Rent gap formation due to public infrastructure and planning policies: An analysis of Greater Santiago,Chile,2008

Abstract

Land upzoning and state investment in public infrastructure are two of the principal factors that increase the rent gap in the city; however, the scale of their impact remains unknown. This paper presents a novel method of rent gap analysis based on multiple linear regressions with controlled fixed factors, tested for Greater Santiago, Chile. Drawing on an administrative dataset of 36,911 transactions for new apartments sold between 2008 and 2011, along with data regarding the size of each apartment and its commercial sale price—but discounting imputed land and construction costs—it can be seen that state investment in the Metro rapid transit network enlarged the potential ground rent (portion of the rent gap capitalized by developers) by 25.6% over the period. Similarly, each additional floor area ratio point (dictated by district zoning guidelines) increased capitalized ground rent by an average of 6.1%. Meanwhile, the portion capitalized by small-scale private landowners through sale of un-redeveloped land increased by only 5.5% over the same period. These results show the high elasticity of the rent gap, the strong influence of land regulations and state-financed Metro infrastructure on rent gap formation, and the need for discussion of taxation of the capitalized ground rent for redistributive purposes.

Keywords

Rent gap private developers metro upzoning

Introduction

Rent gap studies have been receiving increased attention over the past decade. According to its original definition (Smith, 1979), the rent gap is the difference between a capitalized ground rent (CGR)—the present value captured by owners of existing decayed properties that have not been redeveloped—and a potential ground rent (PGR)—the highest rent value accrued by the land once it has been redeveloped.—This is appropriable by the agent that owns that land and has the technical and financial capacity to exploit it, and thus obtain the maximum landed-profit possible. Since its inception, the rent gap theory has been seen as a “supply-side” causal explanation for gentrification, although more recently, it has come to be seen as a generalized form of land dispossession occurring in different settings around the world (Slater, 2015). For this paper, the CGR is referred to as CGR1 (the smaller portion of the rent gap obtained initially by private landowners upon sale of their land to high-rise developers), and PGR is referred to as CGR2 (the larger portion of the rent gap captured subsequently by the developer). This nomenclature draws on López-Morales (2015) and is explained below.

The goal of the present paper is to estimate the effect of land upzoning and state investment in public infrastructure on the enlargement of the CGR2 and thus the formation of the rent gap. This will be done using administrative, repeated cross-sectional data for the Metropolitan Area of Greater Santiago (MAGS) in Chile, and analysis based on multiple linear regressions with controlled fixed factors. The method used is explained in the “Empirical strategy” section below. Results show that for new apartments sold within a radius of 1 km around Metro stations, there is an increase of 25.6% in the CGR2 (the portion captured by the developer) for every square meter, while an additional increase of 6.1% is generated by each additional point of the floor area ratio (FAR) permitted by municipal-level land use and construction guidelines.¹ Meanwhile, the CGR1 (the portion captured by the land owner upon sale to the developer) increases by only 5.5% when the apartments are located within a 1 km radius of a Metro station. This result contradicts the common belief that Metro infrastructure generates significant benefits in terms of land value appreciation for small-scale private land owners (specifically, owners of houses), and is particularly interesting given that land property is widely distributed in Santiago, with around 62% currently owned by residents of the MAGS (MIDESOC, 2015).

There have been only a small number of empirically based and theoretically framed studies of the rent gap to date. The present article attempts to address this problem, and indeed to go beyond the scope of the existing literature in three ways. Firstly, and most importantly, the paper states precisely the extent to which public policies, such as public infrastructure investment, contribute to the creation of CGR2 (i.e. the potential ground rent), and therefore to what degree the rent gap is absorbed by private developers. The paper also substantiates the thesis that the rent gap is formed subsequent to the zoning-induced enlargement of this potential ground rent in intense, rapidly changing and highly liberalized urban economies. This is by no means a trivial issue, and connections between state agency and accumulation of landed-profit by private economic elites are becoming increasingly visible. Furthermore, the state is in effect supporting the generation of large amounts of landed profit, and some form of economic return is appropriate. The issue of land value taxation and other forms of state levies are discussed in the conclusions section. Secondly, although this is not an analysis of gentrification per se (important variables of economic, social or cultural capital of old and new residents are not observed here), the results are extremely relevant to the subject. Thirdly, since the data covers the total time span required by a developer to construct and sell its project, the results rule out the sample selection bias usually present in this type of study.

The next section offers a review of key literature on land rent and rent gap analysis, describes the case of Santiago, and presents the data and methodology used in the study. The final three sections present results, a discussion, and conclusions.

Rent gap: Theory and analytical approaches

As Harvey (1974, 1989) explains, urban land rent is neither a naturally produced nor naturally scarce resource: rather, it is human-produced and made artificially scarce. Land rent is created by a particular social group—often acting as a “class”—which aims to increase its value without making any specific effort to bring this about, instead benefiting from land value appreciation driven by existing positive external conditions. Urban land is a space shaped by power relations, such as planning controls, provision of infrastructure, subsidies, or even political corruption, which allow developers to establish reasonable expectations over the long term, as urban ground rent is dictated by their interests in order to reduce the uncertainty of land-use competition. State institutions and regulations seek to maintain the integrity of this organizational structure, while at the same time encouraging growth and accumulation, avoiding cyclical crises, and mitigating social discontent. Far from having a trickle-down effect, ground rent works from the bottom upwards: it transfers income from land from low or middle income groups to the dominant bourgeoisie, transforming differential rent into monopoly rent, which ultimately ends up in financial spheres. However, this paper deals specifically with policy-led augmentation of the ground rent available for capture by the dominant class by means of maximized high-rise redevelopment. Lefebvre (2003) envisaged that, in the modern era, ground rent would be extracted from rented land or buildings rather than from productive processes, within a complex set of urban-related spaces, activities, and industries. This is what Schafran et al. (2018) refer to as the “urbanization sector.”

Largely based on systematic observations of urban ground rent variations, Neil Smith (1979) provided an explanatory model for both inner city decline and “regeneration,” a definition also highly congruent with the theses of inner city decay due to de-industrialization and subsequent inner city redevelopment (Hall, 1981; López-Morales, 2009). According to Smith’s hypothesis, it was the monopolized urban land rent which produced—in conjunction with the fixity of the capital invested in land (the initial construction and its successive improvements) and its long turnover period—the cycles of devaluation and decline experienced in the inner city. A combination of disinvestment by investors in the inner city, due to its high risk and low rates of return, triggered a long period of deterioration and lack of new capital in these areas (Smith, 1979). After a time, a rent gap appears. This gap represents the disparity between the “capitalized ground rent” (CGR, rent attracted by a piece of land named as CGR1 in the analysis below), devalued by the current dilapidated use of land, and a “potential ground rent” (PGR or CGR2 in the analysis below), boosted by improvements to the surrounding area, including changes to regulations. Thus, PGR (CGR2 as used below) implies the “highest and best use” of land, or at least higher and better use given the central location of the inner city space (Smith, 1996a).

“Realization” of the rent gap is achieved exclusively through development—involving a high degree of fixed capital investment—designed to accommodate this potential use, although always within the limits permitted by zoning laws and construction codes. Therefore, actual ground rent will come to equal potential ground rent as the full resources of the site are redeveloped. For this reason, López-Morales (2015) uses the term CGR1 to refer to the lower ground rent capitalized in a first instance by the landowners upon sale of the land to the developer, and CGR2 for the higher ground rent capitalized in a second instance by the developers. The latter is equal to the potential ground rent once the new building has been completed and its production costs discounted. It is López-Morales’ concept that will be used for analysis in the present paper.

