Taiwan's temple count, checked against the convenience store claim

"Taiwan has more temples than convenience stores" circulates with no year attached and no statement of what a temple is, which is how a sentence can be both true and false at once. On the government's own registered-temple figure it is false by 1,839; add the government's registered churches and it becomes true by 1,016.

Both numbers below come from a single official table, 內政統計年報 06-01 各宗教教務概況, published by the Department of Statistics at the Ministry of the Interior and last updated 30 June 2026 (內政部統計處, 統計表 331030, checked 7 September 2026). This article does not re-derive convenience store density, which is covered separately; it cites one figure from that side of the comparison and moves on.

What counts as a temple?

A temple enters Taiwan's national statistics only by completing a registration process with its local city or county government, so the number is a count of paperwork rather than a count of buildings. The scope note attached to the statistic defines a temple as "任何有僧、道、住持之宗教建築物不論用何種名稱" — any religious building with resident clergy, whatever it calls itself — and states the legal basis as the 監督寺廟條例 (Temple Supervision Ordinance) together with the 寺廟登記規則 (Temple Registration Rules), compiled as of 31 December each year (彰化縣政府主計處, 統計資料背景說明, checked 7 September 2026).

That definition has a constitutional problem the courts have already ruled on. In Interpretation No. 573, announced on 27 February 2004, the Judicial Yuan held that the ordinance's definition — religious buildings with Buddhist or Taoist clergy — applied to some religions and not others, and struck down Article 8 and part of Article 2 for breaching religious equality, with those provisions ceasing to have effect within two years (司法院憲法法庭, 釋字第573號, checked 7 September 2026). The statistic still rests on the surviving remainder of an ordinance whose core definition was found to be discriminatory.

The practical bar is set locally, and it is a building bar rather than a religious one. Taipei City's registration rules require that the property be a public-use building under Article 5 of the Building Act, that "整幢供宗教使用" — the entire structure be used for religion, with an occupancy permit naming a temple as its principal use — and that deity images actually be enshrined and worshipped there (臺北市辦理寺廟登記須知, amended 30 August 2019, checked 7 September 2026). A shrine occupying the corner of a shophouse fails the second condition before anyone asks about the deity.

How many are there, by definition?

The answer depends entirely on which line of the ministry's table you read, so here are all of them with the date each publisher attached. The convenience store row is included only as the benchmark the claim invokes.

Set against the registered population of Taiwan at the end of December 2025, 23,299,132 people (內政部戶政司, 民國114年12月戶口統計資料分析, published 9 January 2026, checked 7 September 2026), the registered temple figure works out at one temple per 1,879 people, or 5.32 per 10,000. All registered places of worship together come to one per 1,528 people. The temple count and the population figure are both dated to the end of 2025, so no dates are being mixed here; the convenience store figure is four months earlier and is the one loose end in the comparison.

The ministry's own bulletins used to make the comparison in per-capita terms, and the ratio has barely moved in more than a decade. Its 2012 bulletin on the 2011 figures reported 6.59 temples and churches per 10,000 people, with Lienchiang County the densest at 71.24 per 10,000 (內政部統計處, 內政統計通報 101年第28週, checked 7 September 2026). The same calculation on the 2025 figures gives 6.55 per 10,000 — essentially unchanged, because the number of registered premises and the population have both been flat. What changed over those fourteen years is the convenience store total, which is why a comparison that may once have gone one way now goes the other.

Which religions are registered?

The ministry breaks the total down by religion, and the distribution is far more concentrated than the number of listed religions suggests. Twenty-three religions appear in the 2025 table; two of them account for 97.6 percent of all registered temples.

Shares are calculated from the ministry's own counts against the 12,397 total; the ministry publishes the counts, not the percentages. On the church side of the same table, Protestantism accounts for 2,373 of 2,855 registered churches and Catholicism for 429, with Tenrikyo (17), Islam (5), Mormonism (5), the Bahá'í Faith (2), Judaism (1) and Scientology (1) making up almost all of the remainder.

Two entries in that table are worth noticing because they are zeros. 天帝教 (the Tienti Teachings), 先天救教, 東正教 (Orthodox Christianity), 真光教團 and 統一教 each record zero registered premises at the end of 2025. A zero here does not mean the religion has no presence in Taiwan; it means no premises of that religion completed temple or church registration, which is a statement about the registry and not about the country.

Which counties have the most?

Tainan does have the most registered temples in Taiwan, and the ministry's figures confirm it rather than folklore doing so. The pattern across the country is that temples concentrate in the agricultural south and west, while churches concentrate in Taipei and along the indigenous east coast.

Taipei City is in that table as the exception rather than as a leader. It holds 280 registered temples against 564 registered churches, the largest church count of any city or county in Taiwan and one of only three jurisdictions where churches outnumber temples — the others being Taitung County (222 temples, 282 churches) and Hualien County (194 and 260), both with large indigenous populations. If your impression of Taiwan is formed in Taipei, you are standing in the least temple-dense major jurisdiction in the country.

Tainan's lead is durable rather than recent. The ministry's 2014 bulletin on the 2013 figures already put Tainan first with 1,629 temples, Kaohsiung second with 1,478 and Pingtung third with 1,096, the three together holding about 35 percent of the national total (內政部統計處, 內政統計通報 103年第33週, published 16 August 2014, checked 7 September 2026). Twelve years later the same three counties sit in the same order.

