The Effects of Economic Sector GDP on Low-Income Housing Supply, Colombia’s Regions Case

Abstract

The regions with the best economy have a greater capacity to develop low-income or social-impact housing, thus contributing to the reduction of poverty and, therefore, to the fulfillment of Sustainable Development Goals. This is observed in fewer people living in extreme poverty and with fewer unmet basic needs. This present article analyzes the correlation between the development in the main economic sectors in different regions (departments) of Colombia and the supply of low-income housing. Nevertheless, the most remarkable relation that was found was between the economic development of the regions (GDP) and the supply of non-social housing (more expensive commercial value) (Spearman’s Rho: 0.9). This means that there is an imbalance between regional economic capacity and the low-income housing supply because the regions with higher economic potential should have a less demanding population, that is, people living in poverty. These correlations are better when they go hand in hand with activities that are mostly developed in an urban environment, such as manufacturing, construction, real estate, and finance and insurance. On the contrary, these correlations are worse with industries such as mining and agriculture that mostly operate in rural areas. The analysis between low-income housing and economic sectors’ GDP yields low correlations, but these are worse for rural industries. Also, the investigation shows a positive change in the correlations’ trend for the year 2021, the beginning of the post-pandemic economic recovery.

1. Introduction

The unmet basic needs (UBN) are indicators that identify the levels of poverty in the regions. The trend in Latin America, and therefore in Colombia, is that in its regions (or departments) with the best economic level (measured from the Gross Domestic Product, GDP per capita), they have the lowest index of extreme poverty or UBN.

The UBN are measured in different socioeconomic aspects, including the material condition of the housing, the lack of basic public services, and overcrowding. These aspects are related to the need to generate low-income housing for the population with the greatest needs, and thus contribute to reducing their UBN and helping them overcome the condition of poverty. Additionally, the generation of low-income housing contributes to compliance with the Sustainable Development Goals (SDG). A review of the 17 SDG [1] shows that one of the common elements between No. 1 “No Poverty”, No. 6 “Clean Water and Sanitation”, and No. 11 “Sustainable Cities and Communities” is to generate satisfactory housing for the poorest communities. Whether it is new home ownership, improved home ownership, or rental housing, it must be provided with the minimum quality standards in order to help improve the indicators for these three SDG.

The construction of housing for low-income households is one of the options to help reduce poverty in a developing country such as Colombia, according to the Colombian Ministry of Housing, City, and Territory (2022) [2]. Low-income or social interest housing (VIS, in Spanish) has the social function of being the lowest-priced: “VIS has the elements that ensure habitability and that meet the quality standards of urban, architectural, and construction design. Their maximum value is equal to one hundred and thirty-five current legal monthly minimum salary (135 SMLM) (SMLM: in spanish: Salario Minimo Legal Mensual, which amount to 1.163 million Colombian pesos (COP) in 2023).” Housing units that, due to their sale price, exceed the VIS value, are called “no-VIS housing” and are not intended to favor low-income households.

The hypothesis posed by this article is that the regions with the highest level of income in Colombia are those with a tendency to develop fewer VIS housing units because they have a lower percentage of the population in extreme poverty and with fewer housing needs. This is supported by high correlation values and negative signs (Spearman’s rho < −0.7) for the 33 departments of Colombia (

Table 1. Rho value for the correlation between Colombian regions’ GDPs per capita, UBN conditions, and inhabitants in extreme poverty.

Percentage of UBN Inhabitants	Percentage of Inhabitants in Extreme Poverty or Misery	Percentage of Overcrowded UBN Inhabitants	Percentage of Housing for UBN Inhabitants	Percentage of Basic Public Services for UBN Inhabitants
−0.84	−0.82	−0.80	−0.75	−0.75

). In order, the highest correlations of GDP are with the percentage of UBN inhabitants, percentage of inhabitants in extreme poverty, percentage of overcrowding UBN inhabitants, percentage of housing for UBN inhabitants, and percentage of basic public services for UBN inhabitants. The rho values for the first three indicators are close, which could mean that the majority of inhabitants with UBN are in extreme poverty and live in overcrowded conditions. Also, another observation is that the coverage of home public services is relatively well-distributed at the national level, which is why the correlation between the GDP of the departments is not so strong with the number of people lacking these services (rho = −0.75).

A huge social problem is the relationship between economic growth and low-income housing. This is exposed by entities such as the Inter-American Development Bank (IDB) by stating that the housing sector, in Latin America and the Caribbean (LAC), has historically been affected by the economic situation in this territory, as recently demonstrated in the COVID-19 pandemic, and is related to the fulfillment of the SDG. For the IDB, it is clear that the development of the housing sector is a fundamental tool for sustainable and resilient growth in LAC [3].

