Lin Chien-cheng's Perspective: How Much Land and Funding Are Needed to Build 130,000 Social Housing Units?

Prior to the 2024 presidential election, Lai Qingde pledged to build 130,000 social housing units during his term. However, after the policy was launched, the actual target was downgraded to 30,000 units, drawing strong criticism from social welfare groups and housing advocates who accuse the government of breaking campaign promises and making abrupt policy reversals. This article does not assess political responsibility but focuses on a frequently overlooked yet crucial technical issue: how much land and funding are actually required to build 130,000 social housing units? It also compares Taiwan’s approach with Singapore’s Housing & Development Board (HDB) system—globally recognized for its high housing supply efficiency—to examine differences in land use intensity.

Land Requirements for Social Housing

In terms of site development intensity, floor area ratio (FAR) and site coverage ratio are two key indicators that determine “how many homes can be built on a given plot of land,” forming the basis for subsequent land demand calculations.

FAR refers to the total floor area divided by the site area. A higher FAR allows for a larger total floor area on the same plot.

Site coverage ratio refers to the building’s footprint area divided by the site area. A lower site coverage ratio means buildings are set back, leaving more space for open areas and greenery.

In practice, common facilities (e.g., stairwells, mechanical equipment rooms) are typically excluded from statutory FAR, and floor area bonuses for urban renewal or social housing projects often result in “permitted FAR” exceeding “base FAR.” This article uses a common empirical multiplier of 1.6 for estimation. This gap, combined with whether public facility land is included and which land acquisition mechanism is adopted, is precisely the key lever determining whether 130,000 units require 270 hectares or nearly 1,350 hectares of land.

The formula for estimating land needed for 130,000 units is: Required site area = (Total units × Average floor area per unit) ÷ Permitted FAR. The following calculations uniformly assume a base FAR of 300% and a permitted FAR of 1.6 times (i.e., 480%).

Assuming 30 ping per unit (including common areas)—corresponding to a mix of slightly larger two-bedroom or three-bedroom units—the total floor area for 130,000 units is approximately 3.9 million ping, requiring about 270 hectares of land. This is equivalent to roughly 10.3 Daan Forest Parks (each about 26 hectares).

These 270 hectares represent only the building site area and do not include public facility land. In practice, for non-urban land development units, public facilities and buffer green belts account for about 50% of the total area; for newly established or expanded urban plans, public facility land accounts for about 40%.

Rough Estimate of Financial Costs for 130,000 Units

With a total floor area of 3.9 million ping, assuming a construction cost of NT$200,000 per ping, direct construction expenses amount to approximately NT$780 billion. Adding 10% indirect costs brings the total to about NT$850 billion. This is a baseline figure common across all scenarios and does not vary with the land acquisition mechanism.

Taiwan’s social housing initiatives have struggled for over a decade, primarily because land acquisition has consistently lacked a systematic solution comparable to Singapore’s HDB model. Therefore, the following estimates assume that housing communities can emulate the holistic planning principles of Singapore’s HDB, adopting large-scale integrated development rather than piecemeal, infill-style construction on small plots.

Two common integrated development approaches in Taiwan lead to different cost outcomes. The first is non-urban land development unit development, involving land expropriation outside urban planning zones, including land compensation, with development costs estimated at NT$120 million per hectare. In integrated development, about 50% of the area is allocated to public facilities and buffer green belts, so the total planned area is approximately double the building land area—i.e., 540 hectares. Land costs would be about NT$65 billion, and with construction costs, the total expenditure would be approximately NT$915 billion.

The second approach is land readjustment under newly established or expanded urban plans, where the government receives a portion of surplus land as building plots. If the surplus land ratio is 20% of the total planned area, the total land readjustment area can be back-calculated to about 1,350 hectares. Development costs are similarly estimated at NT$100 million per hectare, with the government fully funding the integrated development. Land development costs would be about NT$135 billion, and with construction costs, the total expenditure would be approximately NT$985 billion.

Combining both scenarios, the total cost for 130,000 social housing units is roughly estimated between NT$915 billion and NT$985 billion. Construction costs consistently account for over 80% of total expenditure, making it the primary determinant of fiscal scale. While the choice of land acquisition mechanism results in a difference of several hundred billion dollars, it remains a secondary variable in comparison. If budgeted over 10 years, the annual fiscal burden would be approximately NT$91 billion to NT$98 billion.

Adopting an integrated development model and coordinating surrounding land planning allows for the provision of complete public facilities such as roads, parks, schools, and parking lots, enhancing living environment quality and creating communities with full life functions.

