Drivers of Economic Abuse: An Analysis of the Colombian Case

Economic Abuse in Colombia: A Context

Our research proposes a characterization of the prevalence of economic abuse as a form of IPV in Colombia. After decades in which the emphasis of the study of VAW in the country was almost completely centered on armed conflict-related violence, particularly sexual violence, the study of economic abuse and IPV is gaining traction.

Two crucial studies have laid the ground for approaching this subject in contemporary Colombia. Jaramillo's and Anzola's edited book (2018) explores the gender-based aspects of the regulation of child support from a legal perspective. The editors’ shared argument claims that child support's legal design participates in the generation of gender inequalities and economic violence. Similarly, Deere and León (2021) outline the legal institutions that have facilitated, since the foundation of the Republic in the nineteenth century, women's economic abuse in intimate relationships in the country. Their conclusion is that women's exposure to economic abuse has been facilitated by the marital legal authority the law once vested upon husbands to control the person and property of married women. Although this institution was gradually dismounted throughout the twentieth century, its influence is still felt today. To illustrate the pervasiveness of this violence, they used data from the 2015 Demographic and Health Survey from Colombia.

Both studies underline the role legal regulation has had in exposing women to economic abuse and in shaping cultural expectations about them in society. Marital legal authority over women may no longer be the law of the land, but its legal removal does not mean that society is not still shaped under its influence. Feminist struggles in twentieth-century Colombia show the extent of the inequality women have had to overcome. While married women attained the right to administer their property in the 1930s (Congreso de la República, 1932), it was not until 1954 (Asamblea Nacional Constituyente, 1954) that they were enfranchised and 1974 (Gobierno de Colombia, 1974) that they were legally considered equal to their husbands in the context of the family.

Ruling out discrimination against women does not change deeply ingrained beliefs and behaviors immediately. Sometimes, they barely change over time. Even though the objective of legal regulation is to modify conduct and practices, it does not consistently achieve this objective (Bilz & Nadler, 2014). Research has tried to understand whether and how law models human conduct mapping different factors that contribute to or impede rule compliance. Different explanations, ranging from rational choice to the perception of the legitimacy of rules, aim at presenting the conditions that make a rule succeed in this endeavor (Kuiper et al., 2023). Their existence points to the complex network of individual and social determinants of human behavior. Furthermore, they contribute to understanding the production of regulation as just one step in the process of changing attitudes and conduct, one that is not automatically successful and in which a myriad of conditions converges to achieve or undermine the regulation's goal.

Another related issue that contributes to perpetuating gender-based stereotypes is legally forbidding discrimination against women without designing adequate legal mechanisms to effectively dismantle it. For instance, Jaramillo & Anzola (2018) indicate that the Colombian state has been negligent in systematically addressing the economic abuse women face in the context of child support. Another instantiation of this situation is illustrated by the reluctance shown by the Colombian Supreme Court to grant standing to spouses to protect marital property from the fraudulent actions of their partners aimed at diverting assets (Céspedes et al., 2024). Even though the court has started to shift its position on this subject, this new approach has not been adopted without controversy (Corte Suprema de Justicia, 2019).

Enshrining equality between men and women in the Colombian Constitution and legal regulation has not been enough to eradicate VAW. Economic abuse is just one of its expressions. While women legally have the right to equal treatment in public and private life, official data demonstrates otherwise. According to the 2023 Organisation for Economic Co-operation and Development's report on gender equality in Colombia, women spend 22 h per week more than men doing unpaid work, while men spend 23 h per week more than women in paid work. Young Colombian women are over twice as likely as men to be unemployed and uneducated (OECD, 2023). The 2020–2021 National Survey on the Use of Time, administered by the National Statistics Department (Departamento Administrativo Nacional de Estadística—DANE), indicates that women consistently report joint-financial decision-making in the household over men. For example, 14.5% of men said they would decide alone on taking out a loan for the household, as opposed to only 6.9% of women (DANE, 2022).

A life free of violence is still an unfulfilled promise for many Colombian women, notwithstanding the progressiveness of the domestic legal system. This study contributes to a better understanding of one of the most difficult-to-detect expressions of VAW to inform the debate about the public policies needed to realize women's rights and honor Colombia's national and international obligations on this issue.