2. Literature Review

2.1. The Importance of Low-Income Housing Development in Order to Reduce Poverty

It is important to highlight that the good economic situation of the region has contributed to the reduction of the number of households with housing in poor conditions and that Latin America and the Caribbean are the developing regions with the highest level of urbanization in the world. This fact is related to the increase in human settlements and poverty, and that housing development is essential for reducing social, economic, and environmental gaps. Additionally, it is even more worrying when you read the numbers: “If we add to this the high levels of poverty, labor informality and the slowdown in mortgage financing, we conclude that the probable scenario for the coming years is an expansion of precarious settlements in our region, beyond 17.7% of the population urban society that in 2020 lived in marginal neighborhoods” [4].

In another document, the United Nations Economic Commission for Latin America and the Caribbean (ECLAC) highlights the relationship between urban population growth, economic opportunities, and the housing construction sector: “Population growth presents a negative relationship with economic development. That is, accelerated population enlargement presents an opportunity cost in relation to economic development since rapid growth in the labor factor means that more capital has to be used to equip the increase of the labor force, which results in slower growth of capital per worker”, and “The construction sector includes both the creation of new homes and the recovery and rehabilitation of those that are disused and/or deteriorated. Its development not only impacts the most vulnerable population. In addition to alleviating poverty, this is a sector of great relevance within the economy, due to the impact it generates on other sectors. On the one hand, it demands from other industries the inputs used in construction works, inducing dynamism in the latter” [5].

According to the Ministry of Labor (Republic of Colombia) (2023) [6], about 70% of inter-municipal migrants, for lack of work reasons, arrived in the following departments: Bogotá, Cundinamarca, Valle del Cauca, Antioquia, Santander, Risaralda, Meta, Atlántico, and Tolima. Eight of these nine regions are on the list of those ten with the best GDP per capita since 2015 [7].

Figure 1 shows the strong and positive relationship between the growth of GDP per capita and the GDP of the construction sector for a sample of 20 Latin American countries between the years 1990 and 2020. The correlation between these variables is strong and positive. The growth of the construction sector influences the behavior of the GDP since its p-value is less than 0.01. The model exposed in the document finds that a 1 percentage point increase in the growth of the construction sector leads to an increase of 0.06 percentage points in the GDP per capita growth rate [5].

In Colombia’s case, when there is an economic crisis, low-income housing (VIS) is built the most. In order to sustain the construction and real estate sectors, a solvent demand is needed, which requires a population that possesses enough money to buy housing in the middle and upper social stratification segments. When there is a crisis, the housing demand reduces and that is why the State gives subsidies to leverage that demand, and it does so in low-income segments [8].

Now, regarding Latin America and the Caribbean, the importance of producing more quality social housing for the population classified as poor stems from the growth of the urban population coming from rural areas. The ECLAC presents projections for the population increase of the region in each of its countries [9]. According to the general data for Latin America and the Caribbean, the turning point between a higher percentage of the population living in urban areas compared to rural areas occurred in 1960. Since then, the difference has steadily been increasing and is projected to remain this way at least until the year 2050 (Figure 2). In the case of Colombia, the distribution of the population is very similar to the data below (Figure 3). Reviewing the population distribution figures for Colombia in 2023, they differ slightly from the projection in Figure 3 but maintain the same relationship: an urban population of 74% and a rural population of 26%, according to information from the National Statistics Department [10].

The above suggests that the so-called “urbanization phenomenon” will continue in Colombia and Latin America, generating a greater number of migrants from rural to urban areas. Most of this migration occurs under conditions of poverty, and as a result, families arrive in the municipalities to occupy housing with very poor material conditions and without basic public utilities. Unfortunately, many of the causes of migration are linked to violence and lack of economic and job opportunities in many rural areas [11,12,13]. In the case of Colombia, although the population living in poverty has decreased since the early 2000s, indicators continue to be higher in rural areas than in urban areas [14]. It is also noteworthy that in recent years, there have been many people living in poverty who have migrated from Venezuela to Colombia, mostly to major cities [15].

A review of the state of the art to explore the relationship between poverty and social housing shows that, according to Chiodelli (2016) [16], the great proliferation of informal human settlements in the major cities of so-called developing countries began after World War II. This informal growth of cities led to poverty, creating what can be described as a “planet of slums” [17]. For several years, government agencies failed to take significant action to address this issue. Only in the 1960s, and with greater force in the 1970s, did various government and multilateral agencies initiate housing policies consistently and with clear objectives to improve the housing conditions of poor households in the cities to reduce the reality of urban poverty. Over the years, it has been shown that the location of low-income housing in the urban periphery (generated largely by the effects of rural–urban migration) results in transportation issues on work or study days due to increased commute time and costs. It also brings restrictions to urban development [18].