As Eric Clark (1995: 1497) stresses, the tension between actual and potential ground rents is far from being “clinically clean of ties to power in social contexts nor of ties to the imagery of agents.” In fact, the opposite is true. The state and private owners and investors play specific roles in the appropriation of the rent gap. The former creates the economic, legal, infrastructural, and administrative framework, and the latter responds to its private interests over land rent accumulation. Hammel (1999: 1291) observes that potential ground rent (CGR2 here) is produced by city-level factors (i.e. the logic of rent distribution in the metropolis, location within the metropolitan area, development of infrastructure, and land use policies), whereas the capitalized ground rent (CGR1 here) is determined at the neighborhood level. In fact, Hammel broadens Smith’s definition of CGR, widening the focus from the single lot to the whole neighborhood or an even larger unit, and this is also the intention of the present paper. More recently, Liu et al. (2018) and O’Sullivan (2002) confirm the spatial differentiation of both types of rent, and note that they are determined at different scales. Skeptical of the rent gap perspective, Bourassa (1993) argued that differentiation between capitalized and potential ground rent made no theoretical sense, stating that all rent is essentially potential ground rent. However, in response to this, Clark (1995) and Smith (1996b) focused their argument on the temporal aspect in terms of the relative permanence of the built environment, a reality in which land uses and land potential are often not in alignment, thus bringing capitalized ground rent into existence.

A variety of methods and datasets have been used over the years in analysis of the rent gap. Eric Clark (1987) estimated capitalized land rent in Malmö, Sweden based on the nominal tax values of land, and measured potential land rent (equivalent to PGR or CGR) according to the value of undeveloped land immediately prior to development. Hammel (1999) used a similar method to establish that the growth of the metropolitan area in post-war Minneapolis, USA, created a sufficient increase in potential ground rent that rent gaps developed without a concomitant decline in capitalized ground rent. Meanwhile, Badcock (1989) measured potential and capitalized ground rents in Adelaide, Australia, using the average price of a vacant lot and the unit price of all land-use types, respectively. Hackworth (2007) interprets the rent gap in the USA as an increase in PGR (by upzoning and other policies) which causes gentrification, thus agreeing with Hammel (1999). López-Morales (2015) also views capture of the zoning-driven increase in PGR (or CGR2) in Santiago, Chile, as impossible for any agents other than powerful high-rise redevelopers.

In line with Hackworth’s and Hammel’s findings from the USA, the state and local governments in Chile are active agents in the formation of CGR2. Provision of transport and mobility infrastructure, as well as implementation of new zoning and construction codes in certain places (i.e. FAR: the maximum floor area that can be built on a given lot according to its size) significantly raise constructability, generating higher profits for redevelopers (López-Morales et al., 2014). This process differs from classical definitions of rent gap formation (Clark, 1987; Smith, 1979), which identify the devaluation of an elastic capitalized ground rent as the driver.

State urban regulatory agencies are the cause of CGR2 increases in a variety of countries, including Mexico, South Korea, and China (Lees et al., 2016). In some cases, these zoning-driven increases are limitless, and can only be realized through the internalization of what neoclassical economists refer to as the “best and highest land use value” (Marshall, 1961). In the case of Santiago, the CGR2 is ultimately captured by those with sufficient economic and technical power to acquire and accumulate land, demolish existing constructions, and rebuild at the maximum level possible; in other words, large-scale developers. In a context where the level of rent gap captured by other actors—primarily the large number of owners who sell land to high-rise developers—is much lower, the research question of the present study is: to what extent does state action, namely the provision of Metro infrastructure and zoning regulations, increase the portion of the rent gap captured by large-scale developers?

The main objective of this article is to produce a precise estimation of the contribution to rent gap formation made by public investment in Metro infrastructure, local government-enforced FAR, and presence of public and private amenities such as education, health, and shopping centers. To this end, two variables have been defined: CGR1 and CGR2. The CGR1 is the portion of ground rent capitalized by private landowners before high-rise redevelopment occurs; that is, the price received through sale to developers. The CGR2 is the revenue received by the developer (the sale price of the apartment minus all construction costs and the CGR1). These two variables are explained in detail in the “Empirical strategy” section, below.

Santiago: A return to the center “by capital, not people”

Neil Smith’s famous subheading summarizes perfectly the currently problematic urban situation in Chile and other Latin American cities. Traditionally associated with uncontrolled sprawl and an increase in peripheral residential segregation, Chilean cities have become the setting for vigorous “back to the city” movements as intense as Mexico City’s (López-Morales et al., 2016). However, rather than being motivated by cultural preferences and massive post-baby boom middle-class demand for city-center residential opportunities, these Latin American recentralization trends are occurring in response to so-called “rescue” state policies (Lees et al., 2016). These tend to take the form of public incentives—aimed at private investors—which seek high capital returns that in turn are assured by planning guidelines and other state incentives, such as the renowned Chilean housing subsidies. Two significant outcomes of this are an overall increase in housing prices, as demand is never entirely met by supply, and reduced opportunities for remaining households in at least the two lowest income quintiles (López-Morales, 2015). This result accentuates deep socio-spatial inequalities and conflicts sparked by denied rights to housing, and generates a symbolic image of socio-economic fragmentation in the city (Rolnik, 2015). Santiago’s high-rise reconstruction is not evenly distributed; new buildings become concentrated in certain areas, often leaving behind a patchwork of residual spaces. These underdeveloped or abandoned interstitial lots are dotted among the sprouting towers, leaving traditional residents and thousands of vulnerable households to manage in slum conditions. It is in situations like this that Borsdorf and Hidalgo (2013) see gentrification and pauperization occurring side by side.

Santiago is a city of 7 million inhabitants that covers 72,000 hectares—720 million m²—of urban land. Between 2010 and 2016, real estate holdings, in association with banks and retail companies, were permitted to build 337,000 new housing units and 8300 new office units (INE, 2017), enlarging the private housing stock in the 11 central comunas (districts) by 80% between 2002 and 2015. Today, a pronounced “verticalization” of the MAGS is taking place: average height in 2002 was 4.8 stories, and by 2016 had increased to 8.4 stories, with an average of 20 stories in some central comunas (Astaburuaga and Grandon, 2017). Many of the central comunas that have been subject to extreme high-rise redevelopment are also home to traditional neighborhoods—the architectural heritage value of which often remains unprotected—and almost all have experienced increases in FAR and received new state-financed transport facilities (most notably, the Metro network). Santiago is succumbing to verticalization alongside other large cities around the country, with the central skylines of Valparaíso, Antofagasta, and Iquique experiencing a similar transformation. Interestingly, this is in contrast to other major Latin American cities, where zoning regulations are far stricter than those of Santiago (López-Morales et al., 2016).

Residential density has increased sharply in certain areas of the MAGS as a result of alleged attempts by some developers to “democratize” the private housing supply through provision of extremely small-sized residences. These are supposedly more accessible to middle and lower income households, but these occupiers find themselves saddled with disproportionately high rents or mortgages considering the tiny size of the apartments, many of which are under-served in terms of facilities (in some cases, as many as 150 apartments are served by only one elevator). As reported by INE (2002) and MINVU (2017), average apartment size in the MAGS decreased by around 30% between 2002 and 2015, and it is therefore unsurprising that the net density of many projects already exceeds 10,000 inhabitants per hectare. Conflicts relating to overcrowding, loss of privacy, poor services, and diminished quality of housing have been steadily increasing in recent times (Vicuña del Río, 2017), and it is common to find “family” apartments with a floor area as small as 30 or 35 m² in many of these new “verticalized” areas of central Santiago.

