SUTTON, J., delivered the opinion of the court, in which OBERDORFER, D.J., joined.
CLAY, J. (pp. 337-40), delivered a separate opinion concurring in part and dissenting in part.
OPINION
SUTTON, Circuit Judge.
Federal statutory law and constitutional law offer private citizens two avenues to recover damages for constitutional torts committed by employees of the federal government. A plaintiff may sue the employee directly for the constitutional violation under Bivens v. Six Unknown Named Agents of Federal Bureau of Narcotics, 403 U.S. 388, 91 S.Ct. 1999, 29 L.Ed.2d 619 (1971). And a plaintiff may sue the government under the Federal Tort Claims Act (FTCA), on the condition that the “judgment in an [FTCA] action ... shall constitute a complete bar to any action by the claimant, by reason of the same subject matter, against the employee of the government whose act or omission gave rise to the claim.” 28 U.S.C. § 2676.
In this case, Ronnie Harris attempted to use both avenues. The district court dismissed Harris’s Bivens claims on statute-of-limitations grounds. And after a bench trial, the district court rejected Harris’s FTCA claims on the merits. Because a reasonable factfinder could fairly conclude that Harris has failed to show that the individual defendants assaulted or maliciously prosecuted him under Ohio law, we affirm the district court’s rejection of Harris’s FTCA claims. And because that ad[325]*325judication on the merits of Harris’s FTCA claims bars further adjudication of Harris’s Bivens claims against the individual defendants, we also affirm the court’s dismissal of the Bivens claims.
I.
Ronnie Harris, an Ohio resident, is a former Olympic and professional boxer, regarded as one of the best middleweights that the State has ever produced. His boxing career included a gold medal at the 1968 Olympics and victories over Sugar Ray Seales and future world champion Alan Minter.
On January 18, 1997, Harris arrived at Cleveland Hopkins Airport on Continental Airlines Flight 237 from Fort Lauderdale, Florida, where he had traveled to watch his daughter compete in a tennis tournament. He wore a jogging suit and tennis shoes and carried a briefcase but wore no overcoat.
Several federal agents in plain clothes— Debra Harrison, Henry O’Bryant, Raphael Caprez and James Gilchrist — were working at the airport on January 13, and one of their primary duties was to scan the airport for suspicious persons and activity. Because the Drug Enforcement Agency (DEA) has designated Fort Lauderdale as a “source city” for drug trafficking, Caprez and O’Bryant monitored the passengers of Flight 237 as they exited the aircraft.
Upon leaving the plane, Harris walked rapidly to the baggage claim area and immediately went outside the airport to look for his uncle, who was picking him up. O’Bryant noticed Harris because he wore a jogging suit yet carried a briefcase; because he was not carrying an overcoat despite the cold temperature; and because he walked rapidly through the airport while looking around — all of which O’Bryant, a 22-year law enforcement veteran, associated with a drug courier profile. O’Bryant (and eventually Caprez) followed Harris to the baggage claim area.
O’Bryant approached Harris outside the airport and, after identifying himself, asked to speak to him. Harris declined to speak with O’Bryant, and the two men walked away from each other. During this brief encounter, Caprez made eye contact with Harris but did not approach him or overhear the conversation. Harris remained outside the terminal; O’Bryant and Caprez returned to the airport. Harris then returned to the airport without his briefcase (which he had left with his uncle, who had now arrived), walked by the American Airlines baggage claim carousel and glanced at a green piece of luggage. Upon observing this, O’Bryant called for a dog to sniff the bag.
At this point, Harrison joined Caprez and began to observe Harris’s actions. Harris retrieved his bag and, having guessed that Caprez was a DEA agent, approached him and said something like: “I don’t appreciate you and your monkeys following me and if you keep it up I’ll rip your head off.” JA 130-31. Harrison, who overheard the comment, walked up to Harris from behind, tapped his shoulder and asked whether he had just threatened a police officer. Harris asked Harrison who she was; she responded that she was a “police” officer, JA 131, 371; and Harris pushed her in the chest, prompting Harrison to arrest him.
As the officers led Harris to the nearby DEA task force office, they “could feel him resisting and trying to get loose.” JA 131. Once inside the office, they handcuffed him to a chair. O’Bryant entered the room and sought basic identification information from Harris. Harris generally was not cooperative during questioning. Gilchrist also entered the room with a drug-sniffing dog. The dog growled a few times during [326]*326the few minutes that he was in the room, and Harris feared that the dog might “start biting [him],” JA 502. Harris also apparently “bladed” his body (a reference to assuming a sideways, “aggressive” “fight stance” toward the officers), told the officers that he was “fast” and told Harrison that she did not know “what he could have done to [her].” JA 132-33, 436-37. After Harris disclosed where he had left his briefcase, O’Bryant recovered it and discovered no contraband. The officers then took Harris to the police station, where they charged him with aggravated disorderly conduct, resisting arrest and assault on a police officer. An Ohio grand jury indicted Harris on the assault charge on March 4, 1997. And a jury later found him not guilty of that charge.
On the basis of these incidents, Harris filed a Bivens lawsuit against the individual task force officers on January 13, 1998. The officers moved to dismiss the lawsuit on grounds of defective process, a motion that the district court granted without prejudice on October 26, 1999. Harris appealed that dismissal to the Sixth Circuit, and we affirmed on March 26, 2001. See Hards v. City of Cleveland, 7 Fed.Appx. 452 (6th Cir.2001).
On March 22, 2002 (over a year after the district court dismissed the complaint but less than a year after this court’s affir-mance), Harris filed a new complaint against the individual federal defendants under the Federal Tort Claims Act and Bivens. Harris invoked the district court’s jurisdiction under 28 U.S.C. § 2675(a), noting that he had filed an FTCA claim with the DEA on May 12, 1998, and that, because the agency had failed to act on the claim within six months, the claim was presumed denied.