The first initiatives to improve social housing focused on having the community itself actively participate in the construction or enhancement of their precarious housing. However, the 1980s proved that self-built housing was not a good idea due to the quality of the construction process. From then on, housing policies focused on the fact that the State should intervene even more with new housing projects and provide households with financial planning, financial assistance through subsidies, and even rental housing options. The latter option has not been given due consideration by many governments in poor countries, but it is an option for regulating social housing [19]. In the 1990s and 2000s, agencies such as the World Bank, the Inter-American Development Bank, and UN-Habitat have been very active in promoting policies for social housing valuation and land regulation [20]. Such housing policies strengthened the legal ownership of informal housing as a mechanism to reduce the state of poverty of households by increasing their net worth through home ownership [21,22,23].

Between 1995 and 2009, housing conditions in Latin America improved by reducing the number of households in the housing deficit from 8% to 6%. Similarly, the proportion of occupants living in housing built with precarious materials fell from 12% to 8.8%. These improvements in the housing deficit correspond to an era of economic growth, measured in per capita income. However, criticism must be made of the lack of consistency between public policies for urban development and housing development. This disconnect has led to housing projects not being a solution to the quality of life of the inhabitants [24].

Providing new housing or improving existing housing for lower-income households who cannot afford non-VIS housing is not solely about the actual building. Housing also implies providing basic public services and utilities such as water, sewage, roads, energy, and adequate public space [25,26]. This connotes that social housing in a country contributes to achieving SDG.

Some studies conclude that generating housing invigorates the economy by increasing the GDP [27].

2.2. The Role of Low-Income Housing in the Building Housing Sector

Continuing the study of Latin America, the analysis of the evolution of the economy and social housing in Argentina, Brazil, Chile, and Colombia concludes that in the last 50 years, there has been a lack of response from governments to the problem of social housing. Then, there was the expansion of informal settlements and the increase in the social housing deficit. Later, governments wanted to increase financing to reduce the social real estate crisis; finally, the construction sector participated in an important offering of social housing located in the urban peripheries [28].

The Colombian Chamber of Construction (Camacol) (Available in: https://camacol.co/, accesed on 3 December 2023) (considered to be the top consulting agency in Colombia for the building industry) has recently stated that “if the social housing sector is reactivated, the expected growth of the economy could be doubled” [29].

The social housing policy in Colombia in recent years has been based on government subsidies to low-income families for the purchase or improvement of social housing. This financial aid encouraged the supply of social housing by construction companies, mainly in the regions of the country that had the greatest demand, that is, regions with greater economic movement and the presence of migrant households. Furthermore, the regions or departments of Colombia with the best economy (Cundinamarca, Antioquia, Valle del Cauca, and Atlántico) are those that have received the most internal migration due to armed conflict; this is owed to having better employment opportunities, which is why they have developed more social housing projects, reaching low housing deficit rates and the greatest offerings of social housing [30].

The relationship between the construction industry and the economic legal monthly minimum wage has been consistently exhibited in developing countries [31], even in the United States of America [32,33]. Other articles focus on the relationship between the new market housing construction and the low-income housing market [34]. This present study was performed through a statistical validation of the correlation between indicators, employing a nonparametric method. The resulting value of each correlation allowed for establishing whether our hypothesis was satisfied, and it determined most of our conclusions for this study.

Another model suggests that urban growth cannot be solely explained by the per capita income or wage income variable; therefore, housing supply should be correlated with population growth. This study was conducted with information from 2001–2016 for the metropolitan areas of the U.S.A. [35].

Bramley and Pawson (2002) [36] establish that demographics, employment, poverty, income, the attractiveness of the area, and the amount of housing stock or housing for rent are the variables that explain the behavior of urban areas with low demand for housing (United Kingdom, UK).

Other documents relate the impact of policies on low-income housing, like Ha (1994) [37] in Korea, Ikejiofor (1998) [38] in Nigeria, Choguill (1993) [39] in Bangladesh, and Yang et. al. (2021) [40] in the U.S.A.

Housing policies with a focus on subsidies for new and existing (but legalized) housing continue to this present decade (2020) and pertain to cases such as Colombia, Chile, Uruguay, and South Africa, among others [16]. Housing policies in Latin America share similarities and can also be individually analyzed on a global scale [28]. It is also important to note that housing policies in developing countries are often implemented differently from those in developed countries [41].

Summarizing, Latin America mainly results in (a) an analysis of housing policies over the last decades [16,19,20], (b) a comparative analysis between countries in the region or between developed and developing countries [28,41], (c) an analysis of the relationship between housing production and urban public utilities [25,26], (d) a review of the spatial characteristics of informal housing and qualitative deficit [18,42,43,44], and (e) an analysis of the relationship between social housing production and urban poverty reduction [21,22,23,27].

