Urban policies, mobility and gentrification in two neighboUrhoods of belo horizonte 1

This article investigates the relations between urban policies, the residential mobility and the gentrification of urban space, based on research in two neighbourhoods of Belo Horizonte: Santa Tereza and Anchieta. Several types of data were used in the study, including the Origin and Destination Survey, which identifies residential mobility in neighbourhoods, the Demographic Census datasets on households and residents, and municipal data on real estate dynamics and current urban policies. Qualitative data from interviews and local observations has also been used. The results dem-onstrate how the processes are distinct. In Santa Tereza, the urban policies implemented as an outcome of resident mobilization have managed to stop gentrification. In Anchieta, the greater liberality of urban policies, which did not elicit any organized responses from residents, has allowed renovation in some parts of the neighbourhood, which we identified as new-build gentrification.

Urban policies, mobility and gentrification in two neighboUrhoods of belo horizonte 1 sociol. antropol. | rio de janeiro, v.10.01: 561 -586 , may. -aug., 2020 http://dx.doi.org /10.1590 /2238-38752020v10210 gentrification: notes on the concept Today gentrification is considered one of the most relevant and recurrent changes observed in cities, especially in their central areas. As a phenomenon encountered in diverse places, an intense debate exists concerning its causes and effects, but also about the social and historical specificities of the cities where the process occurs, with more than half a century of studies already amassed. Due to these debates, there is no single agreed definition of gentrification. At most, one can speak of some elements over which consensus exists. Among these are a specific type of residential mobility of social groups, which involves the entry, in a neighbourhood or region of the city, of a group with economic and/or cultural capital superior to the group already residing there. The presence of these new residents leads to the valorisation of the district and the displacement and effective expulsion of older residents. It thus constitutes a process of (voluntary) residential mobility of middle-class groups, which in turn provokes the (involuntary) mobility of other groups with less economic power (Lees, Slater & Wyly, 2013;Smith, 1996;Bidou-Zachariasen, 2007). As well as these aspects, some of the studies on gentrification have contributed to making explicit socio-spatial inequalities (Van Criekingen, 2007;Lees & Ley, 2008) and the loss of social cohesion among long-standing residents (Forrest 2016).
Gentrification is not immediately felt since it results from the valorisation of property and the swapping of traditional commerce for a more sophisticated kind, which gradually makes staying in the area infeasible for those unable to afford the new higher residential and commercial costs. 562 urban policies, mobility and gentrification in two neighbourhoods of belo horizonte sociol. antropol. | rio de janeiro, v.10.01: 561 -586 , may. -aug., 2020 562 The first studies focused on the gentrification present in older neighbourhoods of the central areas of cities. In these, the new residents contributed to the maintenance of historical aspects of the area, like its architecture, but not the residents and their social relations (Zukin, 2010;Van Criekingen, 2007;Lees & Ley, 2008). More recently this spatiality was broadened. New studies identified forms of gentrification supported not on the historical characteristics of central neighbourhoods any longer, but on new constructions (Davidson & Lees, 2010;López-Morales, Gasic Klett & Meza Corvalán, 2012), some of them outside the central areas (Pereira, 2017). In these cases, the construction of new buildings can occur in spaces that previously residential or possessing other former uses, such as industrial or commercial.
In Brazil, the majority of the studies have been concentrated in the central areas of large cities where the consumption of well-equipped cultural heritage spaces by the middle classes were identified, but not residential gentrification (Leite, 2004;Rubino, 2009;Frúgoli Jr. & Sklair, 2009;Jayme & Trevisan, 2012). There are historical centres that underwent processes of urban revitalization as a result of public investments, some in partnership with the private sector, which involved the reform of public spaces and the construction or reform of old cultural equipment with the aim of attracting new groups, especially middle class sectors. However, these attempts were not always successful in terms of attracting new residents. Fewer studies exist on gentrification in typically residential neighbourhoods (Furtado, 2014;Reina & Comarú, 2015).
Gentrification involves distinct actors such as the State and the formulators of urban public policies, real estate agents, traders and the population, whether those entering the area, the gentrifiers, or those displaced by the arrival of new inhabitants or traders. Due to the power disparity, the real estate agents are those who most benefit economically from gentrification and the displaced resident population are the most disadvantaged.