Why did the count jump in 2002?

The registered temple series contains one break so large that any year-to-year comparison across it is meaningless, and it is an administrative event rather than a building boom. Temples went from 9,832 at the end of 2001 to 11,423 at the end of 2002, a rise of 1,591 in a single year, against typical annual movements of a few dozen (內政部統計處, 年報 06-01 歷年 1997–2015, checked 7 September 2026).

What happened is documented in the rules. The Ministry of the Interior issued the 未辦理登記寺廟補辦登記作業要點 — operating guidelines for the retroactive registration of unregistered temples — on 28 March 2001, amended it that September, and opened a one-year application window running from 1 April 2001 to 31 March 2002 (內政部, 未辦理登記寺廟補辦登記作業要點, checked 7 September 2026). Temples that could show the building was safe, did not obstruct public facilities or disturb neighbours, and could produce a land transcript plus either an occupancy permit or a landowner's consent, were allowed in without the full paperwork. Roughly 1,600 existing temples became countable in twelve months.

That is why the statistic distinguishes 正式登記 (standard registration) from 補辦登記 (retroactive registration) in its county-level tables. The distinction exists precisely because a large share of Taiwan's registered temples were legalised after the fact, and because plenty more never made the window.

Since then the series has been almost flat. Registered temples moved from 12,207 at the end of 2016 to 12,397 at the end of 2025 — 190 more in nine years, a rise of 1.6 percent — while registered churches fell from 3,044 to 2,855 over the same period, a decline of 189. Taiwan is not building temples at any noticeable rate. It is holding steady, and losing churches.

What is not counted here?

Several things are absent from every figure above, and naming them precisely matters more than the totals do. The most important is the category a visitor encounters constantly: the small unregistered shrine — a 土地公廟 the size of a wardrobe at a lane entrance, an altar occupying the front room of a house, a shrine under a flyover. None of these can satisfy the whole-building occupancy requirement, so none of them can register, so none of them appear in the 12,397. No government body was found publishing a count of them, and this article will not estimate one.

The follower figure also carries a change that breaks it as a time series. Since 2014 the reported follower number has covered temple followers only and excludes church congregations entirely, which is why the church columns for clergy and congregants read zero from that year onward — the yearbook says so in its own notes. The 932,735 figure is compiled from what temples report to their local governments under each religion's own membership rules; it is not a survey of belief and cannot be read as one.

Two things could not be verified for this article. The 全國宗教資訊網 registry at religion.moi.gov.tw, which holds the individual temple records behind the aggregate — the fields include 寺廟名稱, 主祀神祇, 行政區, 地址, 教別, 登記別 and 統一編號 (政府資料開放平臺, 資料集 8203, metadata last modified 23 January 2026, checked 7 September 2026) — was unreachable on every attempt on 7 September 2026, so the number of individual records in the registry could not be compared with the 12,397 aggregate, and the separate counts for registered religious foundations (宗教財團法人) and incorporated religious bodies could not be read at all. Any figure for those categories should be treated as unverified until the site responds.

How do you tell them apart?

A registered temple and an unregistered shrine look different in ways you can read from the street, and the difference decides what you will find inside. The registered ones are the buildings the statistics are about; the unregistered ones are most of what you actually pass in a day in Taipei.

A registered temple occupies its own building end to end, because that is the legal condition for registration. It will usually have a formal name board over the entrance using 宮, 寺, 廟 or 堂, a management committee whose members are listed on a noticeboard inside, published donation and ritual fee schedules, and often a tax number displayed on receipts — the registry records a 統一編號 for each one. Practically, it will have a main hall you can walk into, side halls, a courtyard or a censer, and staff or volunteers present during the day.

An unregistered shrine is defined by what it lacks: no separate building, no committee noticeboard, no receipts. A 土地公廟 (earth god shrine) is often a metre-high masonry box at the head of a lane, sometimes with a red cloth and an electric lamp, sometimes with fruit left that morning. A shopfront altar sits behind a roller shutter in a residential ground floor and may only be open on the first and fifteenth days of the lunar month. Neither is lesser, and neither is countable.

For planning purposes that means two different things. If you want a temple with an address, an opening pattern and enough activity to be worth a detour, you are looking for a registered one, and Taipei has 280 of them across the whole city — a smaller pool than most visitors assume, and one reason Tainan makes a better temple day trip than Taipei does. If you want the texture of daily religious practice, you want the uncounted ones, and the way to find them is to walk lanes rather than main roads: they cluster at lane entrances, beside old banyan trees, at the ends of alleys, and near markets.

One habit worth adopting: check the main deity before you go. The registry field 主祀神祇 exists because a temple's principal deity determines almost everything about it — a Mazu temple, a Guandi temple and a Baosheng Dadi temple differ in festival calendar, ritual, iconography and crowd. Two temples five minutes apart can be completely different experiences, and the name over the door usually tells you which is which.

What is outside this article's scope?

This article does not cover individual temple opening hours, festival dates, etiquette for offerings or divination blocks, or any temple's history, and it does not rank temples for visiting. It also does not re-derive convenience store density, which is treated separately; the single store figure here is cited only as the benchmark the temple claim is measured against.