Camacol analyzed the behavior of the building industry’s GDP, both at the national and regional levels, but they do not relate it to the supply of VIS. They have also studied the participation of 60 subsectors of the economy in the demand for goods and services related to the construction industry; however, they do not relate this information to the number of housing units that are built [45]. More recently, another publication presented a GDP projection for the construction industry but does not correlate other economic industries with housing supply, nor is it analyzed by region [46,47].

The objective of this study was to establish whether there is a significant correlation, statistically speaking, between the supply of social interest housing (VIS) in a region (per department in Colombia) and socioeconomic and demographic variables such as the general GDP of the department, the main industry-specific GDPs, and the general population of the department. This is based on the hypothesis that the regions with the highest level of income, in Colombia’s case, are those with a tendency to develop fewer VIS housing units because they have a lower percentage of the population in extreme poverty and with fewer housing needs.

After a search for technical articles on social housing demand modeling, we found a correlation analysis between housing production, deficit, and regional socioeconomic variables, such as GDP, as well as population.

The state of the art makes clear the relationship between poverty and housing conditions, and therefore, a contribution to the SDG. But there are no publications on the analysis of the relationship between indicators at the regional level in any country in Latin America. This study analyzes whether the areas or departments in Colombia with greater economic development are related to a higher number of VIS housing units in the period between 2015 and 2021, making a reading of the contribution to the SDG from the social building economic sector according to the regional social and economic condition in Colombia.

This study is based on the trend that well-developed regions that receive an important migrant population, such as the migration from rural to urban areas due to social and economic conflicts, are linked to poverty; due to this, it requires a decent housing solution to reduce the UBN of the population and to help improve the indicators of the SDG [13,48,49,50,51].

3. Method

The information used in this study was obtained from the most reliable organizations and governmental agencies in Colombia. All the information used is official for Colombia. It was obtained from the most recent reports published by the National Administrative Department of Statistics (DANE) [7] and the Colombian Chamber of Construction (Camacol) [52].

The data management method used was the statistical correlation between variables.

The correlation procedure used was non-parametric: Spearman’s rho. The correlation establishes the incidence between two variables but does not define whether the principle of causality applies. If the rho value is ≥0.5, there is a statistically significant correlation between the two variables. If the rho value is ≥0.7, the correlation is considered strong.

The analysis of the correlation method used is detailed below. The application of correlation has been used to define whether the behavior in the variation of general and sectoral GDP has an influence on the development of social housing in the regions of Colombia; and thus, establish if it is a basis for modeling future projections.

The first step was to establish whether to use parametric or nonparametric methods to validate the correlations. One of the most commonly used parametric tests is the Pearson correlation coefficient. An important aspect to consider is that in order to apply this method, the normal distribution of the variables to be correlated must be verified beforehand. The resulting coefficient (r) measures the strength or intensity of the linear relationship between two variables. To establish whether or not there is a statistically significant correlation, the normality test was initially performed using the “Shapiro–Wilk” command in R Studio, which calculates the p-value. The normality of the data distribution can be validated if the p-value is ≥0.05 [53].

The test was applied to the following variables (results are shown in

Table A1. Results of the normality test “Shapiro Wilk Test” to the variables studied.

