Isolating the female agency-driven development factor in external sovereign emerging market debt

Abstract

The underlying literature hypothesises and provides randomised evidence for the positive impact of promoting the broad-based inclusion, empowerment, and representation of women on regional ecosystems. This study seeks to isolate a female agency-driven development factor in external sovereign emerging market debt and finds evidence for superior risk-adjusted returns from tilting towards female agency leaders. We propose the female agency factor as an additional scope in the modern investor’s toolbox of holistic credit assessment, allowing investors to isolate the issuers which are the most effective sovereign transmission mechanisms of sustainable development capital. This contribution to the corpus supports the notion of integrating sustainability factors into portfolio construction and reinforces the argument for supporting female-led development from a financial markets’ perspective.

Introduction

Inspired by the underlying literature, this paper explores the relationship between gender equality and sovereign creditworthiness in developing markets, with a focus on the potential for the promotion of female agency to support development. The problématique explores and considers whether the purported benefits of female agency-driven development are internalised by capital markets; we motivate the question: do the bonds issued by economies with a greater degree of female agency outperform those issued by economies with a lesser degree of female agency? We hypothesise that more gender-equal economies allocate resources more optimally, although we do not infer causation, reflecting on future economic development and growth—and that such a halo effect is discounted in traditional assessments of sovereign creditworthiness. As such, we expect the local capital markets of more gender-equal economies, the most effective transmission mechanisms of sustainable development capital, to outperform less gender-equal peers over time, justifying relative allocation to more gender-equal economies from the perspective of risk-adjusted financial returns.

We begin by contextualising the question and discussing the importance of female agency in developing markets, particularly as a source of sustainable development. Next, we provide an overview of our research methodology; the study approaches the problématique by seeking to isolate a female agency factor, the first known study to attempt this, by integrating a framework developed by Henide (2021) into the iBoxx USD Emerging Market Sovereigns Index ("the benchmark”) to construct factor portfolios across three key verticals (politics, economics, society) from leader and laggard portfolios. This is similar to the first-known approach for examining factors in sovereign emerging debt by Brooks et al. (2020) and the first-known approach for examining factors in emerging market credit by Dekker et al. (2021),¹ for further assessment. We conclude by summarising our findings, which support the notion that integrating sustainability risks can be a source of superior risk-adjusted returns and reinforce the argument for supporting female-led development from a financial markets perspective. The findings motivate questions for further investigation, such as whether there is a causal relationship between promoting female agency and reducing the credit risk for sovereign issuers, which has broader and deeper financial markets implications.

Contextual background

The academic literature suggests that rebalancing societies to provide women with a greater degree of inclusion, representation, and autonomy has a favourable influence on local ecosystems, both helping to prevent objectionable social activities and facilitating socially and economically-additive behaviour; the correlation between gender equality and reduced poverty in developing markets is demonstrated clearly by academics (Filmer 1999; Sinha et al. 2007; Strauss and Thomas 1995; World Bank 2001). Esther Duflo, a co-recipient of the 2019 Nobel Memorial Prize in Economic Sciences,² also found in her seminal paper (2012) that "women’s empowerment and economic development are closely interrelated" (pp.1076).

Sowing female inclusion, equality, and empowerment into communities reaps not only rewards for the women of the community but also for men and future generations through more constructive household outcomes (Duflo 2003; Duflo and Udry 2004; Hoddinott and Haddad 1995; Pitt and Khandker 1998; Pitt et al. 2003; Rubalcava et al. 2004; Thomas 1997). This perceived halo effect has given rise to the prevalence of the gender factor in development theory and postmodernist thought; reminiscent of the aforementioned studies, Eastin and Prakash (2013) point to women's intrahousehold bargaining power and hypothesise that the opportunity of developing human capital is a vehicle for conferring greater political and social recognition, providing a rationale for the role of gender equality as a transmission mechanism for economic development, particularly amongst economies in the earlier stages of their economic development.

The bridge between improved gender equality and improved economic outcomes is encapsulated succinctly by the framework of Sinha et al. (2007), adapted in Fig. 1.

