Motivations of tenancy and the diversity of tenants: the case of Ankara, Turkey

ABSTRACT Motivations of tenancy could be related to economic, social and spatial issues, such as, lifecycle/life course, affordability and accessibility. This implies the diversity of tenants living in cities. Ankara, as the capital and the second biggest metropolitan city of Turkey, has a high proportion of tenants. In this study, the diversity of tenants and motivations for tenancy in Ankara is investigated with a socio-demographic, economic and locational perspective, analyzing the reasons for housing mobility and preference of the current house, and the variables affecting the diversification of tenants, according to income level through a survey of 475 households. Considering the evidences of the case study, lifecycle/life course issues have an important role in the choice of tenancy, while economic affordability is a major reason for the housing mobility of tenant households. The diversity of tenants in Ankara is categorized as “temporary” who benefit from educational and employment facilities of the city; “semi-permanent” characterized by the households who prefer to live as a tenant; and “persistent”, characterized by low-income households who do not have the opportunity to purchase a house. As a conclusion, the importance of developing rental housing strategies especially aiming the low-income households is emphasized.


Introduction
Tenancy is increasing worldwide due to many changing socio-demographic, economic, and spatial factors. Trends toward smaller households, more nonfamily households, the aging of the population, greater ethnic diversity, increasing workforce flexibility, labor mobility, etc. transform the social and societal structures and life conditions in urban areas, which reshape the consumer preferences of the households (Peppercorn and Taffin 2013). As an important supplement to owner-occupier housing, rental housing presents alternatives with its peculiar flexible structure in terms of social, economic, and spatial aspects. Private rental sector has become more important in recent years as it has increasingly been seen as part of the solution to the housing supply problem (Hulse and Yates 2017;Crook and Kemp 2014). In Turkey, where there are no social rented housing production and rental housing policies, tenancy appears as an increasing usage form, especially in urban areas.
The housing supply is provided by both the private (individual, cooperatives and build-and-sell housing production) and public sectors [by HDA (Housing Development Administration] in Turkey (Tekeli 1996Türel 1996. The supply of rental housing in Turkey is part of the housing production. In this respect, the supply of rental housing in Turkey occurs in all of these housing supply forms. According to TURKSTAT (2020) private sector had been dominant (more than 90%) in housing production in Turkey. Thus, private sector generates most of the rental housing supply or stock. Institutional investors (-private equity firms, hedge funds, real estate investment trusts (REITs), and publicly listed real estate firms etc.) entered the rental housing markets in large cities of North America, the UK, (Nethercote 2020). In Turkey, this private rental housing is not an institutionalized supply, rather this supply has been constituted from the investments of individual/small households who rent out their dwelling units . In Turkey, the rental housing practice of the public has left only as public housing for government officers. The percentage of this kind of public rental housing (lodging)in total housing stock is very small. Correspondingly, privatization of these public houses, like privatization of social housing seen in most of the developed countries, has gained legal status with the "Change in the by-Law of Sale of Public Houses" in 2011. Other than this, squatter areas present the informal rental housing supply.
In recent years in Turkey, the proportion of renter households has increased, and the proportion of landlord households has decreased. In 2011-2014, it was seen that the rate of homeowners is on an upward trend, while in 2015-2018, when the home support rate tended to increase in 2011-2014, it was seen that this rate tended to decrease. Accordingly, in 2011, 59.6% of households were landlords and 22.2% were tenants, while in 2018 the proportion of landlords decreased to 59%, while the proportion of tenant households increased to 25.2% (Figure 1).
Demand for rental housing has been ongoing in Turkey from past to present. Rental housing demand in urban areas has been higher than rural areas (Table 1). In Turkey, which lacks the production of social-rented housing, the proportion of private-rented housing sector, which is a rare option in most European countries, is especially common in urban areas (large cities such as Ankara, Istanbul, Izmir) (Sarıoğlu 2007). Provinces with the highest proportion of households, who are tenants, being the most urbanized provinces of Turkey such as, Istanbul with a rate of 31.5 %, Ankara with a rate of 30.2%, and Antalya with a rate of 29.9% (TURKSTAT 2011) (Figures 2 and 3). According to Aksoy (2016 in Özdemir Sarı, 2019), the share of the private rented sector in urban areas was 25-30% during the 2000-2014 period. Since 2014 it is not possible to differentiate urban and rural areas from the statistics of TURKSTAT .
In addition to variables such as income, price, credit conditions and interest rates are variables that determine housing demand. High-interest home loans due to high inflation rates in developing countries such as Turkey are not particularly available for sub-income groups (Sarıoğlu 2007). Since its establishment (1984), HDA has played a very important role in housing delivery, but 46% of the houses it produces in 2018 are for the lower income group within the scope of social housing production. These houses are not social rented housing, but residential property housing (Republic of Turkey Ministry of