Analisys Data	p-Value	Normally Distributed
Number_House Units_NoVIS_2015	3.31 × 10−7	NO
Number_House Units_NoVIS_2016	2.52 × 10−4	NO
Number_House Units_NoVIS_2017	2.57 × 10−4	NO
Number_House Units_NoVIS_2018	8.88 × 10−5	NO
Number_House Units_NoVIS_2019	2.52 × 10−5	NO
Number_House Units_NoVIS_2020	7.55 × 10−5	NO
Number_House Units_NoVIS_2021	5.02 × 10−7	NO
Number_House Units_VIS_2015	5.02 × 10−4	NO
Number_House Units_VIS_2016	5.29 × 10−8	NO
Number_House Units_VIS_2017	4.01 × 10−8	NO
Number_House Units_VIS_2018	5.05 × 10−8	NO
Number_House Units_VIS_2019	6.12 × 10−8	NO
Number_House Units_VIS_2020	5.30 × 10−8	NO
Number_House Units_VIS_2021	4.35 × 10−7	NO
Departmental_Manufacturing_GDP_2015	3.52 × 10−5	NO
Departmental_Manufacturing_GDP_2016	3.87 × 10−5	NO
Departmental_Manufacturing_GDP_2017	4.50 × 10−5	NO
Departmental_Manufacturing_GDP_2018	4.73 × 10−5	NO
Departmental_Manufacturing_GDP_2019	4.66 × 10−5	NO
Departmental_Manufacturing_GDP_2020	5.70 × 10−5	NO
Departmental_Manufacturing_GDP_2021	4.99 × 10−8	NO
Departmental_Construction_GDP_2015	6.45 × 10−4	NO
Departmental_Construction_GDP_2016	4.76 × 10−4	NO
Departmental_Construction_GDP_2017	1.93 × 10−4	NO
Departmental_Construction_GDP_2018	1.03 × 10−4	NO
Departmental_Construction_GDP_2019	1.42 × 10−4	NO
Departmental_Construction_GDP_2020	1.89 × 10−7	NO
Departmental_Construction_GDP_2021	1.39 × 10−7	NO
Departmental_Real_State_GDP_2015	2.53 × 10−7	NO
Departmental_Real_State_GDP_2016	2.69 × 10−7	NO
Departmental_Real_State_GDP_2017	2.81 × 10−7	NO
Departmental_Real_State_GDP_2018	2.85 × 10−7	NO
Departmental_Real_State_GDP_2019	2.91 × 10−10	NO
Departmental_Real_State_GDP_2020	2.95 × 10−7	NO
Departmental_Real_State_GDP_2021	3.11 × 10−10	NO
Departmental_Financial_Insurance_GDP_2015	4.79 × 10−8	NO
Departmental_Financial_Insurance_GDP_2016	4.71 × 10−8	NO
Departmental_Financial_Insurance_GDP_2017	4.65 × 10−8	NO
Departmental_Financial_Insurance_GDP_2018	4.70 × 10−8	NO
Departmental_Financial_Insurance_GDP_2019	4.75 × 10−8	NO
Departmental_Financial_Insurance_GDP_2020	4.74 × 10−8	NO
Departmental_Financial_Insurance_GDP_2021	4.72 × 10−11	NO
Departmental_Mining_GDP_2015	3.75 × 10−6	NO
Departmental_Mining_GDP_2016	3.36 × 10−6	NO
Departmental_Mining_GDP_2017	3.31 × 10−9	NO
Departmental_Mining_GDP_2018	3.09 × 10−6	NO
Departmental_Mining_GDP_2019	1.95 × 10−6	NO
Departmental_Mining_GDP_2020	1.25 × 10−9	NO
Departmental_Mining_GDP_2021	2.42 × 10−9	NO
Departmental_Agrobusiness_GDP_2015	2.10 × 10−2	NO
Departmental_Agrobusiness_GDP_2016	1.83 × 10−2	NO
Departmental_Agrobusiness_GDP_2017	2.30 × 10−2	NO
Departmental_Agrobusiness_GDP_2018	2.01 × 10−2	NO
Departmental_Agrobusiness_GDP_2019	1.86 × 10−2	NO
Departmental_Agrobusiness_GDP_2020	1.95 × 10−2	NO
Departmental_Agrobusiness_GDP_2021	1.43 × 10−5	NO
Departmental_Global_GDP_2015	1.24 × 10−5	NO
Departmental_Global_GDP_2016	1.21 × 10−5	NO
Departmental_Global_GDP_2017	1.18 × 10−5	NO
Departmental_Global_GDP_2018	1.08 × 10−5	NO
Departmental_Global_GDP_2019	1.06 × 10−5	NO
Departmental_Global_GDP_2020	9.73 × 10−6	NO
Departmental_Global_GDP_2021	9.13 × 10−9	NO

):

• Number of licensed VIS units reported for Colombia as a whole and for each of the 33 departments, 2015–2021.
• Number of licensed non-VIS units reported for Colombia as a whole and for each of the 33 departments, 2015–2021.
• Departmental GDP for the manufacturing industry, 2015–2021, for each of the 33 departments.
• Departmental GDP for the construction industry, 2015–2021, for each of the 33 departments.
• Departmental GDP for the real estate industry, 2015–2021, for each of the 33 departments.
• Departmental GDP for the finance and insurance industry, 2015–2021, for each of the 33 departments.
• Departmental GDP for the mining industry, 2015–2021, for each of the 33 departments.
• Departmental GDP for the farming, livestock, forestry, and hunting and fishing (agriculture) industry, 2015–2021, for each of the 33 departments.
• Overall GDP or total departmental GDP, 2015–2021, for each of the 33 departments.

It is important to note that the housing unit indicator refers to licensed housing, i.e., housing units that potentially will be built, whether new or improved, according to the need to license the projected construction. This indicator does not factor in rental housing solutions or VIS or non-VIS constructions that have been developed without a license. It also does not disregard those that, after being licensed, have not been partially or completely built.

The result indicates that there is no normal distribution for any of the variables analyzed in the 2015–2020 period in any of the 33 departments (shown in Table A1).

Since there was no normal distribution for the variables, we performed a nonparametric correlation test, the Spearman test. We executed the following command using the R Studio statistical software program:

• Chart.Correlation (selected data group, method = “spearman”).

