Theory

Taxes can apply at the individual or household level, giving rise to various tax-system options, described below (see Annex Table 1 for examples of countries implementing each type of system).

• ■ Individual taxation simply means that each person’s tax is determined based on their own income, without regard to their marital status. Any joint income (e.g., from jointly-owned property) will have to be assigned to individual spouses, for exam ple by splitting it.

• ■ Household-level taxation aggregates a household’s incomes and subjects them to a combined schedule. This combined schedule can be set in many ways. Commonly the same tax rates apply to singles and households, but there are different approaches to setting thresholds. One extreme would be to use the same thresholds. More commonly, thresholds are increased compared to singles to account for the fact the income covers two people. One approach (used for example in Germany) is to double all thresholds⁸ given that two people are covered by the joined income. But countries may also raise them by less than that⁹ (e.g., United States), which can lead to marriage penalties for some incomes, that is cases in which marrying raises the household’s tax rate. In principle they could also be raised by more, but there are no known cases. In some cases, household taxation is optional.

• ■ Mixed systems. There are various forms of mixed systems, which combine aspects of individual and household-based taxation. A common approach is for countries to apply individual-based taxation with elements of a family-based system, such as spouse allowances, which are often contingent on the spouse’s income not exceeding a threshold. Likewise, there could be child allowances, which might be available at the household level or could be allowed to be chosen among spouses. It is also possible that in countries with household-based systems, tax savings from joint assessments are phased out for couples with higher incomes.

From the perspectives of equality or equal treatment, it is not clear whether individual or household taxation is preferable. On the one hand, household-level taxation ensures equal treatment across households, irrespective of how incomes are distributed among spouses. On the other hand, unless the tax schedule is flat, household-level taxation creates inequalities among individuals with the same earnings level, because their tax liability also depends on whether they are married and their spouse’s income. Historically, when the presumption of a family was a working husband with a stay-at-home wife and multiple children, household taxation was seen as way to benefit families, so as to encourage marriage and fertility.

From a labor supply perspective, the key aspect is that household taxation generally raises tax rates for secondary earners and lowers them for primary earners. Exceptions are flat-rate systems or couples in which both partners would be in the same tax bracket even if unmarried. Figure 5 provides a stylized example, for a simple tax system with a tax-free threshold at 30 percent of average income; a rising marginal tax rate that goes from 0 to 50 percent over incomes ranging between 30 and 200 percent of average earnings; and an option for joint filing, in which case all thresholds are doubled. As shown in the left panel, the marginal tax rate faced by someone married to a well-earning spouse (300 percent of average income) increases dramatically compared to a single individual (or someone filing separately). ¹⁰ The impact is particularly strong when the secondary earner has no or a very low income, as significant tax has to be paid from the first dollar earned. The opposite occurs for someone married to a spouse earning little or nothing, as in this example. This reduces marginal tax rates, by doubling the tax-free allowance as well as the threshold for the top rate of tax compared to a single. Consistent with this stylized example, Fabrizio and others (2020), for example, provide a crosscountry calibrated simulation underscoring the importance of the unit of taxation as a driver of female labor force participation.

View Full Size

Marginal and Average Tax Rates for Spouses under Household Taxation

Citation: IMF Working Papers 2022, 026; 10.5089/9798400201127.001.A001

Notes: The tax system has an exemption worth 30 percent of average income, with a marginal tax rate that rises steadily from 0 to 50 percent between 30 and 200 percent of average income and then stays fixed. In case of joint filing, all thresholds are doubled. Source: Authors’assumptions and calculations.

• Download Figure
• Download figure as PowerPoint slide

The change in marginal tax rates created by household taxation can be expected to discourage labor supply of the secondary earner, while boosting labor supply of the primary earner. As the secondary earner’s marginal tax rate is increased by household taxation, their return to each additional hour of work is reduced. This is known as the substitution effect, and it would have a detrimental impact on the labor supply at the intensive margin (i.e.,the decision on how many hours to work).The effect has the same direction at the extensive margin (i.e., the decision to work): for this margin the total additional tax paid by taking up employment counts, and this is also much higher than for a single individual for any job offer paying less than the primary earner’s salary (if it paid the same or more, it would turn the secondary into the primary earner and reduce their marginal tax). In addition to the substitution effect,there is also an income effect. As shown in the right panel of Figure 5, household taxation reduces (or in the worst case leaves unchanged) the household’s average tax rate, and hence boosts its net income. The household therefore can afford more leisure, which would reduce labor supply for both spouses, reinforcing the negative labor supply impact on the secondary earner and counteracting the positive impact on the primary earner. Moreover, household taxation clearly encourages the formation or perpetuation of one earner families, as the tax benefit of household taxation is greatest when one spouse earns nothing, and because the primary earner’s return to labor is boosted, while the secondary’s is reduced.

Optional individual filing does not provide a solution to the incentives created by household taxation. While individual filing would address the issue of raised marginal tax rates for secondary earners, it would not be a rational choice, because it would increase the total tax paid by the household (in the cases where household taxation does not reduce tax rate, it also does not discourage labor supply). Hence this option would not be chosen by income-maximizing couples. It is therefore relevant only for specific cases, such as couples that have difficulties sharing resources. Hence allowing joint filing does not effectively address the detrimental impact of household taxation on the labor supply of secondary earners. Similarly, cohabiting without marriage is not a tax-effective choice. While it may be chosen for non tax reasons, it would not be chosen by couples as a solution to higher marginal tax rates.

The unit of assessment for taxes, other charges, or entitlement to benefits goes beyond the issue of joint versus household taxation. Even where the main tax schedule is applied at the individual level, countries often provide family-based provisions, such as allowances in the tax system, household-level social security, or welfare.

