Car Pulls to One Side in 2026? 7 Causes + the Free 2-Minute Road Test

Your car drifts left or right when you ease off the wheel, or it pulls toward one side under braking. Either way something's off — but here's the fact to start with, because it saves a lot of people a needless trip: roads are built with a crown, sloping slightly toward the edges so water drains, which makes almost every car drift a little to the right on a normal road. That's not a fault. So before anything else, the free two-minute parking-lot test tells real pulling from road slope, and only a real, consistent drift on level ground is worth spending a dollar on.

Once you've confirmed a real pull, one more question decides most of the diagnosis: does it pull when braking or when cruising? A pull only under braking is the brake system — a stuck caliper — which no alignment will ever fix, while a pull at all speeds is alignment, and a pull with clunking over bumps is a worn tie rod or ball joint that has to be replaced before an alignment will hold. That single fork is the difference between authorizing the right $150 fix and paying $120 for an alignment that comes back undone in three days. A worn ball joint is the one suspension cause here that is also a safety item — ball joint replacement cost covers the check and the price.

I built PulsCar after a $380 "full diagnostic" in Phuket in 2025 led to roughly $6,000 over six months across several shops, and the sound never went away (the story is on the About page). Understanding what's actually causing a pull before you authorize anything is how you avoid paying $150 for an alignment when what you needed was a free tire-pressure check or a $25 rotation. This guide gives you the road-crown test, the tire-pressure check, all seven causes ranked from cheapest to most dangerous with verified 2026 costs, and the driveway tests that separate a caliper from an alignment from a bent control arm.

Free tests in order, starting with the road crown

Do the free tests in order before you spend anything: the parking-lot road-crown test first (is it even a real pull?), then tire pressure (the most common real cause, and free), then the one question of whether it pulls braking or cruising. The seven causes below run from cheapest and safest to most dangerous — pressure and alignment first, worn and bent components last — each with a meta card and a driveway self-check. One rule overrides everything here: an alignment can't fix a brake pull, a worn tie rod, a bad tire, or bent geometry — so confirm which kind of pull you have before anyone aligns the car, because an alignment done on a stuck caliper or a worn front end is money spent on a fix that physically cannot hold.

First: is it a real pull, or just road crown?

This one free test rules out the most common false alarm and takes two minutes.

Find level ground — an empty parking lot or a flat, straight road — and drive at 25-35 mph, then gently release the wheel for three to five seconds (hands nearby for safety).

Stays straight? → no mechanical pull. What you feel on regular roads is road crown, and it's normal — nothing to fix.

Drifts consistently left or right? → a real pull. Continue to the tire-pressure check and the causes below.

Only pulls under braking? → a brake system issue (Cause 4), not alignment. The road-crown test will look fine; the pull only shows when you brake.

That single check — does it drift on level ground with the wheel released — is the most valuable thing you can determine, and it's free. Everything below only applies to a confirmed real pull.

The free checks: pressure, heat, and play

The tire-pressure check. Gauge all four and compare left to right against the door-jamb spec — a 5-plus PSI gap explains a pull for free, and about a fifth of pulling complaints end right here.

The braking question. Does it pull only when braking? That's the brake system (a caliper), and an alignment won't touch it.

The wheel-heat test. After a 10-mile drive, carefully feel each wheel from a safe distance — one dramatically hotter wheel is a dragging caliper.

The steering-play test. Turn the wheel a couple of inches quickly left and right — slop before the tires respond means tie-rod play; jack it up and rock the tire at 9-and-3 (tie rod) and 12-and-6 (ball joint) to confirm.

Why does my car pull to one side? Common causes are low tire pressure on one side, wheel misalignment, a stuck brake caliper, worn tie rods or ball joints, or uneven tire wear. But before any of those, rule out road crown: roads slope toward the edges so water drains, which makes almost every car drift a little right on a normal road — that's not a fault. Do the free parking-lot test first: on level ground, briefly release the wheel; if the car stays straight it's just crown, and if it consistently drifts you have a real pull. Then start with the cheapest real cause, tire pressure — gauge all four and compare left to right, since even a small gap explains a pull. If pressure is equal, the next question decides everything: does it pull braking or cruising? Braking-only is a stuck caliper, which no alignment fixes; a pull at all speeds is alignment. PulsCar's free check can't hear a pull; run these tests.

Find your situation: when did it start?

"After hitting a pothole or curb." → A wheel knocked out of alignment or a bent component (Causes 2, 7).

"After a tire rotation or new tires." → Tire-related — pressure or a radial pull (Causes 1, 3).

"Gradually, over months." → Alignment drifting or a slowly worn component (Causes 2, 5, 6).

"Only when braking." → A brake caliper or pad issue (Cause 4); see grinding when braking.

"With clunking over bumps." → A worn tie rod or ball joint (Causes 5, 6); overlaps with noise when turning.

"After a fender bender." → A bent suspension component (Cause 7). Inspect on a lift.

What actually determines a pull to one side cause and its cost

Real pull or road crown — the first gate. The parking-lot test decides whether there's anything to fix at all; a car that tracks straight on level ground has no fault, just slope.