Fig. 1

Gender equality: A pathway for developing local ecosystems

Adapted from Sinha et al. (2007)

Improving the gender balance across economies is core to improving their social and governance profile, as well as underpinning their future growth potential. The underlying framework,³ that we integrate, references three verticals ('Politics', 'Economics' and 'Society') which derive from the pillars of equality identified in the United Nations Sustainable Development Goals (UN SDGs) target 10.2. There is a comprehensive body of academic literature that justifies the pillars identified in the UN SDGs target 10.2 and the rationale behind the sustainable socioeconomic transformations achieved through the improvement of women’s social and economic outcomes' although US SDG 10.2 stipulates the inclusion of "all, irrespective of age, sex, disability, race, ethnicity, origin, religion or economic or other status", we focus in this study only on matters pertaining to gender-based inclusion, relating to the characteristics of women and men constructed by socialisation, rather than the biological and physiological characteristics implied by "sex" (Deaux 1985). We highlight the concepts within the verticals and some key readings that attest to their significance in Table 1. The concepts are based on the framework indicators, which are a subset of the multi-source Tier 1 data ascribed in the UN Minimum Set of [Gender] Indicators (UN MSoGI).

Table 1 Supporting literature.

The framework

The pillars of inclusion, equality, and empowerment

‘Politics’, ‘Economics’, and 'Society' (collectively, abbreviated to 'PES') are the three key pillars of female equality that ensure that women are free, represented, included, and empowered. They align with the spirit of UN SDG target 10.2. To capture sovereign performance across these verticals, we introduce eight concepts built upon Tier 1 data published by the World Bank under Creative Commons 4.0 licensing. The concepts and multi-source indicators are recognised by the United Nations (UN) and are upheld as key developmental indicators, contained in the UN Minimum Set of Gender Indicators.⁴ Based on the UN's evaluation, we provide a mapping of each indicator to the specific strategic objectives outlined by UN Women (1995) in the Beijing Platform for Action (BPA), as well as the goals (and targets, where applicable) set out by the UN (2015) in the Sustainable Development Goals (SDGs).

The categorisation of the indicators as 'Tier 1' is designated by UN Women through the curation of the Minimum Set of [Gender] Indicators. The classification confirms that the indicator 'addresses relevant issues related to gender equality and/or women's empowerment; is conceptually clear and has an internationally established methodology and standards'. Additionally, the classification confirms that the underlying data 'are regularly produced by countries, with sufficient coverage to allow tracking progress over time'.

We reproduce, in Table 2, the indicators referenced in the framework, by vertical, their mapping and the source(s) of the data. A table consisting of all project data source permalinks is available in Table 8 (appendix, Exhibit 4).

Table 2 Indicators referenced in the framework.

Where specified, ‘GAP’ concepts are all proprietary gender-relative derivations that seek to align scores for comparability, equilibrating for structural differences in economic infrastructure, looking specifically to assess gender (dis)parity. Concepts not labelled 'GAP' are strictly measures of absolute outcomes relating to the respective region's eligible female population subset.

The framework also references (inverse) inequality-adjusted GNI per capita, γ, defined:

$$\gamma = \frac{{{\text{g}}({\text{GDP}} + ({\text{F}}_{{\text{I}}} - {\text{ F}}_{{\text{O}}} ))}}{{\text{n}}}$$

where: g = Gini coefficient, GDP = Gross Domestic Product, F_I = Foreign inflows, F_O = Foreign outflows and;n = Population

An extensive discussion of the framework and methodology is provided by Henide (2021). All the data was accessed in November 2020. Annual data was sourced for 2010–2020, inclusive.

Equilibrating sustainability performance

Following the framework, an 'Overall PES Score' is calculated as a weighted average of the scores obtained and manipulated from the underlying data. Introducing a secondary axis, the IIGNI, provides a multi-dimensional scope for the investment universe and a counterbalance for greater comparability. In particular, the space allows investors to introduce a minimum baseline function, which provides an implied sustainability score for each vertical at each IIGNI increment. The sources for the underlying data used to construct the IIGNI are provided in Table 3.

Table 3 Data used to calculate the IIGNI.

The framework prescribes a baseline function:

$$f(x) = \left\{ {\begin{array}{*{20}c} {(\frac{7x}{{8000}}) + 55} & {if\;0 < x < 40,000} \\ {90} & {if\;40,000 < x < \infty} \\ \end{array} } \right.$$

We transplant the default baseline function prescribed by Henide (2021) into our study, as illustrated by the 'baseline' in Fig. 2. The prescribed function is ideal in our setting as it exhibits goodness of fit and is progressive by nature. By inherent design, the baseline requires that countries with a higher degree of GNI per head and/or lower equality of wealth distribution (measured by the Gini coefficient) deliver higher scores in the aggregate across the 'PES' verticals to 'breakeven' with the relative performance of countries with a lower degree of GNI per head and/or higher equality of wealth distribution.