• ■ Tax provisions. In many countries, governments rely on the family’s income to target tax credits, allowances, and deductions. The clearest example is a dependent spouse tax allowance, which is designed to support one-earner families. This allowance is withdrawn when the secondary earner becomes em ployed, hence reducing their incentives to work. Child tax allowances are also typically targeted at households. In individual systems this is often achieved by allowing only one spouse to use them, and with the rational choice being the primary earner, such allowances raise marginal tax rates for secondary earners compared to singles with children. Child tax credits avoid the problem of being more valuable to the primary earner. But if their level depends on the household’s total income, they again raise the secondary earner’s tax payment com pared to being single.

• ■ Social security contributions. Spouses make contributions to the social security system at the individual level (by paying a percent of their individual income), whereas certain benefits (notably health insurance, sometimes also old age insurance through free survivor coverage) are often provided at the family level. Thus, for secondary earners, paying social security contributions provides more limited additional entitlements,¹¹ reducing the incentives for taking up formal employment.

• ■ Welfare benefits. Social welfare is typically targeted at the household level. This clearly affects only households where both spouses earn little, but it still may discourage the secondary earner if the primary earner used up most or all allowable income before steep withdrawal rates set in.

• ■ Fringe benefits. Employment related benefits, even in case of private employers, may also be given at the family level, such as health insurance and pension coverage with free or subsidized survivor benefits. This becomes a tax policy issue where such fringe benefits are tax favored over standard compensation.

Empirical Findings

In practice, tax systems are usually mixed and more complicated than the stylized example, but in most cases, they create work disincentives for secondary earners.¹² Data available for OECD m embers allows for an assesment of the magnitude of these tax disincentives at both the extensive and intensive margins.

• ■ Tax disincentives at the extensive margin. Figure 6 illustrates the tax disincentives faced by secondary earners upon entering employment, assuming they enter full-time jobs. The dark blue bars represent the tax rates secondary earners would face if they were single \left({\tau }_i^E\right), while the orange bars capture the additional tax faced when they are married \left(\Delta {\tau }_{ps}^E\right), which is calculated as follows:

  $$\Delta {\tau }_{p,s}^E=\frac{T_{p,s}-T_{p,0}}{Y_s}-\frac{T_i}{Y_i}=\frac{T_{p,s}-T_{p,0}}{Y_s}-{\tau }_i^E$$

  Y denotes the households’ gross income, and T its tax liabilities net of deductions, credits, and allowances. The subscripts (p , s ) represent the incomes in a couple’ s household, with p indicating the primary and s the secondary earner’s income. The single subscript i denotes a single individual-household with income i. The superscript E simply denotes tax disincentives at the extensive margin. The first term captures the secondary earners marginal tax rate upon becoming employed in a job that pays s: the additional tax she must pay as a percent of the additional income she brings to the household. Panel A focuses on cases of high inequality among spouses, and Panel B on households with low inequality. In most countries, taxes at the extensive margin are higher for secondary earners than for singles, reflecting a host of country specific rules, including full household taxation or family-based tax allowances or credits. The charts indicate that the additional tax burden for married secondary earners is higher among poorer countries and households with high within household inequality in labor earnings.

• ■ Tax disincentives at the intensive margin. Figure 7 shows in orange the additional tax rate faced by married-secondary earners moving from part-time to full-time employment \left(\Delta {\tau }_{p,s}^I\right), relative to the blue bars, which capture the tax increase they would face if they were single \left(\mathrm{\Delta }{\tau }_i^I\right).

  $$\mathrm{\Delta }{\tau }_{ps}^I=\frac{T_{p,s}^F-T_{p,s}^P}{Y_{p,s}^F-Y_{p,s}^P}-\frac{T_i^F-T_i^P}{Y_i^F-Y_i^P}=\frac{T_{p,s}^F-T_{p,s}^P}{Y_{p,s}^F-Y_{p,s}^P}-\mathrm{\Delta }{\tau }_i^I$$

  The variables follow the same notation explained above. The calculations assume that the primary earner works full time. However, super-scripts F and P denote households in which the secondary earner works, respectively, full-time, and part-time. The superscript I denotes tax disincentives at the intensive margin. At the intensive margin, tax rates tend to be higher than for singles, if income inequality among spouses is high (Panel A). With low within household inequality (Panel B), it is less common for tax rates to be higher than for singles, but it occurs in countries with joint taxation (e.g., Ireland, Germany, Luxemburg, Switzerland).

View Full Size

Tax Rates Faced by Secondary Earners Upon Entering Employment, No Children, 2018

Citation: IMF Working Papers 2022, 026; 10.5089/9798400201127.001.A001

Sources: Authors’calculations based on OECDTaxBEN2.3. The country sample includes Australia, Austria, Belgium, Canada, Germany, Ireland, Greece, Slovakia, Iceland, Italy, United Kingdom, Israel, Luxemburg, Switzerland, Finland,Spain, Netherlands, and United States. AW = average wage at the country level.

• Download Figure
• Download figure as PowerPoint slide

View Full Size

Average Tax Rate Increase Faced by Secondary Earner Moving from Partial to FullTime Employment, No Children, 2018

Citation: IMF Working Papers 2022, 026; 10.5089/9798400201127.001.A001

Sources: Authors’calculations based on OECDTaxBEN2.3. The country sample includes Australia, Austria, Belgium, Canada, Germany, Ireland, Greece, Slovakia, Iceland, Italy, United Kingdom, Israel, Luxemburg, Switzerland, Finland,Spain, Netherlands, and United States. AW = average wage at the country level.

• Download Figure
• Download figure as PowerPoint slide

Joint taxation tends to be equally or more attractive than separate filing at the household level, creating a strong incentive to file jointly, even if it raises the secondary earner’s marginal tax rate. As shown in Figure 8, it is quite common (red dots) that lower household taxation is combined with higher secondary earner taxation, with particularly large effect when spouses’ earnings are unequal (see right hand panel).