Environment and Urbanization Housing Development
Administration 2019). According to Türel and Koç (2015) and the Ministry of Development (2018), HDA has a 10% share in meeting existing housing needs. From there, Türel and Koç (2015) and the Ministry of Development (2018) state support states that it is not expected to talk about any contributions provided for those who need the support of the state to acquire property or rental housing, and that the need to offer different housing options for sub-income groups in ways such as organization, financing, etc. continues (Ministry of Development 2018; Türel and Koç 2015).
On the other hand, the existing housing financing system in Turkey is emerged as a system to finance the housing demand of the middle-and upper-income group (Karapinar et al. 2008;Yener Çoşkun 2013) and the monthly reimbursement of the average mortgaged home loan amount according to the current loan system are about 50% above the minimum wage (Ministry of Development 2018). For this sub-income group, tenancy is more than a choice, it is a mandatory choice. Hague (2010) states that the 2008 crisis was caused by banks that were unable to be cautious in using a "high-risk interest mortgage loan (sub-prime)" to make low-income earners housing in America. This has shown that social housing, whose rents are subsided, is a better idea (Hague 2010). Although housing purchasing power has increased in Turkey since the 2000s, the decrease in housing is related to the decrease in housing purchases made for investment/accumulation purposes due to the decrease in real return on house prices and the lack of purchasing power of the lower income group (Yener, 2013;Türel and Koç, 2015). According to TURKSAT 2018 data, 50% of the median income calculated per equivalent individual is on the poverty line, with 13.5% of the population (TRY 7944) in 2017 and 13.9% of the population under the poverty line in 2018 (TRY 8892). Of the groups that are 60% below median income, 30.3% are tenants, 26.8% of groups between 60-120% of median income and 21.8% of those with more than 20% of median income are tenants. As income rises, the proportion of tenants decreases.
In Turkey, housing policies in the past and today have supported home ownership principally. The provision of the housing stock (housing/rental housing market) has been almost left to market forces (Sarıoğlu 2007). The period after the establishment of the Turkish Republic in 1923 up to the 1960s had been characterized by a nation-state building process. The first applications of housing policies in Turkey had been experienced by establishing Ankara as the capital city of the young Republic. In this respect, firstly in 1925, the government officers had been provided with advanced credits to establish housing cooperatives by Law no. 586 (Tekeli 1996). Also, by Law no. 1452(1929, the government officers had been provided with housing amends. In Turkey, until the 1950s the main objective of the housing policies had been to acquire housing for public employees. Social welfare understanding has been accepted in Turkey in the 1960s as in most of the countries in the world. The planned period has begun in Turkey and five-year development plans started to be established. In the first of the four five-year development plans; producing more houses with limited sources, restraining luxury housing production, producing houses for meeting the housing requirement of lowincome groups, removing the problems of low-income groups related to rent-paying by producing low rented houses, providing cheaper urban land, preventing land speculation, encouraging the production of social housing, solving the squatter housing problem, In Turkey, regulations such as control of housing rents, tenure security, rent allowances are evaluated below ( Figure 4). Turkish housing system has been the introduction of a new Code (no. 6098) of Obligations in 2011, superseding Law on Property Rights. Security of tenure and rent control regulations receive widespread support with claims that they improve fairness and the available housing choice mix; enable long-term tenancies; and help low-income groups (Ball 2016). In Turkey, landlords are small investors who rent out their own houses. Before the last regulation, tenants could face eviction under certain situations determined which were subject to misuse by landlords. However, the new regulation improves the situation of tenants against the landlords. According to the new regulation, the landlord cannot dispossess the tenant unless the contract ends. It is aimed to regulate the rent increases which often create conflicts between homeowners and renters in the periods of higher inflation within Turkish Code of Obligations No. 6098. The maximum increase in rent was determined according to the producer price index during the preceding 12 months. However, not to take necessary provisions against the initial rent prices is not a suitable situation for especially lowincome groups, as well as this situation is not a protective for the low-income groups.
It could not be stated that Turkey has efficient rent subsidies as in the European countries. Implementations about this issue in Turkey are quite strict and insufficient. Within the increase of rent prices in the first years of the Republic, government provided rent subsidies named as expensivity raise to public officials in 1927 and provided a payment system named as "muvakkat tazminat" as law no. 1452 which differs depending on the degrees of public officials in 1929. These subsides were abolished by the government in 1951. After this period, there were no such an effective rent subsidy. In the article no: 16 of law no. 6306 in 2012, related with the transformation issue in the areas which include disaster risk, it is approved that to provide rent subsidies from the special budget of transformation projects. Homeowners, renters and people who have restricted real rights which use buildings that are discharged, demolished or publicized with the appointment could benefit from the subsidies in the context of the law. In the context of the law, renters might get subsidy for one time only by considering their moving costs. In the case that renters want to be householders, they might get interest support from the banks which have protocol with the ministry. The period of the rent subsidy is maximum 18 months. Rent subsidy is provided for one time only to the renters considering their moving costs. In the context of this law, slum areas in large cities such as Ankara, Istanbul and Izmir are the subject of urban transformation. Tenants in these informational areas have no choice but to re-tenant within the rental housing market under free market conditions. Tenancy is an alternative to property housing as another form of use; it has led researchers to conduct research on issues related to tenant households, their characteristics and diversity. Based on this, Nelson (1989) classifies tenants for Indonesia as temporary tenants, permanent tenants and rent-payers. According to Nelson (1989), temporary tenants are students, singles, immigrants, etc., while permanent tenants are those who cannot afford to own a dwelling and prefer to be tenants by not wanting to own a dwelling. Those who do not pay rent are those who stay with their relatives or who stay with their relatives without paying a price in the houses belonging to their relatives. Nelson (1989) considers the process in which the individual or household who emigrated from the countryside with the hope of staying in the city stayed with someone else or relative at no cost as the first step in their urban housing career. Gilbert (1993) states that tenants have a long relationship with "property owners" such as classes and family members or friends as "sharers", and Rakodi (1995) also classes tenants as cash payers and those who do not in their work on the rental housing sector in third world countries. De Wandeler (2006 in UN-ESCAP & UN-HABITAT, 2011) described the use as conditional and unconditional in a free rental housing market. These conditions are conditions such as income level, gender, student. The most typical example of this category is low-income tenants in social rented housing. There is also a priority ranking [disaster victim, etc.] for lowincome tenants in this category. It also considers students staying in private dormitories, women's shelters and institutions for the elderly as conditional tenants ( Balamir (1996) categorizes tenants for Turkey in five segments. These are the people who can't think about owning housing; chronic tenants with low incomes who cannot save money and lack assets are the collapsed area and housing stock tenants who prefer historical stock or excommunicative housing areas due to both economic and lifestyle, latent tenants in the extended family, short-term temporary shifting tenants such as students, singles, divorcees, and finally voluntary tenants who do not want to own housing (Balamir 1996).
Adequate and affordable housing for everyone is the result of an ideal housing market. An ideally functioning housing market is expected to pose a wide range of alternatives for all income groups and tenures. Understanding the demand and utilization conditions of tenancy is important to develop appropriate tools and policies for the whole housing market and to perform the spatial planning of metropolitan areas. In this respect, the characteristics of the tenant households, the reasons for tenancy, and the preferences and choices of the tenant households should be investigated in a holistic manner (McKee 2012; Murie and Williams 2015). In Turkey, it is seen that the state do not implement appropriate measures to achieve the production of rental housing, the existence of rental housing demand makes it even more important to develop the urban planning, policies and strategies related to rental housing sector. Through this perspective, the article aims to describe and discuss the socio-demographic and economic diversity of the tenant households according to income level and variables discriminating the tenant households according to income level in Ankara, the capital city of Turkey. The research questions of the study are: How diverse are the tenant households? What are the reasons for tenancy, housing mobility, and choice of current house? Which features discriminate the tenant households in terms of their income?