Braking or cruising — the master variable. A braking-only pull is the brake system (a caliper), a cruising pull is alignment or suspension, and confusing the two is the single most expensive mistake here.

Whether pressure is equal. A side-to-side PSI gap is a free fix; equal pressure sends you to alignment or a component.

Whether it clunks over bumps. Clunking plus a pull points at a worn tie rod or ball joint — parts that must be replaced before an alignment will hold.

Whether it followed an impact. A pull that started after a pothole or crash points at bent geometry, which no alignment can correct.

Whether an alignment held. A pull back within weeks of an alignment means a worn or bent component was preventing the alignment all along.

Why does my car pull to one side when braking? A pull that appears specifically during braking almost always indicates a brake system problem rather than alignment, which matters because it means an alignment won't fix it. The most common cause is a stuck or seized caliper on one side that applies more braking force than the other, so the car lurches toward the sticky side. Other causes in the same family are uneven pad wear, a collapsed brake hose, or uneven rotor thickness. The reason it matters is that drivers and even shops sometimes treat a braking pull as alignment, pay for one, and the pull returns within days because the caliper pulls the car sideways regardless. This is a safety issue to fix within days, since a stuck caliper overheats and can fail. The free confirmation is the heat test: after a drive, one dramatically hotter wheel is your stuck caliper. PulsCar's free check hears a parked engine, not a brake; use the heat test.

The pull to one side price ladder: every cause, cheapest first

Read the ladder with the sequencing rule in mind. The top rungs — pressure, a rotation, an alignment — are where most pulls land and are cheap. But the expensive rungs carry a trap the cheap ones don't: if a worn tie rod or a bent control arm is the real cause, an alignment (a middle rung) won't hold, so you pay for it and then pay again for the component plus a second alignment. That's why the order is always components first, alignment second — fixing the ladder from the bottom up when a bottom-rung cause is present, not the top down. A bent arm is a replace-not-repair item — see control arm replacement cost.

The pull to one side number, by what you drive

Two rules for reading the table: first, match the trigger to the system before anyone aligns the car — a braking-only pull is a caliper, a cruising pull is alignment, a clunk-plus-pull is a worn tie rod or ball joint, and a pull that moves with a tire rotation is the tire, so an alignment is the answer only for the pure cruising pull with a healthy front end; and second, replace worn or bent parts first and align second, never the reverse — an alignment set on a worn tie rod, a bad ball joint, or a bent control arm drifts back out of spec within weeks, so paying for the alignment first just means paying for it twice.

Which pull to one side cause is yours? Five questions

Question 1: Does it drift on level ground with the wheel released? → No → road crown, not a fault. Yes → a real pull; continue.

Question 2: Is the tire pressure equal side to side? → No → low tire pressure (Cause 1). Free fix.

Question 3: Does it pull only when braking? → Yes → a stuck caliper (Cause 4). Alignment won't help.

Question 4: Does it clunk over bumps? → Yes → a worn tie rod or ball joint (Causes 5, 6). Replace before aligning.

Question 5: Did it start after an impact? → Yes → a bent component (Cause 7). No alignment will hold.

The 7 classic causes, ranked by danger

1. Low or unequal tire pressure — $0

An underinflated tire on one side creates more rolling resistance than the other, pulling the car toward the low tire — even a 5-8 PSI side-to-side gap does it. It's the most overlooked cause and costs nothing to check. The gauge check: grab a $5 gauge and read all four, comparing left to right against the sticker inside the driver's door jamb — not the number on the sidewall, which is the tire's maximum, not your target. More than 5 PSI of difference explains the pull. Fix: inflate the low tire to spec and drive normally — the pull should vanish within a mile. If it drops again within days, you have a slow leak (a nail, valve stem, or rim) to find and fix. About a fifth of all "pulling" complaints resolve right here, for free, which is why it's always the first check.

The cheap cause that gets skipped. Where and when you read the pressure. A tire gauged warm after a highway run, or filled from a forecourt machine whose dial is out by 3-4 PSI, reads fine and is not. Every manufacturer procedure for a pull opens here: FCA's bulletin 02-003-15 makes "Are the tires inflated to the recommended placard pressures?" the first of four checks before a technician is allowed to touch the alignment rack, and Nissan's NTB08-097J flow chart starts the same way for every Nissan except the GT-R.

At the shop — say this: "Write down all four cold pressures before you add any air, and tell me the gap between left and right. If one tire was several PSI low, that is my answer and I do not need an alignment today."

2. Wheel misalignment — $80 to $150

Alignment is the precise angles at which the tires meet the road — camber (tilt), toe (pointing in/out), and caster (steering-axis angle) — and when those go off, from a pothole, a curb, or normal wear, the tires fight each other instead of rolling straight. It's the most common cause of a pull that appears gradually, though a hard impact can knock it off suddenly too. The tire-and-wheel check: look at the front tires from the front for inward or outward tilt at the top (camber), note whether the steering wheel sits centered when driving straight, and check for one tire wearing faster on one edge — the classic alignment sign. Fix: a wheel alignment, ideally all four wheels — but insist the shop inspect the suspension first, because worn tie rods or ball joints will pull the alignment right back out of spec, and it can also show up as steering-wheel shaking.