View Full Size

A dditional Tax for Secondary Earner versus Marriage Penalty (percentage point difference)

Citation: IMF Working Papers 2022, 026; 10.5089/9798400201127.001.A001

Sources: Authors’ calculations based on OECD TaxBEN 2.3. A negative marriage penalty implies tax-saving. Hence, dots in the upper-left quadrant (in red) represent countries where marriage entails a trade-off: the overall tax liability of the household falls, but the tax rate faced by the secondary earner upon entering employment increases. These countries tend to have family-based taxation or the option for joint-filing. The blue dots represent countries in which marriage increases the tax rate faced by the secondary earner, without tax savings to the household, or even tax increases.

• Download Figure
• Download figure as PowerPoint slide

Because secondary earners tend to be female, household-based systems are implicitly biased against women discouraging their labor supply. It is not an explicit bias, because a female primary earner benefits as much as a male primary earner, but it is implicit, given the ample statistical evidence of women’s wage gaps, which means that they are far more likely to be secondary earners.

The evidence in the empirical literature suggests that the labor supply of married women is particularly elastic, implying that these higher tax rates on secondary earners have an especially strong impact. Evers et. al (2008) provide a meta-analysis, which summarizes the ample evidence that married women are indeed more responsive to changes in net-of-tax wage rates than men. This might reflect a greater propensity to take on a larger share of home duties (Aguiar and Hurst, 2007; Blau and Kahn, 2007; and Alesina et. al, 2011).

View Full Size

Wage Elasticities of Men and Women at the Intensive Margin

Citation: IMF Working Papers 2022, 026; 10.5089/9798400201127.001.A001

Source: Evers et. al (2008) and authors’ calculations. The markers represent deviations of +/- 2 standard errors of the baseline elasticity. The deviation estimates correspond to the refereed journal results for a linear specification, and the baseline elasticity corresponds to the average estimate for married individuals in the US (as reported in Evers et. al 2008).

• Download Figure
• Download figure as PowerPoint slide

Indeed, countries with household taxation tend to have greater gender gaps within couples. Figure 10 shows how gender gaps at the extensive margin (the ratio of female to male employment probability within couples) and the intensive margin (the ratio of average working hours for couples in which both are employed) relate to features of the tax system. Labor force participation gender gaps are narrower in countries with individual-based tax systems, all else equal. At the intensive margin, gender gaps are also narrower in advanced countries with purely individual-based systems. For developing economies, however, there is not much difference at the intensive margin. Of course, cross-country averages mask important differences related to informality and underreporting (IZA 2014), cultural factors (Besamusca and others 2015) and the more disproportionate burden of unpaid work borne by women, compared to developed countries (Ferrant and others 2014).

View Full Size

Tax Systems and Gender Gaps Within Couples

(country average female-to-male ratios; 1= gender equality benchmark)

Citation: IMF Working Papers 2022, 026; 10.5089/9798400201127.001.A001

Sources: Authors’ calculations based on data from LIS, Schechtl (2020, 2021), Thomas and O’Reilly (2016), I DBF, Ernest and Young, and IRS.Notes: Individual tax system with individual tax filing and no family-based tax provisions (credits, allowances, and deductions). Family= tax system with (optional) joint filing and family-based tax provisions. Mixed= tax system with either (optional) joint filing or family-based tax provisions. See Table 1 in the Annex for more details.

• Download Figure
• Download figure as PowerPoint slide

For advanced economies, it is possible to conduct econometric analysis to gauge the importance of the tax system in explaining labor supply gender gaps. In this way, our analysis differs from what is commonly done in the economic literature, which is to estimate labor supply elasticities by gender and marital status (see Bick and others (2018), for example). Gender gaps in labor supply are constructed using the latest household and individual level data from LIS, covering 22 economies. For each country, the household survey of 2017 (or earlier) is used. ¹³ The analysis focuses only on married individuals, for whom the labor supply choice of the secondary earner is relevant.

• ■ The extensive margin gender gap is computed as the difference in the labor force participation rate of women relative to men, such that a negative (positive) value indicates that women are less likely than men to enter the labor force. The participation rate differential (l_{pskt}^f-l_{pskt}^m) is computed for 400 household types per country-year (k, t), where (p,s) represents, respectively, the primary and secondary earners’ income brackets. Each country-year contains 20 income brackets, ranging from (0–10 percent) of the average wage to 200 percent of the average wage or more. The set of household types is given by all the possible combinations of income brackets between spouses. In other words, l_{pskt}^f-\left(l_{pskt}^m\right) is the female (male) labor force participation rate in country-year (k, t), for households for which the primary and secondary earner are in the p and s income brackets.¹⁴ The relevant tax rate in this case is \mathrm{\Delta }{\tau }_{ps}^E. Other controls that help explain the labor supply gender gap include the hourly wage rates of the primary and secondary earner, w_s and w_p respectively, a dummy c_ps, which takes the value of one if the couple has children, and country-year fixed effects u_kt). Country-year fixed effects interacted with a child-dummy are also included to capture institutional arrangements, such as public provision of childcare, to support families with children. The following equation, which includes country and year fixed effects, is estimated:¹⁵

  $$l_{pskt}^f-l_{pskt}^m={\beta }_0+{\beta }_1\Delta {\tau }_{pskt}^E+{\beta }_2{w}_{skt}+{\beta }_3{w}_{pkt}+{\beta }_4{c}_{pskt}+{\mu }_{kt}+{\mu }_{kt}{c}_{pskt}+{\epsilon }_{pskt}^E$$

• ■ The intensive margin gender gap is com puted as the log difference in the average number of hours worked by women relative to men (h_{pskt}^f-h_{pskt}^m). Again, the differential is computed for 400 household types based on the incomes of the primary and secondary earner per country-year, as explained above. However, the calculations are conducted only for couples for which both partners work, since the intensive margin decision to increase work hours must be conditional on being employed. The equation below is estimated, and the relevant tax rate is \Delta {\tau }_{p,s}^F.