Background: motivations of tenancy, diversity of tenants
Tenancy can be taken as a housing tenure, chosen by individuals or households in their different lifecycle/ life-course phases either depending or not depending on their income level. Tenancy could be a conscious choice for many people because of its flexible features not found in home ownership (Gilbert 2016). Rental housing always has a considerable role in responding to demand from the lower income or mobile households or who did not want to be an owner occupier for some other reasons (Kemp 2011;Durst 2014).
Besides the emphasis on its calibration potential to fluidity and flexibility of the neoliberal and post crises eras (Forrest and Hirayama 2015), tenancy is accepted as being a less secure tenure (Hulse and Haffner 2014). In the related literature, tenure security has been discussed through household characteristics, market and policy factors with affordability and psycho-social aspects together (Lindblad et al. 2017). Soederberg (2018) emphasizes the rising level of low-income rental housing insecurity marked by over indebtedness, evictions and homelessness. Furthermore, security of tenure is linked with the meaning and sense of home (Pawson, Hulse, and Morris 2017) which is related with financial insecurity, housing instability (frequent forced moves) and a lack of autonomy (Bate 2017). Coulter (2017) also suggests that the issues associated with poor condition and high cost accommodation, requires an appreciation of the deeper psycho-social effects of involuntary mobility, insecurity and sociospatial dislocation.
In the following sections, flexibility and insecurity characteristics of private rental housing, affecting the motivations of tenancy and shaping the diversity of tenants in terms of economic, social and spatial aspects will be discussed.

