Ironclad Chaos: Taiwan’s Rooftop Homes and the Logic of Informality

Taipei from above: not a single building is spared from the rooftop add-ons.

Concrete jungle where dreams were made of—a trope so many cities have been described as, usually with either fascination or disdain. But if you were to hover above any major city in Taiwan and look down, it wouldn’t be a sea of neat, slate-gray rooftops. Instead, you’d see a patchwork of colorful, ad hoc rooftop dwellings, known colloquially as Tie-Pi-Wu (鐵皮屋, or “iron-skinned houses”), crowning buildings across the city like party hats littered on the apartment floor.

My little uncle, always considered the “free spirit” of my maternal family, built one of these add-ons with his own hands in the early 1990s, atop his apartment near the southern edge of Taipei, the capital of Taiwan. For a few years, he rented it out to students attending the nearby National Chengchi University and Taiwan Police College. But after discovering in 2004 that his roof had a clear, unobstructed view of the newly completed Taipei 101, he had a new idea. He converted the part of it with the view into a party room, with its own plumbing and floor tiling. Since then, New Year’s Eve became a standing tradition for him: some years, we’d gather and watch the fireworks from there; in other years, he’d charge people an admission fee of $100 NTD (about $3 USD) for a front row seat to the fireworks on Facebook Marketplace. For a time, it was a beloved family fixture, even if everyone quietly acknowledged its legality was… questionable. Yet more than thirty years later, he’s never been penalized for it.

With more than 600,000 of these units existing across Taiwan (and that’s only the officially documented tally as of 2015), surely they can’t all be monetized for their stunning views. Why do they exist then? Let’s peel through the possible explanations.

At 600,000 strong, there are more illegal rooftop add-ons than buildings with elevators in all of Taiwan.

Not Like the Parisians

19th-century Parisians knew well that beneath the breathtaking mansard roofs were Chambres de Bonne, “maids’ rooms” that were so tiny and took so many flights of stairs to get to that only domestic servants could fathom living there. It’s tempting to imagine that, with all the same inconveniences (and being objectively uglier), rooftop add-ons in Taipei are merely the 21st-century spiritual heirs to these Parisian emblems of inequality.

But the similarity ends on the rooftop. Rather than desperate shelters for the urban poor, Taiwan’s rooftop add-ons often serve as rental units, art studios, enclosed balconies, or semi-legal man caves for the moderately well-off. People like my uncle don’t live in these spaces because they must. They maintain them because they can. While they sure command rents somewhat lower than proper apartments, and wealth inequality undeniably exists in Taiwan, the chasm of spatial divide is clearly not being split here – geographical location, namely the relative distance from the city center, is a much stronger predictor of whether someone in Taiwan is rich or not than if their walls were made of concrete or iron sheets. The Parasite narrative simply doesn’t apply here.

With the economic explanation out of the window, their raison d’être seems even more confusing than before. Is there some unspoken cultural logic – an aesthetic, even – that makes these add-ons feel worth keeping, even when the need has passed?

No Beauty Behind the Madness

Anyone well-versed in Taiwan’s internet culture knows about the self-deprecating slang “中華民國美學”, which translates to “Republic of China aesthetics”, a subtle jab at both Taiwan’s ethno-authoritarian past and its unflinching commitment to utilitarianism. Some telltale signs of this aesthetic include obstructive, uncoordinated billboards; motorcycles parked erratically on the sidewalks; betel nut shacks with scantily clad female hostesses; and, most importantly, the rooftop add-ons that Taiwanese people love to hate. Not exactly the most flattering comparisons.

I snapped this photo at a makeshift lunch market in central Taipei – nearly everything in it captures what locals jokingly call the “Republic of China aesthetics.”

