Proximity Is Not Proof: What Dilip Agarwal Tells Us About Vehicle Liability

Introduction

Every so often, the Supreme Court is asked to resolve a question that sounds almost philosophical dressed up as statutory interpretation: when does an act become an "accident" arising from the use of a motor vehicle, as opposed to merely an event that happens to occur near or inside one? In Dilip Agarwal v. Rajshri Agarwal & Ors. (2026 INSC 736), a Bench of Justices Sanjay Karol and Augustine George Masih confronted precisely this puzzle. The case, on its face, looked like a routine compensation dispute under the Motor Vehicles Act, 1988. Scratch the surface, however, and it reveals a killing dressed as a car ride, a criminal acquittal that outlived a civil award, and a Court willing to ask whether liability can survive when the causal thread connecting death to vehicle simply is not there.

The Brief Facts

Anand was last seen getting into a vehicle driven by his acquaintance, Dilip, in November 2009. Three days later, his body was recovered near a village, and his wife Rajshri lodged an FIR alleging kidnapping, murder, and conspiracy. Dilip was tried and convicted by the Sessions Court, but the High Court subsequently acquitted him in appeal, principally because the prosecution's "last seen together" theory — always a fragile evidentiary reed — could not be substantiated.

Meanwhile, a parallel claim petition was pursued before the Motor Accident Claims Tribunal, Raigarh, on the theory that Anand's death occurred inside Dilip's vehicle and therefore fell within the compensatory scheme of the Motor Vehicles Act. The Tribunal, and later the High Court in appeal, accepted this reasoning and awarded compensation, reasoning that the injuries were inflicted while the two men were together in the car and that the dominant intent was not murder as such. Dilip challenged this before the Supreme Court.

The Legal Question

The narrow but consequential question before the Court was whether Anand's death could be said to have "arisen out of the use of a motor vehicle" under Sections 165 and 166 of the Motor Vehicles Act, thereby fastening liability on the driver and, by extension, his insurer.

This phrase — "arising out of" — is deceptively simple. It has generated decades of jurisprudence precisely because it sits between two poles: a strict "caused by" standard demanding direct causation, and an expansive "connected with" standard that could sweep in almost any event with a tangential link to a vehicle. Getting the calibration right matters enormously, because insurers and claimants alike organise their expectations around where that line falls.

The Court's Reasoning

Justice Karol's judgment leans heavily on the recent decision in Bruhat Bangalore Mahanagara Palike v. K.K. Umesh Kumar (2026 SCC OnLine SC 1111), itself tracing back to Shivaji Dayanu Patil v. Vatschala Uttam More, to explain that "use" of a vehicle is not confined to motion — a stationary or even immobilised vehicle can still be "in use." But the more significant borrowing is the distinction, drawn from Australian jurisprudence in Government Insurance Office of N.S.W. v. R.J. Green, between "caused by," which demands a direct and proximate relationship, and "arising out of," which tolerates a looser, less immediate — but still real — causal consequence. The takeaway is that "arising out of" widens the net, but it does not abolish the net altogether. Some causal relationship must still exist; what changes is only its degree of proximity.

The Court then turns to the trickier question of homicide within a compensation framework, invoking the well-known formulation in Rita Devi v. New India Assurance Co. Ltd. (2000) 5 SCC 113, which distinguishes between a "murder simpliciter" — where killing is the primary felonious intent — and an "accidental murder," where death results incidentally from some other felonious act (there, theft of the vehicle). Only the latter, the Court reiterates, can plausibly be said to "arise out of" vehicle use, because the vehicle itself is instrumentally tied to the underlying crime.

Applying this framework, the Bench finds that Anand's case fits neither template comfortably, and more fundamentally, that the causal nexus was never established in the first place. There was no forensic evidence — no blood, no biological trace, no sign of collision — connecting the vehicle to the fatal injuries, which appear to have been inflicted by blunt force. The Tribunal and High Court, the Court observes, appeared to simply assume that the injuries occurred inside the car, without any evidentiary anchor for that inference. Citing National Insurance Co. Ltd. v. Swaran Singh (2004) 3 SCC 297 on the nature of motor accidents, and reaffirming through ICICI Lombard General Insurance Co. Ltd. v. Rajani Sahoo (2025) 2 SCC 599 and N.G. Dastane v. S. Dastane (1975) 2 SCC 326 that civil claims operate on preponderance of probabilities rather than the criminal standard of proof beyond reasonable doubt, the Court is careful to underscore that the acquittal in the criminal trial did not by itself doom the civil claim. Even judged by the more forgiving civil standard, however, the claimants simply had not shown any link, however loose, between the vehicle and the death.