On May 28, 2002, the United States filed a notice of substitution under 28 U.S.C. § 2679, which allows it to replace individual defendants with itself in defending FTCA claims. The individual defendants then filed motions to dismiss the Bivens claims for untimeliness (because they were filed over two years after the incidents), insufficient service of process and lack of personal jurisdiction. Harris responded that the Ohio Savings Statute, Ohio Rev. Code § 2305.19, gave him one year from the date of the Sixth Circuit’s decision to file the lawsuit.
On October 23, 2002, the district court substituted the United States as the sole defendant for the FTCA claims, and dismissed the Bivens claims against all four individual defendants because they had been filed over one year after the district court’s dismissal decision. On December 16, 2002, Harris filed a second amended complaint, this time solely against the United States, which asserted claims of assault, malicious prosecution, false arrest, false imprisonment, battery and abuse of process, all under the FTCA. After a bench trial, the district court ruled in favor of the United States on Harris’s FTCA claims.
II.
On appeal, Harris first argues that the district court mistakenly rejected his FTCA claims on the merits. The FTCA “waives, with certain limitations, [the federal government’s] governmental immunity to suit in tort and permits suits on tort claims to be brought against the United States.” S. Rep. 1400, 79th Cong., 2d Sess. 29 (1946). Under the FTCA, the United States is generally liable for the acts of its employees “in the same manner and to the same extent as a private individual under like circumstances,” 28 U.S.C. § 2674, which means that it is liable for the payment of money damages according to the law of the place where the act or omission occurred, typically state law and [327]*327actually Ohio law in this instance. Before invoking the FTCA remedy, a claimant must present the claim to the federal agency whose activities gave rise to the injuries and allow the agency the opportunity to settle the claim. 28 U.S.C. § 2675. If the agency denies the claim or takes no action (as the United States did here), the claimant may bring suit in federal district court (as Harris did here). See 28 U.S.C. § 2675(a). FTCA lawsuits are tried in federal district court without a jury. See 28 U.S.C. § 2402.
In addressing the dismissal of Harris’s FTCA claims, the district court’s factual findings “shall not be set aside unless clearly erroneous, and due regard shall be given to the opportunity of the trial court to judge of the credibility of the witnesses.” Fed.R.Civ.P. 52(a); Anderson v. Bessemer City, 470 U.S. 564, 572, 105 S.Ct. 1504, 84 L.Ed.2d 518 (1985); Hogan v. United States, 407 F.3d 778, 781 (6th Cir.2005) (noting that fact findings and credibility determinations in this setting are entitled to great deference).
A.
Harris argues that the four federal agents maliciously charged him with assault, resisting arrest and aggravated disorderly conduct. We disagree.
Like the English common law that preceded it, Ohio law has long recognized the right to recover in tort for the misuse of civil and criminal actions. See, e.g., Pope v. Pollock, 46 Ohio St. 367, 21 N.E. 356 (1889). To prove a malicious-criminal-prosecution claim under Ohio law, a plaintiff must demonstrate (1) that the government officials instituted or continued criminal proceedings with malice, (2) that they lacked probable cause and (3) that the proceedings were then terminated in favor of the accused. See Trussell v. Gen. Motors Corp., 53 Ohio St.3d 142, 559 N.E.2d 732, 736 (1990).
Much like federal constitutional law, Ohio defines probable cause as “a reasonable ground of suspicion, supported by circumstances sufficiently strong in themselves to warrant a cautious man in the belief that the person accused is guilty of the offense with which he is charged,” Rogers v. Barbera, 170 Ohio St. 241, 164 N.E.2d 162, 166 (1960), not as the quantum of evidence necessary to convict a criminal defendant, see, e.g., Deoma v. Shaker Heights, 68 Ohio App.3d 72, 587 N.E.2d 425, 428 (1990). Because “[tjhere is no requirement that the defendant [to a malicious-prosecution charge] must have evidence that will ensure a conviction,” id., not every failed criminal prosecution will sustain a subsequent malicious-prosecution suit, see, e.g., Huber v. O’Neill, 66 Ohio St.2d 28, 419 N.E.2d 10, 11-12 (1981). “An indictment,” moreover, “is prima facie evidence of probable cause and a plaintiff must bring forward substantial evidence to rebut this,” Carlton v. Davisson, 104 Ohio App.3d 636, 662 N.E.2d 1112, 1121 (1995) — for example, by showing that “the return of the indictment resulted from perjured testimony or that the grand jury proceedings were otherwise significantly irregular,” Deoma, 587 N.E.2d at 428.
Under Ohio law, malice is “the state of mind under which a person intentionally does a wrongful act without a reasonable lawful excuse and with the intent to inflict injury or under circumstances that the law will imply an evil intent.” Criss v. Springfield Township, 56 Ohio St.3d 82, 564 N.E.2d 440, 443 (1990). “For purposes of malicious prosecution, [the term] means an improper purpose, or any purpose other than the legitimate interest of bringing an offender to justice.” Id. Malice “may be inferred from the absence of probable cause” in appropriate circum[328]*328stances, Garza v. Clarion Hotel, Inc., 119 Ohio App.3d 478, 695 N.E.2d 811, 813 (1997), because “the want of probable cause is the real gist of the [malicious-prosecution] action,” Melanowski v. Judy, 102 Ohio St. 153, 131 N.E. 360, 361 (1921).