Other externalities of large-scale, high-rise rebuilding include the shadows they cast, loss of privacy in neighboring properties, saturation of surrounding streets, and even loss of views, all of which are factors that tend to devalue neighboring properties and/or generate displacement pressure on traditional residents affected by verticalization. In particular, these environmental externalities diminish the sale price that small-scale owners could expect when they come to sell their land to developers (i.e. CGR1), and such low values may not provide them with sufficient compensation to secure equivalent replacement housing within the same neighborhood. According to López-Morales (2015), up to 50% of former small-scale land owners risk being displaced due to low CGR1 in the Santiago Centro comuna.

Santiago’s corporate housing market generates the false impression of a free market for all, and lower-middle income households contemplating their lives in central comunas are presented with only two alternatives: leave, or remain in extremely tight housing conditions.

However, while homes have been decreasing in size over the years, prices have increased. In Santiago Centro, the most central comuna, the maximum sale price for new housing rose by 67.1% between 2002 and 2015, a rate only slightly below that of two neighboring Independencia and San Miguel comunas, and slightly above that of Estación Central. These comunas are also centrally located and are undergoing intense redevelopment, with increases of 89.9%, 79.4%, and 58.2%, respectively. At the metropolitan level, the average price per square meter for a new apartment was close to 37.4 UF in 2008, but that average had soared to 73.4 UF by 2017 (GFK Adimark, 2018).² Prices in the city continue to rise, seemingly unstoppably, and this real estate boom is allowing middle class and upper-middle class segments to access (or invest speculatively in) new and well-located housing. Demand for mortgages has increased, as has the diversity of financing products available (Alarcón et al., 2014). Some 63.5% of the country’s housing mortgages are concentrated in Santiago alone, and almost 27% of total national bank stock is now in the form of housing credits (Daher, 2017).

Large-scale developers are powerful agents. They have unparalleled power to fix the purchase price of the land they acquire, and to impose the sale price of the new apartments they supply. In 2016, there were 524 developer firms supplying housing in Greater Santiago. Sixty per cent of these had built only one project each, and the nine largest developers accounted for half of total production in the city. Meanwhile, at the comuna level, there could be one or two dominant, large-scale developers accounting for 50% or more of housing production (Wainer et al., 2016). Real estate companies obtain the highest possible returns based on maximum exploitation of the land (i.e. they “close” the rent gap), and, as can be seen, they do so in an environment in which competition can be practically non-existent.

At the time of writing, developers are required to pay construction tax (amounting to 1.3% of the declared construction cost), 19% VAT on sale of properties (interestingly, this only came into effect in 2016 as a result of a tax reform that had been highly contested by the corporate sector), and up to 44% of the official tax-assessed value of the land (usually between 20% and 40% of its commercial value), an almost insignificant amount which is considered a form of compensation for the traffic congestion generated around the new buildings. As of mid-2018, however, the latter deduction has yet to be officially enforced. Positive economic contributors to an increase in the rent gap, namely Metro infrastructure, public amenities, green areas, and the widespread public upzoning of land, remain completely untaxed at the time of writing.

Data

The empirical analysis draws on data from all 36,911 new apartment sale transactions in Greater Santiago between 2008 and 2011. Each transaction is defined as a transaction unit. The three databases used are as follows.

The Property Registry Data Base (PRDB) contains administrative data, and records all of the property sale/purchase transactions that took place between 2008 and 2011 in Greater Santiago (including both apartments and houses). The database provides the transaction date, sale price, property location (address, including comuna or district), property floor area, and size of lot (in the case of houses) for each transaction unit. For analysis of CGR2, only apartments built from 2007 onwards were selected, in order to observe properties sold directly by the real estate firm (or, as in a small number of cases, by individual buyers who bought and immediately decided to sell). Although the data corresponds to a limited period of four years, the fact that it was obtained as administrative data (making it hard to acquire, as the PRDB is not publicly accessible) and during an “average” period of real estate exploitation in the MAGS (neither boom nor decline), makes this period historically representative.

The Internal Revenue Service (IRS) contains administrative data, and offers information regarding all existing properties in Santiago. A connection to the PRDB could be established using the IRS codes present in both databases. From the IRS, it was possible to obtain the number of buildings constructed, the number of apartments per block, and the floor area of each, measured in square meters. This data provided the total square meterage of apartments per block.

The Territorial Intelligence Center (CIT) at the Adolfo Ibáñez University contains georeferenced information concerning existing infrastructure in the MAGS. Drawing on this georeferenced database, it was possible to conduct analysis of proximity to Metro stations, green areas, and education and health centers for each of the transaction units.

The PRDB contained a total of 40,059 observations, although the actual number of houses and apartments with all data clearly identified was only 36,911. Figure 1 shows the existing Metro network and stations in operation during the 2008–2011 period, their area of influence (1 km around each station), the spatial distribution of green areas, and the location of education, health, and shopping centers in the MAGS. The figure also shows the FAR for each zone of the MAGS in 2011.

Figure 1.

Metro stations and areas of influence (1 km radius), education, health, and shopping centers, green areas, and floor area ratios in the MAGS.

Table 1 shows the number of transaction units (new apartments purchased) and the average sale price—(CGR1 and CGR2 values measured in UF/m²)—according to their distance from a Metro station. There is no clear negative correlation between distance and either of the two levels of ground rent, or between distance and sale price. In fact, the relationship could mistakenly be taken as being positive, given that the greater the distance, the higher the sale price. However, it is highly probable that this apparent relationship is due to the existence of other, unconsidered variables; for example, the socio-economic composition of the MAGS, which is highly segregated and stratified, and in which the most expensive housing is located further from the centers and sub-centers, and is consequently not connected by Metro stations. As a result, prices of apartments further away from the Metro network are higher.

Table 1.

Average values according to distance from Metro.

Distance from Metro	Number of apartments	Average sale price of apartments (UF/m²)	Average CGR1 (UF/m²)	Average CGR2 (UF/m²)
Between 0 and 500 m	14,100	37.4	13.3	16.1
Between 500 and 1000 m	8869	35.8	10.5	15.3
Between 1000 and 2000 m	8965	36.2	10.8	13.6
Between 2000 and 3000 m	3225	41.1	13.2	17.0
More than 3000 m	1752	45.4	8.1	22.9

All transaction units from the sample are included. Distance ranges include the lower and exclude the upper limits.

Source: Compiled by the authors.

Table 2 shows the average and standard deviation of the variables that are to be used in the empirical analysis. The average value of CGR2 (potential ground rent as net profit to developers) for each square meter of apartment in the MAGS is 15.7 UF/m², while the value of CGR1 (land price estimated to have been paid by developers to land owners) is 11.7 UF/m². The average price of new apartments is 37.4 UF/m². In other words, between 2008 and 2011, the CGR2 represents 42% of the sale price of the average square meter of new apartments sold in Greater Santiago. The average FAR is 3.7, although this varies significantly between comunas. The average distance to the Metro from the apartments studied is 1.07 km, with approximately 62% of units concentrated within a radius of 1 km. Also, the distance to amenities varies, with apartments being located on average 265 m away from health centers, 252 m away from education centers, and 2.3 km away from shopping centers. Finally, there are on average 164 green areas within a walking distance of 500 m from the apartment.

Table 2.

Descriptive statistics for the variables of interest.

Variable	Average	Standard deviation
CGR1 (UF/m²)	11.7	6.7
CGR2 (UF/m²)	15.7	8.7
Sale price of apartments (UF/m²)	37.4	10.6
Floor area ratio	3.7	2.9
Units within Metro area of influence (1 km)	62%	48%
Distance to Metro	10.7	11.8
Distance to education center	2.5	1.8
Distance to health center	2.7	2.5
Distance to shopping center	23.5	13.7
Green areas within 500 m	164	693

CGR1: land price estimated to have been paid by developers to land owners; CGR2: potential ground rent as net profit to developers.