The State with its policies is an important actor in the gentrification of spaces, but this relation is not always evident. Urban policies can facilitate in the sense of 'preparing the terrain' for (state-facilitated) gentrification, as in the cases of renovation of historical centres in which the symbolic image is altered in accordance with the tastes of the middle classes (Hiernaux & Gonzáles, 2014). In other cases, the absence of policies is what allows real estate interests to override the interests of the residents threatened with displacement. In these cases, the municipality favours the urban real estate developers and fails to protect those most affected, the residents. These processes are very often treated as part of a 'market logic,' as if capital did not benefit directly from laissezfaire type urban policies.
In other situations, the State acts explicitly in favour of gentrification, making it state-led or state-sponsored. For some authors, like Lees and Ley (2008), in the current phase, gentrification is fully and affirmatively incorpo-563 article | luciana teixeira de andrade and jupira gomes de mendonça 563 rated into urban policies, whether in places like Shanghai, where gentrification is prompted by the local government, or in countries with a history of social housing policies, recently impacted by more restrictive welfare policies, like the cases of Sweden, Canada and Holland. Whatever the case, it is important to recognize the variety of urban public policies and their relations to gentrification. The theories that connect gentrification to the processes of globalization and neoliberalization tend to leave aside these differences that are far from irrelevant. In the city of Amsterdam, for example, strong housing pressure and a neoliberalizing policy have led to gentrification of parts of the city. However, it is significantly minimized by social housing policies consolidated over many decades (Van Gent, 2013).
Among the actions of public authorities, the so-called urban regeneration or revitalization policies were those prioritized in the analyses of gentrification, but there are others that can both foment and protect a particular area from gentrification. Among these are urban policies that result in a change in zoning or uses, focused on very specific areas and not covering the city as a whole.
Consequently, some regions end up more protected and others less so. Such policies can also result from the population's organization and mobilization.
In these cases, the pressure from social movements can result in the formulation of urban policies that impede or hinder the action of gentrifiers.
Although expulsion or displacement is the most emphasized aspect of gentrification, and frequently associated with the loss of affordable housing, assurance of the latter is not the only factor that enables the continuation of lower-income social groups in a space. The maintenance of forms of commerce, with the offer of products and services accessible to them, are as essential as housing, and both are conditions for maintaining the neighbourhood's forms of sociability, constructed over years of residence in a shared social space. Otherwise, the loss of the local ambience and the sense of place can comprise a third factor of displacement. The existence of social bonds and affective relations with the space are important components for the mobilization and resistance of residents, as can be observed in the anti-gentrification movements. The fact that gentrification has occurred slowly in its initial stages has been one of the factors that prevented or hindered the formation of opposition movements.
Today the first signs of gentrification are already easily recognized, while responses in opposition to the process are also spreading (Just Cause, 2015), in part because of sharing a concept originally formulated in academia with social movements.
In Brazil, studies of gentrification are some way behind the international production, especially in terms of exploring its social indicators like those we have been discussing so far: the social mobility of groups and the urban policies that promote or impede this mobility. It is in this field that the present article seeks to provide a new contribution, anticipating here that the lack of studies 564 urban policies, mobility and gentrification in two neighbourhoods of belo horizonte sociol. antropol. | rio de janeiro, v.10.01: 561 -586 , may. -aug., 2020 564 on these indicators may in part be related to the lack of research on intraurban residential mobility and with the difficulties of working with this data on the scale at which gentrification occurs: the neighbourhood or even smaller regions.
It is the difficulty that we propose to confront here.
