Determinants of urban housing choice in Debre Berhan Town, North Shewa zone, Amhara Region, Ethiopia

Abstract

The main objective of the study is to examine the determinants of urban housing choice in Debre Berhan town, Amhara Regional State, Ethiopia. For this purpose, the study used both primary and secondary data. The data were obtained from 395 household heads selected from four kebeles. A multinomial logit model was used to examine the determinants of urban housing choice. Accordingly, it is found that age, gender, educational level, the price of the housing, years of residence, income of the head and access to credit are the major the determinants of urban housing choice. This study suggests that town administration, zonal and the regional governments should work together with financial institution to provide housing loans to the households so as to construct or purchase house.

Keywords:

• Housing
• choice
• multinomial logit
• Debre Berhan

PUBLIC INTEREST STATEMENT

Nowadays, having residential house is very essential to the residents of urban households to improve the living standards of the livelihoods. This paper examines the determinant of urban housing choice in Debre Berhan Town. The study results confirmed that age, gender, educational level, housing price, years of residence, income and access to credit are the main variables that determine urban housing choice in Debre Berhan Town. This study suggests that the government and stakeholders should take this into account to improve urban housing choice in the study area.

1. Introduction

Housing is basic human need, which has evolved from simple shelters to modern housing units. It is the largest fixed capital investment and acts as a focus of economic activity, a symbol of achievement, social acceptance and an element of urban growth (Malpezzi, 1995). Housing choice can be defined as being able to choose a preferred option from a set of distinctive alternative housing unit (Brown & King, 2005). In many developing countries around 40% of urban residents are living in rental housing units. Accordingly, demographic changes, income level, the availability and location of housing units are some of the factors that prevail all over the world (Council, 2015). Ethiopia is the second most populated African country; it is one of continent’s least urbanized countries. Only 20.7% of Ethiopian residents live in urban settlements. There are 15,103,134 housing units among that 12,206,116 units are found in the rural areas and the 2,897,018 units are found in the towns of the country. In urban parts of Ethiopia, 39% of the owner-occupied housing units and about 40% of the urban housing units are rented from private households. Abnet et al. (2017) indicated that private rental arrangement is prevalent in most urban parts and owner occupancy was the second major tenure arrangement in Ethiopia. There are different scholars, who have been conducted on the determinants of housing demand in developed countries specifically in Africa for instance, Muthoka (2015) and Omtatah (2014) in Nairobi, Kenya, were conducted on household demand for housing. But, in Ethiopia limited studies have been conducted on housing provision and challenges of urban residents in Addis Ababa city. Muleta (2014) on assessment of the housing provision challenge for urban residents in Addis Ababa in Ethiopia. Getachew (2016) conducted an assessment of affordability and living condition of condominium housing in Addis Ababa: the case of Lideta Sub city in Addis Ababa, Ethiopia. The previous study focused on factors affecting the challenges of urban housing demand in Ethiopia. As per researcher knowledge there is not a research conducted on urban housing choice in Debre Berhan Town. However, given the major research and knowledge gap, this study intends to examine the determinants of urban housing choice in Debre Berhan Town, North Shewa Zone, Amhara Region, Ethiopia. The results of this study expected to create awareness about the determinants of urban housing choice, it can help policymakers to formulate appropriate policies and provide meaningful information for urban housing municipality and the surrounding community to take some action with regard to the determinants of urban housing choice in Debre Birhan Town. The rest of the paper is structured as follows: the next section provides “literature review” on the determinants of urban housing choice and related issue which supports the study from different countries. The “Methodology” section describes data and econometric analysis used in the study. The findings and discussion are presented in ‘Results and Discussion”. The final sections offer some “Conclusion and Recommendations”.

2. Literature review

Housing refers to houses or buildings collectively, accommodation of people, planning or provision of accommodation by an authority, and related meanings (Wright, 1983). In Africa, in general and Sub-Saharan African countries in particular, the rapid increase in urban populations, urbanization, the persistently poor financial situation of the urban residents, bad governance and material resources available are the causes of increasing urban housing demand. Ethiopia is one of the poorly developed countries which are characterized by low per capita income, higher population growth rate, rapid urbanization, import-dependent, poor investment in housing because of lack of finance and low supply of serviced residential plot (UN-Habitat, 2014).

Jayantha and Oladinrin (2020) conducted on an analysis of factors affecting homeownership in Hong Kong. It was found that homeownership cut across different age groups and demographic status. Although, too high house prices, general cost of living pressures, and high required upfront deposits are factors affecting homeownership.