Economic flexibility and affordability
Purchasing power is one of the factors that determine the housing tenure choice of individuals or households (Li 2014;Berl 2004;Carne 2004;Tipu 2005). Lowerincome households are more likely to be burdened by housing cost than higher-income households (Davenport 2003;O'Sullivan 2003). Especially lowincome tenant households in large cities encounter housing affordability difficulties (Blanco, Gilbert, and Kim 2016;Dong 2017). The problem of housing affordability is associated with multi-faceted economic, social and political considerations and housing supply constraints caused by land use regulations that increase the rental housing prices (Dong 2017;Li 2014). In the relevant literature, rental property is seen as economically unavailable or unsuitably priced when the rent price is higher than 25-30% of a family's total monthly income (Dong 2017;UN-HABITAT 2008). However, the rent-income ratio measurement is criticized due to several limitations. It does not differentiate between high-income tenants who choose to spend large proportions of their income on rent as a lifestyle choice and low-income households who are forced to spend large proportions of their income on rent (Baker, Mason, and Bentley 2015;Skaburskis 2006;Stone 2006). It also does not take into account differences in housing and neighbourhood quality or the actual financial constraints faced by individual households (Bogdon and Can 1997).
Individuals' or households' preferences are shaped according to the maximization of their satisfaction while choosing among a series of consumption units throughout their lives. People move to cheaper or more expensive houses due to their preferences for increasing income and savings for other basic needs of households, such as education, food, healthcare and management of the household budget (UN-HABITAT 2008; UN-ESCAP & UN-HABITAT 2011). Households can also be able to afford a larger and/or more desirable dwelling in a better location, and a higher standard of living than would otherwise be feasible (Hulse and Yates 2017). The choices made will depend on household income, wealth and household characteristics as well as the relative costs associated with renting compared with owner-occupation. There are also other financial reasons, such as the desire to avoid the long-term commitments arising from purchasing a house or the desire to avoid the long-run maintenance costs of a house (Muyingo 2017;Pawson, Hulse, and Morris 2017;Willem and Altes 2016).

Social flexibility and lifecycle/life course
Housing tenure choice is not simply a consumption or investment decision, but has an inherent relation with lifecycle/life course aspects (Hulse and Yates 2017;De Groot et al. 2011;Clark and Withers 2007;Mehta and Mehta 1989;Rossi 1955). The lifecycle concept is related to the diverse and distinctive phases of a household. These phases include singlehood, marriage, parenthood, the birth or departure of children, divorce, widowhood, etc. (Schiffman and Kanuk 1997). Besides, life course is a more flexible and comprehensive concept that emphasizes the diversity of timing and the linking of life affairs (Coulter, Van Ham, and Findlay 2016;Dykstra and Hagestad 2007). In this perspective, housing decisions and transitions are not unbounded choices but are shaped by micro-level resources and restrictions, as well as macro-scale opportunities and constraints (Coulter 2017). The size of the household, age, education, employment, parentage, retirement, and housing typologies are some of the issues related to individuals' and households' experiences (Fattah et al. 2015;Helderman, Mulder, and Ham 2004;Rabe and Taylor 2010;Tipu 2005). Generally speaking, households become owner-occupiers as their household size, income, and age of the head of the household increase and when events like marriage and childbirth occur. Likewise, young, single households prefer tenancy to home ownership in most cases. Research studies have proven that renting or sharing a house is generally seen at the first stage of the life course, such as, students who are not yet ready to settle in a permanent house, new migrants, workers, single professionals, or people who share their houses for other reasons (Nelson 1989; UN-ESCAP & UN-HABITAT 2011). However, Ford, Rugg, and Burrows (2002) put forward that this orderly life cycle model of housing careers began to break down in the 1980s as structural changes in demography, labour markets, institutions and housing systems came together to reshape housing options and choices of especially young adults. McKee (2012) emphasizes the deepening reliance of young Britons on the private rental sector as the result of interlinked choices and constraints produced by changing life course trajectories and contextual conditions. Furthermore, Coulter (2017) emphasizes the increased diversity, fluidity and uncertainty of housing careers as a key reason for why transitions to adulthood have become more complex, protracted and precarious, by comparing how family structure and socio-economic characteristics shaped rates of private renting amongst young adults heading families.

