Walking vs running for longevity: which one actually buys you more life?

If you could only defend one of them at a dinner party, you’d probably pick running. It sounds more serious. It has kit, and races, and people who talk about their splits unprompted. Walking sounds like something you do on the way to the thing you actually meant to do.

The evidence disagrees, and it disagrees in a specific, slightly annoying way. Because when researchers finally put walkers and runners side by side and matched them on the one thing that actually matters — energy expended, not time spent — the two came out roughly level. Sometimes walking came out ahead.

So here’s the thesis, up front: per minute, running wins by about two to one; per unit of effort, walking matches it or beats it — which means this was never a physiology question. It’s a diary question. How many hours a week do you genuinely have? Answer that honestly and the modality picks itself.

What each one actually does

Both are aerobic work. The difference is intensity per minute, and it’s more or less the whole story.

Physiologists measure this in METs — metabolic equivalents, where 1 MET is you sitting still, mildly bored. Walking at 3 mph on level ground is about 3.5 METs. Brisk walking at 4 mph, the pace where you’re slightly out of breath but could still hold a conversation with visible effort, is about 5 METs. Running at 6 mph — a ten-minute mile, which is a perfectly ordinary jog and not a heroic act — is about 9.8 METs.

That’s the entire mechanical difference. Running is roughly twice the metabolic load per minute of brisk walking, and roughly three times a stroll.

What it isn’t is a different kind of adaptation. Both raise stroke volume, both improve endothelial function and insulin sensitivity, both push mitochondrial density in the working muscle. For most untrained adults a brisk walk lands squarely inside Zone 2 — the same low-intensity aerobic territory that runners spend most of their week in and then talk about as if it were a technical discipline. The runner is just getting there faster and staying there under more load.

There’s one real asymmetry, and it goes against running: impact. A systematic review of long-distance runners (van Gent et al., British Journal of Sports Medicine, 2007) found lower-extremity injury incidence ranging from 19.4% to 79.3% depending on the population and how injury was defined, with the knee the most common site. Walking’s equivalent number is, in practical terms, not a number. That range is wide enough to be almost useless as a prediction, but the direction is not in dispute.

The head-to-head: energy-matched, they tie

This is the study the whole argument turns on, and remarkably few people quote it.

Williams and Thompson (Arteriosclerosis, Thrombosis, and Vascular Biology, 2013) took two enormous parallel cohorts — the National Runners’ Health Study (33,060 runners) and the National Walkers’ Health Study (15,945 walkers) — and did the thing everyone should have done first. Instead of comparing an hour of one against an hour of the other, they compared equivalent energy expenditure, expressed as MET-hours per day, and followed incident hypertension, high cholesterol, diabetes and coronary heart disease over 6.2 years.

Per MET-hour per day, running reduced risk of incident hypertension by 4.2%, high cholesterol by 4.3%, diabetes by 12.1% and coronary heart disease by 4.5%.

The walkers’ numbers, for the same energy: 7.2%, 7.0%, 12.3% and 9.3%.

Read that again, because it’s the opposite of the folk wisdom. Energy-matched, the walkers did at least as well on every endpoint, and rather better on three of them. The authors were appropriately cautious — these are self-reported, physician-diagnosed outcomes in self-selected cohorts, and the walkers were on average older, which changes the baseline risk arithmetic. But the headline conclusion stands: the same energy spent walking and running produced similar reductions in risk. Intensity, in itself, wasn’t buying anything extra.

Now the caveat that matters, and I want to be precise about it rather than let the paragraph above do more work than it can. Those four endpoints are diseases, not deaths. Hypertension, high cholesterol, diabetes and coronary heart disease are excellent proxies for how long you’ll live, and they’re the machinery through which most of the mortality difference would have to run — but the study didn’t count funerals, so on its own it can’t establish that walking and running are equivalent for longevity.