Data and Methodology

The data we use comes from the 2015 Demographic and Health Survey (DHS) (Ministerio de Salud y Protección Social et al., 2017), which collects nationally representative data on health and population indicators. In Colombia, this survey is part of the National System of Population Studies and Surveys and of the National Statistics Plan led by the National Administrative Department of Statistics (Departamento Administrativo Nacional de Estadística–DANE). The overarching objective of the DHS program is twofold: on the one hand, it serves to establish demographic changes over a five-year period, and on the other hand, it seeks to provide proxies regarding generalized knowledge, attitudes and practices surrounding sexual and reproductive health of men and women in reproductive age, 13–49. We use the 2015 survey data, the latest available at the time of writing. For the purposes of our investigation, we only use data from women who manifest being or having been in a relationship in their lifetime, providing a sample size of 24,526 observations. Table 1 gives some general characteristics of these women.

OPEN IN VIEWER

Table 1.

Descriptive Statistics.

	Mean	SD	Min	Max
Age	34.03	8.83	13	49
Wealth index	2.68	9.43	0.31	14.57
Years of education	9.65	3.99	0	17
Has a partner	0.78	0.42	0	1
Head of house	0.24	0.43	0	1
Number of children in household under five	0.57	0.74	0	6
Urban zone	0.78	0.41	0	1
Currently working	0.62	0.49	0	1
Has worked in lifetime	0.70	0.46	0	1

Source: Authors’ calculations based on Colombia DHS 2015 datasets (Ministerio de Protección Social et al., 2017). Estimates using sample weights. The wealth index is a multidimensional indicator created from the reported answers on a range of socioeconomic factors, including household characteristics, the accessibility of health services and education, and ownership of household appliances (Ministerio de Salud y Protección Social & Profamilia, 2017).

The average woman in our sample is 34 years old and has almost 10 years of education. The number of children under five in the household (though not necessarily the respondent's children) varies between zero and six. A total of 78% of women have a partner (married or cohabiting), with the other 22% being divorced, separated, or widowed. Of these, 24% of women recognized themselves as head of the household. Of all women surveyed, 38% were not currently working and 30% had never worked. Around 78% of respondents were from urban areas of the country, roughly in line with the national urban/rural split.

Indicators for economic abuse were grouped as an independent category for the first time in the 2015 survey, in addition to sexual, physical, and psychological violence. These were in the form of five questions to which the women could respond: “yes in the last year”; “yes, prior to the last year”; or “no.” For the purposes of our investigation, we combine the first two options to create binary variables: 1: “yes”; 0: “no.” The first one, “Has your partner monitored how you spend money?,” reveals controlling behaviors, as well as “Has your partner threatened you with withdrawing economic support?,” which likewise reveals behaviors that seek to generate dependency through belittlement. The third indicator, “Has your partner forbidden you to work or study?” allows us to measure levels of economic sabotage, and “Has your partner spent the money necessary for the household?” speaks of economic control and exploitation, in that the partner makes unilateral financial decisions that will affect her and limits her ability to budget. The final indicator, “Has your partner taken away (expropriated) any of your own money, property or goods?,” allows us an insight into levels of property rights violations within the context of IPV, a topic that has not received much attention in the literature. These five questions form the basis of our analysis and allow for a comprehensive, multifaceted characterization of the prevalence of economic abuse in Colombia.

As shown in Table 2, the most common manifestation is that of the partner spending the money necessary for the house, at 16% of the total sample. The least common type is that of property or goods being expropriated, at 4.5%. In total, 7,764 women (32% of the sample) reported at least one type of economic abuse over their lifetime. An interesting feature of this analysis is that we aggregate the measure of economic abuse into a single composite indicator that covers all different dimensions through principal component analysis, which allows for a broad view of the complexity of this concept.

OPEN IN VIEWER

Table 2.

Economic Abuse Indicators.

Partner has:	Frequency	Mean	SD
Spent the money necessary for the household	4192	0.1623	0.3688
Forbidden respondent to work or study	3613	0.1411	0.3481
Monitored how respondent spends money	3231	0.1363	0.3432
Threatened with withdrawing economic support	2624	0.1057	0.3075
Taken away any money, property or goods	1121	0.0447	0.2067
Average		0.121	0.216
At least one	7764	0.317	0.465

Source: Authors’ calculations based on Colombia DHS 2015 datasets (Ministerio de Protección Social et al., 2017). Estimations incorporate sample weights.

