Research Analyst at Climate Bonds Initiative
Candace Partridge, EngD
University College London
Aneil Tripathy, MA
Researcher, Brandeis University
Whether green bonds deliver a cheaper cost of capital to issuers than vanilla bonds has been a contentious issue since the start of the green bond market. In the market’s early days anecdotal statements from green bond issuers that their bonds were being oversubscribed, resulting in a pricing difference against equivalent vanilla bonds, led market participants to argue that green bonds provide a cheaper cost of capital (Harrison 2017b). However, this anecdotal evidence was unverifiable until the market matured to a size sufficiently large enough to provide comparable bonds for analysis. The existence of a “greenium,” a green bond premium over equivalent vanilla bonds, became a key research point for green bond analysts as the market matured (Harrison 2017b; Preclaw and Bakshi 2015). This analysis spread to sustainable finance research centers and bond trading desks and has now become a mainstay topic of green bond conferences and market events (Flammer 2018; Torsten and Packer 2017). Within academic circles, numerous papers have recently focused on looking directly at pricing differences in the U.S. green municipal-bond market (Baker et al. 2018; Larcker and Watts 2019). However, these discussions have yet to provide conclusive evidence for or against the presence of a substantial greenium. The academic debate remains focused on refining a standard methodological approach by which to detect any greenium. Developments such as the green halo effect (Basar and Krebbers, 2019), which blurs the added value of green bonds for issuers by blending it with the issuer’s vanilla bonds, also make the academic search for a greenium insubstantial in relation to the green bond market’s overall dynamics.
Drawing from the social sciences of finance, this paper contextualizes green bond pricing research by examining recent greenium discussions and the role of the Climate Bonds Initiative (Climate Bonds) in these discussions (Beunza, Hardie, MacKenzie 2006; Muniesa 2007). We reflect on the beginning of the first green bond pricing research at Climate Bonds and analyze how these early conversations have evolved among both academics and market participants. Drawing from literature review, quantitative pricing data, and qualitative data from semi structured interviews with market participants, we argue that the iterative nature of pricing discussions is a result of both pricing methodologies and market growth dynamics.
1. Introduction: What’s in a Greenium: An Analysis of Pricing Methodologies and Discourse in the Green Bond Market
Whether green bonds deliver a cheaper cost of capital has been a contentious issue since the start of the green bond market. This debate centers on whether there is a pricing difference between vanilla and green bonds. However, what pricing research coming out of financial institutions and universities misses is that the pricing debate has meaning only if it is contextualized in the conversations between issuers, investors, and other entities in the green bond market. In this paper, we attempt to provide this vital context for understanding pricing conversations in the green bond market. We are a multidisciplinary team consisting of an anthropologist, a climate finance researcher, and a green bond analyst, who have been involved in the green bond market for a cumulative total of 14 years. Here, we lay out the history of pricing analyses and debates in the green bond market. In this endeavor, we expand on the social science study of the “technicality of financial markets” and the performativity of prices (Beunza, Hardie, MacKenzie 2006, 721; Callon 2015; Muniesa 2007) by examining closely both the history of the quantitative analysis that produces a greenium as well as the social and market context around this topic.
2. The Origin of Greenium
In the market’s early days, there were anecdotal statements from green bond issuers that their bonds were being oversubscribed, resulting in a pricing difference in relation to equivalent vanilla bonds and leading some market participants to argue that green bonds provide a cheaper cost of capital. However, until the market matured to a size large enough to amass enough comparable bonds for analysis, this anecdotal evidence had been unverifiable. The pricing difference search between vanilla and green bonds began with a white paper published by Barclays’ analysts (Preclaw and Bakshi 2015). Two years later, the concept of the greenium arose from internal discussions at Climate Bonds. The earliest publication mentioning a greenium to our knowledge is a Climate Bonds’ pricing paper for the organization’s annual conference in March 2017 (Harrison 2017a).
Climate Bonds’ Caroline Harrison coined the term in collaboration with colleagues. According to Harrison, “I met with Sean [CEO of Climate Bonds] in April 2016, he asked me whether I could find any evidence of green bonds pricing differently from vanilla bonds. He had heard market participants talking about green bonds pricing with lower yields than vanilla equivalents, and he thought it could be an interesting hook to encourage more issuers to print green bonds.”
Ordinarily, a bond issuer pays a yield slightly above the “market” to issue a new bond. This is known as a new issue premium, and the price of the bond is therefore slightly cheaper for the buyer. This is a normal feature of the new issue market. Around 2016, however, market talk was that green bonds were being priced with a new issue discount, that is, slightly more expensive than existing debt. This also occurs in the normal new issue market and is contingent on the strength of investor appetite. Investor appetite is dependent on multiple factors from macroeconomics, individual credit preferences, and concurrent bond issues on that day (Harrison 2017b).