The cheap cause that gets skipped. The difference between the two sides, not the numbers themselves. FCA's 02-003-15 says out-of-spec cross camber (left camber minus right camber) and cross caster are "the main sources of lead/pull complaints for vehicles equipped with hydraulic steering systems" — a car can have every individual angle inside its range and still pull because the two sides do not match. The same bulletin warns shops that FCA checks alignment on 100% of vehicles at the plant, so consistent out-of-spec readings on cars under 1,000 miles mean the rack needs calibrating, not the car.

At the shop — say this: "Give me the before-and-after printout and show me cross-camber and cross-caster, not just the individual angles. If both sides are in range and they still do not match, say so."

3. Uneven tire wear or radial pull — $20 to $800

A tire with internal belt separation or a manufacturing "radial pull" steers the car one way no matter what the alignment says, and rotating the tires is the diagnostic: if the pull changes direction or vanishes after moving tires front to rear, the tire is the cause. The tread check: inspect all four for uneven wear, cupping (a scalloped pattern), or flat spots, and check depth with a penny — if you can see all of Lincoln's head, the tire is at or below 2/32" and unsafe in the wet. Fix: rotate first ($20-$50), and if the pull follows the moved tire, replace that tire; if several tires show uneven wear, an alignment plus rotation stops it recurring on new rubber. It's a cheap first move that either fixes the pull or clearly points elsewhere, which is why a rotation is worth doing before an alignment when a tire is suspect.

The cheap cause that gets skipped. Cross-switching the two front tires, which costs a shop minutes and settles the question. It is step four of FCA's preliminary checks ("Does cross switching the front tire/wheel assemblies correct or improve the lead/pull condition?"), it is in Nissan's NTB08-097J flow chart, and GM's 17-NA-087 says the same thing for tramlining: "Swapping tires from a known good vehicle can be used to determine if the condition is tire related." If the pull swaps sides with the tires, no alignment on earth will fix it.

At the shop — say this: "Cross-switch the two front tires and road-test it again before you quote me anything. If the pull changes sides, the tire is the cause."

4. Stuck or seized brake caliper — $150 to $500

A stuck caliper keeps the pad pressed on the rotor even when you're not braking, so when you do brake, that side grabs harder and the car pulls toward it — and under hard braking that can feel like a dangerous sideways lurch. The heat danger is real: a constantly dragging caliper overheats that wheel, boils the brake fluid, causes fade, and can eventually fail on that corner. The heat test: after 10-plus miles of normal driving, carefully feel each wheel from a safe distance — a stuck-caliper wheel runs noticeably, sometimes dramatically, hotter; a car that pulls under braking but tracks straight at cruise is the caliper, not alignment. Fix: caliper replacement — always in pairs on the same axle, since mismatched calipers brake unevenly even after repair — with a rotor inspection in the quote. This is the pull that overlaps with grinding when braking and a dragging-caliper hum, and it's the one an alignment can never fix.

The cheap cause that gets skipped. A brand-new brake job. FCA's bulletin 05-008-15, for 2012-2015 Ram 1500, describes a grey anti-corrosion coating on new rotors that "can stick to the brake pad friction surface" and be felt as a pull or drift during braking — and the instruction is blunt: "Do not replace the brake rotors or calipers for this concern." Clean the rotor friction surface, replace only the pads. Ford's SSM 50718 says the same for 2017-2022 F-Super Duty: a pull or drift during braking "may be due to the brake pads needing to be broken in."

At the shop — say this: "This started right after a brake job. Check for uneven pad transfer on the rotor face before you sell me a caliper, and tell me what the pedal pressure test shows."

5. Worn tie rod ends — $200 to $500 per side

Tie rods connect the steering rack to the wheel hub, and the ball joint inside each tie rod end wears over time, developing play that lets the wheel move slightly independently of your steering — felt as vague, imprecise steering that wanders and drifts. The driveway play test: on level ground, turn the wheel quickly about two inches left and right — any slop before the tires respond is tie-rod play; for a definitive check, safely jack the front on stands, grab the tire at 9-and-3 o'clock and push-pull for outer tie-rod movement, then at 12-and-6 for inner. A note on shop honesty: if an alignment shop asks you to replace tie rods before they'll align the car, that's legitimate, not an upsell — worn tie rods make it impossible to hold alignment angles. Fix: tie-rod replacement followed immediately by a four-wheel alignment; ask to see the alignment spec sheet before and after. It's the same family covered in noise when turning.

The cheap cause that gets skipped. A loose bolt rather than a worn joint. On solid-axle trucks and Wranglers the track bar and its bracket carry the axle sideways, and owners in our complaint set repeatedly describe track bars and steering dampers replaced under bulletins before anyone measured the tie rods. A loose steering-gear mounting bolt produces the same vague, wandering feel for the price of a torque wrench.

At the shop — say this: "Before you quote inner tie rods, put it on the lift, rock the wheel, and show me where the play actually is — and check the steering-gear mounting bolts and the track bar torque first."