Note: All transaction units from the sample are included. Distance ranges include the lower and exclude the upper limits. Distances measured in 100 m units.

Finally, Table 3 shows some of the variables of interest and their evolution over the 2008–2011 period. In general, the variables observed remain relatively constant, presenting only slight increases. Both the price of new apartments and the CGR1 and CGR2 show a slight rise, which is consistent with the general price increases seen throughout the economy. However, the average FAR decreases slightly over the period, resulting from a “trend” among newly elected center-left mayors to impose tighter land and building regulations in several central comunas. The proportion of apartments situated within a 1 km radius of Metro stations is the variable that rises the most, approximately by 15%. This notable increase is mainly due to the opening of new Metro lines during the period. Finally, the number of new apartments sold jumped from 7433 in 2008 to 10,142 in 2009.

Table 3.

Variables of interest per year.

Variable	2008	2009	2010	2011
CGR1 (UF/m²)	11.7	12.4	11.4	11.4
CGR2 (UF/m²)	15.2	15.3	15.7	16.5
Sale price (UF/m²)	36.5	37.4	37	38.3
Floor area ratio	3.8	3.8	3.7	3.6
Units within Metro area of influence (1 km)	59%	59%	63%	68%
Frequency	7433	10,142	9232	10,104

CGR1: land price estimated to have been paid by developers to land owners; CGR2: potential ground rent as net profit to developers.

Note: All transaction units from the sample are included.

Empirical strategy

There are very few regression studies that have scrutinized the rent gap using empirical data. Liu et al. (2018) in Auckland, New Zealand, see “Improvement Value”—namely the quotient (not the difference) between the capital value (the total value of land including all the improvements) and the bare land value—as a proxy for rent gap, and their study shows high correlation between rent gap and most renovation events if viewed at the neighborhood level. The analysis conducted in the present work differs, however, as it was conducted on a case-by-case basis (residential units sold), and comprised repeated metropolitan-level administrative cross-section data over four years, with statistically consistent results.

The study uses a regression model to analyze the impact on CGR1 and CGR2 of two public policies—investment in new Metro infrastructure and the imposition of municipal building regulations (FAR)—and of the distance of apartments from amenities and green areas. CGR1 is understood as being the sale price level of previously empty properties or those occupied by houses or warehouses (observable in the PRDB) during the sample period. CGR2 is equivalent to the potential ground rent, which in the case of Chile is only captured by developers with sufficient construction capacity. This factor was more complex to obtain. Equation 1 shows that the CGR2 of apartment i in comuna (district) c in year t (cgr2_ict) is obtained by subtracting from the sale price (P _ict ) the cost of land (CS _ict ) and the construction cost (CC _ict ) associated with apartment i in comuna c at time t³.

cgr 2_{ict} = P_{ict} - C S_{ict} - C C_{ict}

(1)

The sale price of the apartment (P_ict) is directly observed in the PRDB data. The cost of the land (CS_ict) associated with every square meter of the new apartments sold is not directly observable, since it is the total cost of the land divided by the number of square meters built on it. As this data is not present in any databases, it was estimated based on average purchase price data for houses, warehouses, and empty lots within a 500 m radius of the center of the block, taken from the PRDB. Knowing the total square meterage of new apartments created in one block (taken from IRS data), it was only necessary to estimate the total square meterage of land (unknown) occupied by these buildings. That size was estimated based on the total new square meterage built in the block (IRS data) divided by the FAR established for that block (known data), as it was assumed that developers will aim to take maximum advantage of the building size permitted by the coefficient. Once the total square meterage of land occupied by buildings was known, it could be imputed to each new built square meter (taking into account that 10% of each building would be occupied by communal areas, a figure distributed among the total square meters of all new apartments).

Finally, the direct cost of construction (CC_ict) is assumed on average to be equal to 16 UF/m² for all buildings in Santiago. This is within a spectrum accepted by the real estate sector in Chile for average high-rise residential construction during the analysis period, comprising the direct cost of construction materials plus workforce, but excluding indirect sales and marketing costs. As shown in Table 4, the level of construction cost does not affect the results.

Table 4.

Sensitivity analysis: construction costs and maximum floor area ratio variations.

Panel A: Construction cost variations
Construction costs imputed	12 UF/m²	14 UF/m²	16 UF/m²	18 UF/m²	20 UF/m²
Less than 1 km from Metro station	0.273***	0.220***	0.256***	0.229***	0.189***
Less than 1 km from Metro station	(0.007)	(0.007)	(0.008)	(0.009)	(0.01)
Floor area ratio	0.043***	0.055***	0.061***	0.068***	−0.085
Floor area ratio	(0.001)	(0.001)	(0.001)	(0.001)	(0.001)
Panel B: Maximum floor area ratio variations
Maximum floor area ratios imputed	6	8	10	12	14
Less than 1 km from Metro station	0.248***	0.253***	0.256***	0.256***	0.253***
Less than 1 km from Metro station	(0.008)	(0.008)	(0.008)	(0.008)	(0.008)
Floor area ratio	0.073***	0.071***	0.061***	0.052***	0.043***
Floor area ratio	(0.002)	(0.001)	(0.001)	(0.001)	(0.000)

Note: The table shows only parameters associated with radius from Metro station (dummy that takes the value of 1 if the apartment is within a 1 km radius of the Metro) and floor area ratio (FAR) (a variable that takes the value of the current FAR in the block of apartment i in comuna c in year t) from regressions of equation (2). All controls were included but not shown (distance to education, health, and shopping centers, the number of green areas within a walking distance of 500 m, dummy variables for each comuna, and dummy variables for each year). Standard errors are shown in brackets. *p < 0.05, **p < 0.01, ***p < 0.001.

Table 4 shows the regression results for equation (2) (explained below) for five different levels of construction costs (ranging from 12 to 20 UF/m²) and five different levels of FAR (ranging from 6 to 14 UF/m²). The objective is to establish the effect of the parameter of distance to Metro station and the parameter of FAR on land rent. The higher the construction cost, the lower the effect of distance to Metro on land rent, with the difference reaching 8% when comparing 12 UF/m² with 20 UF/m². Similarly, the higher the construction cost, the lower the effect of FAR on land rent, reaching zero when construction cost is 12 UF/m². On the other hand, variations in FAR have a minor effect on the parameters. The parameter of distance to Metro increases as the FAR becomes higher, but the difference is less than 1% when compared with a lower FAR. The parameter of FAR decreases when FAR increases from 7.3% to 4.3%. Although there are variations in the results depending on whether construction costs are assumed to be low or high, we consider these to be minor, and in line with the general results.

To assess the influence of Metro stations on CGR2, a radius of influence of M meters was determined around each station. Buildings are classified according to whether they fall inside or outside this radius. In order to establish the radius with the strongest effect on valuation, three different distances were selected (250 m, 500 m, and 1 km). These estimations showed that the highest valuation increase occurs within the 1 km radius. Table 5 corroborates this analysis, showing that an increase exists up to a distance of 1 km, on average.

Table 5.

Sensitivity analysis: Radius from Metro station.

	(1) 250 m	(2) 500 m	(3) 1000 m
Radius from Metro station	0.087***	0.121***	0.256***
Radius from Metro station	(0.009)	(0.007)	(0.008)
Floor area ratio	0.057***	0.059***	0.061***
Floor area ratio	(0.001)	(0.001)	(0.001)
N (number of observations)R²	36,911	36,911	36,911
N (number of observations)R²	0.470	0.473	0.484

Note: The table shows only parameters associated with radius from Metro station (dummy that takes the value of 1 if the apartment is within a 250/500/1000 m radius of the Metro) and floor area ratio (a variable that takes the value of the current FAR in the block of apartment i in comuna c in year t) from regressions of equation (2). All controls were included but not shown (distance to education, health, and shopping centers, the number of green areas within a walking distance of 500 m, dummy variables for each comuna and dummy variables for each year). Standard errors are shown in brackets. *p < 0.05, **p < 0.01, ***p < 0.001.