These data reveal some interesting patterns when considering their intersectionality with other socio-demographic variables (see Appendix 1). Considering first marital status, married women seem to be the most sheltered from economic abuse, reporting the lowest incidence for all five indicators. To rationalize this finding, it is important to consider that much economic abuse is reported retrospectively. Therefore, observing a lower incidence among married women may point to the fact that economic abuse impacts the marital status, showing higher incidences among women who are no longer in relationships. Interestingly, divorced women are classified as being the wealthiest and yet have reported the highest rates in all categories, by a wide margin. They indicate economic abuse at extraordinarily high rates, with 64% claiming to have experienced at least one form. We may infer that economic abuse contributed to the divorce, but perhaps also that the legal protection of marriage ensured their financial security when compared to “separated” women, whose average wealth index score is 28% lower. Particularly high is the reported incidence of expropriation by the partner, which is reported at over five times the national average for divorced women. As previously discussed, this is a unique form of economic abuse that tends to manifest during the separation of a couple.

The level of education, however, shows a more complex picture. This may be due to the fact that the educational level is related to the level of awareness of situations related to economic abuse or gender-based violence in general (Kafonek & Richards, 2017; Elboj-Saso et al., 2020). For a woman with low educational attainment who grows up in an environment where women usually depend on their partners, many situations may become normalized even when they constitute economic abuse (Villardón-Gallego et al., 2023), while women with higher levels of education may be more likely to react to tenuous situations of abuse. That is, lower levels of education may be correlated to the probability of underreporting abuse.

What is clear to see is that increasing levels of education walk hand in hand with increasing wealth. Higher-educated women are more likely to work professionally and may earn more, fostering economic stability and independence. Women with “higher education” report the lowest, or at least lower than average, levels of abuse. These women may associate with more highly educated partners, with evidence indicating economic abuse to be more prevalent among men with lower educational levels (Alkan et al., 2021). However, for the rest of the women it is not as clear cut. It seems to peak for women with a secondary education who, interestingly, report the highest rates of being forbidden to work or study by their partner, suggesting perhaps that the ceiling effect tends to be accessing higher education with abusive partners. The poorest women who self-identify as having no education also report the highest rates of property rights violations.

Colombia is a very diverse country, and different regions tend to be ethnically and culturally concentrated, alongside great disparities in wealth, economic and educational opportunities, safety, and the lingering presence of the armed conflict. The ethnicities of the women reveal wide disparities in experiences. Women belonging to the “other” ethnic category (raizal, gypsy, and palenquero) generally report the lowest incidence of economic abuse. However, they do report most highly for the partner spending the money needed for household expenses and for the partner expropriating their belongings. The situation is also complicated for Afro-Colombian women, who overall report higher incidences of economic abuse, yet the lowest for the partner spending household money. Married or cohabitating Afro-Colombian women also manifest greater dissatisfaction in the relationship, with 34% having considered leaving their partner within the last 12 months (the highest of the sample), with the most common reason being domestic violence, at 52%.

Women in Orinoquía and Amazonía, the most rural regions of the country, are not only the poorest but also the most likely to report at least one type of economic abuse. Women in the Atlántico region along the Caribbean coastline report the lowest levels of economic abuse, yet also have the highest rates of having never worked (45%, against the national average of 35%) and also to not currently working at 46% against the national average of 40%. Women from this region are also more likely to have no education and have one and a half times higher than the national average rate of illiteracy, with 21% of women surveyed from this region claiming not to be able to read or write. This indicates that women with fewer educational and work opportunities, and hence less economic empowerment, experience lower levels of economic abuse, but it may also reveal a resignation of women to yield all economic authority to their spouses.

Moreover, much of this may be due to cultural reasons, with Atlántico women generally expressing more misogynistic attitudes towards gender relations, endorsing both male privilege and the minimization of women in much larger proportions compared to their counterparts in other regions. Notably, over 13% of women from this region think women tolerate violence to keep a family together and nearly 7% disagreed with the statement that women are free to decide for themselves if they want to work. Women who do not test the boundaries of social norms are not met with resistance. A woman cannot be forbidden to work if the question of her working never arises.