6. Worn ball joints — $300 to $600 per side

Ball joints are the pivot points between the suspension arms and the wheel hub, letting the suspension move while the wheel also steers — the lower one carries most of the car's weight — and when they wear, the wheel's camber shifts slightly as the car moves, creating inconsistent pull and unpredictable handling. The boot-and-rock check: turn the wheels to full lock and look at the exposed lower ball joint — the rubber boot should be intact and round, not torn or flat (a torn boot means the grease is gone and wear has accelerated); for the definitive test, jack the car at the frame (not the control arm), grab the tire at 12-and-6, and rock it vertically — any perceptible vertical play means lower ball-joint wear that needs immediate attention. Fix: replacement in pairs on the same axle (mismatched new-and-old handles asymmetrically), alignment after; where the joint is pressed into the control arm, ask the shop to check whether the arm itself is bent or cracked from the same impacts that wear joints.

The cheap cause that gets skipped. Confusing ball-joint play with wheel-bearing play. Both show up as movement when you rock the tire at 12-and-6, both make the steering feel loose, and only one of them is a ball joint — the difference is whether the movement is at the joint or inside the hub, and it takes a technician thirty seconds with a pry bar to tell you which.

At the shop — say this: "Rock the wheel at 12-and-6 on the lift and show me whether the play is in the ball joint or the wheel bearing. I want to see it move before I pay for either."

7. Bent suspension component — $300 to $1,500+

A hard impact bends metal designed to flex only within tolerance — a bent control arm permanently changes that wheel's camber, a bent knuckle changes several angles at once — and no alignment can compensate, because the adjustment bolts simply run out of range on bent geometry. The "alignment won't hold" tell: if you had an alignment in the last few months and the pull is back, or the shop says they can't reach spec, ask directly whether every alignment angle is within the manufacturer's adjustment range — any angle at the limit or beyond means a bent part is the block. The post-impact look: stand 10 feet ahead and compare the two front wheels — if one leans in or out noticeably more than the other, suspect a bent component, and any impact that changed how the car drives warrants a lift inspection even with no visible damage. Fix: identify and replace the specific bent part first, then align — never try to straighten bent suspension or steering parts, since the impact has work-hardened the metal and it won't return to factory strength.

The cheap cause that gets skipped. Rust, not impact. Six of the 110 complaints we read in full were pulls that turned out to be a rotted subframe or rear axle carrier — and in several of them the shop found it only because an alignment could not be performed at all. On a salt-belt car over ten years old, a pull that arrives with a clunk deserves a look underneath before anyone quotes a control arm.

At the shop — say this: "If you cannot reach spec, tell me which angle is outside its adjustment range and put that on the invoice. And look at the subframe and the rear axle for rust before you quote me a bent part."

Three more causes the classic list misses

The seven above are the ones every list has. Reading 246 manufacturer communications about pulling and drifting turned up three more that almost none of them mention — and two of the three are free. They are not exotic: one of them is the single largest category in the repair outcomes we read, and the other two are conditions that manufacturers tell their own technicians not to fix with an alignment.

8. Power steering that generates its own pull — $0 to $2,800

Electric power steering does not just reduce effort — it applies torque, and a control unit that applies a little of it in one direction produces a pull that looks exactly like an alignment fault and is not one. Acura's bulletin 17-033, for 2014-2017 RLX, is the clearest statement of it in any manufacturer document we read: the symptom is "the steering intermittently pulls to the right while driving," the cause is that "the steering rack is intermittently self-generating a pull to the right because of an EPS programming issue," and the fix is a control-unit update — with the pointed note that the bulletin "does not fix drift or pulling caused by alignment or tire issues." Ford's SSM 45996 handles 2015-2016 F-150 drift, pull and wander by reprogramming the power steering control module, and its SSM 47589 routes 2016-2018 F-150 complaints to a dedicated pinpoint test rather than the alignment rack. Stellantis went further for the Wrangler and Gladiator: bulletin 08-074-20, revised six times between September 2020 and March 2021, replaces the steering gear and updates the Electric Hydraulic Power Steering software. And NHTSA investigated the category itself in PE10-008, opened in February 2010 into steering wander or drift on 2009-2010 Toyota Corolla and Matrix; the agency closed it in May 2011 finding that the complaints "all relate to the system operating as designed and do not involve failure of any steering system components," after Toyota issued a bulletin releasing an EPS control unit with a re-calibrated on-center feel.

The cheap cause that gets skipped. A software update you already qualify for. In our reading of 110 complaints, steering hardware and software was the largest single outcome — 22 of 110 — and several of those owners had already paid for alignments before anyone looked at the module. A calibration is usually free under warranty and takes under an hour.

At the shop — say this: "Before another alignment, look up open bulletins for my VIN on the steering system and tell me whether there is a control-module update. If the alignment printout is in spec and it still pulls, the module is the next place to look."