The value or result provided by the Spearman correlation test is rho, whose value measures the intensity of the correlation between two variables, and its limits are (−1, 1). If the value is close to one unit, the correlation is good and directly proportional between the two variables. If the value is close to one unit, but is negative (−1), the correlation between the two variables is also good but is inverse. If the value is close to zero, it is considered that there is no significant correlation between the variables [54].

We then proceeded to calculate the value of the Spearman correlation (rho) with the data shown in

Table A2. Spearman test Rho value between GDP and GDP per capita (2015 to 2021).

Department	rho	Department	rho
COLOMBIA	0.9642857	GUAVIARE	1.0000000
AMAZONAS	0.9642857	HUILA	1.0000000
ANTIOQUIA	1.0000000	LA_GUAJIRA	0.9642857
ARAUCA	0.8928571	MAGDALENA	0.9642857
ATLANTICO	0.8928571	META	0.8928571
BOGOTA	0.9642857	NARINO	1.0000000
BOLIVAR	0.9642857	NORTE SANTANDER	0.8928571
BOYACA	1.0000000	PUTUMAYO	0.7857143
CALDAS	1.0000000	QUINDIO	1.0000000
CAQUETA	1.0000000	RISARALDA	1.0000000
CASANARE	0.8928571	SAN_ANDRES	0.9642857
CAUCA	1.0000000	SANTANDER	0.9642857
CESAR	0.8928571	SUCRE	0.9642857
CHOCO	0.9642857	TOLIMA	1.0000000
CORDOBA	1.0000000	VALLE	1.0000000
CUNDINAMARCA	0.6071429	VAUPES	0.7500000
GUAINIA	0.7500000	VICHADA	1.0000000

, and with these results, we correlated the number of VIS units/year and the number of non-VIS units/year with:

• The general economic performance of the country’s departments (measured based on the overall GDP).
• The economic performance for the main economic industries in Colombia (GDP by industry).

The conclusions were established regarding the hypothesis or assumption and concluded that there is a tendency to develop a greater number of social housing units in regions where there is greater economic development, especially in areas where the industrial and service industries are more prevalent.

For this study, the global GDP was used. However, it was previously proven that the general GDP and the GDP per capita would have a similar behavior. For this, the Spearman correlation test was carried out. The relationship between general GDP and GDP per capita in the departments results in a rho value between 0.75 and 1.0, with an average value of 0.93 (Table A2).

4. Results

The Rho values of the correlations between the number of VIS low-income housing units and the global GDP for the economic sectors studied are in Figure 4.

Similarly, Figure 5 contains information for the development of higher commercial value housing or non-VIS housing.

4.1. Correlation between the Finance and Insurance Industry’s GDP and Number of Housing Units

First of all, the name of the economic sector in the corresponding column in Figure 4 and Figure 5 has been abbreviated to “Financial”.

The financial and insurance sector tends to show more significant correlations with the development of non-VIS housing units than with VIS. This means that as the sector’s economy grows, so does the number of high-cost homes (non-VIS) at a very similar annual rate.

4.2. Correlation between the Real Estate Industry’s GDP and Number of Housing Units

First of all, the name of the economic sector in the corresponding column in Figure 4 and Figure 5 has been abbreviated to “Real-St”.

The real estate sector shows a behavior similar to that of the financial and insurance sector. The rho values for VIS housing before 2021 are not significant (rho ≤ 0.5) between the years 2016 and 2020 (71% of the period studied).

In the case of non-VIS homes, the correlation is very high, with rho values greater than 0.9 throughout the period studied. As with the financial and insurance sector, it can be concluded that as the sector’s economy grows, so does the number of high-cost homes (non-VIS) at a very similar annual rate.

4.3. Correlation between the Construction Industry’s GDP and Number of Housing Units

First of all, the name of the economic sector in the corresponding column in Figure 4 and Figure 5 has been abbreviated to “Constru”.

Similar to the two previous economic sectors or industries, the trend of the correlations for VIS housing is that the rho value is not significant in most cases (rho < 0.3 in 57%). Only in 2016 does it reach a value of 0.5, and in 2015 and 2021, the values are close to 0.9.

Therefore, it is determined that the construction sector has no impact on the development of VIS housing units. Although the economic sector is directly linked to housing development, low-income housing is not its major component.

For the development of non-VIS housing units, the rho value increases significantly, with values that exceed 0.9 in 85% of the years studied because, in 2016, it was 0.69, a value that is also considered to have a high statistical correlation.

In general, it can be stated that the economic sector of construction registers a significant correlation with the development of non-VIS housing.

4.4. Correlation between the Manufacturing Industry’s GDP and Number of Housing Units

First of all, the name of the economic sector in the corresponding column in Figure 4 and Figure 5 has been abbreviated to “Manuf”.

In this correlation, the results are similar to those found for the three previous industries. For VIS housing, only the records for the years 2015 and 2021 show a rho greater than 0.5, and the others do not exceed a value of 0.28 in the other years, which means that there is no statistical significance between the manufacturing sector and the development of VIS housing.