Women in the capital of Bogotá overwhelmingly express more progressive attitudes towards gender relations and tend to be richer, while also reporting the higher rates of their partners monitoring how they spend money. This pattern is repeated across other variables, whereby the richest subgroup of women also reports above-average rates for this indicator. This suggests that the more resources at a woman's disposal, the more likely an abusive partner will seek to control them on the one hand, and on the other hand, more awareness of covert controlling behaviors. Bogotá women also have the highest levels of higher education in the country at 33%, which unsurprisingly also coincides with the lowest report rate for being forbidden to work or study.

Finally, looking at house and land ownership, the picture is also muddled. For both house and land ownership, women who jointly own with their partners report the lowest incidence of economic abuse. Unsurprisingly, all of these women are either married (44%) or cohabitating (56%). However, it is for women who own a house solely in their own name and jointly with their partners who consistently report higher levels. For women who own and co-own land, 32% are either divorced or separated. This may be driving the high levels reported of the threat of the withdrawal of economic support, as well as their partners expropriating their belongings, as during and after the dissolution of a union is when these behaviors are most likely to manifest.

Perhaps unexpectedly, women who are sole owners of land are more likely to report at least one type of economic abuse, despite their above-average wealth on top of the temptation to think land ownership may give them more leverage (72% of these women are married or cohabitating). An interesting distinction can be made here between the types of abuse. Threatening to withdraw economic support and expropriating property are reported at lower levels for these women, which may speak of their greater financial independence within the relationship. The indicators they score highly in, the partner spending household money and monitoring how she spends money, may indicate more domineering and controlling attitudes of the partner. Hence, the more a woman has, the more an abusive partner may exert control over her and feel the need to “put her in her place.”

As previously discussed, domestic violence is usually experienced as a continuum with different forms of abuse interacting and reinforcing one another. Our data corroborate this theory. Of all women surveyed, 34% claimed to have experienced at least one form of physical violence at the hands of their partner or ex-partner. The most common form of physical violence reported was that of having been pushed or shaken (30%), followed by being slapped (23%). The most egregious form of violence, being attacked with a knife, gun, or other weapon, was true for 3% of women. For these women who had experienced at least one form of the seven types of violence questioned, the probability of experiencing at least one form of economic abuse was 67%, over twice as high as the survey average.

Likewise, 64% of all women surveyed reported at least one type of psychological abuse, ranging from jealousy (the most common form at 51%) to not being consulted for important family decisions (the least common form at 13%). The close relationship between psychological and economic abuse becomes clear when considering that of all women who reported at least one type of economic abuse, 95% had also experienced at least one form of psychological abuse. If a man threatened to withdraw economic support from a woman, there is also a 58% chance that he has also threatened to leave her and a 52% chance he has threatened to take away the children. Men who have taken property, money, or other goods off their partners or ex-partners are most likely to exhibit controlling behaviors, such as limiting contact with family and/or friends and insisting on knowing where she is all the time. While this form of economic abuse is the least common, it is the one that ranks highest on over half of the psychological abuse indicators, suggesting that more extreme economic exploitation tends to be exhibited by more psychologically abusive men.

Although sexual violence is relatively rarer, with 8% of all women surveyed reporting having been sexually assaulted by their partners or ex-partners, these women, on average, also report economic abuse over four times as much, with all of these differences being significant at the 1% level. Whilst these results cannot allow us any causal inference into the relationship between economic abuse and other forms of IPV, they do provide evidence of the wider context in which economic abuse takes place for women in Colombia.

Measuring an Economic Abuse Index and Its Determinants

Our analysis combines two standard statistical methods. Firstly, to measure economic abuse and other forms of violence against women, we implemented Principal Component Analysis (PCA). This technique allows us to integrate multiple dimensions into a single indicator. Economic abuse manifests in different ways, such as controlling access to money or restricting employment opportunities. Relying on a single measure may provide an incomplete view of the phenomenon. PCA helps summarize multiple correlated variables by constructing linear combinations of the original variables, based on the decomposition of the variance–covariance matrix. Specifically, PCA calculates the eigenvectors and eigenvalues of this matrix. While eigenvectors represent the loadings (weights) of the original variables in the principal components, eigenvalues indicate the relative importance of each principal component. The principal component corresponding to the largest eigenvalue serves as a single measure that best captures the combined intensity of economic abuse.