In order to study the building regulations, the valid FAR were used for the block on which the apartment buildings are built. The higher the ratio, the taller the building can be built over the lot. In the sample used in this study, the FAR vary between 0 and 7.2. However, if the coefficient was not defined in the corresponding Comuna Master Plan (as happens in many comunas of Greater Santiago), a value of 10 is imputed, reflecting total deregulation. The latter coefficient value is currently applied to new buildings located in comunas such as Estación Central and Santiago Centro (López-Morales et al., 2015). Table 4 shows the effect of the maximum parameters of FAR imputed when comuna construction codes are “unlimited” (range of 6 to 14, as variations of this assumption).

To obtain unbiased estimation coefficients, controls relating to the characteristics of the apartment and its environment were used. These controls are: distance to education, health, and shopping centers, and the number of green areas within a walking distance of 500 m. It is believed that these amenities could determine, by way of affecting the apartment price, the CGR2 obtained. Also, as they could be correlated with the variables of interest, failure to include them in the model could bias the estimates. In addition, fixed effects are included in the model both by comuna and by years of construction and sale. The fixed effects by comuna allow control by existing comuna characteristics that remain fixed in time and are directly unobservable, for example, accessibility to place of work or certain desirable or undesirable social characteristics of the comuna. These effects are important to consider in highly segregated cities such as Santiago, which is characterized by stark socio-economic differences between comunas, especially in terms of the availability of and access to locally serviced amenities in low, middle, and upper income comunas. The fixed effects by time allow control by short-term events, for example booms, economic crises, or technological shocks that affect the real estate industry as a whole.

The econometric model specification is shown in equation (2), where r_ict is the natural logarithm of CGR2 obtained from the transaction of apartment i in comuna c at time t. DM_ict is a dummy that takes the value of 1 if said apartment is within a 1 km radius of the Metro. PP_ict is a variable that takes the value of the current FAR in the block of apartment i in comuna c in year t. X_ict is a vector that contains the aforementioned control variables, F_c are dummy variables for each comuna, and F_t are dummy variables for each year. $ε_{ict}$ is the error term which represents the unobserved part of the empirical model, and for each observation is supposed to be identical and independently distributed.⁴ F_c take into account characteristics of each comuna that are fixed over time, and F_t take into account idiosyncratic shocks that affect all comunas each year. F_c and F_t are used only as controls.

r_{ict} = α_{0} + β * D M_{ict} + θ * P P_{ict} + δ X_{ict} + F_{c} + F_{t} + ε_{ict}

(2)

It should be noted that the data sample may be biased, as the analysis only considers those apartment units that were sold, and not those that had not yet been sold. This could be a problem if the unsold units (not observed) are correlated with their distance from the Metro. Indeed, it is more likely that the unsold units are located farther from the Metro, and, if this is the case, the analysis could be underestimating the effect of distance to the Metro on CGR2. To check this, it was noted that the average time for a complete sale of the stock of apartments in a building constructed during the 2008–2011 period is 20 months (Cámara Chilena de la Construcción, 2011a, 2011b); therefore, the apartments built in 2008 should have been sold during the period 2008–2011. Therefore, in order to evaluate this possible selection bias, the same specification (1) was estimated only for new apartments built in 2008, and the results were contrasted with those of the total sample of apartments sold over the whole period. Results were almost the same, as any difference in CGR2 was insignificant.

The advantage of the data and methodology presented here is that it includes all sale transactions for apartments and land plots in the Metropolitan Area, given that the PRDB is administrative data. Likewise, the average real estate revenue after construction—discounting all of the production costs of the buildings—(CGR2) is considered. This is in contrast to Trivelli’s (several years) quarterly report, which considers only the sale price of properties, and to Agostini and Palmucci (2008) who observed only the variation in housing prices (adding the internal attributes of the house).

Results

Table 6 presents the results for the estimates of equation (2) described in the previous section. Column 3 is the best estimate available, since it includes all controls. Column 1 presents the results without considering the fixed effects of time or comuna (district). Column 2 presents the estimate considering only fixed effects by time. Finally, column 3 presents the best estimate, using fixed effects by time and comuna, and allows control of the aforementioned problems.

Table 6.

CGR2 estimates for the MAGS.

	(1)	(2)	(3)
Less than 1 km from Metro station	0.167***	0.166***	0.256***
Less than 1 km from Metro station	(0.007)	(0.007)	(0.008)
Floor area ratio	0.016***	0.016***	0.061***
Floor area ratio	(0.001)	(0.001)	(0.001)
Distance to education center	0.020***	0.021***	0.024***
Distance to education center	(0.002)	(0.002)	(0.002)
Distance to health center	0.004***	0.004***	0.008***
Distance to health center	(0.002)	(0.002)	(0.002)
Distance to shopping center	−0.009***	−0.009***	−0.009***
Distance to shopping center	(0.000)	(0.000)	(0.000)
Green areas within 500 m	0.000***	0.000***	0.000***
Green areas within 500 m	(0.000)	(0.000)	(0.000)
Fixed effect time	No	Yes	Yes
Fixed effect comuna	No	No	Yes
N (number of obs.)	36,911	36,911	36,911
R ²	0.230	0.230	0.484

Note: Standard errors are shown in brackets. *p < 0.05, **p < 0.01, ***p < 0.001. Additional controls per socio-economic characteristic and year of first construction.

According to the specification shown in column 1, being located within the area of influence of a Metro station increases the CGR2 perceived by the real estate developer by an average of 16.7%. It is also found that an increase in the FAR by one unit increases the CGR2 by 1.6%, and if the FAR were modified from the minimum regulation (1) to the maximum (10), the increase in income obtained by the developer would be approximately 16%. The two effects remain relatively constant in the specification shown in column 2. However, if the fixed comuna characteristics are not considered, the coefficient of the results is skewed.

When analyzing column 3, which contains the best available specification, it is noted that being within the 1 km area of influence of the Metro increases the CGR2 by an average of 25.6%. It is also noted that the distance from education and health centers tends to increase CGR2 by 2.4% and 0.8%, respectively. This may be due to a saturation effect, relating to noise and other negative externalities associated with being close to these places. It is also interesting that while proximity to shopping centers has a positive effect on the CGR2, this is relatively small, with the average effect being close to 0.9%. Finally, the availability of green areas within 500 m has a statistically significant effect which is nonetheless close to zero.

As explained in the Empirical strategy section, the same conditional model was estimated for apartments built in 2008 only (table not included) in order to evaluate possible selection bias. The results show that the CGR2 increases on average by 27.3% when the apartment is located within 1 km of a Metro station, indicating a small bias in the selection of the sample illustrated in Table 5 by estimates that are slightly lower than the true effect of Metro infrastructure on the CGR2.

Additionally, the same model was estimated, but only for the comunas of Santiago, Providencia, and Las Condes, as shown in Table 7. These comunas were chosen because, on the one hand, they have the greatest number of apartments sold in the Metropolitan Area of Greater Santiago (above 1000 units per year each), and on the other, they show interesting heterogeneities. Figure 2 shows a GIS spatial concentration analysis of the CGR2 in Greater Santiago (search radius = 500 m), revealing that these three comunas not only had the highest concentration of apartments sold—along with the other longitudinal areas serviced by the Metro network—but also the highest levels of CGR2.

Table 7.

CGR2 estimates per comuna.

