HOUSING DEMAND, TENURE CHOICE AND HOUSING POLICY IN BRAZIL

Maria da Piedade Morais Bruno de Oliveira Cruz Institute of Applied Economic Research, Brazil Fourth Urban Research Symposium 2007 World Bank,Washington, DC, May 14, 2007

Paper Overview I. II. III. IV. V. VI. VII. ƒ

Introduction Survey of the Literature on Tenure Choice Brief Retrospect of Housing Policy in Brazil The Data and Methodology Empirical Models and Results Conclusions and Policy Implications Bibliography APPENDIX- A model of tenure choice: the case of formal and informal housing markets

Paper Motivation ƒ Housing Policy in Brazil and Latin America with emphasis on the promotion of homeownership ƒ Lack of studies on the tenure choice of the Brazilian households ƒ Housing accounts for 30% of the total stock of physical capital in Brazil (IPEADATA, 2004) ƒ Housing corresponds to the main asset in household´s portfolio ƒ Housing tenure conditions vary deeply across countries, irrespective of income patterns, region of globe and levels of development

Ownership Ratios in Seletected Countries of Europe and Latin America 90,0

80,0

70,0

60,0

50,0

40,0

30,0

20,0

10,0

Source: CEPAL, IBGE and Housing European Statistics 2002

G er m an y

ai ca Ja m

la nd Ho l

Fr an ce

Fi nl an d

Ur ug ua y

ad or Sa lv

El

Po rtu ga l

a Bo lív i

Un i

te d

Ki ng do m

l Be lg iu m

Br az i

o Ar ge nt in a

éx ic M

To ba go e

as ad

Tr in id

Ho nd ur

ue la

Ire la nd

Sp ai n

Ve ne z

Ni c

ar ag ua

0,0

Paper Objectives ƒ Analyze the main determinants of tenure choice in Brazil in formal and informal housing markets using micro-econometric techniques ƒ Help in the design of housing policies better adapted to housing demand and income levels, and more effective to meet the housing needs of the Brazilian population

Questions Explored in the Paper ƒ The study tries to elaborate on: 9 1) What are the driving forces behind the housing tenure choice of households? 9 2) What are the tenure options faced by households either in developed or developing countries? 9 3) Do poor households have tenure choice or squatting and precarious rental and sharing arrangements in informal settlements are the only options open to them to satisfy their housing needs? 9 4) What kind of housing policies should be promoted by the government in order to meet housing demand?

Housing Policy in Brazil ƒ ƒ ƒ ƒ ƒ ƒ

ƒ ƒ

Until the 1930s the private rental sector was crucial in the provision of housing for fast growing population in Rio and SP After the 1940s official housing policies start to advogate the benefits of homeownership Desincentive to rental housing and urban renovation projects in central areas, under rapid urbanization, led to the emergence of slums and peripheral settlements in main MAs The official policy for informal settlements until the 70s was erradication and relocation to the periphery Goverment Housing Policies with poor performance and regressive impacts in the reduction of housing shortage Loans to medium and high income class absorbed the bulk of credit to self-owned housing (from 1964-1986 BNH financed only 25% of the increment of the housing stock and less than 20% of beneficiaries have income below 5 minimum wages) Recently, promotion of social housing through slum upgrading, mutualhelp and urban regularization, but accounting for less than 10% of Housing Funding Low income population with little payment capacity what restricts their access to formal finantial, rental and owner-occupied housing markets

ƒBrazil – Tenure Conditions of urban and rural Private Dwellings 1940-2000 80 69,8

70 52,2

60 50

57,4

60,3

74,4

61,7

43,7

40 35,4 24,7

30 20

22,5

19,0

22,5 16,4

20,9

11,4

23,1

10

20,1

20,7

15,8

14,3 13,8

0 1940 Owner

1950 Renter

Source: IBGE – Demographic Census

1960

1970

1980

1991

2000

Other Condition and Without tenure declaration

1 0,9

ƒHousing Expenditures inequality in Brazil is even higher than income inequality

0,8 0,7 0,6 0,5 0,4 0,3 0,2 0,1 0 0

até 400 Reais

400 a 600 reais

600 a 1000

1000 a 1200

1200 a 1600

1600 a 2000

2000 a 3000

3000 a 4000

4000 a 6000

Aquisiçao imovel Housingdepurchase

Ref orma ou melhoria do imovel Housing Improvement

Pagament o de fDownpayment inanciament o de imovel Mortgage

Renda al TotalTotIncome

Renda do Trabalho Labor Income

Renda não monet ariaIncome Non-monetary

Acima de 6000

The Data and Methodology ƒ The data comes from the Brazilian Census Bureau (IBGE) 2005 National Household Survey (PNAD) ƒ Sample selected through an extended concept of “urban areas” (3 urban sectors as classifief by IBGE and rural of urban extension) ƒ After Weightening microdata to be representative of the country as a whole, our sample covers 44,949,283 households

The Data and Methodology ƒ Based on information regarding modes of occupancy of the dwelling unit, property rights over land and type of sector (substandard or not) we have defined 4 categories of tenure: 1) Formal owners: owners owns the house, owns the land and the dwelling unit is not located in a substandard area; 2) Formal Renters: Renters Rents or rent-free outside substandard area 3) Informal Owners: Owners owns the house but not the land or has other tenure condition such as squatter, owns in a substandard area (slum dweller) or both 4) Informal Renters: Renters rents or rent-free in a substandard area.