There is one energy-matched comparison that did count deaths, and it’s narrower than you’d want. Williams (Hypertension, 2013) followed 6,973 walkers and 3,907 runners from the same two cohorts — all of them on blood-pressure medication at baseline — for an average of 10.2 years, recording 1,121 deaths. At 1.8–3.6 MET-hours a day, mortality was 29% lower overall and 34% lower from cardiovascular disease. And on the question this post is about: “results did not differ between running and walking”, with every comparison landing at P > 0.23. The authors also noted, unprompted, the thing this whole post is built around — matching a runner’s energy takes about 50% more distance and roughly twice the time on foot.

So the honest statement of the evidence is narrower than “walking equals running for longevity.” It’s this: energy-matched, the two produce equivalent reductions in the major cardiometabolic diseases in the general cohort, and equivalent reductions in actual mortality in the one subgroup where anyone checked — hypertensives on medication. Nobody has run the energy-matched mortality comparison in a general population, and until someone does, “they’re equivalent for lifespan” is a well-supported inference rather than a demonstrated fact.

What you can say flatly is the negative version: there’s no evidence that running unlocks a superior adaptation walking can’t reach. It just gets you through the energy faster.

The bit where running wins, and it’s only one axis

Running’s advantage is time, and time is not nothing. It’s usually the binding constraint.

Here’s the conversion, done properly, because “walking is as good as running” gets repeated without anyone showing what it costs.

The 2008 US physical activity guideline minimum — 150 minutes a week of moderate activity — works out to about 7.5 MET-hours per week. The pooled analysis by Arem and colleagues (JAMA Internal Medicine, 2015), covering 661,137 adults across six cohorts, found that hitting that minimum was associated with roughly 31% lower mortality versus doing nothing, that the benefit curve kept improving to a plateau at about 3–5× the minimum (around 39% lower mortality), and that there was no excess risk even at 10× it.

Convert that into minutes on your calendar:

That table is the honest version of the comparison. Not “which is better” — what each one costs you in hours to reach the same place. A 30-minute run is worth about an hour of brisk walking and about an hour and a half of strolling. If your realistic weekly exercise budget is 90 minutes, running gets you to the plateau’s doorstep and walking doesn’t.

And if you’re already walking, the marginal maths is worth knowing too: the step-count mortality curve flattens somewhere between 8,000 and 10,000 steps a day, so if you’re already at 8,000, walking more is close to the worst-value 40 minutes available to you. Two easy 20-minute runs in the same 40 minutes add roughly 6–7 MET-hours on top of a base you already have. That’s the actual case for running: not that it’s better, but that it’s the efficient way to buy the volume you can’t otherwise fit.

Where this arithmetic gets suspicious

Now the honest caveat, and it’s a big one — because two of the best running studies flatly contradict the neat MET-hours table above.

Lee and colleagues (Journal of the American College of Cardiology, 2014) followed 55,137 adults in the Aerobics Center Longitudinal Study and found runners had 30% lower all-cause mortality and 45% lower cardiovascular mortality than non-runners, worth about three years of life expectancy. Fine, expected. The unexpected part: the benefit was essentially flat across dose. People running under 51 minutes a week, under 6 miles a week, slower than 6 mph, or only once or twice a week got broadly the same reduction as the people doing far more.

Pedisic and colleagues (British Journal of Sports Medicine, 2020) pooled 14 studies covering 232,149 participants and 25,951 deaths and found the same shape: any running was associated with 27% lower all-cause mortality, 30% lower cardiovascular and 23% lower cancer mortality — with no convincing evidence that more running bought more.

So which is it? Does mortality risk fall smoothly with accumulated MET-hours, or does it fall off a cliff the moment you do anything at all and then flatten?

The tempting resolution is that the running studies are just catching the steep bottom of the curve — sedentary becomes does-something, everything after that is flat. Tempting, and wrong, and worth spelling out why, because it’s the explanation you’ll see offered everywhere. A non-runner is not a sedentary person. Lee’s comparison group could walk, cycle, lift and garden to their hearts’ content, and the analysis adjusted for other physical activity; the same is true across the studies Pedisic pooled. What these designs contrast is running versus not running, not active versus inactive. So they aren’t measuring the bottom of the dose-response curve at all, and the neat story doesn’t hold.