9. Tramlining: wide, low-profile tires following the road — $0

GM's bulletin 17-NA-087 covers Cadillac ATS and CTS, Camaro, Corvette, SS and other wide-tire models, and it exists because owners were being sold alignments for physics. "Tramlining" is the industry word for a car following the longitudinal ruts in worn pavement; GM's document says "all vehicles tramline to some degree, however vehicles equipped with low aspect ratio, wide tires tend to be more sensitive to this condition," and that such cars "can feel 'pulled' in both directions depending on the positions of the tires in the trough." The sentence that should stop an upsell is about the alignment itself: "While alignment may be thought to be a contributing factor, it has NOT been found to affect trough wander or tramlining complaints and therefore a wheel alignment check is NOT recommended for the condition." GM also warns that over-inflating past the placard stiffens the sidewall and can make tramlining worse — the opposite of the usual advice.

Fix it yourself — the two-sides-of-the-lane test. GM tells its own technicians to drive the car within the lane but on each side of it, because ruts are not always visible and are not always in the middle. Do that on a stretch of grooved asphalt, then repeat on smooth concrete. A car that wanders on one surface and tracks perfectly on the other is reading the road, not fighting its own geometry. The honest limit: this tells you what the pull is not. It does not make a tiring car restful, and the only real change is a tire with a different tread and sidewall.

The cheap cause that gets skipped. The tires you chose. A Jeep bulletin quoted back by an owner in our complaint set blames the tread and sidewall design of a specific all-terrain tire for the described condition and concludes "no vehicle repairs are needed at this time." That is an unsatisfying answer, but it is cheaper than three alignments.

At the shop — say this: "Drive a same-model car over the same road before you touch mine. If it does the same thing, I am not paying for an alignment."

10. Lane-keeping assist steering the car — $0

Twenty-two of the 612 complaints we analyzed name a lane-keeping, lane-centering or adaptive-cruise system as the thing steering the car, and manufacturers say the same in their own words. Ford's SSM 46259 tells technicians that for a 2017 Escape with the technology package, a steering pull or vehicle drift complaint may come from the lane-keeping system; SSM 47638 extends it to 2013-2019 vehicles with lane keeping, where steering-wheel vibration, steering pull or drift "may be due to an alert and/or steering aid input from the LKS system." One owner in our set described turning the system back on to reproduce the problem on a highway on-ramp and getting a swerve for it. This is not a defect you repair; it is a setting you test around — but it matters, because a driver who cannot separate the two spends money on the wrong system.

Fix it yourself — the switch-it-off test. Turn lane keeping, lane centering and any steering-assist mode off in the vehicle settings, then repeat the level-ground release test from the top of this page at the same speed on the same road. A pull that vanishes with the system off was never mechanical. The honest limit: some systems re-enable themselves at every start, and adaptive cruise can apply braking on one side without any steering input at all, so run the test with cruise off too.

The cheap cause that gets skipped. The setting menu. It costs nothing, it takes a minute, and it is the only cause on this page you can rule out from the driver's seat before you leave the driveway.

At the shop — say this: "I have already road-tested it with lane keeping and cruise switched off and it still pulls. Do not start with the driver-assist calibration."

Alignment pull or brake pull? How to tell them apart

The classic and most expensive mistake is treating a brake pull as an alignment problem, because no alignment adjustment fixes a brake that pulls the car sideways regardless of where the wheels point. Here's how to tell the two apart before you spend. A brake pull shows only when braking: the car tracks straight at a cruise and lurches toward one side the moment you press the pedal, and the giveaway is the heat test — after a drive, one wheel runs dramatically hotter than the others where the caliper is dragging. An alignment or suspension pull shows at all speeds: it's there at a steady highway cruise with your foot off the brake, and it comes with off-center steering or uneven tire wear rather than a hot wheel. A pull that worsens under acceleration is a third category — often a worn motor mount or a drivetrain issue — and it too is not an alignment problem. So the sequence before any alignment is: pressure, then the braking-versus-cruising question, then a suspension inspection — and only then the alignment, which is the right fix for exactly one of these patterns.

Will a wheel alignment fix my car pulling to one side? An alignment fixes pulling caused by incorrect wheel angles — camber, toe, and caster out of spec — the most common cause of a steady pull that appears gradually, so for many cars the answer is yes. But an alignment will not fix several other causes, and knowing the difference saves you from paying for one that won't help. It won't fix a stuck caliper (that's a brake pull), worn tie rods or ball joints (those must be replaced first, or the alignment drifts back out quickly), a tire with belt separation or radial pull (replace the tire), or a bent component (the adjustment bolts run out of range on bent geometry). This is why a good shop inspects the suspension and checks the brakes before aligning — an alignment on a worn or bent front end is money on a fix that can't hold. PulsCar's free check can't hear a pull; the braking-versus-cruising question does.

What 612 NHTSA complaints about pulling actually show

We pulled every owner complaint NHTSA published between 1 January 2015 and 8 September 2026 — 777,040 of them — and searched the narratives for a car that pulls, veers or drifts to one side. The raw keyword search returned 2,551 complaints; reading 30 of them at random showed that 57% were not this problem at all but one-off events, a wheel coming off or a driver describing a deliberate maneuver. Tightening the query — the complaint has to be tagged to steering, suspension, tires, wheels, brakes or structure, has to contain a steadiness cue like an alignment attempt or "constantly" or "let go of the wheel", and cannot be a crash narrative — left 612 complaints, and a second reading of 30 fresh ones put the residual error rate near 10%. Every number below comes from that set. The full method, the exact phrases and both sample readings are in method.txt.