The second method is linear regression models. Since we aim to identify the factors that significantly explain economic abuse, linear models allow us to simultaneously control for a variety of variables and quantify their statistical relevance. We use the Ordinary Least Squares (OLS) method to estimate the coefficients corresponding to the set of variables included in the model.

In our first phase of analysis, to summarize the diversity of factors that the survey collects regarding the different types of violence, we implement PCA (see Appendix 2). In addition to the economic abuse index, indices of physical violence, psychological abuse, attitudes regarding gender roles in a relationship, and attitudes regarding the legitimization of domestic violence are built using the original indicators included in the survey.¹

We then implemented regression analysis to identify the factors that determine economic abuse, using the first principal component of the economic abuse index as the dependent variable. This accounts for 46% of the total covariance of the five economic abuse indicators, making it a suitable representative of this set of behaviors. Our model includes two groups of explanatory variables. First, the socioeconomic and demographic characteristics such as education, age, marital status, ethnicity, household structure (head of household and the presence of children), and employment status. Additionally, the first principal component indices, as well as a binary indicator of sexual violence, are included to quantify the interdependence between different types of violence.

The results shown in Appendix 3 reveal some factors to be more influential than others in the probability of experiencing economic abuse, providing evidence for both the composite indicator and for each of the dimensions. Regarding socioeconomic variables, and in line with previous findings, education was associated with a higher incidence of economic abuse, while age and ethnicity did not seem to affect the incidence. However, Afro-Colombian women are at greater risk for various indicators, such as being prohibited to work or study. The marital situation and the level of responsibility in the home (head of household or other member), on the other hand, influence the occurrence of economic abuse, with heads of households reporting more economic abuse at a statistically significant rate. As previously discussed, this may be due to abuse experienced during or post separation, with this model including both women currently in a relationship or having been in one. The threat of withdrawing economic support is the most significant outcome, which further lends support to this theory. Living in an urban zone reduces the risk, indicating these types of behaviors are more common in rural areas.

Regarding the socioeconomic and family context, both household income and the presence of children in the home matter in conditioning economic abuse. The presence of children in a home is often a determining factor as the greater care burden that children represent in the home reduces labor participation, limiting economic independence (Sousa-Poza et al., 2001; Ospina-Cartagena & García-Suaza, 2020). In this sense, our results support the hypothesis that the presence of children is a factor that positively influences economic violence. Although some of the results seem intuitive, they do show the relevance of socioeconomic characteristics in determining the occurrence of economic abuse.

After controlling for a large number of women's characteristics, our results show a close relationship between types of gender-based violence. That is, manifestations of physical, psychological, and sexual violence are related to a higher incidence of economic abuse. Therefore, the latter is only one of the forms of violence, showing that women in violent environments usually face a multiplicity of situations that affect not only their self-esteem but also their economic independence. This means that actions to mitigate gender-based violence require complementarity that responds to the multifaceted nature of violence. However, it is difficult to establish causality mechanisms with respect to the other dimensions of gender-based violence due to their interdependence. It is impossible to know, based on these regressions, if, for example, physical violence leads to economic abuse or if economic abuse is a precursor to violence.

Regarding the individual dimensions of economic abuse, interesting heterogeneities emerge. Focusing on those statistically significant results, it is observed that the types of economic abuse that are positively influenced by education have to do with monitoring the use of money and economic support. An expected result is that the employment condition, i.e., being employed, reduces being prohibited from working or studying. It is also revealed that the total number of children a woman has and owning a house become significant risk factors when considering the outcome of being threatened with the withdrawal of economic support. This is to be expected with the increased number of children and the limited financial independence this often brings for women. The woman currently working increases the likelihood of her reporting her partner having spent the money needed for the household, as well as having had property, money, or other goods taken from her. Finally, consistently, other types of gender-based violence are the only variables that are positive and significant for all manifestations of economic violence.

