Large-scale Urban Interventions

The Case of Faria Lima in Sao Paulo

Large-scale urban redevelopment projects (termed grandes projectos urbanos or GPUs in Spanish) raise many questions about the impacts of subsequent urban development induced by the intervention. GPUs are characterized by an impact in a significant part of the city, often with the use of some new fiscal or regulatory instruments and the involvement of a large network of agents and institutions. These projects are expected to affect land prices, recycle existing or create new infrastructure and facilities, and attract other new buildings.

GPUs as an urban policy instrument have been the object of considerable controversy and debate throughout Latin America. It is often argued that they promote social exclusion and gentrification, have limited effects in stimulating real estate activities, and require large (sometimes hidden) public subsidies that often draw fiscal resources from other urban needs. In spite of their increasing popularity in Latin America, there is little empirical evidence to support these criticisms.

This article presents the case of a GPU introduced in São Paulo, Brazil, in 1996 as an “urban operation” to redevelop a middle-income area of mostly single-family homes that was to be traversed by the extension of the Faria Lima Avenue. The project is known as the Faria Lima Urban Operation Consortium (OUCFL). We examine economic principles that affect the fiscal performance of the project and its opportunity for value capture, evaluate changes in residential density, and analyze changes in income distribution and ownership structure. Finally, we offer some policy suggestions on how and when to use this kind of instrument based on these assessments.

What is an Urban Operation?

An urban operation is a legal instrument that seeks to provide local governments with the power to undertake interventions related to urbanistic and city planning improvements in association with the private sector. It identifies a particular area within the city that has the potential to attract private real estate investments to benefit the city as a whole. The proper city planning indexes (i.e., zoning and other regulations on construction coefficients, rates of occupation, and land uses) are redefined in accordance with a master plan, and investments are made in new or recycled infrastructure.

An urban operation allows the municipality to capture (through negotiated or mandatory means) the land value increments associated with the subsequent land use changes. In contrast to other value capture instruments, these funds are earmarked or internalized within the perimeter of the project to be shared between government and the private sector for both investments in urban infrastructure and subsidies to private real estate investments to support the project itself.

Each urban operation in Brazil is proposed by the executive and approved by the legislative branch of the jurisdiction. In the case of São Paulo, this authority was created in the Lei Organica Municipal (Constitution of the City) in 1990, which was later inserted in the new Brazilian urban development law (Statute of the City of 2001). The first proposed projects were the Operation Anhangabaú (subsequently expanded as a part of the Downtown Operation and renamed Center Operation) and Água Branca, followed by the Água Espraiada and Faria Lima operations. After the approval of the city’s new Master Plan in 2001, nine other urban operations were generated. These thirteen projects are expected to affect 30 to 40 percent of the buildable area of the City of São Paulo.

Financing Faria Lima

The Faria Lima urban operation (OUCFL) was proposed and approved in 1995 with the aim of obtaining private resources to fund the public investments necessary to purchase land and install infrastructure in order to extend Faria Lima Avenue. These costs were deemed at the time to be approximately US$150 million, two-thirds for land acquisitions and one-third for the avenue itself. The project was heavily contested by many stakeholders on grounds ranging from the source of the funds (i.e., advanced out of the local budget through new debt) to neighborhood concerns (one of which managed to keep the floor-area-ratios [FARs] unchanged and legally excluded from the OUCFL zoning) and technical design issues.

Technical studies carried out at the time indicated that it would be possible to take advantage of an additional potential 2,250,000 square meters beyond what was already permitted by the city’s zoning legislation, and the FARs were changed accordingly. These additional building rights were granted against a payment of a minimum of 50 percent of their market value using the existing “Solo-Criado” (Selling of Building Rights) instrument. OUCFL aroused great interest on the part of real estate entrepreneurs. This instrument nevertheless was also questioned for its lack of transparency, its project by project approach, and the arbitrariness in the way relevant prices were established and then used to calculate the value of the additional building rights.

By August 2003 a total of 939,592 square meters, or nearly 42 percent of the available total of these 2,250,000 square meters, had already been licensed. More than 115 real estate projects were approved, including nearly 40 percent commercial buildings and 60 percent high-quality residential buildings. Nevertheless, the resources (approximately US$280 million) obtained from these approved projects had not fully compensated for the expenditures (US$350 million, including principal plus interest) associated with the expansion of the avenue, considering the high interest rates prevailing in Brazil for the nearly eight years since the realization of expenditures. Thus, about 80 percent of the cost (albeit more than anticipated) has been recovered through the Selling of Building Rights process. Since July 2004 the compensation for these advance funds was obtained through an ingenious new value capture mechanism known as CEPAC, an acronym for a Certificate of Additional Potential of Construction. One CEPAC represents one square meter.