Two things in that table are worth sitting with. The first is that a pull is, statistically, the least dramatic thing owners report to NHTSA: it gets towed and crashes less often than the average complaint in the database, which fits the everyday experience of it — annoying, expensive, rarely an emergency. The second is the alignment column. Two out of three complaints mention an alignment, 104 owners say one was actually done, and more than a third of those say the car pulled anyway. That is the whole argument of this page in one ratio: the alignment is the default answer, and for a large minority of cars it is the wrong one.

Mileage is the surprise. Of the 612, 362 report a usable failure mileage (we drop zeros and anything over 400,000). The median is 24,000 miles. The 25th percentile is 1,098 miles. And 87 of the 362 — 24% — report the pull at 1,000 miles or less, which is to say a quarter of the pulls owners complain about arrived with the car. That is not a worn tie rod or a pothole; it is delivery condition. It also explains why so many manufacturer bulletins on this subject exist to tell dealers how to fix a car that pulls before it is sold.

Which models and years show up most — and why that is not a ranking

Counts of complaints are not failure rates. There is no denominator in this data: nobody publishes how many of each model are on the road, how old they are, or how likely their owners are to file with a federal agency. A model can lead this table because it sells in huge numbers, because its owners are unusually organized, or because it genuinely pulls. With that said, here is where the 612 land.

By make: Honda 79, Jeep 72, Ford 61, Toyota 60, Hyundai 49, Ram 32, Kia 30, Chevrolet 29, Nissan 25, Subaru 21. Model years cluster in 2013 (51), 2015 (53), 2018 (55) and 2020 (50) — the years when electric power steering became near-universal, which is consistent with what the repair outcomes show, and which is also exactly the kind of correlation that a complaint database cannot prove. Two model-level patterns are worth naming because the manufacturer's own paperwork backs them: the Wrangler and Gladiator cluster sits alongside Stellantis bulletin 08-074-20 and its six revisions on steering gear and EHPS software, and the 2019 Honda Passport and Pilot appear in 27 separate dealer messages about steering found off center or pulling at pre-delivery inspection.

What owners say actually fixed it — 110 complaints read in full

Counting keywords tells you what people complained about. It does not tell you what the car turned out to need. So we took the 285 complaints in the set that contain a diagnosis or repair cue, drew 110 at random, read every one, and assigned each a single primary outcome: what the owner says the shop found or did.

Three findings there are worth more than the counts. Steering hardware beat alignment two to one. The single largest outcome was not the wheel angles at all but the steering system itself — a rack, a gear, a module, a pump. Brakes were almost absent: 4 of 110. That does not make a braking pull unimportant — it is still the one with a safety deadline — but it does mean a brake pull is a specific, testable minority, not the default. And 31% of the complaints we read (34 of 110) end without a named cause, either because the shop could not reproduce it or because the owner was still going back. That is the single most useful number on this page for setting expectations: a third of the people who write to a federal agency about a pull do so because nobody has been able to tell them what it is.

One honesty note on that last figure. A complaint is filed at a moment in time, often while the car is still at the shop, so "unresolved" can mean "not yet resolved" rather than "never fixed"; nobody files a follow-up saying it all worked out. Read it as a floor on frustration, not a repair-success rate.

What the recalls and bulletins say — 246 documents and 37 campaigns

The same flat files carry manufacturer communications and recalls. Searching bulletin summaries for a vehicle that pulls, leads, drifts, wanders or sits off center — then hand-filtering the false positives that phrase catches, such as camera images and exhaust tips being off center — leaves 246 distinct documents dated 2016 to 2026. Jeep files the most (48), then Honda (33), Chevrolet, Ram and Dodge (15 each), Audi (14), Toyota (13) and Mazda (12).

Document counts are not problem counts, and this set proves it: 27 of Honda's 33 are the same dealer message, about 2019 Passport and Pilot steering found off center or pulling at pre-delivery inspection, re-issued week after week from April to September 2019. Strip the repeats and Honda has one issue, not 33. Jeep's total contains 08-074-20 in Revisions A through F. Any article that tells you a brand "has 48 bulletins about pulling" is counting paperwork.

What the documents actually tell technicians to do is more useful than the totals, and four of them are worth reading yourself. Stellantis 02-003-15 is the most complete diagnostic procedure published for this symptom by anyone: it insists on placard tire pressures, even tire wear, a damage check and a front tire cross-switch before the alignment rack, then defines the pull by stopwatch — drive a right-crowned road at a steady 50 mph, let the car find its own direction, and time how long it takes the front tire to reach the painted lane line, in both directions. Under 7 seconds average, you have a lead/pull condition. Over 7, "this bulletin does not apply." Stellantis 05-008-15 explains a braking pull caused by nothing more than the anti-corrosion coating on new rotors. GM 17-NA-087 tells dealers not to sell an alignment for tramlining at all. And Nissan's NTB08-097J opens with a plain-English explanation of road crown for every Nissan except the GT-R, ending: "This bulletin does not address road crown issues because they are not vehicle related, although the customer may incorrectly perceive them to be."