A possible concern about the validity of the results is related to the level of aggregation in the reporting of economic abuse. For example, women who suffered economic abuse over several years may have changed their job conditions and even marital status. In this sense, analyzing recent reports in a more homogeneous group may reduce biased results. As a robustness exercise, we repeated the analysis but this time limiting the sample to women who were currently in unions and who had claimed to have suffered at least one form of economic abuse in the last year. This provided a sample of 19,038 women, around 78% of the original sample. The results in Appendix 3b show them to be broadly similar to the main analysis with the same variables presenting as significant risk factors. The main difference lies in the relationship with wealth, which in this case appears to reduce the likelihood of the partner spending household money. This is in line with previous findings showing that in dual-earner households, women have greater negotiating power regarding household financial decisions (Carlson and Hans, 2017). Moreover, being the head of the household is no longer significant, suggesting that economic abuse is experienced more commonly as a consequence of a separation, not that being the head of the household necessarily leads women vulnerable to economic abuse.

Concluding Remarks

The emerging interest in understanding economic abuse contributes to better characterizing VAW. While physical abuse has been more widely studied since it is easier to identify and generally poses a more imminent threat to women's lives, comprehending economic abuse is of paramount importance to fully grasp IPV's dynamics and consequences. Women's economic disempowerment in the context of intimate relationships could be a part of a wider pattern of abuse that includes physical and/or psychological manifestations. Curtailing women's economic independence could have lasting negative consequences for them and their families. This study analyzes economic abuse as a form of IPV in Colombia using information from the 2015 Demographic and Health Survey, which has allowed for the characterization of this phenomenon based on multidimensional indicators. Our main objective was to determine which factors are associated with higher levels of economic abuse; in short, which women were more likely to suffer from it. Although a low incidence is evident in comparison to other types of VAW (which may be explained by misreporting) relevant factors associated with higher levels of economic abuse are found, primarily related to the household context, e.g., the number of children. This is coupled with other types of physical and psychological violence, showing that actions to mitigate economic abuse should be comprehensively viewed alongside other forms of violence. That is, there are clear policies and laws that protect against physical violence, but they may overlook other forms of violence that have equally adverse impacts on the well-being of women.

Marital status and wealth are dimensions that provide a more granular picture of economic abuse. Divorced women are the wealthiest and the ones who reported the highest incidences of economic abuse. Even though it is possible that this abuse contributed to their divorce, given the retrospective nature of the survey report, it is also reasonable to conclude that they may have been exposed to economic abuse or discovered it during the negotiation or litigation of the division of the marital property. This could explain why the reported incidence of expropriation by the partner is five times the national average for this group. Moreover, this finding is in line with the literature that indicates that economic empowerment does not necessarily prevent economic abuse, but it sometimes boosts it. This is evidence that interventions to eradicate economic abuse need to go beyond improving women's economic conditions and tackle other factors that could be driving this type of abuse, such as negative gender-based stereotypes and regulations that impede discrimination against women to find a prompt and adequate legal remedy.

This 2015 Demographic and Health Survey is currently the strongest instrument to measure VAW in Colombia. At the time of writing, this is the latest version available in the country. Although the data could be considered dated, our analysis makes a strong case to continue its application and illustrates how it contributes to an in-depth understanding of economic abuse, a type of violence that has been understudied and faces issues of underreporting. Also, our examination demonstrates the importance of exploring the intersection between sex and other demographic variables to determine which women are more exposed to this kind of abuse. Furthermore, our study proposes a methodology to unpack all the information the data of the survey offers and constitutes a baseline against which future versions of the survey could be analyzed.

More quantitative and qualitative research on this area is needed. Our article offers possible areas of exploration for this purpose. Making visible the intersections between economic abuse and other demographic variables, such as education, marital status, or children, is crucial to contribute to a better understating of this phenomenon. Identifying the nuances of economic abuse is essential to designing effective interventions. An important input to inform the design of these interventions is data that allows for a permanent and standardized measurement of forms of economic abuse. This makes it essential to periodically conduct the Demographic and Health Survey. This survey has several advantages. Among others, it can establish a baseline with a comprehensive characterization of different types of violence, which can be scaled to other countries when they ask the same set of questions. This can allow for comparative analyses and tracking of the behavior of these indicators over time, and even evaluating the impact of interventions aimed at mitigating different types of violence.