Which leaves the flat dose-response genuinely unresolved, and I’d rather say that than invent a mechanism. The candidate explanations are unglamorous: participation is a much cleaner variable than dose, so a real gradient can be blurred into flatness by self-reported mileage and pace; the high-dose strata are small, so there’s less power to detect a difference exactly where you’d want one; and people who run at all differ from people who don’t in ways no covariate list fully captures. Any of those would produce the observed pattern without the underlying biology being flat.

The practical consequence is the same either way, and it’s the only thing you need from this section: the MET-hours table is a budgeting heuristic, not a law of nature, and it will overstate what your fifth hour a week buys relative to your first. What Lee’s 51-minutes-a-week finding demonstrates is narrower than “51 minutes is all you need” — it’s that the low-dose runners in that cohort were not measurably worse off than the high-dose ones. That’s a genuinely encouraging thing to know when you’re deciding whether a small amount is worth bothering with. It is not a licence to treat volume as irrelevant.

And underneath all of this sits the strongest argument against the whole framing of the comparison. The gap between a walker and a runner is small. The gap between either of them and someone who does neither — which is most people, most of the time, and something I’ve written about in the sitting-versus-smoking post — is enormous.

Pace is the free variable nobody uses

One more thing before the verdict, because it’s the cheapest upgrade available to a walker and it costs no extra minutes.

Stamatakis and colleagues (British Journal of Sports Medicine, 2018) pooled 50,225 walkers across 11 British population cohorts. Compared with slow-pace walkers, average-pace walkers had 20% lower all-cause mortality and brisk-or-fast walkers 24% lower — after adjusting for total volume and intensity of all physical activity. In over-60s the cardiovascular effect was larger still: 46% and 53% lower risk respectively.

Same route, same duration, different intent. If you’re a walker who thinks the running crowd has something you don’t, this is where to look first — and it’s the same argument I made at length in the post on walking pace as a longevity marker.

So who should pick which

A framework, rather than an answer, because the answer depends on you:

Pick running if: your honest weekly budget is under about two and a half hours; you’re already walking 8,000+ steps a day and have hit the flat part of that curve; your joints and current body weight tolerate impact; you’ll actually go. Start with two 20-minute run-walks a week and stay there for a month before adding anything — the injury range in that van Gent review is dominated by people who added volume too fast.

Pick walking if: you have more time than joint tolerance; you’re returning from injury, deconditioned, or over 60; you’d like the activity to be transport rather than an appointment. Then spend your effort on pace, not distance. Brisk, not ambling.

Pick both if you’re being sensible. This is a false binary in practice — most of the runners in those cohorts also walk everywhere, and run-walk intervals are the single most underrated format for anyone rebuilding a base. Twenty minutes of alternating two minutes easy jog and three minutes brisk walk is a legitimate session and nobody needs to know you did it.

And don’t pick either as an argument for skipping strength work. The four pillars don’t collapse into cardio, however good the cardio is.

The verdict

On the evidence we actually have, running is not better than walking — it’s faster than walking, which people persistently mistake for the same claim. Energy-matched, the walkers in the largest head-to-head we have did at least as well on every measured endpoint: hypertension, cholesterol, diabetes and coronary heart disease in the full cohort, and death itself in the one subgroup where mortality was checked. The general-population, energy-matched mortality trial doesn’t exist, so hold that last step as a strong inference rather than a settled fact — but nothing in the literature points the other way.

What running genuinely offers is compression: roughly twice the metabolic return per minute, which matters enormously if minutes are what you’re short of, and matters not at all if you have a lunch hour and a river path. What walking offers is a near-zero injury tax and the ability to hide inside your existing day.

The choice between them is a scheduling decision dressed up as a physiology one. Pick the one you’ll still be doing in fifty weeks, do it briskly, and stop reading comparison posts about it. Including this one.