  $$h_{pskt}^f-h_{pskt}^m={\alpha }_0+{\alpha }_1\Delta {\tau }_{pskt}^I+{\alpha }_2{w}_{skt}+{\alpha }_3{w}_{pkt}+{\alpha }_4{c}_{pskt}+{\theta }_{kt}+{\theta }_{kt}{c}_{pskt}+{\epsilon }_{pskt}^I$$

The empirical results suggest that the additional tax burdens faced by secondary earners in non-individualized, progressive systems, help explain labor supply gender gaps. The cross-county average tax increase at the extensive margin ranges between 5 and 18, depending on the primary and secondary earner’s incomes. When joining the labor force, the tax increase secondary earners face relative to singles is associated with a reduction in women’s labor participation rate relative to men’s that ranges between 1 and 3.6 percentage points (Figure 11.A). Tax disincentives play a larger role among lower income families and spouses with high income/ability disparities and, within this group, taxation accounts for 10 to 50 percent of the observed labor force rate differential between women and men. At the intensive margin, cross-county average tax increases are between 2 and 16 percentage points higher than for singles, leading to a shortfall of 2 to 20 percent in the number of hours women work relative to men (Figure 11.B). When considering moving from part-time to full-time employment, tax disincentives help explain between 20 and over 90 percent of the observed gender gap in hours worked among lower income families and families with high within-household inequality.¹⁶

View Full Size

Labor Supply Gender Gaps: Implications of the Personal Income Tax

Citation: IMF Working Papers 2022, 026; 10.5089/9798400201127.001.A001

Source: Source: LIS, TaxBEN 2.3, and authors’ estimates.The diamonds represent point estimates of the predicted value implied by the cross-country average tax disincentive of each household type, where the latter is determined by the income level of the primary and secondary earner. The ranges represent the 95 percent confidence interval.

• Download Figure
• Download figure as PowerPoint slide

Policy Options

The most obvious tax policy choice countries can take to support female labor supply is to move toward individual taxation. In countries with full household taxation, this means a move to an individualized system. It is insufficient to allow optional separate filing, because under most circumstances it would not be rational for households to choose this option, and hence the negative effects of joint taxation would remain. In countries with only some family-based elements, such as child allowances, minor adjustments, such as replacing a child allowance (which is worth more to the primary earner) with a flat child tax credit will strengthen married women and mothers’ incentives to work. Moving to individual taxation also achieves equal treatment across diverse household types, irrespective of whether they are married.

Moving to individual taxation comes at a cost as it raises many households’ average tax rates, including those of poor one-earner households, but there are possible remedies. Earned-income tax-credits, for example, can address poverty, and can be designed in a way that avoids or minimizes negative labor supply effects.¹⁷ They can, for example, be conditional on or rising in hours worked before reaching the threshold for withdrawal. For a proposal on how one could protect the subsistence level of couples while minimizing distortions from joint taxation, see Perry, Klemm, and Hebous (2019). Finally, given that the move to individual taxation would by itself raise revenues, a revenue-neutral reform could include reduction in tax rates, which would further strengthen the positive labor supply impact for secondary earners and mitigate the negative impact for primary earners. An example of a country that moved from household to individual taxation is the United Kingdom, where this occurred in 1990.¹⁸

While reforms toward a more individualized system would only support married women, measures that increase progressivity by reducing the net marginal tax rate at the lower end of the distribution, would support female labor supply regardless of marital or family status. The reason is that females tend to be over-represented at the bottom of the distribution. Policies such as earned-income tax-credits and childcare tax credits lift the poor and raise female labor force participation, including by encouraging single poor women and lone mothers to join the labor force or work more hours.

Explicit Differentiation in Tax Codes

As noted, explicit forms of gender differentiation through specific provisions of the law or regulations that identify and treat men and women differently are much rarer than implicit biases. There are, however, many historical examples of such explicit differentiation, and there are also some examples in today’s tax systems.

When explicit gender differentiation exists, it is most commonly found in labor income taxation. As discussed above, tax codes often have elements of family or household-based taxation with the attendant implicit biases. However, biases can be explicit, when there is outright different treatment of husbands/wives, or when there is reference to a “head of household” and that head is determined by some other laws to be the man under most circumstances.

Explicit gender bias in the tax code may take several different forms. It could be found in the allocation of income, provision of exemptions, deductions, allowances, or credits, as well as in the setting of tax rates, thresholds, or the responsibility for filing and paying the tax (Stotsky 1997). The following provides an overview of examples encountered currently or historically. They are drawn from European Commission (1984), Bettio (2009), Williams (2019), Spencer (1986), Grown and Valodia (2010), with current information from mostly from 2021 IBFD tax guides, unless otherwise noted.

Tax free allowances were differentiated by gender, with both examples of higher and lower allowances for women. The Netherlands, for example, used to grant a higher tax-free allowance to a married man than to a married woman until 1984 (European Commission 1984). In India, however, women had higher basic exemptions than men prior to tax year 2012/13, when they were aligned. Similarly, in Pakistan, the basic exemption threshold for working women was higher than for working m en until 2010 (Williams 2019).

Some countries provided different specific tax allowances or credits by gender, which may again either favor or disfavor women. Until 1993, the United Kingdom had a “married man’s allowance” (almost 1.6 times a single person’s allowance). This was later replaced by a transferable married couple’s allowance, and then abolished (except for some grandfathering) (Stotsky 1997). Greece also gave the husband an extra allowance whether or not the wife had income of her own (Spencer 1986).