In their early research, Harrison and the Climate Bonds team wanted to know whether this discount was consistently present in green bonds, or if the green label could influence pricing. This novel focus on pricing in the green bond market needed a name, and the name emerged through deliberation at Climate Bonds. In Harrison’s words, “I had some conversations with Andrew Whiley, Head of Communications at Climate Bonds, about labeling this difference (real or perceived). This was around the time that the UK voted to leave Europe, and that process had been termed Brexit, which we liked, and which had been incorporated into the vernacular. . . . Andrew loved the greenium term and immediately began to use it in communications. We were thrilled when we noticed the term had been used by an independent third party.”
Focusing on the greenium, Harrison spent the first couple of months informally looking at spreads of green bonds in the secondary market. Most of the large green bonds were issued either by energy/utilities and financials or by supranational, sub-sovereign, and agency issuers (SSAs). She looked at EUR, GBP, and USD denominated bonds and compared them with vanilla bonds of the same issuer. Instinctively, the Climate Bonds team had not expected to find any differences, given that the bondholder would be facing the same entity irrespective of whether the bond was labeled green. SSA bonds were roughly trading in line, and for certain EUR denominated green-labeled corporate credits, spreads were tighter (Harrison 2017b). The logical explanation for this was the green label, along with the yield, also gained from moving further down the credit curve.
In 2017, Climate Bonds began to produce a market monitor, looking at the primary market performance of the largest bonds issued in three-month windows (Harrison 2017b; Harrison 2018). The reports began with a focus on bonds with a minimum size of USD 300 million, including only those denominated in EUR and USD to establish a critical mass of bonds for analysis. Metrics of analysis were evidence-based, offering insight into supply and demand dynamics of green bonds. Book building, spread compressions, and yield curves were analyzed to determine the presence of the greenium and the performance against matched indices in the immediate secondary market (pricing + 30 days). The reports later added comparisons to vanilla bond baskets and the amount of each green bond sold to investors declaring themselves as green.
The secondary market observations of the Climate Bonds’ pricing report series are based on baskets of bonds sharing similar characteristics to the green bond in terms of credit rating, sector, maturity, seniority, and emerging/developing market. The indices used are the broad iBoxx indices in which green bonds currently sit. The motivation for these comparisons was to see how new green bonds perform against non-green bonds issued in the same three-month window, and to see how the new green bonds performed against the secondary markets.
3. Growing Green Bond Market Pricing Analysis
Alongside Harrison’s and the Climate Bonds’ work, the search for greenium initially started by looking for differences in yields for corporate green bonds compared with non-green corporate bonds (Climate Bonds Initiative and IFC 2017; Preclaw and Bakshi 2015). Recent studies that explore this pricing dynamic include Hachenberg and Schiereck (2018), Bachelet, Becchetti, and Manfredonia (2019), Kapraun and Scheins (2019), Wulandari et al., (2018) and Zerbib (2018). All of these studies focus on using the matched pair analysis method to compare the yields of green bonds with their closest equivalent non-green counterparts, some of which may be synthetic. This is also the approach taken in the Climate Bonds Initiative (2017 and 2018) pricing reports. Overall, there has been a wide variation in greenium results reported so far in the literature for corporate green bonds.
In searching for a greenium, bond pricing research needs consistent, reliable prices. Bonds are often priced using theoretical prices (sum of discounted cash flows), or using spread techniques, where bonds are set at a spread against a liquid benchmark and move in parallel. This does not strictly reflect market demand or activity for individual bonds. Since a sum total of green bonds were scarce in 2017, the Climate Bonds team was skeptical about the accuracy of data and reluctant to put too much emphasis on longitudinal studies in such a nascent market.
In the last two years, this relative paucity of data available for matched-pair analysis of corporate green bonds has led several green bond researchers to focus their analysis on the U.S. green municipal bond market. As of March 2020, total issuance in the U.S. green municipal market has reached $40.4 billion (Climate Bonds Initiative 2020). The U.S. green municipal bond market is unique in that it is largely tax exempt and has smaller green bonds issued more frequently, which enables more direct comparisons. One of the first analyses of this market was published by Karpf and Mandel (2018), whose dataset included 1,880 municipal bonds that were labeled green by Bloomberg, and which were compared with 36,000 conventional bonds by the same set of issuers from 2010 to 2016. Their results indicate no clear greenium until 2016, where they subsequently found a mean spread of 23 basis points (bp) (Karpf and Mandel 2018).