Fig. 2

Source: World Bank, Authors

Introducing a progressive baseline scoring function.

The equilibration concept and the derivation of relative performance versus the defined baseline scoring function allows investors to better compare issuers within the universe. An assessment of relative performance and the absolute 'Overall PES Score', indicates a consistent trend in scoring and relative performance, but allows us to isolate clusters of inferior issuers whose scoring is in-line with 'less wealthy' countries or those with a greater equality of wealth distribution. Conversely, we can identify countries with a superior 'Overall PES Score' performance relative to peers of a similar level of (inequality-adjusted) 'wealth'. We can also isolate clusters across the curve of sovereigns that under/outperform on an absolute level (the quadrants to the left of the universal minimum and those to the right of the universal minimum, respectively) and on a relative level (sovereigns that demonstrate a negative relative performance and those which demonstrate a positive relative performance, respectively), as highlighted inFig. 3. ⁵ The PES performance curve can also be used to calibrate and therefore reverse-engineer the baseline function; users may wish to 'fit' or benchmark baseline PES performance to a given set of country profiles, to which they can align the universal minimum and/or the PES performance curve.⁶

Fig. 3

Source: World Bank, Authors

Plotting the PES performance curve.

Quantifying the relative performance of sovereign issuers across the key pillars of female empowerment provides us with an additional socially conscious scope with which to assess the development of developing countries and to determine portfolio capital allocation. For the purposes of our study, our focus will be solely on the relative performance of issuers.

Introducing a stratification that is progressive and relates to economic wealth overcomes some of the challenges with existing forms of ESG factor integration, such as negative exclusion based on objective ESG laggards, which contributes to a set of economic incentives that perpetuate a self-enforcing capital bifurcation cycle and entrench ESG laggards, particularly developing market countries, in a development trap (Henide 2021).

Isolating the female agency-driven development factor

Constructing long-only portfolios

We apply the framework methodology to output relative performance,⁷ the difference between the actual and implied scores divided by the implied score, for the sovereign issuers represented in the iBoxx USD Emerging Markets Sovereigns Index, "the benchmark".

We construct two sets of long-only portfolios in Fig. 4. The first set consists of market-value weighted portfolios containing all of the bonds in the benchmark issued by the top quintile (Q1) of sovereign issuers based on an issuer ranking by 'performance margin' for each individual factor. Conversely, the second set consists of market-value weighted portfolios containing all of the bonds in the benchmark issued by the bottom quintile (Q5) of sovereign issuers based on an issuer ranking by 'performance margin' for the Overall PES score as well as for each individual factor.⁸ In-line with the universe, issuers are re-assessed and re-ranked periodically, at every month-end rebalancing, to eliminate survivorship bias.

Fig. 4

Source: IHS Markit, World Bank, Authors

Cumulative log returns of the long-only leader and laggard portfolios.

Deriving excess return portfolios

To better illustrate the performance of our leader and laggard portfolio sets, we calculate a cumulative log excess return relative to the cumulative log return of the benchmark in Fig. 5. We find that the portfolios of leaders have a higher cumulative log return from inception than the benchmark across the majority of months, whilst portfolios of laggards have a lower cumulative log return than the benchmark across the majority of months in the study window.

Fig. 5

Source: IHS Markit, World Bank, Authors

Leader and laggard excess return versus the benchmark.

The high degree of pairwise correlation demonstrated in Table 4 may suggest some redundancy among the framework pillars and data inputs, but all elements of the framework are retained for the fullness of the assessment; the female agency framework does not look to prescribe variables that provide significance and additionality in determining asset pricing, but rather a scorecard for measuring gender equality holistically. Said differently, the selection of the pillars looks to satisfy conceptual wholeness and the prescribed pillars together constitute the definition of the female agency factor; the ‘combination’ of the factors in this context is unencumbered by multicollinearity as they otherwise would be had they been selected as independent variables in a model that sought to relate to a given dependent variable, for example. Regardless, the factors are isolated across each individual vertical for context and, as explored in the study, the high pairwise correlation is a result of, and a testament to, the regional consistency among sustainability hegemons across key female agency concepts. The co-relationship between the pillars is further illustrated in Fig. 8 (appendix, Exhibit 2) (Table 5).

Table 4 The factors demonstrate a high degree of pairwise correlation.

Table 5 Calendar returns of the Overall PES leader and laggard portfolios.