Spatial flexibility, housing mobility and accessibility
Housing mobility refers to the intentions, decisions, and actions associated with moving from one residence to another (Coulton, Theodos, and Turner 2012). According to Clark and Huang (2003) housing mobility is seen as a function of life-cycle characteristics, quantifying the role of age, education, marital status, income level and housing tenure in configuring residential mobility.
The demand for rental housing is connected to not only economic and social factors but also spatial factors, e.g. job/school accessibility, urban services, safety, dwelling quality/infrastructure, public transportation facilities etc. (Ballesteros 2004;Bartlett et al. 2002). Some tenants stay for extended periods, while others prefer to remain mobile because they are looking for different job opportunities or do not want to stay in the same place for too long. Housing mobility is conceptualized as a spatial adaptation process depending on changing internal and external conditions, which determines the tenure choice of a household that regulates its housing consumption according to location, size, and typology. Flexible mobility implies that households can move to another place easily without being tied to a certain house (UN-HABITAT 2008; UN-ESCAP & UN-HABITAT 2011).
In the lifecycle/life-course related literature, different variables are emphasized as affecting housing mobility (Rabe and Taylor 2010;Smith and Olaru 2013). Some of the studies focus on the pushing factors affecting the housing consumption processes depending on the structural changes of the households in different phases (Mulder and Lauster 2010;Shelby 2017). A special focus is on the position of tenants in urban renewal, especially since urban renewal may result in the relocation of current tenants through the force of law (Willem and Altes 2016). Correspondingly, the attractiveness of an area, quality of a house, and proximity to urban functions and services are emphasized as pulling factors (Smith and Olaru 2013). Housing market conditions could function as both pulling and pushing factors and affect lifestyle preferences during the different lifecycle/life-course phases.
Hulse and Yates (2017) put forward that, tenancy has become a choice for households with a strong preference for, but an inability to purchase in, an innercity location. They found that relative to all tenant households, there have been an increasing percentage of higher income tenants who are couples both with and without children. It appears that these couple households are increasingly making a lifestyle choice to live close to high-quality jobs increasingly located in and around CBDs. These findings are consistent with the study of Pawson, Hulse, and Morris (2017), which suggests that although long-term renting mainly reflects adaptation to increasingly unaffordable home ownership, lifestyle choices are also significant where renting in a "desirable" area may be preferred to owning elsewhere.

Methodology, study area, and sample case of Ankara
As mentioned in the introduction, national housing policies have encouraged home ownership in Turkey. However, tenancy is a widespread urban tenure, especially in metropolitan areas. In the capital city of Turkey, Ankara, while 58.6% of the households are owner-occupiers, 31.3% are tenants. In Ankara, 3.7% of the households are living in public housing, the percentage of households that do not own the house in which they live but also do not pay rent is 5.5%, and the others constitute 0.9%. In this respect, it is important to study the motivations of tenancy and the diversity of tenants for developing appropriate tools and policies to meet the rental housing demand in the private rental housing market of Ankara.
The study is based on data obtained from a survey 1 that was conducted with 475 households in the Ankara Metropolitan Area between the months of October and December 2012 ( Table 2). The seven central districts of Ankara, Altındağ, Çankaya, Etimesgut, Keçiören, Mamak, Sincan, and Yenimahalle, were determined as the area of the survey ( Figure 5). In the sample area selection, the average rental values obtained within the framework of the research conducted for the current rental housing market were taken into view to represent the diversity of the areas within the sub-markets and sub-markets. Then, the development of residential areas in the historical process of Ankara and the data of the analyses/studies carried out according to Ankara's income groups in the literature were used and determined within the neighborhoods, the south, the southwest, the north, the east, the center and close environment, western planning zones in a way to represent different income groups. To reach more tenant households within the framework of this determination, the most up-to-date statistical data at the county and neighborhood level by TURKSAT is from 2000. Then, no such work has been carried out. Neighborhoods with high number of tenant households have been selected. Questionnaires had been distributed to the randomly selected households in the selected neighborhoods. At the beginning of the research, 611 questionnaires with a 3% rate of sampling of existing tenant households in the neighborhoods had been planned. However, the return number was 475 tenant households.
In the first step, an analysis of the locational and socioeconomic characteristics of the sampling area and a descriptive analysis of the sampled tenant households in terms of their socio-demographic and economic characteristics were undertaken to explore the diversity of tenant households. Reasons for housing mobility and reasons for preferring the current house according to income level were put forward in the second step using a chi-square analysis. In the third step, a discriminant function analysis was used to explain the differences in tenants' households, indicating which socio-demographic, economic, or spatial characteristics contribute most to group separation in terms of income level.