Some of these gender-specific tax allowances remain to this day. In Gibraltar, an additional allowance is available for women taxpayers over 60 while for men the age limit is 65. In Ukraine, special allowances are granted to single mothers. In Uzbekistan, a monthly allowance is granted to single women with many children. In Indonesia, if the wife derives income and the household opts for joint assessment, the husband is entitled to an additional allowance for his working wife on top of a spousal allowance. In Singapore, married women are entitled to additional allowance for children if they opt for separate assessment and have passed at least three GCEs (General Certificate Examinations) or have a higher education certificate (Williams 2019). In Argentina, a directors’ fee has a higher threshold for taxation if the recipient of the fee is a woman (and an even higher one, if the taxpayer is transgender). This measure may also indirectly act as a tax incentive to encourage the appointment of non-male board members. In Israel, female taxpayers get an additional 0.5 credit point on top of the individual taxpayer credit point entitlement of 2.25 credits. The tax code also stipulates differential entitlements for a working father in a single-parent family as opposed to a working mother in a single-parent family in favor of the woman. Time thresholds to be eligible to avail of a tax credit for a soldier discharged from the defense forces is also different for men and women with preferential treatment for the woman. Further, a working wife separately assessed with dependent children have special differential credits. Similarly, in Spain, special additional credits are available for working mothers according to the income tax law. In Martinique, a tax credit is granted to married or single women who return to employment provided certain conditions are satisfied. In Cambodia, resident individuals who are solely occupied as housewives are eligible for a monthly tax rebate. In Tunisia there is a special allowance f or the head of the household who based on jurisprudence is by default the man.

Even tax rate schedules differed across gender in some countries. For example, until 1995, South Africa levied higher taxes on “married women” than “married persons” (a woman only exceptionally qualified for the married person treatment, for example in case of widowhood), with the former taxed more and the latter less than an unmarried person.²⁰ The system thus achieved something similar to household taxation, even when married partners were taxed separately, but unlike the typical implicit bias which results from women being more often secondary earners, in this case, it is achieved through a an explicitly higher tax payment imposed on married women.

Some countries provide outright tax exemptions under certain circumstances. In Equatorial Guinea, for instance, single women with more than 3 dependents under the age of 18 are exempt from the individual tax (Article 336 of Equatorial Guinea’s Tax Code). In Mozambique, municipal personal tax is levied on resident individuals between 18 and 60 but, women in domestic service are exempt from it.

The responsibility for filing taxes was often explicitly the husband’s. Until 2018, Greece required that even if spouses were taxed separately on the basis of their respective incomes, they nevertheless had to file a joint tax return. Despite there being individual level taxation, the husband was responsible for submitting the tax return and was the recipient of any refund or any claim for outstanding tax balances. Thus, if the wife was entitled to a refund, she had to claim it from the husband (Bettio 2009, PwC 2021). France placed the responsibility of signing the joint tax return for the couple on the husband, akin to Malaysia in the 1980s (European Commission 1984, Spencer 1986). Similarly, Ireland’s tax code required that the man file and sign the joint tax return even if the wife earned the sole income in the household (Bettio 2009).

Responsibility of filing taxes remains tied to gender in some countries. In Guernsey, the husband is responsible for filing the joint return of a household and paying the tax. Jamaica also requires that joint assessment of their combined income is in the husband’s name. In Kenya and in the Democratic Republic of Congo, the income of a married woman living with her husband is deemed to be the income of the husband’s and the tax is assessed on the husband. In Tunisia, there is no joint taxation of spouses, but deductions for children are granted only to the head of the household who is by default the man.

Some countries assign responsibility for filing taxes to a head of household²¹ which could by law or custom be constricted to be the man or at least consider the husband as the default head of household. In the Republic of Congo and in Gabon, income is taxed on the household as a unit and according to the tax code comprises of the head of the family, his spouse and dependents with separate assessments being made only in exceptional cases. This is also the case in Comoros, Jersey, and Chad, but exceptions were more broadly available.

There are also examples of tax biases beyond labor taxes. In Argentina, for example, income deriving from joint property is to be considered in the husband’s tax filing (Grown and Valodia 2010, Table 1.1) . In India, property tax rebates are available for women in municipal corporations in states when the property is registered in the women’s name.

While explicit tax bias against women still exists, the number of such provisions has been declining, and tax provisions favoring women have proportionally risen. While there is no exhaustive database of tax laws with explicit gender biases throughout history, the examples collected are shown in a treemap both for tax biases that existed in the past and for existing differentiation (Figure 12). As this figure reveals, while in the past most discriminatory provisions created disadvantages for women, more recently explicit differentiation against and in favor is equally common globally (though there is no balance within countries). Most remaining explicit biases against women today are found in the responsibility of filing taxes and signing them and on whom the taxes are assessed. Apart from a handful instances, most explicit biases in terms of allowances, credits, thresholds, are in favor of women. These might reflect intentions to improve female labor force participation or reduce gender income inequality—a point discussed in the following section.

View Full Size

Comparison of Explicit Differentiation in Countries’ Tax Codes Over Time

Citation: IMF Working Papers 2022, 026; 10.5089/9798400201127.001.A001

Source: IMF staff compilation using European Commission (1984), IBFD (2021), PwC (2021), Bettio (2009), Williams (2019), Spencer (1986), Grown and Valodia, (2010), and websites from national authorities.Note: This data is not an exhaustive list of all countries that have historically and today had explicit biases in the tax code but is meant to portray trends from those countries discussed in this section. The data represented in the charts start from 1984.

• Download Figure
• Download figure as PowerPoint slide

Explicit Differentiation to Address Gender Inequities

Some economists have made the case for outright gender differentiation in taxation, relying on key principles of public finance theory. Two strands of optimal taxation literature have been used to study gender issues. One focuses on tax systems that raise revenue, minimize distortions, and redistribute (Ramsey or Mirrlees motive),²² and the other focuses on taxes that correct externalities associated with market failures (Pigou motive).²³ In both cases, arguments have been developed for an explicit differentiation in the taxation of men and women, or so-called gender-based taxation.

Differences in labor supply behavior across m en and women, provide a potential reason for gender-based taxes on both efficiency and (horizontal) equity grounds. Ramsey (1927) argues that optimal taxes should be inversely proportional to the labor supply elasticity of taxpayers. It follows that, since the labor supply of women (especially married ones) is more elastic than men’s (Pencavel, 1986; Evers and others, 2008), tax rates for women should be lower. This type of gender-based taxation addresses distortions in the intra-family bargaining process that favor the husband and induce women to pursue home duties relative to men, thus reducing horizontal gender inequality (Alesina and others, 2011 and Apps and Rees, 2011).