The Introduction of CEPACs

Although CEPACs were defined in Brazil’s Statute of the City of 2001, they were not approved by the CVM (Brazilian equivalent to the U.S. Security and Exchange Commission) as freely tradable in the Brazilian Stock Exchange until December 2003. The regulation establishes that the price of each certificate is defined by public auction and that the corresponding square meters of building rights (which also include use changes and occupation rates) expressed in each certificate may be executed at any time. The regulation also states that new batches of certificates can be issued (and sold through auction) only upon confirmation that the resources captured by the previous sale have been effectively earmarked to the project. To ensure this designated use, the revenues are deposited in a special account, not in the municipal treasury. From the perspective of the private investors this designation ensures the acceptability of this value capture instrument at its own valorization. By issuing a lower number of certificates than potential building rights—that is by managing their scarcity—the public sector may benefit from the valorization and thus be able to capture value “ex-ante” (Afonso 2004, 39).

The final approval of CEPACs for OUCFL and all the necessary steps for launching them in the financial market occurred in mid-2004, and the first auction at the end of December 2004 generated nearly R$10 million (about US$4 million), corresponding to the sale of approximately 9,000 CEPACs out of an authorized stock of 650,000 square meters. The OUCFL certificates were sold at a face value of R$1,100 (about US$450) per square meter with no observed premium pricing as a result of the bidding process.

This situation contrasts with that of the Água Espraiada urban operation, which was expected to be fully funded by CEPACs from its start. In its third auction, the certificates were already capturing R$370 per certificate against a face value of R$300 set for this operation. A more recent auction in Água Espraiada sold 56,000 CEPACs and captured R$21 million ($US9.5 million), reflecting a certificate price of R$371. This pricing contrast reflects the different original face values in the two projects. In the case of OUCFL developers bought (and stocked) building rights in advance, to benefit from the more flexible rules prior to the CVM approvals. The certificate price in Faria Lima started at more than R$1,100 because it is a more valued area. In Água Espraiada developers were willing to pay more than the original face value because the certificates were less expensive and thus in greater demand.

Land Price Implications

The prices of vacant land and developed areas experienced a considerable increase in some blocks within the perimeter of OUCFL during the 1990s, but decreased in other blocks. Yet, the average square meter price of new real estate development fell throughout the Metropolitan Region of São Paulo (RMSP) in all price bands, when comparing the average price from 1991 to 1996 with those of 1996 to 2000.

After controlling for a number of attributes associated with the changing character of the developments and their location, the price estimations showed an unequivocal relative increase after the operation was launched. The average price per square meter within the OUCFL perimeter increased from R$1.68 thousand in the 1991–1996 period to R$1.92 thousand in the 1996–2001 period, a 14 percent increase, while prices in RMSP decreased from R$1.21 thousand to R$1.06 thousand, a 12 percent decrease in the same period (R$1.95/US$1.00 in December 2000). Thus, the price per square meter in OUCFL was higher than that of RMSP by around 26 percent. The price per square meter in OUCFL was 38 percent higher than the average price in the RMSP in 1991–1996, and it increased to 81 percent higher in 1996–2001.

Was this increase captured by the municipality as anticipated? Considering that the cost of construction in average is around R$1,000 per square meter, the 2004 auction (the only one so far) captured almost all of the value added at current prices. The previous pre-CEPAC system captured about 50 percent or more, depending on the capacity and success of municipal negotiators, and the correctness of the reference price. CEPAC now changes this percentage and the face value of the instrument may capture all the value increment or even more, depending on the relation of this face value to market prices, and on the results of future auctions. Comparing a redevelopment project financed totally by construction bonds (like CEPACs) and one financed totally with general property taxes, there is no doubt that the former is less regressive than the latter. Even with a progressive property tax, with rates increasing according to values, part of the costs would be paid by poorer households.

This evidence that about 80 percent of the total cost of the project has already been recovered, combined with the auctioning of the remaining building rights through CEPACs and the impact of the property appreciation on the current property tax revenues, indicates that the project should not only pay its own way but actually generate a fiscal surplus for the city as a whole over the next five or seven years.

In effect, the changes caused by substituting older single-family houses with new residential and commercial buildings resulted in a substantial change in property tax collection in the OUCFL area. Many lots and even entire blocks had been occupied by single- and two-story houses constructed since the 1950s. Many of these structures were eligible for a discount coefficient for obsolescence of up to 30 percent of the property tax. They were replaced with new, taller and higher-quality buildings for which the discount was null. Our estimates indicate that the differences in property tax collection by square meters constructed may have increased by at least 2.7 times and up to 4.4 times. That is, the average property tax per square meter increased to a minimum of R$588.50 up to R$802.50 from R$220.95 if the house was 25 years old, or from R$179.70 if the house was 30 years old.

Social Implications

The OUCFL case offers a unique opportunity to quantify changes in resident characteristics before and after the intervention, since data at the census track level is available for 1991 and 2000, and the intervention began in 1996. Our analysis of gentrification and displacement of poorer residents mainly confirms the findings of Ramalho and Meyer (2004) that the average income has increased relatively in most of the blocks inside the OUCFL perimeter. By Brazilian standards, the upper-middle class was displaced from the region by the richest 5 percent of households in the metropolitan area. The census data also showed that residential density fell between 1991 and 2000, from 27 to 22 residences per hectare, although these figures may be distorted because they reflect the ratio of total residences in the entire area, not an average of the ratios per plot where land use was converted.