Karpf and Mandel’s work was followed up by Baker et al. (2018), who performed an analysis of 2,083 municipal bonds defined as “green” by Bloomberg. Their comparison bond data was comprised of 643,299 conventional municipal bonds, also issued from 2010 to 2016. In this paper, their focus was on the primary market, and their regression analysis found an average greenium at issue of 6 bp. The bonds used for this analysis included taxable and tax credit muni bonds along with the tax-exempt bonds, so they took the step of adjusting the equivalent yields before doing the regression analysis, in contrast to Karpf and Mandel, who did not adjust their equivalent after-tax yields. As a result, Baker et al. assert that the reason that Karpf and Mandel failed to find a greenium in the early years was because “early green bonds were disproportionately taxable,” and state that “our results suggest that this conclusion is incorrect” (Baker et al. 2018).
Further narrowing the focus to a sample of 640 pairs of matched green and non-green municipal bonds issued from 2013 to July 2018, Larcker and Watts (2019) found a nominal green discount of 0.45 bp, with the difference in price at issue being zero in 85% of the matched cases. They also found negligible greenium when their analysis was expanded to include neighboring bonds issued by the same issuers but at separate times. This work also found no significant difference in liquidity or institutional ownership levels and no pricing difference for certified green bonds. Overall, they state that “our results suggest that municipalities actually increase their borrowing costs by issuing Green bonds,” and further, they claim that regression findings from previous works are “insufficient to effectively control for non-linearities and issuer-specific time variation which ultimately leads to spurious inferences” (Larcker and Watts 2019). One key aspect that these analyses have neglected is the potential change over time in the behavior of greenium in the municipal bond markets.
Building from these greenium analyses, a paper by Partridge and Medda (2020) performs a matched-pair analysis, but for 453 matched pairs of green and vanilla bonds issued from 2013 to 2018. The paired bonds in their sample were issued at the same time and under the same official statement, such that they have the same issuer, use of proceeds, issue date, maturity date, and coupon. This analysis looks at greenium in both the primary and secondary markets, and furthermore breaks down the pricing differences into yearly averages to detect trends in greenium as time progresses and the market grows. Partridge and Medda observe a greenium that grows to nearly 5 bp in the secondary market by 2018. While no statistically significant differences in greenium were observed in the primary market, they found that during 2017 and 2018, in the cases where paired bonds were issued, there are pricing differences.
All of these studies have relied on some form of regression and/or matched-pair analysis, but even the municipal bond studies still experience small sample sizes, largely due to the fact that the green municipal bond market in the U.S. only started in 2013, and its issuance has remained relatively low in comparison to the overall market. It may be that there is simply not enough data to be able to support the assertion of any greenium in the primary market yet, but this could change as the market grows. This could also be related to the finding of several studies (Partridge and Medda 2020) of evidence of a greenium in the secondary market, where there are many more data points than in the primary.
4. Contextualizing Pricing Research in Market Talk
The growing number of greenium analyses and debates among academics and green bond analysts since the start of Caroline Harrison’s research coincides with continual market commentary on the existence of a greenium. Among green bond market participants, perspectives on a greenium rest on market positions. On the issuer side, green bond issuers claim to be beyond, and others claim to be at par. As the market has developed, commentary on the investment side has changed. Some investors are arguing that they are investing in green bonds at par with vanilla bonds, while others say that they give financial preference to green bonds.
A panel titled “Pricing deep dive: greenium, halos and trajectories” at the Climate Bonds Conference 2019, which focused on pricing discussions in the market, demonstrates the difference between academic pricing discussions and market participant debate (Climate Bonds Initiative 2019). This panel was made up of members of “bond syndicate and origination” at UniCredit, SEB, Credit Suisse, JP Morgan, and other market entities. Pricing dynamics in the green bond market were discussed at length, but in a very different and much more pragmatic manner than in the pricing research we have
The panel recognized the proliferation of the greenium conversation in both academic and market circles, with one participant noting that “the term greenium had been adopted to mean many different things.” According to the participant, greenium has been used to describe bonds being priced underneath the issuer yield curb, while issuers have adopted the term to mean they shaved a couple of basis points off what they thought they would get from the market for an issuance. While the bond syndicate and origination heads were aware of and noted respectively assessed academic pricing research, their description of the decision making that goes into actual green bond deals differed greatly from the focused pricing description of academics. As one syndicate head noted, “[t]here is still a lot of art rather than science in syndicate pricing on desks.” In constructing green bond transactions, syndicates bring in market feeling and conversation into their analysis, as opposed to the overall market analysis of academic research. In this vein, when asked whether there is a greenium in the green bond market, a panelist stated, “intuitively there should be a greenium.” When pressed to elaborate on what this intuition is based on, the panelist argued that “the thing is, economics is not like physics, ultimately it’s about people.” In the green bond market, the process by which prices are put forth into the market is as much from current pricing analysis as it is from human relationships and discussion (Guyer 2009; Muniesa 2007). This pricing that aggregates in a potential yield curve is both “an epistemic and affective object” (Zaloom 2009). Economists and academics working on the greenium highlight its epistemic use, while the syndicate desk members on the Climate Bonds panel highlight its affective role in the green bond market itself (Çalışkan 2005). Prices are a synthesis of negotiations between people and institutions at distinct moments (Ferry 2016). The academic focus on pricing conversations is a result of the lens of economic analysis, while green bond prices themselves are generated by the motivations of the issuer, underwriters, verifiers, and investors involved in each issuance.