The data used here have come from diverse sources. The main source is the data on residential mobility coming from the Origin and Destination Survey, 2 which shows the changes of residences of inhabitants from the Belo Horizonte The choice of the neighbourhoods of Santa Tereza and Anchieta is related to the localization of both in the city's central area (where the processes of gentrification take place more intensely), though in distinct regions. Despite having been formed in different periods, today the two areas shared the same condition as completely urbanized neighbourhoods, with a good infrastructure, close to the city centre and occupied predominantly by middle-class groups, albeit from distinct strata, as will be explained later. These differences also guide the choice. Santa Tereza is one of the oldest neighbourhoods of Belo Horizonte, one of the reasons that justified its listing as the city's cultural heritage, a condition it shares with another two adjacent neighbourhoods. Anchieta, a more recent occupation, has no kind of special urban planning protection and, for this precise reason, has been undergoing a rapid process of substitution of its older residences by apartment buildings. This type of urban dynamic is what approximates it to various other neighbourhoods located in the same region, the Centre South, where the upper-middle income strata are concentrated (Mendonça, Andrade & Diniz, 2019). Over the course of the twentieth century, Santa Tereza grew in density.
However, given the few access roads and the barriers constituted by a river and a railway line, its growth was relatively slow, in part because it was contained within these limits. As happened in other parts of the city, a favela formed close to the river and the margin of the railway line. In the 1980s, this favela, like various others in the city, was removed after heavy flooding destroyed most of its dwellings (Nazário & Andrade, 2010). There remain just two small parts, today known as Vila Dias and Vila São Vicente, which distinguish themselves from the neighbourhood by the irregular morphology of their occupation and by the social condition of their residents.
Among the central area neighbourhoods, Santa Tereza is one of the most studied (Westin, 1998;Baggio, 2005 In the social map of the city, Santa Tereza is situated in the East Zone, initially occupied by workers and which over time transformed into a region predominantly inhabited by middle class sectors, but with a different social composition to the neighbourhood located in the South Zone of the city, dominated by upper-middle-class sectors. The latter zone is where the other neighbourhood studied in this article, Anchieta, is located.
Unlike Santa Tereza, Anchieta has not attracted the attention of researchers.
It is known that its occupation began fairly timidly in the 1920s with denser growth occurring only later, between the 1950 and 1970s, when various roads were opened up in the surrounding area and within the neighbourhood itself (Arreguy & Ribeiro, 2008). Over its history, it has not been home to any important public institution, unlike the case of Santa Tereza.
The information available to us on the neighbourhood is contained in the primary sources like the Demographic Census, the Council Property Tax Registry, and field observations. The sources available on the internet, many of them relating to estate agents, consider Anchieta a residential neighbourhood, with the presence of families who have lived in the area for a long time, but which has been taking in new residents over the years due to the supply of new residents in apartment blocks. Anchieta contains an important local commerce of small shops and diverse service providers, such as footwear and clothing repairs, bars, small grocery stores, bakers, beauty salons and so on. More recently, one avenue has seen a concentration of various bank branches, commercial store buildings, retail outlets and a shopping centre, as well as restaurants and bars that attract residents from surrounding areas to the neighbourhood. It also has the Paróquia de São Mateus  Both neighbourhoods have a predominance of residents who self-declare white, but this percentage is higher in Anchieta (87%) compared to Santa Tereza (66%). The latter data relates solely to residents of the Santa Tereza neighbourhood, however, since in the two favelas situated inside its boundaries, the presence of black and brown residents (76%) is higher than whites (24%). One of the principal indicators of gentrification is residential mobility in space, related to the place that each family occupies in the social hierarchy, given that "inhabited (or appropriated) space functions as a kind of spontaneous symbolization of social space" (Bourdieu, 1997, p. 160). The residential mobility characteristic of gentrification involves an inhabited space losing lower status social groups due to the entry of higher status groups (Bidou-Zachariasen, 2007;Zukin, 2010;Smith, 1996). The investigation of these movements and urban policies can shed light on the socio-spatial processes under way in the two neighbourhoods and their potential relations to the gentrification processes.
The Origin and Destination (OD) Surveys conducted in the Belo Horizonte Metropolitan Region since 1972 provide information on residential mobility: that is, who enters and leaves the neighbourhood over the decade, as well as the socioeconomic data on these people. However, given the size of the sample for the scale of the neighbourhood (3.5%), it does not present statistical consistency. The option was to use a larger unit, the data unit delimited by the IBGE 6 . The possibility of using these larger areas when our interest resided in just part of them -the neighbourhoods of Santa Tereza and Anchieta -was related to the fact that this is the only source available, but also because it presents a degree of social homogeneity, higher in Santa Tereza and lower in Anchieta.
