Japan has more than 33,000 businesses over 100 years old and roughly 3,000 over 200 years old — the highest concentration of ancient family businesses on earth — including a hotel called Nishiyama Onsen Keiunkan that has been run by the same family since the year 705

A hot spring at the foot of the Akaishi Mountains in Yamanashi Prefecture has been feeding the same bathhouse since the year 705. Nishiyama Onsen Keiunkan, the inn built around it, opened during the reign of Emperor Monmu, seven years before the Kojiki, Japan’s oldest surviving chronicle, was completed in 712. It is still taking bookings. In 2011 Guinness World Records certified it as the oldest hotel on earth, a title it holds by a margin measured in centuries.

Keiunkan is the most photographed example of a much larger pattern. A nationwide survey by the corporate research firm Teikoku Databank, published in 2019, identified more than 33,000 Japanese companies that had been trading for a hundred years or more, of which around 3,100 had passed the two-century mark. No other country comes close. When the Bank of Korea studied firms older than 200 years across 41 countries in 2008, it found that more than half of them were Japanese.

What the numbers actually count

Japanese speakers call these firms shinise, literally “old shop”. The label covers sake breweries, confectioners, temple carpenters, ryokan, soy sauce makers and paper mills, most of them small or medium-sized and family-held. That 33,000 figure is worth handling precisely, because it moves. Teikoku Databank and other trackers count centenarian firms year on year, and the total has climbed as more companies founded in the early twentieth century cross the line. By Teikoku’s later tallies the number of hundred-year-old firms had risen past 45,000. The headline claim is not that Japan has exactly 33,000 ancient businesses forever, but that at the time of the 2019 count it had that many, far more than anywhere else, and that the figure keeps growing.

Some of them are genuinely old. Toshio Goto, a professor at the Japan University of Economics who directs the Research Institute of Centennial Management, has helped compile the most commonly cited breakdown: around 140 Japanese firms have passed five hundred years of operation, and roughly twenty claim histories longer than a thousand, though foundings that old cannot be documented and rest on the companies’ own accounts.

The construction firm that ran out of road

For most of the modern era the title of world’s oldest company belonged to Kongō Gumi, an Osaka temple-building firm founded in 578 to work on Shitennō-ji, one of the city’s principal Buddhist temples. It built and repaired temples for fourteen centuries. Then, in 2006, weighed down by debt taken on during Japan’s asset bubble, it was absorbed by the Takamatsu Construction Group and lost its independence. The craft continues under new ownership. The unbroken corporate line did not.

Kongō Gumi’s end is a useful correction to the romance. rLongevity here is not magic. Firms that lasted tended to share a set of habits that read almost as the opposite of a modern growth strategy: stay close to the core trade, keep cash reserves, avoid debt, favour continuity over expansion. Those habits limit how fast a company can grow. They also help it survive the wars, fires, earthquakes and financial shocks that flattened its more ambitious competitors.

Kongō Gumi lasted 1,400 years by building temples and came undone when it borrowed against property.

How a family stays in charge for fifty generations

Behind these long runs sits a quieter mechanism, one that explains how a single family can appear to hold a business for a millennium without ever running short of capable heirs. Japanese firms have long used adult adoption, or mukoyoshi, bringing a promising employee or son-in-law into the family as a legal heir when the bloodline offers no obvious successor. The practice keeps the name, the ownership and the sense of continuity intact even when biology does not cooperate.

Keiunkan itself is the clearest illustration of both the strength and the limit of the system. The inn was run by 52 generations of the same family, adopted heirs included, for more than 1,300 years. That chain ended in 2017. No relative was willing to take over, so the general manager, Kenjiro Kawano, became president despite having no blood or adoptive tie to the owners. Ownership passed to a new holding company and the old entity was dissolved. The hotel is as old as ever and still trading. The family that carried it is not.

What is easy to overstate

The temptation with a story like this is to turn it into a lesson, as though Japan had discovered a formula the rest of the world simply forgot. It is worth resisting. Japan’s concentration of ancient firms reflects a specific history: relative political continuity over long stretches, an island geography that limited conquest, cultural weight placed on inherited trade and reputation, and legal tools such as adult adoption that most other countries lacked. These are conditions, not instructions.

A café in Melbourne cannot adopt its way to the year 3325.

What the shinise do offer is a plain counterexample to the assumption that a company’s natural ambition is to get as large as possible as fast as possible. An average firm on the S&P 500 now lasts under two decades. In the same window, a soy sauce maker in Chiba has been fermenting to a recipe its founders would recognise for close to four hundred years.

Where to look next

The interesting figure to watch is not the record-holder but the trend line. Japan’s shinise count keeps rising even as the country’s population falls and rural towns empty out, which means the succession problem that ended Keiunkan’s family line is spreading. A hundred-year-old firm with no willing heir has three options: adoption, sale, or closure. Which of those the next wave of shinise chooses will say more about the phenomenon than any single anniversary. The record for the oldest hotel is safe for a long time. The habits that produced it are the part actually in question.