There is no point in pretending people consider them beautiful: compared to Taipei’s spotless MRT stations and pristine sidewalks (seriously, you’ll be hard-pressed to find chewing gum stains), these corrugated rooftop add-ons that seem like they’ll be easily toppled over by a small typhoon are an odd visual glitch in the city’s otherwise neat urban fabric. But most Taipei residents don’t seem to register any contradiction. In fact, if Reddit threads like this one are any indication, locals and expats alike treat these add-ons with a kind of bemused tolerance. Maybe it’s because we’ve all just quietly agreed to stop looking at them. After all, if they don’t obstruct your view or collapse during a typhoon, what’s the harm?

So, we’ve established that rooftop add-ons aren’t beautiful – not in the traditional sense. We’ve also established that they don’t seem to alleviate or even attempt to address high housing costs, especially when cities like Taipei have been losing people since the 1990 population peak. Together, the arguments against their existence seem to outstrip the usefulness that someone like my uncle might vouch for. Let’s investigate whether their legality (or lack thereof) strengthens or weakens the case that they should exist.

The Art of Looking Away

The law is, at least on paper, unambiguous: under Taiwan’s Building Act, any construction without a formal permit is considered illegal – even if it’s on a rooftop you own, and even if it doesn’t seem to harm anyone. One more strike against the add-ons’ existence. But the reality of enforcement is far more stratified, and far less absolute.

The crucial dividing line is time. Structures built after January 1, 1995 are considered “new illegal constructions” and, in theory, should be removed on sight. But if your add-on predates that threshold, even by a hair, it often falls into a softer category: a “列管” case, meaning it’s been documented, photographed, and then largely ignored. This tiered enforcement approach, outlined in Taipei’s own housing bureau documentation and summarized by local urban law firm HB Housing, shows that the bureaucracy quietly acknowledges its own capacity limits and it might be better to leave things be.

As much as the structures remain illegal, I think of this dilemma as a framework instead of a loophole. Entire buildings operate under the assumption that age equals immunity, even though nominally they’re still due to be torn down at some point in the indefinite future. Some landlords even market their units as “列管的” (documented & photographed) on real-estate listings – as though being informally tolerated by the state is actually another amenity. And in a way it is, because – like what my little uncle built – just being old enough becomes its own kind of shield.

Nevertheless, the rooftop add-ons’ immunity from removal still doesn’t fully explain why people choose to create them in the first place, when bureaucratic annoyance and potential demolition risks are still real. Why, then, are they a thing?

Rooftop add-ons are so de-stigmatized that home improvement contractors openly list it as one of their most popular products available.

The real answer might lie deeper: rooftop add-ons survive because formal systems – building permits, zoning laws, planning departments – don’t properly address how people want or need to live. In short, rooftop dwellings exist precisely because something formal planning promises hasn’t materialized.

To explore just how deeply these dwellings have been integrated into urban life, it’s worth imagining the hypothetical: What if they were removed all at once?

If They Were Gone Tomorrow

Let’s say the city wakes up one morning and decides it’s had enough. Soft enforcement? Photo documentation? No more. Bulldozers are sent out. The party’s over.

It might feel satisfying to imagine all those corrugated iron sheets finally gone. The skyline would’ve been restored to a clean, uniform ideal – no more rusted add-ons, no more precarious stairways welded to buildings, just the sleek silhouette of a “civilized” technocratic state, as Taiwan often markets itself globally. What a tempting thought! But even if every rooftop structure vanished in a single sweep, it wouldn’t touch the core misalignment between formal housing systems and how people actually live.

That misalignment becomes painfully obvious the moment you zoom in on the ground. Families who’d been using these spaces as storage rooms, rental units, or de facto second kitchens would suddenly be displaced – not from their homes exactly, but from the extra square meters that had quietly become part of their everyday lives.

A book that discusses how to design and maintain corrugated iron rooftop dwellings? Of course it would become a bestseller in Taiwan.

My uncle would lose his rooftop lounge and New Year’s tradition in one go. Tenants renting out these units – some legally, but the vast majority informally, as investigative reports from platforms like Business Weekly and SETN estimate that over 90% of Taiwan’s rental agreements go unregistered – would be scrambling for alternatives in a housing market that isn’t exactly known for its generosity. And, not to defend them from their privileged position, but landlords who had come to rely on the small but steady rental income might suddenly find themselves pushed into official channels they had spent years avoiding.