As can be seen by the maps on the next page, the data unit in which Santa Tereza neighbourhood is located is highly homogenous in terms of income.
The data unit containing the Anchieta neighbourhood is more heterogenous: there is a set of neighbourhoods with higher incomes, indicated in dark red, and Anchieta and the adjacent neighbourhood of Cruzeiro, with a higher variation of income sectors, although all of them higher than 10 minimum wages.
This is one of the reasons explaining the fact that the results for Anchieta do not allow a conclusive reading, as we discuss later. The area coloured white in the map is an environmental protection area and thus uninhabited.

571
article | luciana teixeira de andrade and jupira gomes de mendonça To know which social groups entered and left the data unit, the OD Survey records both income and educational level. However, we had to opt for educational level since a significant proportion of respondents did not declare income. Taking educational level as a proxy for socioeconomic status, the data contained in Table 2 (next page) shows that the data unit containing Santa Tereza did not reveal any significant entry of groups with a higher status than those already residing there: the majority of those entering were household heads with a status very similar to existing residents. In this case, therefore, a gentrification process was not observed.
The reading of the data on entry is complemented by the data on leaving.
The departure of people with lower educational levels would be indicative of displacement, that is, the effective expulsion of people with a lower status.
However, this is not what is found. What calls most attention in this data is the departure of 56.6% of household heads with higher educational levels, another indicator that strongly contradicts gentrification. What would explain this fact is the lack of property on offer with better conditions for people who are socially upwardly mobile in the neighbourhood, as we discuss subsequently.
Generally speaking, what this data shows us is that there is a turnover of residents in the area where Santa Tereza is located, but this movement does not significantly change the social makeup of the area. Our hypothesis resides in the fact that approximately 70% of this area is protected by heritage policies.
As well as Santa Tereza itself, protected by an ADE and by its listing as heritage, the Floresta neighbourhood and part of the Colégio Batista were also listed as municipal heritage, which heavily reduced the possibilities for new constructions and, therefore, the activities of the property market. 7 The working-class origin of some of the neighbourhoods composing the data unit of Santa Tereza has left its marks on many properties. Small lots and buildings, a simple construction style that today does not match the consump-573 article | luciana teixeira de andrade and jupira gomes de mendonça 573 tion patterns of the higher levels of the middle classes. With the heritage listing, the possibility for new developments became practically null. At the same time, the buildings that were constructed under the regulatory control of the ADE follow the pattern of small properties without luxurious finishings or design, more appropriate for rental by professionals starting their career or students who share residence of these properties, as attested by a real estate broker we interviewed: "Here there are a lot of students renting. A lot of healthcare professionals, because it's close to the hospital area. The accessory dwelling units at the back of the houses also have residents, but very few. Really there is a lot of renting to outside people coming to study" (interview conducted 16/05/2016). 8 In terms of a public with a higher income, the heritage listing posed, in this broker's view, many difficulties, principally for people who want to internally adapt their homes to contemporary standards of accommodation, as well as have car parking spaces. As the lots are small and many houses have been listed, it is difficult to find properties that cater for this As remarked earlier, the data for the Anchieta data unit, given the area's heterogeneity, are not so easy to interpret. Differently from the Santa Tereza data unit, in the Anchieta data unit the population turnover (entering and leaving) indicates a slight social change in relation to the social group with a lower educational level, and stability in relation to the group with a higher educational level. The most significant data showing a change is the departure of household heads with an educational level between complete primary education and incomplete secondary education (9.8%), with no entries of this demographic sector (Table 3). There is a small entry of household heads without education or incomplete primary education (1.7%), whose destination was probably the favelas inside the data unit, given that it is difficult to suppose that this happens outside the favelas where the cost of homes is very high. Almost 575 article | luciana teixeira de andrade and jupira gomes de mendonça 575 all those who entered presented an educational level equal to or higher than secondary schooling, with 68.7% possessing higher education, reinforcing the social characteristics of the data unit. The data indicating this stability is the equivalence of entering and leaving at the higher educational level. In the context of the discussion on gentrification, the data unit can be seen to be expelling people with a lower socioeducational level. However, it would not seem to present significant changes in terms of higher educational level. The hypothesis for this case is that the data unit in question already presents a high density of household heads with higher educational levels and, for this reason, there is less possibility of increased densification of this educational stratum.