Harris has failed to satisfy these requirements. As an initial matter, the district court, in its role as a factfinder in this bench trial, declined to infer malice from the officers’ actions in this case and specifically noted that “there is no evidence that the proceedings were instituted with malice,” JA 138 — a finding that is fatal to Harris’s claims that he was maliciously prosecuted. See, e.g., Williams v. Cambridge Bd. of Educ., 370 F.3d 630, 643 (6th Cir.2004) (rejecting plaintiffs Ohio-law malicious-prosecution claim when plaintiff “failed to allege any facts that demonstrate[d defendant] acted with malice toward him in commencing [the] prosecution”); Coker v. Summit County Sheriff’s Dep’t, 90 Fed.Appx. 782, 786 (6th Cir.2003) (rejecting Ohio malicious-prosecution claim because the complaint “failed to set forth any facts in support of a showing of malice”).
But even if we were to ignore the district court’s factual findings on this score and even if we were to consider Harris’s invitation to presume malice under these circumstances, Harris has not established another key element of Ohio’s malicious-prosecution claim — the lack of probable cause — for any of the three counts with which he was charged.
First, a reasonable factfinder could fairly conclude that the officers had probable cause to charge Harris with criminal assault under Ohio Revised Code § 2903.13(a), which proscribes “knowingly causfing] or attempt[ing] to cause physical harm to another.” Ohio Rev.Code § 2903.13(a); see also Ohio Rev.Code § 2901.22(b). State prosecutors brought the assault charge against Harris in state court only after obtaining an indictment, which establishes “prima facie evidence of probable cause,” Carlton, 662 N.E.2d at 1121, that Harris has not tried to rebut by showing that the grand jury proceedings (or any other aspect of the indictment) were irregular. And cases interpreting Ohio law in analogous circumstances confirm that the officers had probable cause to believe that Harris knowingly attempted to cause physical harm by pushing Officer Harrison. See, e.g., Palshook v. Jarrett, 120 F.Supp.2d 641, 649 (N.D.Ohio 2000); Stillwell v. City of Xenia, No.2000CA-41, 2001 WL 127880, at *4, 2001 Ohio App. LEXIS 573, at *11-12 (Ohio Ct.App. Feb.16, 2001) (noting that “evidence existed to support a criminal assault charge” under Ohio law when plaintiff “intended to and did strike [the officer] in the chest with both hands”); Matlock v. Ohio Dep’t of Liquor Control, 77 Ohio Misc.2d 13, 665 N.E.2d 771, 773, 775 (1996) (holding that “plaintiff did not prove by a preponderance of the evidence that her arrest [for assault under § 2903.13] was unlawful or that her prosecution, was malicious” when she “grabbed [the officer] from behind,” the officer “pushed [her] away,” and she then “grabbed [the officer’s] shirt”); cf. Hopkins v. City of Westland, 21 F.3d 427 (6th Cir. Apr. 4, 1994) (holding, in a non-Ohio-law case, that probable cause existed where plaintiff shoved police officer in the chest).
In attempting to escape the reach of these cases, Harris principally relies on the observation that Officer Harrison did not fear for her own safety during the encounter and on a trio of cases — Watkins v. Millennium Sch., 290 F.Supp.2d 890 (S.D.Ohio 2003), Smith v. John Deere Co., 83 Ohio App.3d 398, 614 N.E.2d 1148 (1993), and State v. Weber, No. H-98-005, 1998 WL 700676, 1998 Ohio App. LEXIS [329]*3294822 (Ohio Ct.App. Oct. 9, 1998) — that deal with the standard for proving the civil tort of assault under Ohio law. But these cases and Officer Harrison’s frame of mind do not advance a malicious prosecution claim for criminal assault, where the relevant mental state is the assaulter’s, not the victim’s. Also unavailing is Harris’s reliance on cases — such as State v. Weber, No. H-98-005, 1998 WL 700676, 1998 Ohio App. LEXIS 4822 (Ohio Ct.App. Oct. 9, 1998), State v. Kelly, No. CA2001-12-104, 2002 WL 31414931, 2002 Ohio App. LEXIS 5738 (Ohio Ct.App. Oct. 28, 2002), and In re Mark M., Nos. E-99-028/E-99-046, 2000 WL 125800, 2000 Ohio App. LEXIS 311 (Ohio Ct.App. Feb. 4, 2000) — that pertain to Harris’s culpability and the standard required to prove criminal assault in a criminal trial, not the standard required to establish probable cause to arrest an individual for criminal assault. See Lyons v. City of Xenia, 417 F.3d 565, 575 (6th Cir.2005) (recognizing that, “[i]n ascertaining whether a constitutional violation occurred, the only question is whether [the officer] had probable cause — not whether the evidence would be sufficient to support a conviction”). The officers, in short, had probable cause to arrest Harris and to charge him with criminal assault.