A Critical Appraisal

The judgment is, in this author's assessment, a necessary corrective. Motor accident compensation law is deliberately claimant-friendly, and rightly so — it exists to cushion families against the arbitrary tragedy of road mishaps without forcing them through the rigours of full-blown tort litigation. But that generosity has a natural limit, and this ruling reasserts that a vehicle's mere physical proximity to a death cannot substitute for actual causal connection. Had the Tribunal's approach been left undisturbed, virtually any death that occurred to have a car somewhere in its factual vicinity might have been swept into the compensatory net meant for genuine accidents — an outcome that would strain the insurance framework and dilute the very category of "accident" the statute is meant to address.

At the same time, the judgment leaves some discomfort in its wake. The Court accepts that the "accidental murder" doctrine from Rita Devi could, in principle, extend liability to homicides connected to vehicles, but it does not fully map out how future tribunals should distinguish borderline cases where the vehicle plays a more ambiguous instrumental role — say, where it is used to transport, restrain, or trap a victim before the fatal act. There is a risk that overly literal readings of this judgment could make it harder for future claimants to succeed even in cases where a vehicle is genuinely implicated in a chain of criminal causation, simply because forensic evidence from the vehicle itself is unavailable. Given how heavily the Court leans on the absence of forensic traces, it is worth asking whether that evidentiary bar sets too demanding a standard for claimants who, by definition, are litigating years after the fact and rarely control the criminal investigation's thoroughness.

For insurers, the ruling is unambiguously favourable, offering a clearer basis to resist claims where causal proof is thin. For claimants' counsel, it is a reminder that pleading and evidence in MACT proceedings cannot simply piggyback on a criminal chargesheet's narrative — independent proof of the vehicle's causal role remains indispensable, especially once a criminal acquittal has already undermined the underlying factual story.

Conclusion

Dilip Agarwal is unlikely to make front-page news, but it performs the quiet, important work of disciplining a phrase — "arising out of" — that could otherwise be stretched to cover almost anything. By insisting on a demonstrable, even if attenuated, causal link between vehicle use and death, the Court has preserved the coherence of motor accident jurisprudence while leaving the door open, through the accidental-murder doctrine, for genuinely vehicle-connected homicides to be compensated. Whether future benches will draw that line with equal rigour, or whether claimants in genuinely ambiguous cases will find themselves squeezed by an evidentiary standard calibrated for cleaner scenarios, remains to be seen. For now, the judgment stands as a useful, if imperfect, guardrail against the quiet expansion of statutory liability beyond its intended reach.

References

1. Dilip Agarwal v. Rajshri Agarwal & Ors., 2026 INSC 736 (SC).
2. Bruhat Bangalore Mahanagara Palike v. K.K. Umesh Kumar, 2026 SCC OnLine SC 1111.
3. Shivaji Dayanu Patil v. Vatschala Uttam More, (as cited in Bruhat Bangalore Mahanagara Palike).
4. Government Insurance Office of N.S.W. v. R.J. Green, (1965) 114 CLR 437 (High Court of Australia).
5. National Insurance Co. Ltd. v. Swaran Singh, (2004) 3 SCC 297.
6. Rita Devi v. New India Assurance Co. Ltd., (2000) 5 SCC 113.
7. ICICI Lombard General Insurance Co. Ltd. v. Rajani Sahoo, (2025) 2 SCC 599.
8. Mathew Alexander v. Mohd. Shafi, (2023) 13 SCC 510.
9. N.G. Dastane (Dr) v. S. Dastane, (1975) 2 SCC 326.
10. N.K.V. Bros. (P) Ltd. v. M. Karumai Ammal, (1980) 3 SCC 457.
11. Bimla Devi v. Himachal RTC, (2009) 13 SCC 530.
12. Dulcina Fernandes v. Joaquim Xavier Cruz, (2013) 10 SCC 646.
13. Motor Vehicles Act, 1988, ss. 165, 166.
14. Indian Penal Code, 1860, ss. 120B, 201, 302, 365.