A gender-based tax system can mitigate the adverse effects of market failures that, implicitly or explicitly, discriminate against women. Absent radical labor market and childcare reforms that would achieve a first-best outcome, a gender-based tax system can help address gender inequities. As shown by Alesina and others (2011), if women face lower marginal tax rates than men, they will have stronger incentives to work and invest in skills, and they will likely be supported or encouraged to do so by their male spouse (or family members). In addition, there may be a case for introducing a gender-based tax system to correct employment and gender wage differentials arising from the market failures described below.

• ■ (Perceived) higher costs of hiring female workers. Employers may expect women employees to be more costly than men, because they are more likely to interrupt their career for full- or part-time childcare after becoming mothers. This type of implicit bias tends to occur in sectors with high-paid jobs in which permanence in the labor force is highly valued, but it can also spread to other sectors, leading to pervasive employment and gender wage gaps. Moreover, due to asymmetric information about workers’ intentions to become parents, women who chose not to bear children are also discriminated against. Blau and Kahn (2017) and Cremer and Roeder (2019), show that while explicit gender bias appears to be declining, wage differences across genders cannot be explained by schooling, experience, and job-type differentials alone. Waldfogel (1997) and Kleven and others (2018) show that women who have interrupted their career for childcare suffer from a wage penalty that lingers over the long-term. As shown by Cremer and Roeder (2019), unwarranted gender wage gaps can be addressed through a gender-based tax system, with lower marginal taxes for women.

• ■ Incomplete job markets resulting from non-linearities introduced by family-based features of the tax and social security system. These non-linearities make part-time employment unattractive for both employers and employees, thereby thinning the part-time job market. For employers, part-time jobs tend to be more costly, as fringe benefits generally don’t depend on hours worked. On the employee-side, part-time jobs may not be feasible when after-tax earnings are not enough to cover the fixed costs of entering employment, such as childcare. A collective action problem may also arise, because even if some families individually recognize the benefits of having one or both parents work part-time, their individual decision does not help create a market. The absence of part-time opportunities encourages the formation of one-earner households, which tend to favor male over fem ale labor force participation, given women’s higher level of readiness to engage in housework. While a gender-based tax system cannot help complete markets, lower marginal tax rates for women compared to men may encourage the formation of one-earner households with stay-at-home husbands and working wives. An increase in the number of such households would reduce aggregate gender gaps in employment and wages.

• ■ Myopic expectations. Economic agents may fail to internalize the impact of their choices in a distant future, including through the perpetuation of gender biases across generations. For example, anticipating the possibility of exiting or re-entering the labor force into a low-paid job after motherhood, young women may think that becoming stay-at-home wives is the best option, which in extreme cases, may even discourage investing in education. By making choices which, in one way or another, contribute to the formation of a single-earner family, women may fail to internalize that such decisions do not serve their own interest, notably in the case of widowhood or divorce. That is, (married) women inefficiently discount too much the possibility of widowhood and divorce, and they fail to maximize their own utility. By encouraging female-labor participation, a gender-based tax system, could help offset the work disincentives to women created by myopic expectations.

There are conceptual and practical difficulties in the implementation of gender-based tax systems, however. The influential work by Alesina et. al (2011) advocating for a gender-based tax system with lower marginal rates for women, focuses on a representative husband and wife. In such a setup, the transfers system between rich and poor is irrelevant. Bastani (2013) shows that when individuals differ in their productivities and marital status, optimal taxation must weigh the effects of redistribution across households against the effects of redistribution within households, yielding outcomes in which it is not optimal to tax women less than men. Any gender-based system may be difficult to implement, because it may be perceived as unfair by homosexual couples, and difficult to apply in the case of nonbinary genders. Moreover, a system in which women face lower marginal tax rates creates avenues for tax avoidance, as families could shift income towards the spouse paying the lower rate (Stotsky, 1997 and Grown and Valodia, 2010).

Against this backdrop, while the case for gender-based taxation can be made, it seems less controversial to begin with more neutral reforms that create a more efficient and equitable tax system for women, such as reforms toward individualized and progressive tax systems. Moving from joint to individual taxation powerfully reduces marginal tax rate for women and is in line with the Ramsey taxation principle (Boskin and Sheshinski, 1983; Apps and Rees, 1999; Meier and Wrede, 2013). Individualization also eliminates the non-linearities of he tax and social security systems that make job markets incomplete. To address the market failures associated with the (perceived) higher cost of hiring female employees, the cost of paying salaries during parental leave could be made tax (or social security) financed²⁴—although this would still not cover any administrative costs (such as from finding and training workers for coverage during absences). A more marginal reform could at least reduce payroll taxes during parental leave.

Wealth

Notwithstanding the gradual reduction of explicitly discriminatory legal environments in most countries since the 1970s, the current wealth gender gap is substantial, well documented, and wider than that for income in most countries. In the United States, the gender wealth gap is largest between men and women who have never been married, at 71 cents on the dollar, according to recent work by the St. Louis Federal Reserve Bank (2021) based on the Survey of Consumer Finances, after conditioning on contributing factors unrelated to gender. In Europe, while household finance and consumption survey data do not seem to show a difference in unconditional average net wealth levels between genders, Schneebaum and others (2018) find a large gender gap at the top of the distribution. Labor market characteristics and participation in asset and debt categories are found to explain some differences between male and female single households, but the remaining unexplained gap in gross wealth is still as large as 45 percent in Germany and 48 percent in Greece.