In the case of non-VIS housing, the rho values exceed 0.9 in 85% of the years studied because, in 2015, it was 0.63, a value that is also considered to have a high statistical correlation.

From the data for the manufacturing industry, it is deduced that this economic sector shows a relatively significant incidence with the development of non-VIS housing.

4.5. Correlation between the Mining Industry’s GDP and Number of Housing Units

The mining industry presents non-significant correlations with rho values less than 0.42 in all cases.

For low-income housing, or VIS, the highest rho is 0.18 in 2021. But in the other years’ records, there are even negative or inverse correlations, and some that are practically equal to 0.0.

In the case of non-VIS homes, the rho values are closer to a correlation, but it is not considered significant because their values are between 0.16 and 0.42.

From the data, it is interpreted that the development of the mining industry has no relevance or impact on the behavior of VIS social housing in Colombia. It is important to remember that the mining industry is developed in rural areas, and it is probably one of the reasons why it does not show a significant correlation with either VIS or non-VIS housing.

4.6. Correlation between the agrobusiness Industry’s GDP and Number of Housing Units

First of all, the name of the economic sector in the corresponding column in Figure 4 and Figure 5 has been abbreviated to “Agrobus”.

In the case of the agrobusiness sector development, the correlation with the behavior of the generated housing units is better than with the mining sector but does not reach the levels of the previous economic sectors.

In the case of VIS housing, the highest rho values are 0.38 and 0.49 for the years 2015 and 2021. But they are not significant (≤0.5) and only represent 57% of the data. The others are close to 0.0.

For non-VIS housing, the trend is that the rho values are between 0.6 and 0.7, presenting a significant correlation with this type of home, but not as much as with other sectors, such as financial and insurance, real estate, construction, and manufacturing.

4.7. Correlation between Global GDP and Number of Housing Units

Having developed the data analysis for Colombia’s 33 departments, there are only significant rho values (≥0.5) in the years 2015 and 2021. This means that in the period analyzed, there is no tendency towards a correlation between the production of VIS units and the general behavior of the economy, presenting a very strong and positive change in 2021 (rho = 0.85).

The correlation between overall GDP and the number of non-VIS units between 2015 and 2021 presents significant rho values for the entire period. Values were above 0.88 for 86% of the years analyzed. This validates the correlation between higher economic growth and a higher number of high-cost housing units (non-VIS). Unlike the analysis for VIS housing, there is a good correlation between economic development and the behavior of the number of non-VIS housing units.

Figure 6a–g complements the analysis and allows us to observe the dispersion of the points representing the 33 departments in each of the 7 years under study. Graphically, there is no trend towards linear behavior. Throughout the analyzed period, year by year, most departments (80%) show a tendency towards a GDP below COP 25,000 thousand million and less than 5000 low-income housing units. Only 20% of the departments register GDP values between COP 20,000 and 25,000 thousand million and between 5000 and 100,000 units of low-income housing. This means that there are only a few regions in Colombia that have a significant number of low-income housing production, and they are those with the best economy.

The relationship between the GDP of each department and non-VIS housing is presented in Figure 7a–g. The graphs show a greater linear trend in the data than the graphs for VIS units. Only some departments (20%) register high values for non-VIS housing units and a high GDP.

The rho values to correlate the variables in the graphs in Figure 6 and Figure 7 are available in the last columns of Figure 4 and Figure 5.

4.8. Low-Income (VIS) Housing

According to Figure 4, the rho values with some significance are for the years 2015 and 2021. For the period from 2016 to 2020, no significant correlation is proposed for any of the GDPs, neither for the global nor for the sectoral ones.

From the above, it is concluded that the development of low-income housing in Colombia does not depend on the economic behavior of the country.

4.9. Non-VIS Housing

From Figure 5, it is concluded that for the development of non-VIS housing, there is a good correlation with the economic performance of the country, both globally and by sector. The agrobusiness sector presents a good correlation with rho values close to 0.7. The mining sector is the only one that does not register statistically acceptable correlations (rho ≤ 0.5). The other economic sectors (manufacturing, construction, real estate, and finance and insurance) tend to register high correlations with the development of this type of housing (rho ≥ 0.9).

In conclusion, economic development, measured by GDP, either overall or with industry-specific data, shows a better correlation with the development of non-VIS housing units. This is because non-VIS housing units are the most built-in Colombia in terms of units and saleable areas. They are also the most economically profitable for the construction industry, considering that they are not social housing and are more expensive.

4.10. Correlation between the Number of Victims of Internal Displacement in Colombia and the Number of VIS Housing Units

In this case, the number of internal displacement victims in Colombia was correlated for each of the 33 departments with the respective number of licensed VIS housing units per department between 2018 and 2020.