	(1) Santiago	(2) Providencia	(3) Las Condes
Less than 1 km from Metro station	0.457***	0.197***	0.527***
Less than 1 km from Metro station	(0.014)	(0.032)	(0.021)
Floor area ratio	0.057***	0.084***	0.016
Floor area ratio	(0.008)	(0.015)	(0.002)
Distance to education center	−0.017***	−0.023***	−0.012
Distance to education center	(0.006)	(0.007)	(0.004)
Distance to health center	0.044***	0.026***	−0.017**
Distance to health center	(0.005)	(0.007)	(0.005)
Distance to shopping center	−0.019***	−0.024***	0.014***
Distance to shopping center	(0.001)	(0.002)	(0.001)
Green areas within 500 m	0.000	0.000	−0.015***
Green areas within 500 m	(0.000)	(0.002)	(0.000)
Fixed effect time	Yes	Yes	Yes
Fixed effect comuna	Does not apply	Does not apply	Does not apply
N (number of observations)	11,239	4659	2365
R ²	0.350	0.322	0.336

Note: Standard errors are shown in brackets. *p < 0.05, **p < 0.01, ***p < 0.001. Additional controls per socio-economic characteristic and year of first construction.

Figure 2.

Spatial concentration analysis of CGR2 in Greater Santiago.

There are different and heterogeneous effects of distance to education center, health center, and shopping center, as well as of availability of green areas. Being close to an education center is positive for CGR2, while living close to a health center is negative in Santiago Centro and Providencia but positive in Las Condes. This may be because in Las Condes the health centers tend to be very exclusive private clinics, while in Providencia and Santiago Centro they are mainly hospitals. Being close to a shopping center is positive in Providencia and Santiago Centro, but negative in Las Condes. Finally, the availability of green areas within 500 m has a statistically significant positive effect only in Las Condes.

In the comuna of Santiago Centro, the CGR2 of properties within a 1 km radius of a Metro station is an average of 45.7% higher than for those that are further away. In the same comuna, each additional unit of the FAR generates an increase in the CGR2 of 5.7%. In terms of the other variables, for every 100 m increase in the distance between the apartment and a shopping center, the CGR2 reduces by 1.9%. Increased distance to health centers increases CGR2 by a significant 4.4% per 100 m, while a greater distance to education centers reduces it by 1.9% per 100 m. In other words, developers are able to capture a higher CGR2 in the Santiago Centro comuna because buyers ascribe value to living near education, health, and shopping centers, and the presence of many high-ranking public and private schools in the comuna add to this. The opposite is true at the level of Greater Santiago, where living closer to an education center is considered undesirable, thus reducing CGR2. Finally, the presence of green areas within a radius of 500 m has an effect that is not statistically significant.

Meanwhile, in the upper-income comuna of Providencia, apartments located within a 1 km radius of a Metro station yield a CGR2 that is higher by 19.7%, although this value is below the metropolitan average. This is because Providencia is a highly central, affluent comuna, where access to different modes of public and private transport is far more evenly distributed than in the average MAGS comuna. However, in Providencia, every additional unit of FAR increases the CGR2 by 8.4%, a value considerably higher than the MAGS average. This disparity is explained by the much stricter building regulations in place in Providencia, a scenario which means that any increase in the construction coefficient is highly appreciated by the real estate market due to high demand for housing from affluent households. The other variables behave similarly to the previous case of the Santiago Centro comuna, in that a greater distance from education centers produces lower CGR2, as does a shorter distance to health centers.

The third specific case is the extremely affluent Las Condes comuna, in which proximity to the Metro has a far greater effect on CGR2 than the MAGS average. According to the Casen survey (MIDESOC, 2015), Las Condes is second in the national ranking of comunas with the lowest number of households living in multidimensional poverty, after neighboring Vitacura. These two comunas, together with four others, comprise Santiago’s so-called “Barrio Alto,” namely the self-segregated, affluent eastern wedge of the city, situated in the foothills of the Andes mountain range. Although gentrification in Las Condes is highly unlikely, the CGR2 here increases by around 52.7% within a 1 km radius of Metro stations. However, unlike the previous two comunas analyzed, changes in FAR seem not to have a statistically significant effect. The coefficient associated with the distance to education centers is not statistically significant either, while distance to shopping centers shows that for every 100 m further away from malls, the CGR2 increases by 1.4%. This is probably because shopping centers in this comuna are of the large, suburban variety, which tend to be less attractive to live near to. Finally, for every additional 100 m from health centers, the CGR2 decreases by 1.7%, making Las Condes the only case where this effect was identified. Being a highly affluent comuna, the private health clinics located in this part of the city are valued more highly than elsewhere.

It was possible to estimate the effect of each Metro line on CGR2 separately, and the regressions are shown in Table 8. Proximity to a Metro station on Line 1 (the oldest line, opened in 1975 and only extended to the “Barrio Alto” in 2009) generates an increase in CGR2 of 17.8%. Line 2 (opened in 1978) increases CGR2 by 17.3%, and Line 4 (opened in 2005) by 16%. Line 4a (a branch of Line 4 opened in 2006, and running through traditionally deprived southern comunas) generates a higher increase in CGR2 of 22.2%. Finally, Line 5 (opened in 1997) generates the lowest increase in CGR2, at 11.1%. In conclusion, high-rise real estate development driven by the numerous stations on Metro lines 2 and 4—particularly in the traditionally deprived southern part of Santiago—has benefited from public investment during the 2008–2011 period, leading to increases in CGR2 that are considerably higher than along the other Metro lines.

Table 8.

CGR2 estimates for each Metro line.

	(1)	(2)	(3)
Radius from Metro station Line 1	0.223***	0.231***	0.178***
Radius from Metro station Line 1	(0.009)	(0.009)	(0.009)
Radius from Metro station Line 2	−0.063***	−0.065***	0.173***
Radius from Metro station Line 2	(0.008)	(0.008)	(0.011)
Radius from Metro station Line 4	0.454***	0.457***	0.160***
Radius from Metro station Line 4	(0.013)	(0.013)	(0.012)
Radius from Metro station Line 4a	−0.229***	−0.233***	0.222***
Radius from Metro station Line 4a	(0.025)	(0.025)	(0.027)
Radius from Metro station Line 5	−0.022***	−0.026***	0.111***
Radius from Metro station Line 5	(0.008)	(0.008)	(0.008)
Floor area ratio	0.022***	0.022***	−0.055***
Floor area ratio	(0.001)	(0.001)	(0.001)
Fixed effect time	No	Yes	Yes
Fixed effect comuna	No	No	Yes
N (number of observations)	36,911	36,911	36,911
R ²	0.259	0.261	0.483

Note: The table shows only parameters associated with radius from Metro station (dummy that takes the value of 1 if the apartment is within a 1 km radius of Metro stations on lines 1, 2, 4, 4a, and 5) and floor area ratio (a variable that takes the value of the current FAR in the block of apartment i in comuna c in year t) from regressions of equation (2). All controls were included but not shown (distance to education, health, and shopping centers, the number of green areas within a walking distance of 500 m, dummy variables for each comuna and dummy variables for each year). Standard errors are shown in brackets. *p < 0.05, **p < 0.01, ***p < 0.001.

The very high appreciation along Line 2 is interesting, because this line crosses two of the most active centers of high-rise development in the MAGS in the last 10 years: the San Miguel and La Cisterna comunas. Meanwhile, the 4a branch line connects La Cisterna with the comunas of Puente Alto and Ñuñoa. Puente Alto, a traditionally working class area which has in recent years become more middle-class, showed increases in land prices of more than 1000% between 1990 and 2010 (see Trivelli, 2011). By comparison, the “middle class” Ñuñoa comuna, situated on the border of Santiago’s “Barrio Alto,” has been seeing active residential real estate production over the past 20 years (Figure 2).