The real challenge in land acquisition is not “total quantity” but “fragmented acquisition”: the need to divide into dozens of sites scattered across densely populated urban fringes in counties and cities nationwide, each undergoing urban planning adjustments, land expropriation, and public communication procedures—often time-consuming and protracted.

Historical Experience and Social Costs

The choice of land acquisition mechanism must also consider social costs from historical experience, beyond financial differences. In the 1970s, the government frequently used direct expropriation of private land to build large-scale public housing, such as Keelung’s Yongle Public Housing and Taipei’s Wanfang Public Housing. Discrepancies between compensation and landowners’ perceived rights led to intense protests. Since then, policy has shifted toward land readjustment, allowing original landowners to receive a proportionate share of land in return and share development gains, significantly increasing public acceptance. However, some landowners remain skeptical about compensation methods and the allocation ratio of return land, and individual protests occasionally occur.

This history reminds us that the 270–1,350 hectares of land required for 130,000 social housing units—regardless of the financially favorable acquisition mechanism chosen—involve not only engineering and budgetary issues but also landowners’ rights and social coordination costs across dozens of dispersed sites. These costs are difficult to quantify in financial estimates but are often the key variables affecting actual progress.

There is another more practical path: within planned areas where land readjustment has already been completed, the government, which receives surplus land by ratio, originally planned to auction it off to subsidize development costs. Instead, if this land is transferred at cost to be used for social housing construction, it enables direct development on already established, uncontested public land—avoiding a new round of expropriation procedures, protests, or disputes over return land allocation. The social costs for such land were already borne in the past; using it now for social housing may be the least controversial and fastest implementation option, worthy of being prioritized.

Another highly feasible path is transforming Taiwan Sugar from a mere land resource provider into a long-term operating “housing corporation,” establishing institutionalized cooperation with the government to continuously invest its own land in large-scale social housing construction. Taiwan Sugar owns vast land reserves widely distributed across Taiwan with simple ownership structures. If it moves beyond case-by-case collaborations like Taipei’s Huajiang and Huaqiang Public Housing or Tainan’s Dalin Public Housing (where Taiwan Sugar provides land and local governments build public housing), and instead establishes a permanent, specialized housing development corporation, it could gradually accumulate experience similar to Singapore’s HDB. This would transform land acquisition from one-off negotiations into a regular supply source—potentially the most significant structural change enabling the 130,000-unit goal to be achieved within a reasonable timeframe.

Comparison with Singapore’s HDB Model

Singapore is globally recognized as one of the most successful examples of large-scale, high-efficiency public housing supply: over 80% of the population lives in government-built HDB flats. Compared with Taiwan’s social housing challenges, several key differences emerge.

First, in terms of unit size, Singapore does not succeed by “building smaller.” The main unit types (4-room and 5-room flats) are approximately 27–34 ping—sometimes even larger. In fact, Singapore does not reduce land demand by compressing unit size but wins through integrated land development.

Second, in overall floor area ratio intensity, Taiwan is not inferior to Singapore. Typical HDB new towns have FARs of about 210–280%, rising to over 350% in prime MRT-adjacent areas. Taiwan’s typical social housing sites have base FARs of 200–300%, with actual permitted FARs reaching 320–480%, which in most cases is comparable to or even exceeds Singapore’s general levels.

Third, both regions follow a “high FAR, low site coverage” design approach. Singapore’s HDB flats commonly use high-rise slab or tower designs with wide building spacing and large open ground floors, resulting in low site coverage per plot—similar to Taiwan’s context. There is no fundamental difference in architectural design logic; the real gap lies in the next point.

Fourth, the key difference lies in land reserve and integrated planning mechanisms. Since the 1960s, with the establishment of HDB, Singapore has long pursued state-led large-scale land expropriation, reserve, and integrated planning, enabling massive HDB construction through new town development. In contrast, Taiwan’s social housing projects are mostly “infill” developments on scattered, fragmented plots within existing urban plans across counties and cities, lacking institutional support for long-term land reserves and integrated planning. As a result, small-site developments must bear externalized social costs and face greater constraints from local resistance due to site-specific FAR, building spacing, and NIMBY effects.

Singapore’s experience shows that through long-term, integrated land reserve and urban planning mechanisms, even with unit sizes similar to Taiwan’s and FARs not particularly higher, it is still possible to achieve a public housing supply scale and efficiency far exceeding Taiwan’s. The implication for Taiwan’s social housing policy is not to debate whether unit sizes should be further compressed, but rather to re-examine land reserves, integrated development mechanisms, and long-term fiscal planning.

Conclusion

Based on assumptions such as 30 ping per unit (including common areas) and a permitted FAR of 480%, the land required for the pure building sites of 130,000 social housing units is approximately 270 hectares. If planned units or land readjustment