The fact that those entering and leaving were equivalent may be related to a market strategy of constantly offering new real estate products, which leads to a periodical change of properties among this social group, but always in the space already occupied by these higher classes.

576
This conclusion of stability perceived in the data from the OD Survey, together with the knowledge that this data unit encompasses two distinct groups of neighbourhoods, combined with the field observations in Anchieta and the Urban Land and Building Tax (IPTU) Register on new constructions, which indicate a significant change in some parts of the neighbourhood, led us to a more individualized analysis of Anchieta. In the field research it was possible to perceive that the neighbourhood has been undergoing a change for some years already, involving the substitution of houses not by small threestorey buildings, as happened in Santa Tereza, but by large luxury apartment blocks, some of which are among the most expensive in the city. The OD Survey was conducted in 2011, but since the research forming the basis of this article was begun, in 2016, to the present, the substitution of residence type has been readily perceptible.
As well as these field observations, the data from the property register of the Belo Horizonte City Council, used to collect IPTU land and building tax, show that since the 1980s there has been a sharp drop in the construction of lower-end property developments and a growth in more sophisticated apartments, standards 4 and 5, which are the highest building levels in the city council's ranking. In this hierarchical classification, 1 corresponds to the lowest level and 5 to the highest (Graph 1). Until 1980, the construction of mid-level P3 properties predominated. In the following decade, P4 superseded the construction of P3, and the construction of P5 increased considerably, continuing to rise in the 2000s. This increased supply of high-end apartments is related to the real estate dynamic and the national economic context. The economic crisis of the 1980s and the drastic decline in funding for housing provoked a decline in the construction of residential developments for low and lower-middle income sectors. In these years, self-funded production predominated in which the purchasers paid for the construction work (Gomes, 2008). In the case of Anchieta, this more general and selective dynamic impelled the production of high-end buildings.
Field observation corroborates a piece of information that appears in  tion, what this data shows us is that the Anchieta neighbourhood has been undergoing a process of gentrification through new construction initially concentrated in one of its areas, but whose tendency has been one of expansion.
In these cases, gentrification takes place through the supply of new properties to a population with a purchasing power higher than that of the population residing in the neighbourhood previously (Davidson & Lees, 2010).

final considerations
This article confronted the challenge of exploring one of the main indicators of residential gentrification, the mobility of social groups in space, with all the difficulties that this task presents given the lack of data at the neighbourhood scale. We hope to have contributed thereby to the methodological discussion over the limitations and possibilities of the data, as well as to a discussion of the phenomenon of the residential gentrification of these empirical reference points.
The two cases shed light on the need to think about gentrification in Brazilian cities through other reference points than just those linked to cultural heritage, as has been done until now, by also including the production of new constructions, as observed in Anchieta. They also show the importance of urban policies and the role of the State in legislating on the city, either by contributing to the fixing of residents, or by favouring real estate interests. And, finally, it shows how the action of the State and private actors can meet with resistance in some spaces. This data corroborates the idea that gentrification is not an ineluctable force that imposes itself on all places and one single way.
On the contrary, the study of gentrification must contemplate the differences between the neighbourhoods in terms of their histories, their social and cultural compositions, as well as in their forms of organizing and acting in the public sphere.  Ribeiro (2000).

The borders to the Census Sectors in the 2000 and 2010
Censuses are not the same. For this comparison, they had to be made compatible so that the areas equivalent to each neighbourhood in one decade and the other would be as similar as possible.
6 The data unit used here is an aggregation of census sectors realized in order to ensure confidentiality and statistical consistency. For this work, these data units were compatibilized with the regionalization used by the OD Survey.
7 Santa Tereza's listing in 2015 came after the OD Survey, but the neighbourhood was already protected by the ADE.
The other neighbourhoods were listed in 1996 as urban complexes, which means a protection that goes far beyond the individually listed properties. Among the restrictions is the reduction in the height of buildings constructed near to listed properties.