Second, the officers had probable cause to charge Harris with resisting arrest under § 2921.33 of the Ohio Revised Code, which proscribes “recklessly or by force, [ ] resisting] or interfering] with a lawful arrest of the person or another.” Ohio Rev.Code § 2921.33. As we have already held that the officers lawfully arrested Harris, the sole question is whether the officers had probable cause to believe that he recklessly resisted or interfered with that arrest. A reasonable factfinder could conclude that Harris’s actions — his uncooperative attitude, attempts to “resist[] and try[] to get loose,” unwillingness to accompany the officers to the DEA office and confrontational and threatening statements and gestures — satisfy this requirement, despite the lack of an “actual struggle.” JA 131; see Hansen v. Westerville City Sch. Dist., Bd. of Educ., Nos. 93-3231/93-3303, 1994 WL 622153 at*6, 1994 U.S.App. LEXIS 31576, at *15 (6th Cir. Nov. 7, 1994) (holding that plaintiff had failed to show officers lacked probable cause to arrest and prosecute him for resisting arrest under Ohio law, when he “pulled his arm away from [the officer], told him to ‘get your paws off me,’ and struggled with the arresting officers”); see also Palshook, 120 F.Supp.2d at 651. Making Harris’s contrary position more tenuous, Ohio appellate courts have upheld jury verdicts convicting criminal defendants of resisting arrest in circumstances not far different from these. See State v. Keegan, 67 Ohio App.3d 824, 588 N.E.2d 928, 930 (1990) (holding that “physical activity which prevents or delays an arrest,” including “going limp,” is enough to sustain a conviction for resisting arrest under Ohio law); State v. Thomas, No. 2910, 1995 WL 39402, 1995 Ohio App. LEXIS 397 (Ohio Ct.App. Feb. 1, 1995) (raising one’s fists and backing away is sufficient); State v. Rose, 75 Ohio App.3d 656, 600 N.E.2d 382, 384 (1991) (moving arm for twenty or thirty seconds to avoid handcuffing is sufficient); State v. Williams, 84 Ohio App.3d 129, 616 N.E.2d 540, 543 (1992); State v. Bird, No. CA 6836, 1981 Ohio App. LEXIS 13173, at *7 (Ohio Ct. App. Jan. 8, 1981) (noting, in the course of holding that defendant’s unwillingness to be handcuffed amounted to resisting arrest, that “passive, indirect or circuitous impediments to the exercise of the official duties of a police officer” can be sufficient). That such factual circumstances have supported a conviction for resisting arrest under Ohio law necessarily establishes that [330]*330the officers had probable cause to charge Harris with resisting arrest.
Lastly, the officers had probable cause to charge Harris with disorderly conduct. Under Ohio law, a person engages in disorderly conduct if he “recklessly cause[s] inconvenience, annoyance, or alarm to another” by, among other things, “[e]ngaging in fighting, in threatening harm to persons or property,” or by “[ijnsulting, taunting, or challenging another, under circumstances in which that conduct is likely to provoke a violent response.” Ohio Rev. Code § 2917.11(A); see also Ohio Rev. Code § 2917.11(E)(3) (making it aggravated disorderly conduct to violate the section “in the presence of any law enforcement officer”). A factfinder could reasonably conclude that Harris, by threatening the officers with physical violence and by making threatening gestures, provided them with the necessary probable cause to charge him with disorderly conduct, see, e.g., Fields v. Johnson, No. 74375, 1998 WL 827610 at *3, 1998 Ohio App. LEXIS 5575, at *8 (Ohio Ct.App. Nov. 25, 1998) (coarse, threatening and grossly abusive language establishes probable cause to charge individual with disorderly conduct under § 2917.11), not the least because similar actions have in the past sustained convictions under § 2917.11, see, e.g., State v. Shumaker, No. 1332, 1994 WL 47676 at *5, 1994 Ohio App. LEXIS 571, at *14 (Ohio Ct.App. Feb. 18, 1994) (loud and abusive language may suffice to sustain a conviction under § 2917.11).
We recognize and can certainly understand our dissenting colleague’s view that the unfortunate facts of this case indicate that the agents wrongly suspected Harris of acting as a drug courier. Yet we cannot agree with our colleague’s statement that “no evidence of physical activity which prevents or delays arrest is present in the instant case.” Infra at 15. The district court found that the officers, as they attempted to place Harris under arrest, “could feel him resisting and trying to get loose.” JA 131. Under Ohio precedent, those actions are sufficient to support a finding of probable cause to charge Harris with resisting arrest. See Thomas, 1995 WL 39402, 1995 Ohio App. LEXIS 397; Rose, 600 N.E.2d at 384; Bird, 1981 Ohio App. LEXIS 13173, at *7.
B.
Harris next argues that the fouifederal agents committed assault and battery during the course of his detention at the airport. We again disagree because a reasonable factfinder could conclude otherwise.
Like other States, Ohio provides civil-law causes of action for assault and battery. The State defines assault as “the willful threat or attempt to harm or touch another offensively, [where that] threat or attempt reasonably places the other in fear of such contact.” Stokes v. Meimaris, 111 Ohio App.3d 176, 675 N.E.2d 1289, 1296 (1996); Vandiver v. Morgan Adhesive Co., 126 Ohio App.3d 634, 710 N.E.2d 1219, 1221 (1998). Proof of physical injury is not required. Id. The State defines battery as an intentional uninvited contact with another. See Snyder v. Turk, 90 Ohio App.3d 18, 627 N.E.2d 1053, 1057 (1993); Love v. City of Port Clinton, 37 Ohio St.3d 98, 524 N.E.2d 166, 167 (1988).
The totality of the circumstances does not support Harris’s argument that the agents committed assault or battery. The only physical contact was the shoulder-tapping, hand-holding and handcuffing, all of which agents may lawfully do when arresting an individual. Leichtman v. WLW Jacor Communications, Inc., 92 Ohio App.3d 232, 634 N.E.2d 697, 699 (1994). The district court also found that Harris had failed to demonstrate his own [331]*331reasonable apprehension of fear. Finally, the district court rejected Harris’s assertions that the interaction with the dog in the DEA task force office amounted to an assault or battery because the district court did not believe Harris’s testimony regarding these events.
Harris aptly notes that the district court incorrectly described his burden of proof as “beyond a reasonable doubt” at one point, but also admits that the court mentioned the correct proof burden as “a preponderance of the evidence” two times in the very same section. JA 135-36; Matlock, 665 N.E.2d at 775. On this record, the district court’s single reference to a “beyond a reasonable doubt” standard was harmless.
III.