The data from 1991 indicated that the population was already leaving the OUCFL area before the approval of the urban operation, but this exodus intensified after 1996, generating vacant plots in the process of site-assembly to accommodate the new high-rise developments. At the same time, building density increased. The average number of floors per new building in the area increased from 12.6 in the 1985–1995 period to 16.7 in the 1996–2001 period. The number of housing units per building increased from 37.1 to 79.6 over the same periods.

This apparent contradiction between decreased residential density and increased numbers of housing units is explained in part by the construction of commercial buildings that replaced many single-family residencies on small and average-sized lots. OUCFL induced considerable real estate concentration as the new commercial and residential buildings replaced the houses and required greater land areas for high-class architectural projects. The 115 projects approved between 1995 and August 2003 that requested increases in the utilization coefficients required a total of 657 lots, or an average of 5.7 lots per project.

The combination of the increase in income level and the reduction in household density indicates that the gentrification process advanced in and around the OUCFL region during the 1990s. Nevertheless, this is not a classic case of gentrification, where poor families are driven out of an area due to various socioeconomic pressures. In this case mostly upper-middle classes were displaced. Except for the small nucleus of remaining favelados (Favela Coliseu), the region was already occupied by people belonging to the richest segments of society.

Some Policy Observations

This article contributes to the debate about the social management of land valuation by furnishing real data assessments and economic elements. These elements have been missing from most analysis, and we believe that this gap in the literature has contributed to an incomplete interpretation of the implications of an urban operation and to mistaken public policy recommendations.

Our conclusion is that the CEPAC funding mechanism itself does not increase the regressive characteristic of urban operations, since without those building rights bonds all the investment in redevelopment would be financed by general taxes. If the OUCFL project were inadequate in terms of income distribution, it would have been even worse without the value capture mechanism. Instead, CEPAC and the value capture mechanism used previously offered two desirable characteristics of any public investment: charging the new landowners is at least neutral in terms of income distribution; and the primary beneficiaries end up paying for the project.

Furthermore, the urban operation mechanism offers incentives for redevelopment. Given that most projects increase land prices and drive out the poor from the region, it would be better to invest the entire municipal budget in small-scale projects. This is the opposite of what happened with the redevelopment of the adjacent high-end Berrini area where developers decided how to concentrate their investment, resulting in even more income concentration than in the OUCFL area. Because of inaction by policy makers in that case, the municipality did not capture any value from Berrini, yet paid the entire cost of infrastructure.

The use of building rights bonds may diminish the regressive aspect of land development, but to make a project truly progressive requires attention on the expense side, by funding all the investment through instruments like CEPACs. The main limitation on distributing benefits to the poor is that the law establishes that all funds collected through value capture (CEPACs or other instruments) must be invested within the perimeter of the intervention. One way to make these interventions more progressive is to invest in activities that will furnish spillovers to the poor, such as public transit, education, and health. Moreover the relevant legislation allows the administration to select an area inside the perimeter of an urban operation and declare it a zone of special social interest (ZEIS) where lots can be used only for low-income social housing.

Another alternative is to establish social housing areas within the perimeter of the urban operation. By subsidizing low-income housing with money from developers and new landowners, there would be no distortion in prices outside of the housing industry. The subsidy results from segmenting the market and transferring the extra rent to poor households. This is real social management of land valuation.

Ciro Biderman is affiliated with the Center for Studies of Politics and Economics of the Public Sector (Cepesp) at the Economic and Business School at the Getúlio Vargas Foundation in São Paulo, Brazil. He is a visiting fellow in international development and regional planning in the Department of Urban Studies and Planning at Massachusetts Institute of Technology, Cambridge.

Paulo Sandroni is an economist and professor at the Economic and Business School at the Getúlio Vargas Foundation.

Martim O. Smolka is senior fellow and director of the Lincoln Institute’s Program on Latin America and the Caribbean.


(These publications are available only in Portuguese.)

Afonso, Luis Carlos Fernandes. 2004. Financiamento eh desafio para governantes (Financing is a challenge to government). Teoria ane Debate No. 58, Maio-Junho: 36–39.

Ramalho, T., e R.M.P. Meyer. 2004. O impacto da Operação Urbana Faria Lima no uso residencial: Dinâmicas de transformação (The impact of the Faria Lima Urban Operation on residential use: Transformation dynamics). Mimeo. São Paulo: Lume/FAUUSP.

Biderman, Ciro, e Paulo Sandroni. 2005. Avaliação do impacto das grandes intervenções urbanas nos precos dos imoveis do entorno: O caso da Operação Urbana Consorciada Faria Lima (Evaluation of property price impacts near large-scale urban interventions: The case of Faria Lima Urban Operation Consortium). Lincoln Institute of Land Policy Research Report (April).

Economics, Housing, Infrastructure, Land Market Monitoring, Land Value, Property Taxation, Public Policy, Urban Development, Value Capture
Back to top