Is it worth buying a new tractor?

A money question we cannot answer with money — so here is what the data does settle.

We cannot tell you, and the first useful thing this article does is explain why rather than pretend otherwise.

We hold no price data. Not list prices, not used values, not auction results. None of it is published freely at any scale: list prices are a dealer matter and vary by specification and territory, used values sit inside paid trade guides, and auction results are scattered across hundreds of sales with no common format. Every figure on this site carries a source, so rather than print a plausible number we have nothing to attach to, we leave the column empty and say so.

What our records can settle is the question underneath the money one: how long these machines last, and how well the old ones are documented.

How old is the documented fleet?

Much older than the showroom suggests, and the figures are blunt.

Queried 2 October 2026, across 3,232 published machines, every one with a sourced start year:

  • The median machine was introduced in 1983 — 43 years ago.
  • 2,810 are at least 25 years old, and 1,846 are at least 40.
  • Only 422 were introduced within the last 25 years.
  • The range runs from 1913 to 2023, and the busiest decades are the 1990s (680), the 1980s (610), the 1970s (531).

Our coverage skews old, and that is partly an artefact of where documentation exists — but only partly. A tractor is a capital item with a working life measured in decades, and the machines that got documented are the machines that stayed around long enough to be worth documenting.

Why is there no free price data?

Because of how the trade is structured, and it is worth understanding before you trust any site that claims to know.

New list prices are set by distributors and move with specification, territory, finance deal and whatever the dealer needs to shift that quarter. A single headline number for a model is a fiction before you have chosen a transmission.

Used values are the product of paid trade guides. Those organisations collect auction and retail data, clean it, and sell access — which is a legitimate business, and it also means the numbers are licensed rather than public. A free site publishing them at scale is either paying for the right or taking what it has not paid for.

Auction results are the one genuinely public strand, and they are scattered across hundreds of sales with no shared format, no consistent condition grading, and heavy selection effects: what reaches auction is not a random sample of what exists.

So the honest position is that a free, comprehensive, sourced price dataset does not exist. We could assemble a bad one. We would rather have an empty column than a column of numbers whose provenance we cannot print — which is the standard every other figure here is held to.

What does “well documented” buy a buyer?

The ability to check what you are being told, which is worth more on a used machine than on a new one.

A new tractor comes with a specification sheet from the people selling it. A thirty-year-old machine comes with whatever the previous owner remembers. The difference between those two situations is exactly what an independent record is for.

We publish 232 independent test reports behind pages you can reach, covering 189 machines. These are laboratory measurements, not manufacturer claims, and they cluster on the 1990s and 2000s — 64 and 117 machines respectively in the set where we hold both a claim and a measurement. The newest machines are the least likely to have an independent figure on record, because the publishing programmes move slowly and the machines are still current.

That inverts the usual assumption. If what you want is a verified number rather than a brochure figure, the older machine is more likely to have one.

What about parts and the size of the production run?

Here our data says something concrete, and it is the closest thing to a financial argument we can make honestly.

Across the 1,022 models where we hold a production figure, only 17 passed a hundred thousand units — and those are the machines whose parts are still made, whose manuals still circulate, and whose quirks are known to every mechanic over fifty. The Ford 8 N alone was built 500,000 times between 1947–1952.

The median production run, by contrast, is 1,557. A machine built fifteen hundred times is a different proposition from one built half a million times, whatever its condition, and it is a difference you can look up before you travel to see it.

Documentation density follows the same logic, and it is worth knowing which machines in a band are the well-covered ones before you start looking. In the 120 to 200 hp band we publish 439 machines, and the three documented most completely in our records are the John Deere 6910, SAME Laser 150 DT and Case-IH MX 170. In the 60 to 120 hp band, across 1,162 machines, it is the Deutz-Fahr Agroplus 70, Case-IH MX 100 and Lamborghini 674-75 (DT). Above 200 hp, across 177, the Case-IH MX 230, Case-IH MX 240 and John Deere 8320.

Those are not recommendations and we are careful not to dress them up as any. They are the machines whose figures you can most thoroughly check against an independent record before you spend money, which is a different and more defensible claim than “best”.

That is not advice about whether to buy new. It is the variable people most often fail to check.

From our database

Here is the honest summary of what this site can and cannot contribute to a buying decision.

What we hold: 3,232 machines across 85 makes, every stored value carrying a source. Power on an explicit standard. Mass, dimensions, hydraulic capacity, transmission, PTO speeds. 232 independent test reports. 3,214 pages with a picture.

What we do not hold: any price. Any dealer inventory. Any service history. Any failure rates. Any fuel consumption measured in the field. We also hold no figure at all for most machines of the last few years, because the documented record lags the market by a decade or more — the newest machine in our records is from 2023.

The one number we would press on anyone comparing machines, new or used: across the 185 machines where we hold both a manufacturer’s claimed engine figure and an independent measurement at the power take-off, at the engine’s own rated speed, the measured figure is 11.4% lower on average, and it was lower on 180 of them. Compare a claimed figure with a measured one and you will mis-rank two machines without ever making an arithmetic error. See PTO horsepower versus engine horsepower.

For sizing rather than pricing, which tractor is best for farming works through it by job and acreage. To browse by scale, start at the tractor index, or compare utility tractors against the high-power machines.

And if a site does quote you a confident used value, ask where it came from. The answer is usually a trade guide that site is not licensed to republish, or nowhere at all.

Common questions

Why don't you publish tractor prices?
Because no comprehensive price data is published freely. List prices are dealer matters, used values sit behind paid trade guides, and auction results are scattered. We would rather say we do not have it than print a number whose source we cannot name.
Do older tractors hold their value?
We cannot answer that from our data, and anyone answering it confidently from free sources is guessing. What we can show is that older machines remain in service and documented in very large numbers, which is a precondition for a resale market rather than proof of one.
How old is a typical working tractor?
Older than most people expect. The median machine in our records was introduced in 1984, which is 43 years ago, and only 422 of 3,232 were introduced within the last 25 years.
Is a new tractor better documented than an old one?
Not in our experience of the sources. Older machines are often better documented, because the independent test programmes published their results — we hold 232 test reports behind published pages, and they cluster on machines from the 1990s and 2000s rather than the newest ones.
What should I actually compare, then?
Figures on the same basis. The single most common mistake is comparing a manufacturer's claimed engine power with another machine's measured figure at the shaft, which across our records differ by 11.4% on average.

Sources

  1. Traktorenlexikon, the model articles cited on each tractor's page — Wikibooks, CC BY-SA 4.0. Retrieved . Original
  2. OECD standard codes for the official testing of agricultural tractors — OECD. Figures from these test reports are cited per machine on each model page, from the report records in our database; we hold no fetched copy of the OECD programme's web pages.