Harris next argues that the district court erred in dismissing his Bivens claims as untimely. Under the landmark 1971 Supreme Court decision of that name, private citizens may “recover money damages for any injuries [they have] suffered as a result of [a federal agents’s] violation of the [Constitution].” 403 U.S. at 397, 91 S.Ct. 1999. In addressing the timeliness of a federal constitutional damages action, “the settled practice has been to adopt a local time limitation as federal law if it is not inconsistent with federal law or policy to do so.” Wilson v. Garcia, 471 U.S. 261, 266-67, 105 S.Ct. 1938, 85 L.Ed.2d 254 (1985). That practice applies both to § 1983 actions (because “Congress did not establish a statute of limitations or a body of tolling rules applicable to actions brought in federal court under § 1983,” Bd. of Regents of the Univ. of the State of New York v. Tomanio, 446 U.S. 478, 483, 100 S.Ct. 1790, 64 L.Ed.2d 440 (1980)) and to Bivens actions (because the federal Constitution does not list, any more than the § 1983 statute does, the timeliness rules that govern implied damages actions). See McSurely v. Hutchison, 823 F.2d 1002, 1005 (6th Cir.1987); see also Sanchez v. United States, 49 F.3d 1329, 1330 (8th Cir.1995); Industrial Constructors Corp. v. United States Bureau of Reclamation, 15 F.3d 963, 968-69 (10th Cir.1994); Van Strum v. Lawn, 940 F.2d 406, 410 (9th Cir.1991); Bieneman v. City of Chicago, 864 F.2d 463, 469-70 (7th Cir.1988); Chin v. Bowen, 833 F.2d 21, 23-24 (2d Cir.1987). Not only the length of the limitations period, but also “closely related questions of tolling and application,” thus are “governed by state law.” Garcia, 471 U.S. at 269, 105 S.Ct. 1938; Johnson v. Ry. Express Agency, Inc., 421 U.S. 454, 464, 95 S.Ct. 1716, 44 L.Ed.2d 295 (1975); cf. Hardin v. Straub, 490 U.S. 536, 544, 109 S.Ct. 1998, 104 L.Ed.2d 582 (1989).
Among the tolling provisions interrelated with the statute of limitations, we have held, is the Ohio Savings Statute. See Harris v. Canton, 725 F.2d 371, 377 (6th Cir.1984). The statute provides:
In any action that is commenced, ... if the plaintiff fails otherwise than upon the merits, the plaintiff ... may commence a new action within one year after ... the plaintiffs failure otherwise than upon the merits or within the period of the original applicable statute of limitations, whichever occurs latey.
Ohio Rev.Code § 2305.19. The Savings Statute applies where “an action is timely commenced and is then dismissed without prejudice after the applicable statute of limitations has run.” Lewis v. Connor, 21 Ohio St.3d 1, 487 N.E.2d 285, 287 (1985); Int’l Periodical Distribs. v. Bizmart, Inc., 95 Ohio St.3d 452, 768 N.E.2d 1167, 1168 (2002); Chadwick v. Barba Lou, Inc., 69 Ohio St.2d 222, 431 N.E.2d 660, 663 (1982). A “plaintiffs failure” under § 2305.19 generally occurs when an appellate court affirms a district court’s dismissal of a case [332]*332“otherwise than upon the merits.” See Clones v. Kohli, No. 02 CA 121, 2003 WL 21500203 at *1, 2003 Ohio App. LEXIS 3153, at *3-4 (Ohio Ct.App. June 26, 2003); Van-American Ins. Co. v. Schiappa, Nos. 97-JE-42/97-JE-46, 1999 WL 260904, at *4, 1999 Ohio App. LEXIS 1976, at *10 (Ohio Ct.App. Apr. 29, 1999) (savings provision commences to run only on date of appellate court’s dismissal order because “[t]o hold otherwise would be to force appellants to risk losing the benefit of the savings statute in order to appeal the trial court’s decision”); McDermott v. Lynch, No. 71131, 1997 WL 253168 at *3, 1997 Ohio App. LEXIS 2087, at *7 (Ohio Ct. App. May 15, 1997); Greene v. Hawkins, No. CA 93-27, 1993 WL 544350 at *2, 1993 Ohio App. LEXIS 6417, at *5 (Ohio Ct. App. Dec. 10, 1993); LaBarbera v. Batsch, 5 Ohio App.2d 151, 214 N.E.2d 443, 448-49 (1966), rev’d on other grounds, 10 Ohio St.2d 106, 227 N.E.2d 55 (1967); Colello v. Bates, 88 Ohio App. 313, 100 N.E.2d 258, 260-61 (1950); cf. Burnett v. New York Cent. R.R. Co., 380 U.S. 424, 435, 85 S.Ct. 1050, 13 L.Ed.2d 941 (1965) (“[T]he limitation provision is tolled until the state court order dismissing the state action becomes final by the running of the time during which an appeal may be taken or the entry of a final judgment on appeal.”).
In accordance with these precedents, Harris adequately preserved his claim. Harris filed his original claims within the two-year Ohio statute-of-limitations period. See Ohio Rev.Code § 2305.10. When the district court dismissed those claims without prejudice, Harris unsuccessfully appealed that decision. And Harris refiled his claims within one year of the Sixth Circuit’s affirmance of the district court’s dismissal of his claims without prejudice. That timeline and Ohio tolling principles establish that Harris’s Bivens claims fall within the Savings Statute’s compass.
In reaching a contrary conclusion, the district court held that the Savings Statute began to run once the district court dismissed Harris’s action without prejudice, not when the Sixth Circuit affirmed that decision. But this view rests upon a mistaken reading of Gruber v. Kopf Builders, Inc., 147 Ohio App.3d 305, 770 N.E.2d 598 (2001). In Gruber, the plaintiff voluntarily dismissed his first complaint on April 1, 1998, but the trial court continued adjudicating the case, ultimately granting summary judgment to each defendant. See id. at 600. The plaintiff then appealed this decision, and the Ohio Court of Appeals “dismissed the appeal ... because the trial court did not have subject-matter jurisdiction ... after the April 1, 1998 voluntary dismissal.” Id. When the plaintiff filed a new complaint more than one year after he voluntarily dismissed his first complaint, both the trial court and the Ohio Court of Appeals held that “a voluntary dismissal ... constitutes a failure otherwise than upon the merits within the purview of the statute,” id. at 601, and distinguished prior Ohio law stating that the limitation period is tolled during the pendency of an appeal because “none of the cases ... involved a plaintiffs voluntary dismissal,” id. at 602. By contrast, when an involuntary dismissal is at issue, Ohio case law indicates that “an action fails” on the date of the appellate decision, if any, from the dismissal. Cf. Stine v. Kansas City Terminal Ry. Co., 564 S.W.2d 619, 622 (Mo.Ct.App.1978).