Despite scarcer data on the distribution of wealth by gender in developing countries, there are some limited survey-based studies that find similar gaps across geographical areas and various cultural settings. Deere and Doss (2006) and Deere and others (2013) report such evidence, with some variation across countries. There is some evidence that for single individuals this gap may actually be larger at the lower tail of the distribution in some lower income countries due to the higher correlation with gaps in the return to labor (e.g., Anglade and others, 2017 for Ecuador). Globally, surveys focusing at the top tail of the wealth distribution (e.g., Wealth X, 2019; Credit Suisse, 2019) suggest even in recent years women account for a small minority of the richest individuals, representing as little as 12 percent of the world billionaire population, with only 16 percent self-made.²⁷

Furthermore, unexplained gaps in asset holdings are found across asset types. For example, with few exceptions globally (such as Malawi),²⁸ women constitute a much smaller share of all landowners, and often acquire that land via inheritance, rather than market purchases. This share is as low as 20 percent in countries such as Nigeria, Tajikistan, and Peru (FAO 2018). In low-income and emerging market economies, women are generally far less likely to own a home (IFC 2019). Even in countries such as the United States where single women are now more likely to be homeowners than single men, men’s home values are 10 percent higher and appreciate 16 percent faster than those of women (Guerrero 2020). Finally, in part due to accrual consequences of the gender pay gap over women’s lifetime, they tend to be financially underprepared for retirement globally. This is illustrated even in the United States, where women have median retirement savings equal to about two-thirds that of men’s (Transamerica Center for Retirement Studies, 2018).

On average, for advanced economies with available household surveys, women have both lower gross wealth and net worth, despite men being more highly leveraged, as illustrated on Figure 14.

View Full Size

Net Worth of Women Relative to Men

(latest available year)

Citation: IMF Working Papers 2022, 026; 10.5089/9798400201127.001.A001

Source: Authors’ estimates based on LIS Data Access Research Tool (DART), Luxembourg Wealth Study (LWS).

• Download Figure
• Download figure as PowerPoint slide

Entrepreneurial Income

While entrepreneurial income has increased as a contributor to women’s capital income relative to inherited wealth (Edlund and Kopczuk 2009), its potential for income mobility is still limited by a significant gender gap. Notably, women have lower access to finance to start or expand businesses (see Figure 15 for an illustration of this pattern), particularly at lower costs. In part, this difficulty can be attributable to supply side constraints (e.g., lower assets/collateral with which to apply for credit), but primarily to self-selection out of credit markets due to higher risk aversion or lower financial literacy. For example, Ongena and Popov (2016) find that in European countries with high gender bias (measured through a survey indicator), female entrepreneurs (themselves a lower share of the female population to start with than their male counterparts) are more likely to opt out of the loan application process and resort to informal finance, unrelated to active bank discrimination against them or differential credit risk between male and female-led firms. Morsy and others (2019) find similar patterns widespread in Sub-Saharan Africa, where women’s low perceived creditworthiness contributes to self-selectio n out of the credit market. In addition, worldwide female entrepreneurs may also face higher interest rates to obtain loans, even controlling for business characteristics, borrower credit history and bank chosen, as shown by Alesina and others (2013) for Italy. Thus, even in the absence of wage income or personal income tax differences across genders, these biases could hamper women’s ability to grow their representation in the capital income distribution.

View Full Size

Percent of Adult Population that Borrowed to Start, Operate, Expand a Farm or Business

(percent of age 15+ population in respective gender, 2017)

Citation: IMF Working Papers 2022, 026; 10.5089/9798400201127.001.A001

Source: World Bank, Global Financial Inclusion Database, 2017.

• Download Figure
• Download figure as PowerPoint slide

Political Economy of Taxation

Legal differences in inheritance, asset ownership or credit access rules between genders are receding, but still far from extinct. As of 2020, nearly 40 percent of all econom ies still lim it wom en’s property rights (World Bank 2021). Though increasingly a minority, as of 2020 23 percent of countries in the world still did not have full equality of inheritance rights between men and women. Most of these countries are in West Africa, the Middle East and Pacific Islands. Additionally, inheritances might be unequally split across genders due to cultural factors, even in the presence of equal legal treatment. For example, the payment of a dowry upon the marriage of a daughter constitutes a transfer of parental property to the men’s family and thereby shifting future inheritance from the bride to the bridegroom. In some cases, longstanding custom or social norms favoring men’s right to property may lead to lax enforcement by courts of statutory laws protecting gender equality (e.g., Tanzania or Bangladesh, described in Rabenhorst and Bean 2011).

Unfavorable inheritance rights for women are found to be associated with lower inheritance tax revenues. Combining data from the World Bank’s 2020 Women, Business and Law score database and the Tax Foundation (2015), we find a positive correlation between the number of legal provisions ensuring equality of inheritance rights for women and the top inheritance or estate tax rate to the lineal heir: averaging under 4 percent in countries with no equal inheritance rights, but nearly 13 percent in countries will full equal rights. This suggests that countries that still discriminate against women from an inheritance rights perspective also tend to tax inheritances received by men particularly lightly on average.

A similar pattern arises with property tax revenues, which tend to be higher where property rights for women are stronger. As shown in Figure 16, there is some correlation between the lack of robust property taxes and equal treatment of gender in terms of property rights. Namely, countries which do not provide equal asset ownership provisions in their legal framework³⁰ (35 percent of our sample) also frequently do not tax property or other wealth at all, likely because of underlying poor institutional capacity (which explains the disproportionate representation of low-income economies among such countries). Among countries which do, there is some survey-based evidence that the incidence of property taxes may be heavier on female-headed and female-adult-only households (see the example of Ethiopia, analyzed by Komatsu and others 2021). On average, countries with weaker legal gender equality in asset ownership could collect over 0.5 percent of GDP in property tax revenues if approximating the frontier of countries with equal rights. But even among countries which do provide for legal equal property ownership, many collect very little revenues from existing property taxes (excluding inheritance and estate taxes), suggesting more could be leveraged, and in doing so the gap in effective tax rates between genders attenuated.