Districts and sampling neighbourhoods
The capital city of Ankara has a population of nearly 5.5 million. The sample districts can be assumed to represent the metropolitan area comprising 85% of the total population. Çankaya and Keçiören are the most crowded districts (16% of the population of the city lives in each income-level groups prefer the A.Taner Kışlalı and Koru neighbourhoods, which are accepted as the most prestigious residential areas in the city. The households living in these areas are mostly of high social status in terms of education level, profession, income, etc. The Etimesgut district in the western part of the city has become a developing area where the population growth rate is high. After the 1950s, the area was characterized by squatter areas due to the rapid rural-urban migration. However, today, the whole area is planned, and the houses are authorized. The Kazım Karabekir neighborhood from the Etimesgut district is one of the oldest neighborhoods where mostly the lower income-level groups choose to live. The Sincan district located in the western part of the city has the lowest socioeconomic development level and the highest poverty rate of all the districts. The 1990 master plan of Ankara focused on developing the western corridor of the city. The Sincan area has been developed based on land use decisions, resulting in an organized industrial district and a squatter prevention area. The Fatih neighborhood from the Sincan district has been affected by the squatter prevention area decision, which includes planning implementations related to low income-level groups. The Keçiören district in the north part of the city has been the focus of population growth since the 1950s, witnessing the rapid development of residential areas characterized by both squatters and apartment-type buildings. The İncirli and Tepebaşı neighbourhoods were developed mostly by squatter housing during this process. Since the 1990s, authorized apartment buildings have been developed by the build-and-sell system, the dominant housing production system in the city.
The Altındağ district, in the northeast part of the city, is the first settled area of Ankara and the location of the historic city centre. From the past to the present, low-income groups have mostly lived in squatter areas. Recently, HDA investments and urban transformation projects have stirred the housing sector in the district. Aydınlıkevler in the Altındağ district is known as a white-collar employees' neighbourhood, whereas Güneşevler is the area where the low-income groups can be found. However, the status of the neighbourhood is changing nowadays because of its location, being on the way to the city airport, which is witnessing big urban transformation projects. Mamak, a district of Ankara, was one of the important subjects of squatter housing during the rapid urbanization process of Ankara after the 1930s. In recent years, urban renewal projects have been implemented by HDA in the district. Generally, low and middle-income groups of households prefer to live in Mamak. The Abidinpaşa neighbourhood in Mamak is close to the important universities, hospitals, and public institutions of the city centre. While it was a squatter area in the 1960s, today, it is a residential area with authorized apartment buildings. Another neighbourhood, Mutlu in Mamak, however, is an area where both authorized and unauthorized buildings can be found and very low-income groups live.

Diversity of tenant households -sociodemographic and economic characteristics
The demand for rental housing in the housing market is shaped by the socio-demographic and economic diversity of the tenant households. Through the survey in Ankara, 475 questionnaires were completed by the sampled tenant households. The socio-demographic and economic characteristics of the sample households are summarized in Table 3. Of the household survey respondents, 51% were male and 49% were female. While all of the male respondents were the heads of their households, only 10% of the female respondents were the heads of their households. Regarding the marital status of the households, 15% of the tenant households were in the single phase, 15% were in the honeymoon phase (married but without children), 55% were in the parenthood phase, 6% were in the elderly parenthood phase (departure of children), and 9% were in the separation phase (divorced/widowed). Regarding house sharing, 10% of the households contained a single person, 8% contained two or more unrelated people, 73% contained a nuclear family, 3% contained an extended family, 3% contained a person who was separated but with children, and 3% contained a person who was widowed but with children. The age groups of the household survey respondents were as follows: 10% were between 18 and 26, 42% were between 27 and 35, 32% were between 36 and 44, 8% were between 45 and 54, and 8% were older than 54. Regarding the educational level of the household survey respondents, 8% were primary school graduates, 17% were middle school graduates, 35% were high school graduates, 33% were university graduates, and 7% had master's/doctoral degrees. As for working status, 87% of the heads of households were working. Moreover, 84% of the working head of households were wage earners, 14% were selfemployed, and 2% were employers. Roughly 55% of the non-working heads of households were retired, 29% were students, and 16% were homemakers. Regarding the income level of the households, 13% presented the lowest income level, 22% had low income, 22% had middle income, 10% had high income, and 33% presented the highest income level. This proportional pattern of income level of the households can be evaluated as the result of the intentional sampling of the neighborhoods to represent all income groups in the metropolitan area. Lastly, looking at the rent-income ratio of the households, only 8% of households had a rent-income ratio that was over 30% of their income, whereas 92% had a rent-income ratio that was less than 30% of their income.