The recalls are a different animal. Thirty-seven campaigns describe a defect or consequence that makes the vehicle pull, veer or drift to one side, covering 1,401,412 vehicles. But 663,823 of those units are heavy trucks — Freightliner, Western Star, International, Mack, Hino — and nearly all of those are air-brake campaigns where one wheel's brake fails and the truck pulls. Among light vehicles the largest by far is 19V-761: 652,967 Chevrolet Suburban, Tahoe, Silverado 1500, GMC Yukon and Sierra 1500 with a 5.3-liter engine, 3.08 rear axle and four-wheel drive, where a failed wheel-speed sensor could let the brake-control software activate driveline protection and brake the opposite wheel, "causing the vehicle to pull to one side unexpectedly." The remedy was a reprogram. Across all 37 campaigns, five of them — 984,585 units — are fixed with software. A pull is no longer necessarily a mechanical fault.

Three pull to one side traps, dissected

Trap one: the alignment sold for a brake pull. The situation: a car pulls, a shop aligns it, it drives straight for three days, then the pull returns — because a stuck caliper was the cause. What's real: the pull was braking-only, and the heat test showed one hot wheel. The price: $120 for an alignment that couldn't hold, plus the caliper still to fix. The defense: ask "does it pull only when braking?" and do the heat test before any alignment.

Trap two: the alignment on a worn front end. The situation: an alignment is set on a car with worn tie rods, and the pull is back within weeks. What's real: worn tie rods can't hold alignment angles, so the alignment drifted straight back out. The price: an alignment paid twice, plus the tie rods. The defense: insist the suspension be inspected first, and replace worn parts before aligning.

Trap three: the $25 rotation skipped for a $150 alignment. The situation: a pull from a radial-pull tire gets an alignment that doesn't fix it. What's real: rotating the tires would have moved the pull and named the bad tire. The price: $150 for an alignment when a $25 rotation was the test. The defense: rotate first when a tire is suspect — it's the cheapest diagnostic there is.

Three pull to one side situations, decoded

These are composite scenarios, not individual customers: each combines details that come up repeatedly in owner reports and shop quotes.

Scenario 1: the $120 alignment that lasted three days. A driver's car pulled, got aligned, and pulled again within the week. The pull was braking-only — a $300 stuck caliper. Lesson: a braking pull is brakes; an alignment can't touch it.

Scenario 2: the free fix that ended a "steering problem." A driver was quoted an alignment for a mild drift. A gauge showed the right front 8 PSI low; airing it up fixed it for $0. Lesson: check pressure before you spend anything.

Scenario 3: the alignment that wouldn't hold. A driver's alignment kept drifting back out within weeks. Worn tie rods were the block; replacing them ($300 a side) let the alignment finally hold. Lesson: replace worn parts first, align second.

Four pull to one side playbooks for where you are right now

"It pulls only when I brake." A stuck caliper — the safety one. Do the heat test (one hot wheel), and replace the caliper in pairs on that axle within days, since a dragging caliper can overheat and fail. An alignment won't help; see grinding when braking.

"It pulls at all speeds, steering's off-center." Alignment — but check pressure first and have the suspension inspected, or the alignment won't hold.

"It pulls and clunks over bumps." A worn tie rod or ball joint. Jack and rock (9-and-3, then 12-and-6), replace in pairs, align after — see noise when turning.

"It started right after a pothole." Possibly a bent component. Compare the two front wheels by eye, get a lift inspection, and if a shop "can't reach spec," a bent part is the block — replace it before any alignment. See signs of overcharging if the quote balloons.

Read your car's own pull to one side record before you pay anyone

Everything above came from public files, and the same files answer questions about your specific car in about five minutes.

1. Find your VIN — driver's door jamb sticker, the base of the windshield on the driver's side, or your insurance card.
2. Run the VIN through NHTSA's recall lookup. Open recalls are free to fix at any franchised dealer, with no time limit and no ownership requirement. Recall 19V-761 above is the reason this matters for a pull specifically.
3. Search manufacturer communications for your year and model. NHTSA publishes the summaries; our bulletin CSV is a filtered copy of every one that mentions pulling, drifting, wandering or an off-center wheel, with the document id you can quote to a service advisor.
4. Read other owners' complaints for your exact year. Filter for steering and suspension and read what the shops found. It is the closest thing to a second opinion that costs nothing.
5. Take the document number with you. A service advisor who hears "bulletin 02-003-15" runs a different playbook than one who hears "it pulls."

Should you buy a used car that pulls on the test drive?

Usually yes — at a discount, and only after you know which kind of pull it is, because the same symptom is worth $0 and $2,800 on this page. Run the level-ground release test during the test drive, then ask the seller the one question that matters: does it pull under braking or at a steady cruise? A steady cruising pull on a car with equal tire pressures and clean tire wear is the cheap version; budget an alignment and move on.