Finally, Table 9 shows the estimate of the proportion of CGR1 (the portion of the rent gap captured by land owners when selling their properties to developers) generated by the Metro and the FAR. This analysis also sought to test the common assumption that landowners are the main beneficiaries of public investment in Metro infrastructure. The results are surprisingly low, with an increase of only 5.5% in the sale price of lots and houses within a 1 km radius of a Metro station compared with those properties located further away. Compared with the at least 25% increases in CGR2 received by developers, this result clearly indicates that the latter are the biggest beneficiaries of public investment in the Metro. Landowners, who represent more than 60% of MAGS residents (MIDESOC, 2015), benefit far less. Table 8 also illustrates that the common belief that changes in FAR have positive effects on CGR1 is also unsubstantiated; it is primarily the high-rise developers who benefit from increases in FAR as set within comuna master plans.

Table 9.

CGR1 estimates.

	(1)	(2)	(3)
Less than 1 km from Metro station	0.031***	0.040***	0.055***
Less than 1 km from Metro station	(0.006)	(0.006)	(0.006)
Floor area ratio	−0.000***	−0.000***	−0.000
Floor area ratio	(0.000)	(0.000)	(0.000)
Distance to education center	0.055***	0.053***	0.006***
Distance to education center	(0.001)	(0.001)	(0.001)
Distance to health center	0.012***	0.011***	0.002*
Distance to health center	(0.001)	(0.001)	(0.001)
Distance to shopping center	−0.008***	−0.008***	−0.003***
Distance to shopping center	(0.000)	(0.000)	(0.000)
Green areas within 500 m	0.000***	0.000***	0.000***
Green areas within 500 m	(0.000)	(0.000)	(0.000)
Fixed effect time	No	Yes	Yes
Fixed effect comuna	No	No	Yes
N (number of observations)	63.498	63.498	63.498
R ²	0.148	0.186	0.415

Note: Standard errors are shown in brackets. *p < 0.05, **p < 0.01, ***p < 0.001.

Conclusions

Results at the Metropolitan Area of Greater Santiago (MAGS) level show that apartments sold within a 1 km radius of a Metro station yielded a CGR2 that was 25.6% higher than for apartments located farther away. According to López-Morales’s (2015) method, the CGR2 is the net profit—or the maximized portion of the rent gap generated by “public investment”—that developers can capture following the sale of the residential units they produce. The analysis also shows that each additional point of FAR in comuna master plans (upzoning—the “regulation” portion of the rent gap created) produces an increase in CGR2 of about 6.1%. These results partially confirm Hammel’s (1999) thesis, as potential ground rent (CGR2) is not only determined at the metropolitan level (Metro), but also at the local level (zoning). All in all, the issue of the scale of observation in rent gap creation is shown to be highly important in the present paper. Separate analysis by comuna shows that altogether, both of these generators of increases in the rent gap result in enormous profits for developers. By contrast, the increase in CGR1 (namely the portion of the rent gap that homeowners can capture) resulting from proximity to a Metro station is marginal, being barely above 5.5%. A range of results were obtained through separate analyses based on Metro line and comuna.

On the whole, these results suggest three conclusions relevant to Santiago and elsewhere.

Public transport investment and construction guidelines are two important rent gap creators. Similar results may be seen in large cities elsewhere, where extensive zones have traditionally lacked transport infrastructure, and subsequently been subject to considerable changes in zoning ordinances. These two factors, while extremely relevant, have received little attention in the gentrification literature (Lees et al., 2016).

The rent gap is mainly enlarged not by devaluation of the CGR1—as the first versions of rent gap theory propose (Clark, 1987; Smith, 1979)—but by inflation of the CGR2. The results of the present study are not particularly surprising, as they confirm previous accounts by Hammel (1999), Hackworth (2007), and López-Morales (2015). The CGR1 does in fact increase, but to a far lesser degree than the CGR2.

In cities where land is highly unregulated, the rent gap tends to be absorbed by large-scale developers, as only these have sufficient economic power and capacity to operate on the enormous scale that the local-level planning codes permit. The primary social effect of this trend is greater inequality between those in control of high-rise housing production (the dominant “class,” according to Harvey, 1974, who have access to financial resources, superior market information, and, of course, greater capacity for land acquisition) and those whose only commodity is land, usually in a tiny portion.

The conceptual separation between CGR1 and CGR2 (or the “potential ground rent”) is a powerful theoretical instrument that helps to resolve the important question of who benefits from appreciation in land value triggered by state action. Given that those who benefit belong to the upper segments of society, a possible corollary (within the limits of market regulation, and in the case of Chile) is to establish a tax on gains in real estate income (CGR2) resulting from state investment in the city. A simple measure may be to consider an “average” CGR2 tax levied across the whole MAGS, potentially being at least 30% of the average profit generated per square meter of new residential units sold as a result of Metro-related benefits plus one FAR point. The remaining 70% of profit—which is still a considerable proportion of the CGR2—could then be enjoyed by the developer.

Variants of this high-rise tax are discussed for a large number of international cases by Smolka (2013) and Mirrlees (2011), and there are several other exhaustive studies that have been conducted in Brazil, Colombia, Argentina, Ecuador, and Uruguay (Blanco et al., 2016; Bocarejo et al., 2013; Mendieta and Perdomo, 2007; Muñoz-Raskin, 2010; Sandroni, 2011) confirming that the taxation of land value appreciation following public transport investment is not only a progressive policy, but also highly implementable. Furthermore, the opportunities that this type of policy could generate for the state are considerable, including potential for financing more spatially inclusive social housing policies (on central or improved—and hence more valuable—land), or new and much needed public transport systems.

Footnotes

Acknowledgments

Thanks to Matías Garretón, Luis Valenzuela, Elvin Wyly, Juan Pablo Cid, Daniel Meza, Ignacia Saona, Pablo Trivelli, and Paul Salter for their invaluable comments on earlier versions of this paper.

Declaration of conflicting interests

The author(s) declared no potential conﬂicts of interest with respect to the research, authorship, and/or publication of this article.

Funding

The author(s) disclosed receipt of the following financial support for the research, authorship, and/or publication of this article: This work was supported by the Centre for Social Conflict and Cohesion Studies (COES, Conicyt Fondap #15130009) and Fondecyt Project #1151287.

Notes

References

Agostini

Palmucci

(2008) Capitalización heterogénea de un bien semipúblico: El Metro de Santiago. Cuadernos de Economía 45: 105–128.

Alarcón

Demaestri

Piedrabuena

(2014) Financiamiento de la vivienda en Chile. Washington, DC: Inter-American Development Bank.

Astaburuaga

Grandón

(2017) Cambios habitacionales de la capital de Chile entre 1995 y 2015. Santiago: Astaburuaga y Grandón Consultora.

Badcock

(1989) An Australian view of the rent gap hypothesis. Annals of the Association of American Geographers 79(1): 125–145.

Blanco

Fretes

Muñoz

(eds) (2016) Expandiendo el uso de la valorización del suelo: La captura de plusvalías en América Latina y el Caribe. Washington, DC: Banco Interamericano de Desarrollo.

Bocarejo

Portilla

Pérez

(2013) Impact of Transmilenio on density, land use, and land value in Bogotá. Research in Transportation Economics 40(1): 78–86.

Borsdorf

Hidalgo

(2013) Revitalization and tugurization in the historical centre of Santiago de Chile. Cities 31: 96–104.

Bourassa

(1993) The rent gap debunked. Urban Studies 30(10): 1731–1744.

Brueckner

(1980) A vintage model of urban growth. Journal of Urban Economics 8(3): 389–402.