The membership and respective weightings of the top and bottom quintile portfolios for the most recent month in the study and further descriptive statistics are produced in Table 6 (appendix, Exhibit 1) and Table 7 (Appendix, Exhibit 2), respectively. An additional assessment of the leader and laggard portfolios’ sensitivities to key benchmarks is also provided, in Fig. 9 (Appendix, Exhibit 3).

Deriving factor portfolios

We combine the long-only leader and laggard portfolios, by subtracting the cumulative log returns of the laggards from the cumulative log returns of the leaders (Q1–Q5), to derive factor performance in Fig. 6, deconstructing the average monthly log returns of the overall bottom and quintile portfolios in Table 5.

Fig. 6

Cumulative log returns of the factor portfolios. Source: IHS Markit, World Bank, Authors

Triangulating leader and laggard portfolios' social and governance profile

Triangulating the academic findings that societies with a greater degree of female agency have a stronger social and political infrastructure, we introduce a dataset based on an independent methodology and sources: the Freedom House 'Global Freedom Total Scores'.

A 'Global Freedom Total Freedom Score' is a combination of a country or territory's political rights and civil liberties and is referenced as a proxy for the protection of citizens across the socioeconomic spectrum from discrimination, repression, and the infringement of their fundamental rights. We calculate the hypothetical 'Global Freedom Total Score' for the overall leader and laggard portfolios based on the portfolio constituents for each year of history (2012–2021, inclusive) and plot the results against the portfolios' hypothetical IIGNI.

Corroborating the literature, the leader portfolio consistently captured 'freer' countries, as illustrated in Fig. 7. Furthermore, the leader portfolio displayed 'stickiness' (less year-on-year variability); leader countries demonstrated hegemony across all three factors and were also likely to maintain their hegemony over time, the reverse was true for laggard portfolios.

Fig. 7

Source: IHS Markit, World Bank, Freedom House, Authors

Leaders are, consistently, freer.

Conclusion

In this paper, we integrate a female agency-driven development framework in the systematic construction of factor portfolios to assess whether bonds issued by economies with a greater degree of gender equality outperform those issued by economies with a lesser degree of gender equality from within the iBoxx USD Emerging Markets Sovereigns Index universe. We hypothesise that this approach to integrating sustainability considerations should contribute to financial gains for investors, from a halo effect that is otherwise discounted by traditional assessments of creditworthiness. Furthermore, we propose that the consideration of female agency-driven development widens the investor toolbox for assessing sovereign creditworthiness and allows investors to construct portfolios more optimally with respect to isolating the most effective sovereign transmission mechanisms of sustainable development capital.

Through developing factor portfolios, we find that a portfolio of issuers in the top quintile of performers⁹ across either of the three verticals (Politics, Economics, Society, collectively abbreviated to 'PES') outperformed their peers in the bottom quintile of performers across each respective vertical in the long-run.

Additionally, we find that the top quintile performers across all of the verticals and over time tended to be from a narrow cluster, causing the top quintile performers to register a similar performance and demonstrate a high degree of pairwise correlation. This cluster was narrower than the laggard cluster (although laggard portfolios were also subject to a high degree of pairwise correlation). The hegemony of the top quintile of performers across time contributed to a 'stickier' portfolio and a lower cumulative portfolio turnover over the study window.

Triangulating the top and bottom quintile portfolios with Freedom House 'Global Freedom Total Scores', an external dataset with an independent methodology and data sources, we corroborate the underlying academic literature, proposing the intuition that citizens whose countries have a higher degree of female inclusion, equality, and empowerment provide a benefit to and benefit from a higher degree of political rights and civil liberties.

The findings are constructive in contextualising the behaviour of the debt issued by more gender-equal economies versus their less gender-equal peers but do not necessarily establish an explicit causal relationship between gender equality and factor outperformance. There is scope for further research to consider and establish a causal relationship for the impact of female agency on the credit risk of top quintile versus bottom quintile sovereign issuers over time.

Appendices

Appendix

Exhibit 1

See Table 6.

Table 6 Composition of leader and laggard portfolios (a/o 30 Jun 2021).

Exhibit 2

See Table 7 and Fig. 8.

Table 7 Descriptive statistics

Fig. 8

Source: IHS Markit, World Bank, Authors

Monthly log returns by factor pairings.

Exhibit 3

Table 8 and Fig. 9.

Table 8 Data permalinks

Fig. 9

Source: IHS Markit, World Bank, FRB, IMF, U.S. Bureau of Labor Statistics, Authors

Overall PES leader and laggard portfolio sensitivities.