The Data and Methodology Tenure Condition Valid Formal Owner Formal Renter Informal Owner Informal Renter Total Missing System* Total

Frequency

Valid Percentage Percentage

Cumulative Percentage

2,999,3897

66.7

7.9 72.9

7.9 72.9

8,067,093

17.9

19.6

92.5

2,907,590

6.5

7.1

99.5

191,112

0.4

0.5

100.0

41,159,692 3,789,591 44,949,283

91.6 8.4 100.0

100.0

Source: the authors, based on PNAD microdata

The Data and Methodology ƒ We have used four main blocks of explanatory variables: 1) Life Cycle and Household Characteristics: Characteristics Age of the Household Head, Household size, Marital Status 2) Wealth and Permanent Income: Income Per capita income, Household income, Years of schooling of head, Wealth proxy 3) Social Vulnerability and Credit Constraint: Constraint Gender (single women with kids under 14), Migrant (up to 4, 4 to 9, 9 or + years), race (non-afrodescent), Economic dependency (head income/total income), Labor Market status (formal employees, public servants, employers) 4) Location variables : metropolitan areas, large cities, macroregions

The Data and Methodology ƒ We have constructed a proxy for household wealth based on housing conditions and access to durable goods ƒ Adequate housing conditions (UN-Habitat definition)decrease by 1 point for each attribute missing : ƒ ƒ ƒ ƒ ƒ ƒ

piped water from public network, sewage network or septic tank electricity permanent walls Permanent ceiling Not overcrowded (less than 3 people per dormitory)

The Data and Methodology ƒ Access to durable goods : increase in 1 point for each desirable good ƒ Fridge (1.5 point for 2 doors) ƒ TV ƒ Freezer ƒ Computer ƒ Internet ƒ Washing Machine

ƒ Size/quality of the dwelling unitunit number of

bathrooms per person :decreasing 0.5 point per each quartile, 1st 1, 4th -1

Social Vulnerability

Locational Variables

Income Wealth

TENURE CHOICE

Life Cycle

Formal Owner

Formal Renter

Informal Owner

Informal Renter

Empirical Model and Results ƒ Cross-Section Logit e Multinomial Logit Models: ƒ Logit (owners vs. renters) ƒ Multinomial Logit 1 (formal owners, formal renters and informal dwellers) ƒ Multinomial Logit 2 (formal owners, formal renters, informal owners and informal renters)

Empirical Model and Results ƒ The forecast performance of the Multinomial Logit Models that include informal arrangements in the dependent variable is superior to the simple Logit Model ƒ Life cycle variables such as age of head, household size and marital status increase the probability of homeownership ƒ Migration has a negative impact on homeownership but this impact dissipates over time

Brazil 2005: Impact of Age on tenure choice 100,0% 90,0% 80,0% 70,0% 60,0% 50,0% 40,0% 30,0% 20,0% 10,0% 0,0%

Less than 18

18 a 24

formal owner

24 a 30

30 a 35

formal renter

35 a 45 Informal Owner

45 a 55

55 a 65 Informal Renter

more than 65

Logit Model: Impact of Age and Migration over homeownership

1,2

Probability

1 0,8 0,6 0,4 0,2 0 age Native or other condition

Recent Migrant

Empirical Model and Results ƒ Locational variables are significant and present the expected signs: living in large cities decreases the probability of homeownership ƒ Education increases the probability of being in formal housing markets, either as an owner or a renter ƒ An employment in the Public Sector increases the probability of becoming a homeowner in formal housing markets. Formal employees present a negative sign, contra-intuitive result

Impact of Education on tenure choice 80,0% 70,0% 60,0% 50,0% 40,0% 30,0% 20,0% 10,0% 0,0% 0,00

1,00

2,00

3,00

4,00

Formal Owner

5,00

6,00

7,00

Formal Renter

8,00

9,00 10,00 11,00 12,00 13,00 14,00 15,00

Informal Owner

Informal Renter

Empirical Model and Results ƒ Wealth is a good predictor for formal ownership ƒ Current income is statistically significant but has little impact ƒ More vulnerable households such as the poor, the afro-descendents or single women with children under 14 have a higher probability to beeing in informal sector, showing that they have limited tenure choices, confirming results in the literature (Gilbert, 1993; Cocatto, 1996; Nenochea, 1987)

Logit Model: Impact of Wealth on the probability of homeownership

0,6

Probability

0,5 0,4 0,3 0,2 0,1 0

-7,5

-5

-2,5

0

2,5

5

7,5

10

wealth

Policy Consequences ƒ Policy makers in Brazil, should consider wealth and income distribution and point in the life cycle in the design of housing programmes, such as incentives for first homeownership or subsidies to rental housing for young people ƒ The pattern of expenditures of the Brazilian households on home improvement shows the opportunity for microfinance programmes for progressive housing

Policy Consequences ƒ Policy makers should not focus exclusively on owner-occupied housing ƒ A wider range of housing options with different modalities, prices, qualities and locations should be available ƒ Renting must be recognized as both a respectable and necessary housing option ƒ Rental housing can offer good housing solutions for young people and newly arrived migrants in search of employment,with importance of rental markets to aleviate housing shortage increasing in urban agglomerations and fast growing urban areas

Recommended Future Research ƒ Estimate separate models for each metropolitan area or estimate tenure choice at intra-city level, taking into account location and tenure choice decisions based on Gibb (2000) or Elder and Zumpano (1991) ƒ Developed a dynamic analyses of tenure choice based on pseudo pannels from the 1992-2005 PNAD series, to take into account effects of inflation on tenure user cost of owing vs. renting. The effects of tenure insecurity can also be better captured on a dynamic framework. ƒ Improve information on tenure security and property rights in PNAD and further refining of social vulnerability and credit constraint variables ƒ Study the impacts of precarious insertion in the labor markets in new households formation rates (Sharing, Late Stayers)

Contacts ƒ IPEA Homepage: https//: www.ipea.gov.br ƒ Maria da Piedade Morais E-mail: [email protected] ƒ Bruno de Oliveira Cruz E-mail: [email protected]