Spalkova and Spalek (2014) have been conducted on housing tenure choice and housing expenditures in the Czech Republic using probit regression model based on the sample data. The results of the analysis show that household income, marital status of the household head, household size and residence are factors that affect tenure choice. Olanrewaju and Woon (2017) stated that general, financial, building, income, accessibility, market and location factors are determinants of affordable housing choice in Lokoja, Kampar, Malaysia.

The residential choices are influenced by age, income, and size of the household, as well as by the rent-to-income ratio. An increase in any of these variables decreases the probability of choices of all the alternatives other than the most often chosen alternative. Moreover, the distance to work systematically influences the housing choice for single-parent families and two-earner households in Lille, northern France (Flambard, 2017).¹

Studies on the determinants of housing choice have been carried out in Kenya and in Ethiopia. For example, Muthoka (2015) revealed that people are changing from rented tenures to purchase only due to an increase in household size and age of the household head and also household size and age of household head are a key and positive effect to constructing their own house. Omtatah (2014) reported that income proxied by GDP, the number of households and housing prices are the major determinants of demand for housing in Nairobi city. Mulu et al. (2017) conducted the performance of condominium housing program in Jimma Town, Ethiopia. The study reveals that condominium housing is not affordable to the poor section of the town. Getachew (2016) revealed that the lack of water supply, open space, the size of the rooms and domestic noise are considered as the main challenges of living in condominium housing in Addis Ababa city, Ethiopia. Regassa and Regassa (2015) conducted on housing and poverty in southern Ethiopia examining affordability of condominium houses in Hawassa city. The study revealed that full-time employment may have better affordability status than that of the part-time worker.

3. Research methodology

3.1. Study area profile

Debre Berhan or Birhan, formerly spelled Debra-Birhan or Bernam, is a city and woreda in central Ethiopia. Located in the North Shewa Zone in the Amhara Region, about 130 kilometers northeast of Addis Ababa on the paved highway to Dessie, the town has a latitude and longitude of 9°41′N 39°32′E and an elevation of 2,840 meters. It was an early capital of Ethiopia and afterward, with Ankober and Angolalla, was one of the capitals of the kingdom of Shewa (Chisholm, 1911). Today, it is the administrative center of the North Shewa Zone of the Amhara Region. Debre Birhan town has nine kebele with a total population of 103,450 whom 46, 553 are men and 56,897 women (CSA, 2014) Figure 1.

Figure 1. Geographical location of Debre Birhan Town of North Shewa

3.2. Research approach

The study used both quantitative and qualitative research approach. In analyzing the quantitative data presented with the help of tables, percentages, frequencies and figures. Furthermore, it measures many characteristics, which are naturally numeric in nature (such as years of education, age, income, etc.) Whereas, qualitative data are also analyzed by using the narrative mechanism.

3.3. Sampling method and size

For this particular study, a multi-stage sampling procedure was employed. In the first step, out of twenty-seven woredas in the zone, Debre Berhan Town were purposively selected for this study. The reason for purposive sampling was, related to the opening of different industries and factories housing problems aggravated. There are nine kebeles in Debre Berhan Town. In the second stage, out of nine kebeles four kebele were selected randomly. In the third stage, the sample size for each selected kebeles was determined proportionally to the number of households within each kebele. Finally, 395 sample households were selected from the total households of four kebeles by using random sampling technique. This study applied a simplified formula provided by Yamane (1967) to determine the required sample size at 95% confidence level and 5% the margin of error. Yamane formula is expressed as;

$n=\frac{N}{1+N\left(e^2\right)}=\frac{32,579}{1+32,579\left[0.0025\right]}=395$

Where: n = sample size; N = the total number of household lives in four kebeles and ε = error tolerance or margin of error. Sample respondent distribution from each kebele is depicted as follows;

3.4. Data collection methods

Cross-sectional data from primary and secondary sources were collected as well as structured and pre-tested questionnaire was applied to collect primary data. Before data collection, pretesting of the questionnaire have been carried out, and then depending on the results, there were some adjustments that would be made to the final version of the questionnaire. Secondary data were gathered from documented and published sources including books, journals, government reports, articles, reports from North Shewa urban municipality Development Bureaus and other publications.

3.5. Methods of data analyses

The data were analyzed using descriptive, inferential statistics and econometric methods. In this study, descriptive statistics were employed to gain a better understanding of the socio-economic, institutional and demographic characteristics of the respondents. Inferential statistics are used to determine the relationship between a variable and making a prediction. Descriptive and inferential statistics such as mean, percent, standard deviation, chi-square test and likelihood ratio test were used to analyze the data that was collected from sample household heads. To examine the determinants of urban housing choice were carried out using econometric methods. Particularly multinomial logit model. For this analysis, STATA software version 13 and SPSS software version 25 were used.