Spearman’s rho yielded values of −1.0 and 1.0 for 45% of the departments, 12% of which were negative values. Only 9% (three out of thirty-three) of the departments registered a null rho value, proving that there was no VIS development in those regions. The other 54% of departments yielded a rho of 0.5 and 0.5, 21% of which were negative.

We can conclude from this assessment that there is no correlation that could indicate that the departments with a greater number of migrant victims have had a greater increase in the number of VIS housing units.

As a further analysis, the Spearman and Pearson correlation coefficient was performed using the aggregate data of all 33 departments for the available years 2018–2020. The results were rho = 0.5 and r = 0.02. These data suggest that it is not possible to establish a significant correlation between the number of licensed VIS housing units in the country and the number of forcibly displaced victims.

5. Discussion

As stated by various authors [5,6,7,8], in circumstances in which there is a greater economic boom, the production of social housing should be greater to alleviate the condition of poverty that is generated by the job opportunities generated by cities with better economies. However, the results obtained show that there is a greater relationship with the development of housing for families with higher incomes.

Some studies conclude that generating housing invigorates the economy by increasing the GDP [27], which, according to the results obtained, is related to the generation of non-VIS housing, validating that statement since the reference does not emphasize low-income housing.

Other studies [35,36] establish that demographics, employment, poverty, income, the attractiveness of the area, and the amount of housing stock or housing for rent are the variables that explain the behavior of urban areas with low demand for housing and that housing supply should be correlated with population growth. Nevertheless, this study analyzes economic performance through GDP, not urban population growth. However, the approach of the references can be supported because the correlation was validated with sectors of economic activities with an urban tendency and by the fact that population growth is generally related to economic growth.

The relationship between social housing production and urban poverty reduction [21,22,23,27] in Latin America’s case is supported by the results obtained since economic development is related to the generation of housing (not VIS), and economic development can be linked to the reduction of poverty in cities.

Several of the studies on housing development in Colombia [45,46,47] do not specify the figures for VIS nor do they correlate general housing values with different economic sectors. Consequently, the results obtained in this present study are a contribution to the state of the art on the topic analyzed for Colombia.

As can be seen in the development of the results, the hypothesis is not fulfilled globally for the period analyzed. Strong correlations (rho ≥ 0.7) are constant for the development of non-VIS housing units. For the VIS units, although the trend is towards positive correlations that indicate that if the global or sectoral economy improves, so does the number of housings being developed, the value of the coefficient indicates that it is not in similar proportions.

Although the review of the literature expresses the hypothesis that has been raised, in the case of Colombia and its regions or departments, for recent years, the hypothesis must be reconsidered towards a relationship between economic development and the number of non-VIS housing, but not for all economic sectors, only for activities related to urban development.

6. Conclusions

The economic standing of the departments in Colombia, measured through the overall GDP for the 2015–2021 period, has a very good statistical correlation with the number of non-VIS units, which hold the highest commercial value.

Conducting a correlation analysis using the general GDP of each department as an economic indicator should have similar results if replaced by the GDP per capita since the correlation between these two indicators is very high.

At the economic-specific level, this correlation is higher with activities that are mostly developed in an urban environment, such as manufacturing, construction, real estate, and finance and insurance. The correlation is lower with industries such as mining and agriculture, which mostly operate in rural areas.

Conversely, the correlation of economic performance over the same period with the number of VIS social housing units offered is not good, except for the year 2021, which shows a significant increase in the number of registered housing units compared to previous years.

The industry-specific analysis for VIS housing yields low correlations, but these are especially dismissible for more rural industries, such as mining and agriculture.

Given their low correlation, the above confirms the hypothesis that more low-income housing units are offered in departments with a stronger economy due to their low level of poverty and unmet basic needs (UBN).

It is also clear that regions with better economies develop the non-VIS housing market more (high rho values), especially in those that have better numbers in urban economic sectors like manufacturing, construction, real estate, and finance and insurance, and have less relationship in those regions where the economic development has been better in rural activities like mining and agriculture.

Then, the behavior of the non-VIS housing market (more expensive) is linked to the main economic sectors, but VIS social housing develops constantly, regardless of the situation of the main economic sectors.

The development of this research had as limitations the dependence on secondary information, which, despite being from official sources, has a degree of uncertainty in the data. Additionally, the databases are not updated to the year 2023.

The next step of this research is to analyze the correlation of the variables over a period of time, that is, correlating GDP for a 3-year period with low-income housing production for the following 3 years to identify if there is cause and effect in that period. Likewise, the period can be extended to a period of 5 years to know if the correlation improves or not.

Another future analysis that could be important is to repeat this same procedure a few years later to find out if the behavior of these correlations is maintained, or if there have been changes that can be explained.