The individual defendants seek to defend the district court’s decision on a second ground — that this reading of Ohio law conflicts with federal policy and accordingly should not be followed. In making this argument, the defendants point to Lewellen v. Morley, 875 F.2d 118, 121 (7th Cir.1989), where the court chose not to invoke an Illinois state law permitting the service-of-process time to “be extended indefinitely by unsuccessful attempts to serve defen[333]*333dants who live outside Illinois” because the rule was at odds with Federal Rule of Civil Procedure Rule 4(m)’s requirement that the plaintiff obtain service within 120 days absent “good cause.” Id. Attempting to follow suit, the individual defendants argue that here, too, the extension permitted by the Savings Statute allows a plaintiff to circumvent the service-of-process requirements of Rule 4. But, in this instance, no similar direct conflict between Ohio law and federal policy exists. Rule 4 governs service of process after a complaint has been filed and before dismissal, see Fed. R.Civ.P. 4(m), not what happens after a dismissal for lack of process. The appropriate touchstone for the latter inquiry is state law, not federal law, and accordingly this alternative ground for upholding the district court’s decision must be rejected.
IV.
Even though the district court incorrectly dismissed Harris’s Bivens claims, we do not reinstate them because they are barred by the court’s adjudication of his FTCA claims. The FTCA’s judgment bar provides that:
The judgment in an action under section 1346(b) of this title [the FTCA] shall constitute a complete bar to any action by the claimant, by reason of the same subject matter, against the employee of the government whose act or omission gave rise to the claim.
28 U.S.C. § 2676; see generally United States v. Gilman, 347 U.S. 507, 509, 74 S.Ct. 695, 98 L.Ed. 898 (1954) (“The Tort Claims Act does not touch the liability of the employees except in one respect: by 28 U.S.C. § 2676 it makes the judgment against the United States ‘a complete bar’ to any action by the claimant against the employee.”); Benbow v. Wolf, 217 F.2d 203, 205 n. 4 (9th Cir.1954) (“The Congress has the apparent intention that the individual be not pursued if the United States be liable.”). • No statutory bar, by contrast, prohibits suits against the government following a suit against or settlement with the employee; in that scenario, customary rules of preclusion and the terms of the settlement govern whether an additional lawsuit may be filed. See Branch v. United States, 979 F.2d 948, 951 (2d Cir.1992) (looking to “local law to determine the effect, as to a successor FTCA action against the United States, of a release provided by a government employee”); Friday v. United States, 239 F.2d 701, 703-04 (9th Cir.1957) (“[I]t is significant that the [FTCA] provides with particularity that a release of the United States releases the employee. The obvious inference from this specific provision is that a release of the tort feasor employee does not release the tort feasor employer, the United States, at least where rights against the United States are reserved.”); United States v. First Sec. Bank of Utah, 208 F.2d 424, 428 (10th Cir.1953) (“Congress with meticulous care provided that recovery of a judgment against the government shall constitute a bar to any action against the employee whose negligence gave rise to the claim, but for reasons satisfactory to itself, it failed to provide, directly or indirectly, that a satisfaction of a claim against an employee should bar an action against the government.”).
Courts, including the Sixth Circuit, have consistently read § 2676 to bar a Bivens claim against a government employee “arising out of the same actions, transactions, or occurrences” as an FTCA claim. Serra v. Pichardo, 786 F.2d 237, 239 (6th Cir.1986); see also Arevalo v. Woods, 811 F.2d 487, 490 (9th Cir.1987); Andrews v. Gee, 599 F.Supp. 251, 253 (D.S.C.1984); Armstrong v. Vogel, 424 F.Supp. 445, 447 (D.S.C.1977). Against this backdrop, Har[334]*334ris does not dispute that the district court entered a “judgment” on the merits of his FTCA claims and that those claims “aris[e] out of the same actions, transactions, or occurrences” as his Bivens claims. And he does not dispute that § 2676 applies to Bivens claims as well as FTCA claims even though the Congress that enacted this statute in 1946 had no way of knowing that it would cover the Bivens implied-constitutional claim announced in 1971.
Harris argues instead that the judgment bar does not apply “where plaintiff has from the outset alleged his Bivens claims and sought a jury trial in the same lawsuit alleging FTCA causes of action, where the Bivens claims have been erroneously dismissed, and where the later bench trial of plaintiffs common law tort causes of action results in a judgment in defendant’s favor.” Harris Reply Br. at 12. We disagree.