View Full Size

Property Tax Revenue and Women’s Ownership Rights (percent of GDP)

Citation: IMF Working Papers 2022, 026; 10.5089/9798400201127.001.A001

Sources: GFS 2019; OECD Global Revenues Database 2017; World Bank Women, Business and Law,2020.

• Download Figure
• Download figure as PowerPoint slide

Notwithstanding, greater legal equality alone does not necessarily equate to immediate equal property ownership, nor higher tax revenues directly due to that balance, as shown in Figure 16 cross-sectionally among a sample of low-income and emerging market economies. Recurrent property taxes are typically capitalized into property prices, and therefore first-order neutral from a perspective of influencing an increase in property ownership by women. However, high property transfer taxes can on the contrary hamper increased property ownership by women even in countries with favorable legal environments. Some local governments in India have explicitly instituted lower stamp duty rates for property purchases by women in order to mitigate that disincentive (e.g., New Delhi and Punjab).³¹ Simultaneously, transitioning to a more gender balanced property ownership system could lead to various second order effects on the economy and fiscal externalities not accounted for in this survey. They include greater property market demand leading to higher market-clearing prices and property values in the short run (especially given evidence of a gender gap in housing returns, where women tend to buy relatively overvalued properties because of differences in the choice of initial list price and negotiated discount relative to the list price (Goldsmith-Pinkham and others 2020)—leading to potential property market bubbles in the short-run, but also higher property tax bases if values are assessed at current levels. However, lower concentration of ownership would have a dampening effect on property values in the long run. In addition, increased availability of collateral in the hands of female entrepreneurs could enable increased corporate productivity, thereby indirectly expanding the potential tax bases for corporate and personal income tax.

View Full Size

Land Ownership Gender Ratio and Property Tax Revenue in Low-Income and Emerging Market Economies

Citation: IMF Working Papers 2022, 026; 10.5089/9798400201127.001.A001

Source: World Bank Gender Statistics, latest years; Women, Business and Law, 2020; IMF staff calculations.

• Download Figure
• Download figure as PowerPoint slide

Excises

As in the case of differentiated VAT rates, excises can entail implicit bias as a result of differences in consumption patterns across genders. The analytically simplest case is an excise that is not addressing an externality but is purely imposed to raise revenue. In that case, any difference in consumption patterns across genders would constitute an implicit gender bias. In case of goods only or predominantly used by one gender (e.g., female hygiene products), the bias is arguably almost explicit. Such implicit excises can be avoided by refraining from using excises that do not address externalities. In any case, when the aim is raising revenue rather externalities, broad-based consumption taxes are preferable to excises on specific goods or services.

In case of goods or services with externalities, gender bias can occur if the excise does not fully internalize them. In the case of an excise that matches the externality (Pigou tax), so that consumers bear the full true cost of good or service, there is no gender bias, even if consumption patterns differ across genders. If, however, the excise is set too low (or even at zero), then it is similar to a subsidy, because the true cost including externalities is not borne by consumers. If, however, the excise exceeds the externality, then the excessive part is effectively a tax that can entail implicit bias as discussed in the previous paragraph.

Tobacco taxes are an example of an excise that is often set below the likely externality and thereby benefits men financially who on average smoke more. Worldwide women consume on average less tobacco (Crawfurd and Le Nestour 2019) and alcohol than men (Nelson 2014). The World Health Organization recommends that taxes make up at least 75 percent of tobacco retail prices, a level that is often not reached, as shown in Figure 19. Therefore, even though tobacco taxes have a higher incidence on men given their greater smoking rates, they effectively subsidize them, as they are likely undertaxed compared to the externalities/internalities generated. From an optimal commodity taxation perspective (to reduce deadweight loss), one could argue this bias should in fact be the reverse, as men’s tobacco consumption is typically less tax elastic (thus justifying a higher tax rate).³²

View Full Size

Tobacco Taxation and Relative Gender Smoking Prevalence

Citation: IMF Working Papers 2022, 026; 10.5089/9798400201127.001.A001

Source: World Health Organization and IMF staff calculations.

• Download Figure
• Download figure as PowerPoint slide

Interestingly, the implicit subsidy for tobacco products is highest in countries where relative female smoking rates are lowest. Figure 19 depicts this interesting relationship, which has some similarity to the finding on property taxes shown in Figure 16 above, in that both relationships are marked with high taxation when women are more likely to be liable to such tax. Though determining the source of this correlation is beyond the scope of this paper, it is possible policymakers may be more inclined for socio-cultural reasons to introduce heavier excises when women represent a relatively larger share of the smoker population.

Of arguably even greater relevance to the policy discussion space today, there is a dearth of literature on carbon footprint by gender, and nothing on the incidence of carbon taxes on women specifically, to the best of our knowledge. There are several channels through which carbon taxes could be borne differentially by gender: consumption of carbon-intensive goods and services, employment in carbon-intensive industries (production-angle), and exposure to negative environmental externalities. There is some evidence men consume relatively more carbon-intensive transportation and food than women—informally termed the “eco gender gap” (e.g., Raty and Carlsson-Kanyama 2010, Hunt 2020, OECD 2020b). In addition, carbon taxes raise the relative price of goods in sectors where more men are employed than women (e.g., mining, oil, and gas, see inter alia Cohen (2015) and Catalyst (2019)), increasing production costs and potentially lowering associated wages. On the other hand, the “green” economy stimulated by climate change mitigation measures (including carbon taxes) creates new jobs in currently male-dominated industries such as renewables, manufacturing, and construction (ILO 2018). Furthermore, mortality rates from environmental-related occupational risks and ambient exposure are higher for men worldwide (OECD 2020b). The true relative net burden of carbon taxes to men should thus offset the associated welfare cost of these premature deaths prevented by the introduction of carbon taxes against any increased tax burden due to higher carbon-intensive consumption and employment. Future research could also helpfully combine existing household consumption survey data across countries with greenhouse gas emission figures and carbon tax data to determine the gender split of the effective tax burden across countries.