Walk away, or spend $100-$200 on a pre-purchase inspection first, when any of four things is true. The tires wear unevenly — feathered, cupped, or worn on one edge — because that is a car whose geometry has been wrong for a while and a used set of tires is now part of the bill. It pulls under braking, which is a brake repair the seller has already decided not to make. It clunks over bumps, which means the front end is worn and an alignment cannot hold until parts are replaced. Or the seller volunteers that "it just needs an alignment" — 104 owners in our data had exactly that done and 37 of them say the pull came back. On anything ten years old in a salt state, have the subframe and rear axle looked at from underneath: six of the 110 complaints we read in full ended in rusted structure, and that is the one outcome on this page that can total a car.

After the fix: keep it tracking straight

Check tire pressure monthly. Equal pressure left to right prevents the most common pull, and it catches a slow leak early. Rotate tires on schedule, which heads off the uneven wear that pulls the car. Align after new tires or suspension work. Avoid potholes and curbs where you can — those impacts are the leading cause of a sudden pull. And don't ignore a braking pull, the one with a safety deadline. When a big repair is quoted, an independent shop, a look at the signs of overcharging, and the noise-diagnosis guide beat a guess.

Your pull to one side plan: right now, today, this week

Right now (free):

1. Do the parking-lot road-crown test — drift on level ground means a real pull; straight means it's just road slope.
2. Gauge all four tires and equalize left to right against the door-jamb spec.
3. Note whether it pulls only when braking (brakes) or at all speeds (alignment/suspension).

Today: 4. If it's braking-only, do the wheel-heat test for a stuck caliper. 5. If it clunks over bumps, do the jack-and-rock test for tie rods and ball joints.

This week: 6. Match the fix to the test — pressure or a rotation for a tire pull, an alignment for a clean cruising pull, a caliper for a braking pull, components before any alignment if the front end is worn. 7. Before any alignment → check pressure, rule out a braking pull, and have the suspension inspected. 8. If an alignment "won't hold" or a shop "can't reach spec" → suspect a worn or bent component and fix that first.

For the related steering, brake, and cost guides: wheel alignment cost, noise when turning, steering-wheel shaking, grinding when braking, humming noise, clicking noise, and brake pad pricing. For diagnosis and money: strange car noises, what a diagnostic should cost, an honest mechanic, and signs you're being overcharged. And our story explains why PulsCar exists, tie rod replacement cost.

Sources and documents

Primary documents, downloaded from NHTSA's server and read in full. Counts are our own analysis of NHTSA public flat files. Technical documents verified September 2026.

1. Stellantis 02-003-15, 27 January 2015 — "Vehicle Pulls Or Leads Right Or Left": the four preliminary checks, the 7-second lane-line test, cross camber and cross caster. PDF
2. Stellantis 05-008-15, 29 August 2015 — "Vehicle Pulls Or Drifts While Applying The Brakes": rotor anti-corrosion coating, pads only. PDF
3. Stellantis 02-001-20, 23 April 2020 — 2020 Ram 1500 lead/pull from plant alignment settings, released as Rapid Response Transmittal 20-028. PDF
4. GM 17-NA-087, March 2017 — tramlining and wide low-aspect tires; alignment explicitly not recommended. PDF
5. Acura 17-033, 15 September 2017 — EPS self-generating a pull to the right on 2014-2017 RLX. PDF
6. Nissan NTB08-097J, 18 November 2021 — pull and off-center diagnosis for all Nissan except GT-R, including the road-crown explanation. PDF
7. Ford SSM 50718, 2022 — 2017-2022 F-Super Duty pull or drift during braking, pads needing break-in; Ford SSM 47589, 2018 — pinpoint test H for steering pull, drift and wander. 50718 · 47589
8. NHTSA recall 19V-761 — 652,967 GM trucks and SUVs, driveline-protection software, remedy a reprogram. Recall record
9. NHTSA investigation PE10-008, opened 18 February 2010, closed 4 May 2011 — steering wander or drift, 2009-2010 Toyota Corolla and Matrix; quoted from the ODI investigations flat file. Field definitions
10. Our analysis — 612 complaints from 777,040 published 1 January 2015 to 8 September 2026; 110 read in full; 246 bulletins; 37 recall campaigns. Method · complaint field definitions · CC BY 4.0

What we looked for and did not find. No recall in the file covers a passenger car that pulls because of ordinary alignment or suspension wear — every light-vehicle campaign here is a specific broken or mis-programmed part. And we could not re-verify the repair prices in this pass: the estimator sites we would check refuse automated access, so the dollar ranges on this page stand as previously published estimates, not September 2026 re-checks.

How these numbers were built: cross-checked against 2026 estimator and shop-survey data (tire pressure $0; tire rotation $20-$50; wheel alignment $80-$150, premium near $200; two tires $400-$800; stuck caliper $150-$500; tie rod $200-$500 per side; ball joint $300-$600 per side; bent component $300-$1,500+), at 2026 independent-shop labor rates. The figures assume the pull is confirmed real with the parking-lot test, tire pressure is ruled out first, and worn or bent components are replaced before any alignment — since an alignment set on a worn front end or a brake pull cannot hold. Prices verified July 2026 and reviewed quarterly — the estimator sites that carry them refused our checks in this pass, so the ranges are unchanged.

Have a pulling pattern we didn't cover? Email [email protected] with whether it pulls braking or cruising, whether it clunks over bumps, and whether it started after an impact, and we'll add it to the next version of this guide.