Tenant households' housing mobility, reasons for moving
The mobility of tenant households is accepted as an indicator of the flexibility of tenancy or the tenure security-related issues and it reveals the motivations for tenancy. Thus, the housing tenure and housing mobility of the sampled households are examined in this part. First, the changing housing tenure choices according to mobility are examined, then the reasons for the mobility are discussed, and lastly, the reasons for preferring the current house according to income level are put forward Table 4 shows the changing housing tenure choices of the households according to housing moves. Of the respondent tenant households, 35% were in their first house that they had lived in with their parents before they became an independent household. Of the respondent households, 23% stated they had moved twice, 30% stated they had moved three times, 12% stated they had moved four times, and 1% stated they had moved five times in their whole life. Of the households, 9% either had been owner-occupiers once before or still owned a house but stayed as tenants due to various factors.
The reasons for the tenant households' housing mobility are important to display the flexibility or security features of tenancy. Table 5 presents the reasons for respondent households' housing mobility. For first moves, 84% of the households put forth marriage as the reason, while 16% proposed life course reasons (6% related to employment/appointment issues, 6% related to the desire to live independently, and 4% related to education issues). For second moves, however, the lifecycle reasons decreased to 12% (1% related to marriage, 8% related to parenthood (i.e., childbirth), and 3% related to divorce). Employment/appointment comprised 6% of the second move as the only life course reason. On the other hand, 41% of the reasons for second moves related to housing conditions (28% related to size, amenities, location, etc.), 13% related to compulsory conditions (e.g., disposal of the homeowner, demolition of the building), and 41% related to economic conditions (e.g., affordability). In third and later moves, 74% of the reasons were related to affordability conditions, 16% related to housing conditions, and 3% related to compulsory conditions. Economic levels of households are an important determinant of preferences. Table 6 presents the reasons for the preference for the current house according to the income levels of the households. The statistically significant variables affecting the preference for the current houses according to income level are as follows: (1) economic factors and housing quality: appropriate rent values, adequate house size, security, assets (playgrounds, parking areas); (2) accessibility: adequate transportation services, proximity to shopping centers; (3) social factors: prestigious neighborhood, proximity to relatives and fellow citizens (hemşehri), close neighborhood relations. For higher income-level households, the most stated reasons for preferring the current house were adequate house size and prestigious neighborhood. For middle income-level households, they were adequate house size, adequate transportation services, and appropriate rent values. For lower income-level households, they were appropriate rent values and adequate transportation services.

The desire of tenant households to become homeowners
11.7% of tenant households with the lowest income are considering to own a home, while 59.8% of tenant households with the highest income are planning to be homeowners. In other words, as the income rises, the desire to become homeowners increases ( Table 7).
50% of the lowest-income tenants are considering being homeowners as they are tired of being a tenant. 46.9% of low-income tenant households, 38.8% of middle-income tenant households, 32.4% of highincome tenant households, and 56.2% of the highest income tenant households are considering to be homeowners because they find the rents high (Table 8).
In the lowest income group, the high rents and the lack of satisfaction with being a tenant strengthen the desire to have a home. In all of the other income groups, the high level of rents strengthens the desire to have a home. In other words, each income group may find rents high due to their own conditions or the market they are in and therefore considers it to be a homeowner.
Low, medium and high tenant households who plan to become home owners plan to become homeowners by using housing loans with the highest rate. On the other hand, 42.9% of the tenant households in the lowest income group considering to be home owners are planning to become a home owner by applying to the Mass Housing Administration. On the other hand, the monthly loan repayment power of tenant households in the lowest income group considering to be homeowner is less than TL 500, while the monthly lending power of lowincome households is 500-749 TL. However, when loan interest rates and monthly payments in Turkey in the period when the survey took place are taken into consideration, it does not seem very realistic that tenant households from this income group could to become homeowners, in particular those who lack a separate accumulation of a certain amount of wealth. In addition, considering the limited opportunities based on draw-off or application priority of the Mass Housing Administration, it is not very realistic for lowincome tenant households to have housing.
Apart from the above, based on the responses of tenant households who are thinking of home ownership to the question of where you would like to live if you become a homeowner, there is a statistically significant relationship between income  T  35  35  165  165  T-T  18  23  87  109  O-T  5  22  T-T-T  26  30  124  140  T-O-T  2  8  O-T-T  2  8  T-T-T-T  11  12  53   and the desired neighborhood/area. All the tenant households with the lowest income responded "it does not matter", but the tenant households with the highest income stated their choices clearly. For the tenant households who responded "it does not matter" being homeowners takes precedence other   criteria such as social prestige, properties of the house, environment and so on.

Diversity of tenant households according to income level -discriminant analysis
Different socio-demographic and economic characteristics contribute to the diversity of tenant households. The question regarding which variables explain the differentiation of tenant households with different income levels was investigated by discriminant analysis. After eliminating the nonexplanatory variables, the analysis was done with 14 variables. As shown in Table 9, eight variables out of 14 were statistically significant. The following variables were statistically significant in the diversification of the tenant households according to income level: affordable amount of monthly credit, monthly rent paid, monthly transportation cost, annual rent increase rate, age of dwelling, number of houses owned by the household, number of people working in the household, number of rooms in the dwelling. The household's affordable amount of monthly credit was the most effective variable discriminating the tenant households according to income level. According to the results of the discriminant analysis (Table 9), the function of "the diversity of tenant households according to income groups" is as follows:

Discussion
The results of the study indicate that interrelated sociodemographic and socioeconomic variables explain the motivations of tenancy and the diversity of tenants. As marriage is still a pervasive social institution that orients the relations in the society, parenthood is the dominant lifecycle phase in which the members of the core family are sharing the house. Mostly, the heads of the tenant households are working as wage earners are educated and run approximately middle incomelevel households, where the rent-income ratio is satisfactory. In addition, there are tenant households in the single and honeymoon phases where the heads of households are retired or students at lower or higher income levels. Marriage, young family members' desire to live independently, education, and employment/ appointment issues present the temporary situation of tenancy. The study indicates that lifecycle/lifecourse issues have an important role in the choice of tenancy, especially in the first housing moves of tenant households. In the later moves caused by changes in the size of the household or changes in its objectives and requirements, housing condition facilities (size, facilities, location) become important. However, economic affordability is becoming the major reason at an increasing rate. Depending on the economic situations of the households, rental housing is either a step on the way to owner-occupier housing or a chronic housing tenure without an alternative. Reasons for preferring the current house differ according to income level. While the size of the house is the most stated reason for middle and higher income-level tenant households, rent values is the most stated reason for lower income-level tenant households. While the second most important reason is adequate transportation services for lower and middle income-level tenant households, it is This also shows that tenancy is a housing tenure choice of all income groups, but as income levels rise, rental housing becomes an alternative to owner-occupier housing.
In the light of all evaluations, the demand and supply of rental housing can be summarized as depicted in Table 10.

Conclusion
Despite the desire to own property housing in developing countries, the choice of housing, especially for households with low incomes, takes place in a very limited environment. Because the cost of housing production individually is very high. For this reason, especially in developing countries, households with low incomes resort to informational regulations (Ballesteros 2004). However, the demand for rental housing increased markedly in the 1980s and 1990s due to factors such as increased urbanization with the global trend (Peppercorn and Taffin 2013), especially in developing countries, due to factors such as land shortages, increased land and house prices, and reduced tolerance to informational settlements (Peppercorn and Taffin 2013).
This forces governments in metropolitan cities to solve the housing problem in a more realist way. Colombia, South Africa and Indonesia, for example, are working to develop the rental housing market as an alternative housing option in urban areas (Ballesteros 2004). Because rental housing is thought to offer great opportunities to improve the living standards of narrow income group individuals and households (UNHCS 1993).
In order to prevent threats to rental housing provision and demand in Turkey (especially for lowincome households), new policies and regulations are needed in the rental sector. The implementable and functional policy alternatives should involve the The housing finance system for housing ownership is not for the low income group Limited number of property housing production by the public for the low income group restructuring of roles and responsibilities of central government and municipalities. Central governments and Housing Development Administration should be in a position of controlling, regulating and subsidizing agencies. This study shows that many demand-side factors play important roles in tenancy, rental housing mobility, and rental housing affordability, which implies the importance of housing policies on the demand side. Private rental housing can improve the functioning of the housing market by promoting housing mobility, increasing housing options for households, and generating a competitive supply and thus affordable prices. In Turkey, the affordability problem has been partly reduced in urban housing markets by supplying regular housing at vastly different prices in different parts of the urban area (Alkan, 2017). However, while the growing size and quality of new houses means more and better housing services per housing unit, they are also reducing housing choices available for low-income households and forcing them to spend greater shares of their limited income on housing (Dong 2017). Many people cannot afford to buy a house. For them, unsubsidized tenancies in the lowest-priced sections of the housing stock are the only option. While owneroccupier housing is stimulated, the private rental housing market for low-income groups suffers the negative consequences. It is therefore essential to promote private rental housing in favor of the lowand middle-income groups. In this respect, the tenure neutrality approach (De Boer and Bitetti 2014) could significantly improve the position of the private rental housing market. The challenge is to develop public policies that see tenures as complementary, enabling households to meet their housing needs more flexibly according to their socio-demographic and economic circumstances.

Disclosure statement
No potential conflict of interest was reported by the author(s).

Notes on contributors
Aysu Uğurlar, graduated from Dokuz Eylul University, Faculty of Architecture, Department of City and Regional Planning in 1999. She received her M.R.P (2006) and her Ph.D (2013) from City and Regional Planning, Gazi University. She worked as Research Assistant in City and Regional Planning Department of Gazi University. Recently, she works as Assoc. Prof. Dr. in the Department of City and Regional Planning of Van Yuzuncu Yil University. Her research interests are; housing policies, urban planning, tourism, sustainability indicators, local economic development.
Tanyel Özelçi Eceral, received B.CP (1990) from City and Regional Planning Department, Faculty of Architecture, Middle East Technical University; M.RP (1993) from the Regional Planning Program and Ph.D. (2002) from the City and Regional Planning Program of Middle East Technical University. She had worked in Ministry of Culture and GAP Regional Development Administration. Recently works as Prof. Dr. in the Department of City and Regional Planning, Faculty of Architecture at Gazi University. Her research interests are housing studies, urban economics, and regional and urban planning.