In accordance with the consistent application of the judgment bar over the fifty years since its enactment, we have held that the provision applies even when “the claims were tried together in the same suit and [ ] the judgments were entered simultaneously.” Serra, 786 F.2d at 241; see also Clifton v. Miller, No. 97-2342, 1998 WL 78992, at *2, 1998 U.S.App. LEXIS 2794, at *5-6 (7th Cir.1998); Rodriguez v. Handy, 873 F.2d 814, 816 (5th Cir.1989); Arevalo, 811 F.2d at 490 (“The moment judgment was entered against the government, then by virtue of section 2676, Woods was no longer answerable to Areva-lo for damages.”); Aetna Cas. & Sur. Co. v. United States, 570 F.2d 1197, 1201 (4th Cir.1978) (“[A] judgment against the United States would automatically bar the entry of any contemporaneous or subsequent judgment against [the government employees].”); Gilman v. United States, 206 F.2d 846, 848 (9th Cir.1953) (“[T]he moment judgment was entered against the Government, then by virtue of Sec. 2676 ... the employee was no longer primarily answerable to the claimant, — he was not answerable at all.”) (footnote omitted); United States v. Lushbough, 200 F.2d 717, 721 (8th Cir.1952) (“The District Court, having awarded a judgment in favor of [plaintiff] in his action against the United States, could not in the face of the explicit provisions of [§ 2676] order judgment against [the government employee] in favor of [the plaintiff] in the same action.”); Popal v. United States, 1996 WL 185731 at *3, 1996 U.S. Dist. LEXIS 4967, at *7 (S.D.N.Y. Apr. 18, 1996) (“Having elected to proceed against both the Government and the individuals and having received judgment against the Government on the issue of liability, the plaintiffs cannot now choose to continue their action against the individual defendants and elect at some later time between the FTCA and Bivens claims.”); Satterwhite v. Bocelato, 130 F.Supp. 825, 829 (E.D.N.C.1955) (entering judgment “only against the United States” and not against the federal employee because of § 2676); Hopper v. United States, 122 F.Supp. 181, 190 (E.D.Tenn.1953) (holding that, because a judgment was “awarded the plaintiffs against the United States, no judgment will be announced against [the government employee]”); cf. Prechtl v. United States, 84 F.Supp. 889, 890 (W.D.N.Y.1949).
By making “[t]he judgment” in an FTCA claim the bar to a future Bivens claim, the statute also fails to draw a distinction between a decision for or against the government. Nothing in the common interpretation of the word “judgment” suggests that a judgment in favor of the United States may be treated differently from a judgment against the United States. See Black’s Law Dictionary (7th ed.1999) (defining “judgment” as “[a] court’s final determination of the rights and obligations of the parties in a case”). And indeed [335]*335courts have generally held — in accordance with the plain terms of § 2676 — that the FTCA operates as a bar even when a judgment is adverse to the claimant. See Estate of Trentadue ex rel. Aguilar v. United States, 397 F.3d 840, 858 (10th Cir.2005) (“[T]he judgment bar in § 2676 precludes plaintiffs from bringing a Bivens claim regarding the same subject matter regardless of whether the final FTCA judgment is rendered in favor of a plaintiff or the government.”); Farmer v. Perrill, 275 F.3d 958, 963 (10th Cir.2001) (“It is wholly irrelevant that [prior Tenth Circuit case law] addressed an FTCA judgment against the United States while Farmer’s case involves an FTCA judgment in favor of the United States. Section 2676 makes no distinction between favorable and unfavorable judgments — it simply refers to ‘the judgment in an action under section 1346(b).’ ”); Hoosier Bancorp of Ind., Inc. v. Rasmussen, 90 F.3d 180, 184-85 (7th Cir.1996) (“There is no indication that Congress intended Section 2676 to apply only to favorable FTCA judgments.”); Freeze v. United States, 343 F.Supp.2d 477, 481 (M.D.N.C.2004) (“[T]he court’s judgment on Plaintiffs FTCA claim prevents Plaintiff from asserting an action against [the] defendant [ ] based on a constitutional claim.”); cf. Hallock v. Bonner, 387 F.3d 147, 155 (2d Cir.2004) (distinguishing between FTCA actions dismissed on subjective-matter-jurisdiction grounds, as opposed to merits grounds, but implicitly accepting that § 2676 reaches both judgments adverse to and in favor of the government), cert. granted, — U.S. -, 125 S.Ct. 2547, 162 L.Ed.2d 274 (2005); Leaman v. Ohio Dept. of Mental Retardation & Developmental Disabilities, 825 F.2d 946, 955 (6th Cir.1987) (en banc) (noting, while interpreting an analogous Ohio provision, that, under § 2676, “one who pursues his statutory remedies against the United States to the point of judgment— even an adverse judgment or a judgment for only a small part of the amount claimed — bars himself from any recovery against federal employees”).
One court of appeals decision, it is true, has distinguished between favorable and unfavorable merits judgments in applying § 2676, concluding that an FTCA judgment in favor of the government did not bar a contemporaneous Bivens action arising from the same events. See Kreines v. United States, 959 F.2d 834, 838 (9th Cir.1992); compare Gasho v. United States, 39 F.3d 1420, 1437 (9th Cir.1994) (noting that “the holding in Kreines was narrowly confined to its facts” and rejecting plaintiffs’ argument that “Congress intended to permit a claimant to have a second chance after losing his FTCA action”). In doing so, Kreines reasoned that “[t]he text is ambiguous on the question of whether an FTCA judgment favorable to the government bars a contemporaneous Bivens judgment.” 959 F.2d at 838. But the word “judgment” does not have the ambiguity that the Ninth Circuit ascribed to it. The statute refers to “[t]he judgment in an action under section 1346(b) of this title”; it says nothing about who prevailed in “the judgment” — a statutory fact that engenders certainty rather than uncertainty that the bar applies to all “judgments,” whether for or against the government.
Nor is there any background principle regarding judgment bars that supports applying the bar only when one party, as opposed to another, prevailed in the first judgment. Calling the word “judgment” ambiguous, at any rate, hardly leads to the conclusion that the Ninth Circuit embraced. Why is it that a “judgment” bar should apply only where the government loses? When the government loses the initial case, one could argue that a judgment bar is not necessary to prohibit dual recoveries because the law of remedies [336]*336already would sharply curtail the risk of a double recovery and already would sap the plaintiffs incentive to pursue another claim. In other words, it makes just as much sense to apply the bar when the government wins the initial case. Better under these circumstances, we submit, to follow the language of § 2676 where it leads • us, which is to the conclusion that the identity of the victor in an. FTCA claim does not delimit the reach of a bar that applies to all “judgment[s].”