3.6. Model specification

The most frequently used model in studies of housing choice is the multinomial logit model. That model can be derived from the assumption that the error terms are independently and identically distributed (Timmermans et al., 1994).Thus, in this study housing choice variable mainly categorized as constructed, purchased and rented housing choices in the study area. But the public houses aren’t included in this study; since the construction of condominium and kebele of house stagnant in the study area, consequently the households are forced to live in his/her own house by constructed, purchased, otherwise rented house. This study assumes that the urban housing choice can be used for three (3) mutually exclusive housing units. At a particular time, the household could be only living in the privately rented house, only living own constructed house, and only in the purchased house. This gives rise to a polychotomous choice framework. Hence, the probability of home ownerships (j = 0 private rented house only; j = 1 own constructed house only; j = 2 purchased only) is given by the following multinomial logit model. The multinomial probability model assumes that the possible distinct states are exhaustive in that they cover all possibilities. In the multinomial case, the rented house is considered the base level and all the logit are made relative to the base category. When category k is taken as a base category, and let be the multinomial probability of an observation falling in the jth category, then the multinomial logit model is specified as follows:

(2) $Pr(Yi=j)=\frac{exp(Xi\beta j)}{1+{\mathrm{\Sigma }}_{k=1}^{N}exp(Xi\beta k)},\mathrm{F}\mathrm{o}\mathrm{r}\mathrm{j}=1,2\dots (\mathrm{k}-1)\mathrm{a}\mathrm{n}\mathrm{d}\mathrm{i}=1,2,3\dots \mathrm{n}$(2)

Where, Yi is housing choice, βi is the vector of parameters and Xi is a vector of all explanatory variables those are age of household head, gender of household head, marital status head, family size, educational level of the household head, employment sector, income of household head, access to credit, years of residence and housing price and ε_i is the disturbance term of the equation. The interpretation of the multinomial logit model is relative to the reference or base category group is difficult, even if this study used rented only as a base category. The coefficients need to be adjusted to be marginal effects in the case of the logit model.

3.7. Description of the model of variable and its measurement

4. Results and discussions

4.1. Descriptive statistics

The results of the study focused on the determinants of urban housing choice. The finding has been presented and analyzed the demographic and socioeconomic characteristics of the respondents and to examine the determinants of urban housing choice in the study area.

Table 3 shows that, the mean age of the household head is 48.352 years with standard deviation of 12.075; a minimum age of 20 years and the maximum age 79 years. The mean family size is 3.929 with a standard deviation of 1.847, a minimum one and maximum nine family members. The mean income of the household is 6781.772 ETBs with standard deviation 4375.311 and minimum 650 and maximum 23,000 ETB. The mean educational level of household is 12.97 years with 3.821 standard deviation and min 0 and max 21 years enrollment. On the subject of the mean years of residence is 40.306 years with standard deviation 18.573 years and min 4 and 79 years. Finally, the mean price of housing is 197,000 ETB with standard deviation 276,000 and the min 300 and max 1,500,000 ETB Table 1Table 2.

Table 1. Selected kebele and sample distribution in Debre Birhan Town

Name of sampled kebeles	Household size	Sample size
03	7640	93
04	8331	101
06	8319	101
09	8289	100

Table 2. Measurement and description of variables

Variable Name	Description and measurement of variables	Expected sign
Housing Choice	It is a categorical variable (rented house = 0, own constructed = 1, and purchased = 2).	Dependent variable
		Construct	Purchase
Age	Defined as the number of years from the date of birth to the day of the survey interview date in the full year. This is measured to in years.	+ve	+ve
Gender	It is a dummy variable which can be male = 1 and 0 otherwise).	+ve	+ve
Marital status	It is a categorical variable 0 if single, 1 widowed, 2 divorced and 3 married.	+ve	+ve
Family size	One or more persons occupying a housing unit and have the same eating arrangement. Measurement in number	+ve	+ve
Educational level of household head	It is a continuous variable. Measured in years of formal schooling.	+ve	+ve
Employment sector	It is a dummy variable which is categorized ‘0ʹ public sector, ‘1ʹ private sector.	+ve	+ve
Income	Permanent disposable income of a household measured in Ethiopian Birr (ETB)	+ve	+ve
Access to credit	It is a dummy variable which can be an individual 1 if access to credit, 0 otherwise	+ve	+ve
Years of residence	It is a continuous variable, which means an individual how many years stayed in the study area, which can be measured in year.	+ve	-ve
Housing price	It is a continuous variable, which means the quantity of payment given by one party to another in return for goods or services. Measurement in Ethiopian Birr (ETB)	+ve	+ve

