gibbs, 383 U.S. 715, 726, 86 S.Ct. 1130, 1139, 16 L.Ed.2d 218 (1966). The result was right, but not the reasoning. The Supreme Court had held several months before the judgment in this case that there is no pendent party jurisdiction in cases brought under the Tort Claims Act. Finley v. United States, 490 U.S. 545, 109 S.Ct. 2003, 104 L.Ed.2d 593 (1989). The district court never acquired jurisdiction over Farringer (for he and the decedent, whose citizenship is what counts for diversity purposes although his personal representative is the actual party, 28 U.S.C. § 1332(c)(2), are both citizens of Indiana) — against whom, by the way, the plaintiff has filed a parallel suit in state court. Although Finley was decided after the present suit was brought, there is no reason not to apply it to this case; the harm to the plaintiff is slight, since she has another and solvent defendant (the United States) to pursue in this case, and a suit against Farringer in state court.
The dismissal of Farringer (without prejudice, of course) is affirmed, albeit on a different ground from the district judge’s; but the dismissal of the United States is reversed and the case remanded for further proceedings consistent with this opinion. Costs in this court to the plaintiff.
Affirmed in Part, Reversed in Part, and Remanded.