Nor does the Supreme Court’s decision in Carlson v. Green, 446 U.S. 14, 100 S.Ct. 1468, 64 L.Ed.2d 15 (1980), support a contrary interpretation. In Carlson, the Court refused to condition the availability of .a Bivens claim on the federal government’s waiver of sovereign immunity through the FTCA. See id. at 20, 100 S.Ct. 1468 (holding that plaintiffs, “[i]n the absence of a contrary expression from Congress, ... shall have an action under FTCA against the United States as well as a Bivens action against the individual officials alleged to have infringed their constitutional rights”). Nothing about our decision, however, conflicts with this principle. Although “[cjonstitutional rights may not be extinguished by any statute, state or federal, [ ] this truism does not mean that suits or potential suits for alleged violations of such rights may not be compromised or waived.” Leaman, 825 F.2d at 956 (en banc). Nothing in Carlson prohibits the federal government from conditioning its waiver of sovereign immunity on a plaintiffs willingness not to pursue a constitutional remedy — which is precisely what the terms of § 2676 do and which explains why we rejected a similar argument in Serra. See Serra, 786 F.2d at 241 (reasoning that “the instant case, unlike Carlson, deals with the effect of a FTCA judgment on a plaintiffs power to continue to pursue a Bivens remedy,” not with the availability of a Bivens remedy “even though plaintiffs allegations also could have supported a suit against the United States under the FTCA”); see also Arevalo, 811 F.2d at 490 (“Unlike Carlson, in the case now before us the plaintiff Arevalo sued both the United States under the FTCA and the individual federal officer under Bivens.”)-, Sanchez v. Rowe, 651 F.Supp. 571, 575-76 & n. 2 (N.D.Tex.1986) (“[TJhe holding in Carlson does not alter the plain meaning of § 2676 because Carlson did not deal with the effect of a FTCA judgment on a plaintiffs power to continue to pursue a Bivens remedy.”).
Neither does the statutory history of § 2676 advance Harris’s position. The predecessor statute read as follows:
(a) Subject to the provisions of this title, the United States district court for the district wherein the plaintiff is resident or wherein the act or omission complained of occurred, including the United States district courts for the Territories and possessions of the United States, sitting without a jury, shall have exclusive jurisdiction to hear, determine, and render judgment on any claim against the United States, for money only, accruing on and after January 1, 1945, on account of damage to or loss of property or on account of personal injury or death caused by the negligent or wrongful act or omission of any employee of the Government while acting within the scope of his office or employment, under circumstances where the United States, if a private person, would be liable to the claimant for such damage, loss, injury, or death in accordance with the law of the place where the act or omission occurred. Subject to the provisions of this title, the United States shall be liable in respect of such claims to the same claimants, in the same manner, and to the same extent as a private individual under like circumstances, except that the [337]*337United States shall not be liable for interest prior to judgment, or for punitive damages. Costs shall be allowed in all courts to the successful claimant to the same extent as if the United States were a private litigant, except that such costs shall not include attorneys’ fees,
(b) The judgment in such an action shall constitute a complete bar to any action by the claimant, by reason of the same subject matter, against the employee of the Government whose act or omission gave rise to the claim.
Legislative Reorganization Act of 1946, Pub.L. No. 79-601, 60 Stat. 812, 843-44 § 410 (emphasis added). Relying on this earlier version of the statute, Harris argues that the judgment previously referred to was the liability of the United States and that, by retaining the phrase “the judgment” instead of using “any judgment” or “a judgment,” Congress did not change this meaning when it altered the statute. But the text of the predecessor statute is at best unclear. In using the phrase “the judgment in such an action,” Congress could just have easily meant all judgments as it could have meant those judgments in which “the United States shall be liable.”
Nor, lastly, do we believe that this holding works an injustice on Harris or other plaintiffs like him. The equitable punch of Harris’s argument comes from the fear that the district court erroneously dismissed his Bivens claims and that he has therefore never had his day in court on those claims. But the fact of the matter is that, even if the district court had not mistakenly barred Harris’s Bivens claims on timeliness grounds, the outcome would have been no different: Fifty years of largely consistent case law would have precluded the district court from entering a judgment favorable to Harris on his Bivens claims upon entry of the merits judgment on his FTCA claim “arising out of the same actions, transactions, or occurrences.” Of course, a plaintiff still may bring a Bivens claim and pursue it to judgment. See Carlson, 446 U.S. at 20, 100 S.Ct. 1468. But having chosen to bring an FTCA claim as well, the statute provides that a “judgment in [that] action ... constitute^] a complete bar to any action by the claimant ... against the employee of the government whose act or omission gave rise to the claim.” 28 U.S.C. § 2676. See Engle v. Mecke, 24 F.3d 133, 135 (10th Cir.1994) (“A decision to sue the government ... affects the availability of a Bivens action against the federal officer. Although the plaintiff may elect initially to bring his action against either defendant, a judgment against the United States under the FTCA constitutes ‘a complete bar to any action by the claimant.’”); Ting v. United States, 927 F.2d 1504, 1513 n. 10 (9th Cir.1991) (“The FTCA ... imposes an election of remedies. While a plaintiff may maintain both a FTCA and a Bivens action, ... recovery against the United States bars recovery against the employee.”); cf. Leaman, 825 F.2d at 956 (en banc) (noting that “the Federal Tort Claims Act ... offers claimants a better deal than they would have without it” because it does not “force[ ] any claimant to accept the government’s statutory offer” but “afford[s] claimants a superior mechanism for vindicating their rights”).
V.
For these reasons, we affirm.