Newark, LaGuardia or JFK: which New York airport actually works

New York is the only US market where a traveller routinely picks between three major airports, and the three do not perform alike. Over June 2025 to May 2026: Newark Liberty International arrived on time 77.2% of the time, LaGuardia 72.6%, John F. Kennedy International 75.1%.

Those are the grades B−, C− and B− respectively on this site's scale. Where two of the three land on the same letter, read the underlying percentages rather than the letter: a grade band is four points wide and hides differences that matter over a year of trips.

The ordering above persists across years, for reasons that have nothing to do with which airline you booked.

The airspace is the constraint, not the airports

The three fields sit inside roughly twenty miles of each other and share one of the most congested pieces of controlled airspace on earth. Departure and arrival paths cross. A runway configuration change at one airport forces sequencing changes at the other two. When the ceiling drops, arrival rates fall everywhere at once.

This matters more than it sounds, because it means airline choice cannot rescue you. A carrier operating out of Newark faces the same flow control as every other carrier there. The 21% of national delay minutes charged to the National Airspace System is heavily concentrated in markets like this one, and no amount of operational excellence prices its way out of a ground delay program.

What you can change is which of the three you use.

What the gap is worth

Cancellations diverge as sharply as punctuality: 3.10% at Newark, 4.60% at LaGuardia, 3.09% at Kennedy.

For a connection, the number that decides the outcome is the share of arrivals landing an hour or more late: 9.3% at Newark against 9.8% at Kennedy. A 60-minute layover survives one of those distributions considerably better than the other.

Averaged across a year of monthly trips, the difference between the best and worst of the three is several extra ruined evenings. That is the honest way to think about a few percentage points: not as a probability you will feel on any one flight, but as a count of bad days over a hundred trips.

What each one is actually good for

Newark is the strongest option for anyone based in New Jersey or western Manhattan, and it is a United hub with the deep schedule that implies. Its weakness is structural: the airport sits under some of the most constrained airspace in the country and its own on-time figure reflects that. If you use it, take a morning departure and mean it.

LaGuardia has the shortest ground journey from Manhattan and the thinnest margin for error. Short runways rule out most long-haul aircraft, and the perimeter rule limits how far flights can go, so the schedule is dominated by short domestic hops with tight turns. Short flights have the least cruise time in which to recover a late pushback.

Kennedy carries the most international traffic and the longest average stage length. Longer flights carry more scheduled block time, which is part of why its arrival figures read the way they do. It is also the furthest from most of Manhattan, and the ground journey is the variable travellers most consistently underestimate.

The decision, in order

Does one of the three offer a non-stop the others do not? Take it. A connection costs more reliability than any airport difference recovers.

Is the ground journey honestly comparable? Ninety minutes to Kennedy against thirty to LaGuardia is not a fair trade for a few points of on-time rate. Measure door to door, at the hour you will actually travel.

Are you connecting onward? Then weight the hour-or-more-late share above the headline rate, and give yourself ninety minutes rather than sixty.

Is it winter? All three sit in the same weather. The airport choice does not help; the date might.

What the data cannot settle

Frequency matters and does not appear in any of these figures. An airport where your carrier flies your route six times a day recovers from a cancellation in hours. One where it flies twice recovers tomorrow. Before choosing on reliability alone, count the departures.

The per-airport pages here carry each field's own hour-block curve, monthly strip and carrier table, which is where the choice actually gets made.

The hour matters as much as the airport

Nationally the 5am–8am departure window runs 89.3% on time against 67.7% for 5pm–8pm, a spread of 21.6 points. In the New York market that curve is steeper than average, because the airspace never gets a chance to drain. A delay absorbed at ten in the morning is still being paid off at seven in the evening.

The practical consequence is that a morning departure from the weakest of the three airports can beat an evening departure from the strongest. Before treating the airport as the decision, check the hour-block chart on each field's page and compare like with like: the 7am from Newark against the 7am from Kennedy, not the 7am from one against the 6pm from the other.

Frequency is the column nobody reads

Reliability figures tell you how often a flight goes wrong. They say nothing about what happens next, and in this market the answer varies enormously by route.

Work it through. Your carrier flies your route six times a day from one airport and twice from another. Both fields post the same cancellation rate. On the six-a-day route a cancellation costs you three hours; on the twice-daily one it can cost you the night, because the later flight is already sold out and the next morning's is the first real option.

That difference is larger than anything in the on-time column, and it is not in the DOT data at all. Count the daily departures on your actual route before letting a two-point gap decide anything.

What would change this conclusion

Two things, and it is worth naming them rather than presenting the ranking as permanent.

A runway reconstruction or a new approach procedure at any of the three would move its figures materially, and these projects run for years. And a carrier substantially rebuilding its schedule at one field changes that field's average without changing anything about the airport itself, because the average is weighted by who flies there.

The figures on this page are recomputed monthly from the raw filings, so they follow those changes. The reasoning above is what you should re-check if the numbers ever stop matching it.

---

Compare the three directly: Newark (EWR) · LaGuardia (LGA) · Kennedy (JFK)