More than anything, seeing Taiwanese cities suddenly without these rooftops would feel dissonant. For all their illegality, rooftop dwellings don’t really register as disruptive. I personally pass by them without looking up. They don’t block traffic, they don’t attract protests, and – much to the chagrin of western NIMBYs – they are not detrimental to neighborhood character. Their quietness is part of why they’ve lasted.

And beyond the dissonance, the incentives that created them would still be there, possibly even intensify with a vengeance. In densely packed Taiwanese cities, space is tight, demand is persistent across social classes, and informal logic always finds a way. Unless the formal system offers something better – something as flexible, cheap, well-situated, and immediate – people will want to rebuild them again. Maybe not next week, but probably no later than the next New Year’s Eve. Because if there’s one thing rooftop add-ons reveal, it’s that planning might try to draw the map, but people are always going to redraw it in the margins.

What We Really Built Up There

When my uncle first built his rooftop space, I’m certain he wasn’t thinking about pushing back against a rigid housing system, challenging city officials, or making some ideological statement. It was simpler than that: the rooftop was there, it was his, and he saw no logical reason why he couldn’t do what every other homeowner in Taipei was doing at the time and convert it into something more useful than just a flat, empty slab of concrete. And honestly, isn’t that what most people are really doing when they construct these add-ons? Beneath the surface of tangled regulations and half-hearted enforcement, rooftop add-ons quietly tell a story of people who saw possibility where planning departments saw nothing – people who chose utility and immediacy over formalities and red tape.

My twin nephews genuinely enjoy spending time in my uncle’s rooftop nook – now so well-outfitted that it even boasts a glass (!) roof.

I vividly remember one New Year’s Eve spent on my uncle’s rooftop. The air was unusually brisk, the kind of cold that creeps up once the sun drops behind the buildings, and we had covered ourselves in sweaters and jackets, huddling under a row of makeshift lights he’d strung up years ago. The iron sheets overhead creaked a little whenever the wind picked up, but otherwise the vibe was calm, patched together with familiarity. Someone spilled tea, someone else brought party-sized packs of the Kuai-Kuai snack, and as Taipei 101 began its countdown, we shuffled closer to the railing for a clearer view. When the first round of fireworks lit up the skyline, I remember thinking how absurdly perfect it was: our plastic stools, our unheated rooftop shack, and a billion-dollar skyscraper erupting in light just for us. The experience wasn’t glamorous, or even that comfortable – but it was unmistakably ours. That’s the thing about these rooftops: their value doesn’t lie in the materials or legality, but in how fully they belong to the people who built them.

If these rooftops vanished overnight, you could probably replace the physical space somehow: I could’ve stood at the railings on my uncle’s roof to get the same view anyway. Or even just establish a permitting process that lets people build something better looking and more fortified. But you’d lose something far less tangible, yet equally valuable: a gentle reminder of what urban life truly demands: not uniformity and compliance, but flexibility, creativity, and a willingness to embrace messiness. I don’t think these informal add-ons are trying to break the rules; they’re simply living proof that many of the rules we’ve inherited might be due for a thoughtful update.

The Rooftops Aren’t the Problem

To wrap this all up, we should all abandon the “planner” instincts for a second to understand the true genesis of rooftop add-ons. It’s not defiance for defiance’s sake, and it’s certainly not a glitch in the system. It’s a quiet, persistent vote by actual residents for more flexible, livable cities – places that acknowledge the full range of how people actually use space, not just how they’re told to. The corrugated iron sheets aren’t the problem. They’re just what happens when planning isn’t “human” enough and people start to fill in the gaps.

And maybe that’s the best place to end this. Because whether you see my uncle’s rooftop as an eyesore, a loophole, or a lovely accident, it tells a kind of truth: cities aren’t just made by plans and permits. They’re shaped by people, who will always find a way, especially when the official ones aren’t enough.