Table 3. Summary statistics for continuous variables

Variable	Mean	Std.Dev.	Min	Max
Age	48.352	12.075	20	79
Family size	3.929	1.847	1	9
Income	6781.772	4375.311	650	23,000
Education	12.97	3.821	0	21
Years residence	40.306	18.573	4	79
Housing Price	197,000	276,000	300	1,500,000

Table 4 shows that the classifications of urban household housing ownership out of 395 households 209 (52.91%) privately rented house, 108 (27.34%) own constructed house and 78 (19.75) households are living in the purchased house.

Table 4. Percentage distribution on choice variable

Housing Choice	Freq.	Percent
Rented house	209	52.91
constructed house	108	27.34
Purchased house	78	19.75
Total	395	100.00

Table 5 displays that housing choice has strong relationship with gender of household head, marital status, access to credit and employment sector with p-value zero (0.000).

Table 5. Pearson’s Chi2 tests of the categorical variable against housing choice

Housing choice	Pearson chi2	p-value
Head gender	81.0383	0.000*
Marital status	91.5814	0.000*
Access to credit	115.8918	0.000*
Employment sector	27.2640	0.000 *

4.2. Econometric analysis

In econometric analysis, the study applies a method of analysis of maximum likelihood estimation technique for the purpose of estimating the multinomial logit functions. For the purpose of effective estimation of the model several diagnostics tests were tested. For instance, multicollinearity test, independence of irrelevant alternative test, Likelihood ratio test for independence of the variables, combination test among the alternatives of the model and goodness-of-fit test were seriously conducted for multinomial logit model.

4.2.1. Multinomial logit estimation results

Table 6 reveals that the parameters of the age of households are positive and significant at 10% for own construct houses and 5% for purchased house. The marginal effect indicates that as age of the household increase by one a year, then the probability of constructing own house and purchased increases by 0.423% and 0.23%, respectively, relative to renting a house, holding all other variables constant. This implies the older household more income and have more wealth (asset and income) which helps them to construct and purchase housing. This result is consistent with recent studies conducted in Kenya, which indicated that age of household head was significant variable in explaining housing choice (Muthoka, 2015).

Table 6. Multinomial logit coefficients and marginal effect rented house is a base level

Multinomial logistic regression				Number of obs = 395.000
				Wald chi2 (18) = 1631.89
				Pseudo R-squared = 0.723
Log pseudo likelihood = −110.51046				Prob > chi2 = 0.000
Housing Choice	Constructed house			Purchased house
	Coef.	Std.	Mfx(dy/dx)	Coef.	Std.	Mfx(dy/dx)
Age	0.0974*	0.0533	.0042299	0.234***	0.0648	.002298
Gender	10.95***	1.717	.2668742	9.202***	1.861	.041541
Marital Status	0.590	0.459	.0259704	0.664	0.530	.006366
Family size	0.297	0.377	.0131678	0.156	0.398	.001422
Education	0.226**	0.109	.0099412	0.238**	0.112	.002276
Employment status	1.268*	0.653	.0451711	2.238***	0.799	.016153
Income	0.000219	0.000203	9.43e-06	0.000710***	0.00022	7.00e-06
Access to credit	3.565**	1.751	.3882483	3.568**	1.801	.084075
Years of residence	0.0661** *	0.0229	.002952	−0.0129	0.0225	−.000160
Housing price	1.48e-05***	2.99e-06)	6.52e-07	1.34e-05***	3.09e-06	1.27e-07

Note: *, ** and*** are level of significance at 10%, 5 % and 1 % respectively²

Gender of household head is both positive and significant at the 5% level of significance. If household are male headed, then probability of constructing and purchasing a house increased by 26.68% and 4.15% with relative to rented respectively, all other things being equal. This is due to the fact that, males never leave the workforce for such expected events as childbearing and rearing. They have the opportunity to gain more experience in the workforce (by working continuously over their work life) and even more with a particular company. As a result, males often have higher incomes than female. This result is consistent with recent studies conducted in the Swedes, which indicated that gender of household heads was a significant variable in explaining housing choice (Niedomysl, 2008).

The educational level of household head is also positive and significant at the 5% level of significance. The marginal effect indicates that as household’s head educational level increase by one year, the probability of constructing and purchased house increase by 0.99% and 0.23% relative to rented house respectively, holding all other variables constant. This says as household heads with a high level of educational attainment will often have a good job with a generous salary. This result is consistent with recent studies conducted in Malaysia which indicated that education of the head is significant variable in explaining housing choice (Bujang et al., 2010).