10.

Cámara Chilena de la Construcción (2011a) Informe Mach (Macroeconomía y Construcción) 32. Santiago: CCHC. Available at: http://www.cchc.cl/uploads/archivos/archivos/MACh-32.pdf (accessed 31 July 2017).

11.

Cámara Chilena de la Construcción (2011b) Informe Mach (Macroeconomía y Construcción) 34. Santiago: CCHC. Available at: http://www.cchc.cl/uploads/archivos/archivos/MACh-34.pdf (accessed 31 July 2017).

12.

Clark

(1987) The Rent Gap and Urban Change: Case Studies in Malmö 1860–1985. Lund: Lund University Press.

13.

Clark

(1995) The rent gap re-examined. Urban Studies 32(9): 1489–1503.

14.

Daher

(2017) Centralismo financiero en Chile. In Vial

Hernández

(eds) ¿Para qué descentralizar? Centralismo y políticas públicas en Chile. Santiago: Ichem, pp.235–250.

15.

GFK Adimark (2018) Informe Trimestral Mercado Inmobiliario Gran Santiago - 1 er Trimestre 2018 y Series Históricas por Comunas y Tipologías 1T 2018. 15 August. Available at: https://www.gfk.com/fileadmin/user_upload/country_one_pager/CL/GFK_Adimark_IOI__1T_2018.pdf (accessed 15 August 2018).

16.

Hall

(1981) The Inner City in Context: The Final Report of the Social Science Research Council Inner Cities Working Party. London: Heinemann Educational Books.

17.

Hammel

(1999) Re-establishing the rent gap: An alternative view of capitalised land rent. Urban Studies 36(8): 1283–1293.

18.

Harvey

(1974) Class-monopoly rent, finance capital and the urban revolution. Regional Studies 8(3–4): 239–255.

19.

Harvey

(1989) The Urban Experience. Baltimore, MD: Johns Hopkins University Press.

20.

INE (2002) Censo Nacional de Población - Chile 2002. Santiago: Instituto Nacional de Estadísticas.

21.

INE (2012) Resultados Preliminares Censo de Población y Vivienda 2012. Santiago: Instituto Nacional de Estadísticas.

22.

INE (2017) Tabla Permisos de Edificación Nacional 2010–2017. Santiago: Instituto Nacional de Estadísticas. Available at: http://www.ine.cl/herramientas/galeria-de-mapas/permisos-de-edificacion (accessed 21 November 2017).

23.

Hackworth

(2007) The Neoliberal City: Governance, Ideology, and Development in American Urbanism. New York: Cornell University Press.

24.

Hammel

(1999) Gentrification and land rent: A historical view of the rent gap in Minneapolis. Urban Geography 20(2): 116–145.

25.

Lees

Shin

López-Morales

(2016) Planetary Gentrification. Cambridge, UK: Polity Press.

26.

Lefebvre

(2003) The Urban Revolution. Minneapolis and London: University of Minnesota Press.

27.

Liu

O’Sullivan

Perry

(2018) The rent gap revisited: Gentrification in Point Chevalier, Auckland. Urban Geography 2018: 1–26.

28.

López-Morales

(2009) Urban Entrepreneurialism and Creative Destruction: A Case-study of the Urban Renewal Strategy in the Peri-centre of Santiago de Chile, 1990–2005. PhD Thesis, University College London. Available at: http://discovery.ucl.ac.uk/18707 (accessed 20 May 2019).

29.

López-Morales

(2015) Assessing exclusionary displacement through rent gap analysis in the urban redevelopment of inner Santiago, Chile. Housing Studies 31(5): 540–559.

30.

López-Morales

Meza

Gasic

(2014) Neoliberalismo, regulación ad-hoc de suelo y gentrificación: el historial de la renovación urbana del sector Santa Isabel, Santiago. Revista de Geografía Norte Grande 58: 161–177.

31.

López-Morales

Arriagada

Gasic

, et al. (2015) Efectos de la renovación urbana sobre la calidad de vida y perspectivas de relocalización residencial de habitantes centrales y peri centrales del AMGS. EURE 41(124): 45–67.

32.

López-Morales

Shin

Lees

(2016) Latin American gentrifications. Urban Geography 37(8): 1091–1108.

33.

Marshall

(1961) Principles of Economics. London: Macmillan.

34.

Mendieta

Perdomo

(2007) Especificación y estimación de un modelo de precios hedónico espacial para evaluar el impacto de Transmilenio sobre el valor de la propiedad en Bogotá. Bogotá: Universidad de Los Andes and CEDE Centro de Estudios sobre Desarrollo Económico.

35.

MIDESOC (2015) Resultado Encuesta de Caracterización Socioeconómica (CASEN). Santiago: Ministerio de Desarrollo Social de Chile.

36.

MINVU (2017). Estadísticas Observatorio habitacional. Santiago: Ministerio de Vivienda y Urbanismo.de Chile. Available at: http://www.observatoriohabitacional.cl/opensite_20080220163307.aspx (accessed 21 November 2017).

37.

Mirrlees

(2011) Tax By Design, The Mirrlees Review. London: Institute for Fiscal Studies (IFS).

38.

Muñoz-Raskin

(2010) Walking accessibility to bus rapid transit: Does it affect property values? The case of Bogotá, Colombia. Transport Policy 17(2): 72–84.

39.

O’Sullivan

(2002) Toward micro-scale spatial modelling of gentrification. Journal of Geographical Systems 4(3): 251–274.

40.

Rolnik

(2015) Guerra dos lugares. A colonização da terra e da moradia na era das finanças. São Paulo: Boitempo.

41.

Schafran

McDonald

Lopez-Morales

, et al. (2018) Replacing the services sector and three-sector theory: Urbanization and control as economic sectors. Regional Studies 52(12): 1708–1719.

42.

Slater

(2015) Planetary rent gaps. Antipode 49: 114–137.

43.

Sandroni

(2011) Recent experience with land value capture in São Paulo, Brazil. Land Lines 23(3): 14–19.

44.

Smith

(1979) Toward a theory of gentrification: A back to the city movement by capital, not people. Journal of the American Planning Association 45(4): 538–548.

45.

Smith

(1996a) The New Urban Frontier: Gentrification and the Revanchist City. London and New York: Routledge.

46.

Smith

(1996b) Of rent gaps and radical idealism: A reply to Steven Bourassa. Urban Studies 33(7): 1199–1204.

47.

Smolka

(2013) Implementing Value Capture in Latin America: Policies and Tools for Urban Development. Policy Focus Report Series. Cambridge, MA: Lincoln Institute of Land Policy.

48.

Trivelli

(2011) La propuesta de modificación del Plan Regulador Metropolitano de Santiago PRMS 100 requiere una justificación más sólida. EURE 37(111): 179–184.

49.

Trivelli

(several years) Análisis Trimestral del Boletín de Mercado de Suelo Urbano, Área Metropolitana de Santiago. Santiago: Consultora Trivelli.

50.

Vicuña del Río

(2017) Impacto de la densificación residencial intensiva en la segmentación del tejido urbano de Santiago: un enfoque cuantitativo. Revista 180(40): 78–93.

51.

Wainer

Gasic

Bauzá

(2016) Grandes Promotores Inmobiliarios. Oferta y Mercado Objetivo. Santiago Metropolitano. Santiago: Inciti.

Rent gap formation due to public infrastructure and planning policies: An analysis of Greater Santiago,Chile,2008–2011

Abstract

Keywords

Introduction

Rent gap: Theory and analytical approaches

Santiago: A return to the center “by capital, not people”

Data

Empirical strategy

Results

Conclusions

Footnotes

Acknowledgments

Declaration of conflicting interests

Funding

Notes

References