Employment sector of household head is positive and significant at the 5% level of significance. It points out that households who were employed in the private sector more likely to construct, own house and purchasing a house, than counterparts, holding all other variables constant. The marginal effect indicates that as households who are employed in the private sector, then the probability of constructing and purchased house increase by 4.51% and 1.61% relative to rented house respectively, ceteris paribus. This implies that households are working in the private sector earn higher income than public sector. This result is consistent with recent studies conducted in the Hawassa city which indicated that employment sector is a significant variable in explaining housing choice (Regassa & Regassa, 2015).

Access to credit is positive and significant at 5% significance level. The marginal effect indicates that as household heads access to credit then, the probability of constructing own house and purchased has increased by 38.82% and 8.41% relative to rented house respectively, holding all other variables constant. This result is consistent with recent studies conducted in Malaysia which indicated that access to credit was a significant variable in explaining housing choice (Olanrewaju & Woon, 2015).

The relationship between the income of household head and housing choice is positive and significant for purchased house at the 5% level of significance. The marginal effect indicates that, as the income of a household increase by one ETB the probability of purchasing house increases by (7.00e-06) = 0.0007% relative to a rented house, while other things remain constant. This implies higher incomes mean greater demand for properties, as the general purchasing power of households increases and they choose to spend a greater proportion of their consumption on living, other things remain constant. This result is oppositely consistent with recent studies conducted in Kenya, which indicated that income was significant variable in explaining housing choice (Muthoka, 2015).

A year of residence was positive and significant for constructed house at 5% significance level. If household head years of residence have increased by one year, then, the probability of constructing own house increase by 0.295% with relative to a rented house, ceteris paribus. This implies that as households are staying for long years, then households fulfill some of the criteria to obtain land in the form of lease to construct own house. This result is consistent with recent studies conducted in the Czech Republic which indicated that years of residence were a significant variable in explaining housing choice (Spalkova & Spalek, 2014).

Lastly, the general price of house is both positive and significant at the 5% level of significance. The marginal effect indicates that, as the general price of a housing is increased by one Birr (ETB) the probability of both constructing own house and purchasing has increased by (6.52e-07) = 0.0000652% and (1.27e-07) = 0.0000127% with relative to rent, while other things remain constant. This implies that even if the price of housing increases consumers are willing to construct and purchase house, related to the income of household heads and fulfilling all services related to housing is increasing household pleasure results rise up willing to pay, lead to rising house prices. This result is consistent with recent studies conducted in Kenya, which indicated that general hedonic price was significant variable in explaining housing choice (Muthoka, 2015).

5. Conclusion and recommendation

The issue of urban housing choice is a major concern for developing countries, particularly in Ethiopia. The main objective of the study to examine the determinants of urban housing choice. A multi-stage sampling technique was employed to select 395 sample household heads and used structured questionnaire. From descriptive statistics, 27.34% own constructed house, 19.75 purchased house and 52.91% households are living in rented house in Debre Birhan Town. Multinomial logistic regression model was employed to identify the determinants of housing choice in Debre Berhan Town. As a result, the study found that, age, gender, income, access to credit, employment sector, educational level, years of residences and general housing price of the housing are the major determinants of urban housing choice, which prohibited from changing rented to construct and purchase. Finally, I recommended that households should work in private sector rather than waiting for employment in the public sector. The administrative of town must work with financial institutions to provide housing finance in the form of loan to be returned in the long run.

Acknowledgements

I would like to thank the editor and anonymous reviewers for the supportive comments and suggestions.

Additional information

Funding

The authors received no direct funding for this research.

Notes on contributors

Abebaw Hailu Fikire

Abebaw Hailu Fikire is a lecturer at Debre Berhan university, Ethiopia. His current research interests are development issues like urban and rural development. He has one paper under review dealing with the “Effect of COVID-19 on food security in North Shewa Zone”

Notes

1. Housing refers to the use of a house, plus associated services for a specific period of time the house or any buildings which are principally built to serve a single household or family for the residential purpose (Bethel, 2003). Kebele is (Amharic word) which means the lowest administrative unit in Ethiopia (Robi, 2011). Districts, or woreda (in Amharic) are the third-level administrative divisions of Ethiopia. They are typically collected together into zones, which form a region (Yilmaz & Venugopal, 2008).

2. Pseudo R2 tells us our model as a whole fit significantly better than an empty model (i.e., a model with no predictors).