Along with the direct discussion on pricing, the panel also discussed the numerous side effects and complementary impacts of green bonds that complicate a singular search for a pricing difference. The green halo effect, the tendency of a green bond issue to positively impact both the bond and equity pricing of a green bond issuer, also complicates the quantification of a greenium (Basar and Krebbers 2019). As one bond syndicate head on the panel argued, “a green bond is of course a loudspeaker; it is the best way for me to communicate directly to the market on my sustainability.” This communication impacts multiple relationships between an issuer and its market relationships to investors and underwriters, which in turn influences multiple prices. In this context, issuing a green bond is a performative act by an issuer, which, when recognized by investors and underwriters, can produce a greenium (Callon 2015; Beunza, Hardie and Mackenzie 2006). The existence of a green halo effect complicates the search for a greenium in that the halo extends to vanilla bond issuance as well (Basar and Krebbers 2019).
Reflecting on the weight of arguments and statements on green bond pricing from issuers, underwriters, and investors, Harrison from Climate Bonds argues that “this is the thing I cannot measure. . . . When issuers say this and there is no benchmark for comparison, I am skeptical. I trust the syndicate response because they are the ones doing this work . . . I buy what the syndicates say, but I cannot prove it either.”
With regard to green bond pricing dynamics, discussions from panels, such as the one at the Climate Bonds’ Annual Conference, provide a reflection that is closest to direct-pricing decision making in the green bond market. The consensus of the panel that “intuitively there should be a greenium,” illuminates how the concept of a greenium already exists for green bond dealmakers. Echoing this, from the results of investor surveys carried out by Climate Bonds, investors seem to overweigh green bonds in bond holdings regardless of a pricing difference (Almeida, Harrison, and Sette 2019).
The comments presented here from this conference panel are important not only in relation to academic greenium debates but also in relation to official statements from green bond market entities. While much more aggressive green bond buying seems to be the norm from off-the-record or Chatham House rules events, official statements are much more conservative. As Marilyn Ceci, Head of Green Bonds at JP Morgan, states in her official statement for this paper, “Green Bonds price on market. Generally, they price in line with traditional bonds, but occasionally demand outstrips supply and they can price a few basis points tighter.”
This statement needs to be taken at face value in signifying the type of institution JP Morgan is. In its green bond work, JP Morgan has been actively bringing new green bond issuers to market, and this added supply impacts the pricing dynamics of the market, but for its role as an underwriter, conservative stances are required.
The initial creation of the greenium through conversations at Climate Bonds and through Caroline Harrison’s pricing research has generated both academic and market discussions, propelling critical reflection on what green bonds are and what characteristics they offer issuers, investors, and other market participants. The value of a greenium and the continuation of the academic debate in published papers on its existence has to be placed in the wider context we have traced throughout this paper if it is to have any relevant meaning beyond academic embroiling. The greenium’s production is a consequence of market activity and subsequent academic analysis impacting these market dynamics (Mackenzie 2006; Zaloom 2009). Neither a greenium nor the green bond market is static—and as an increasing amount of academic study focuses on green bonds, this reality must be taken into account alongside attempts to analyze certain objects in the market, such as the existence of a greenium.
Caroline Harrison formally joined the Climate Bonds Initiative in April 2016 to work on research projects, including green bond pricing. Caroline has a fixed income background. She worked at ASSET4 Thomson Reuters to develop and market a model to classify the nonlisted portion of a bond portfolio according to ESG observations. Previously, Caroline spent four years in the Index and Portfolio Strategy team at Merrill Lynch and eight years as a Research Analyst at Morgan Stanley. Caroline has a BA in Italian and Business Studies from UCL.
Dr. Candace Partridge completed her doctorate at University College London, where her research focused on green bonds in the U.S. municipal bond market and ways to help finance more sustainable infrastructure. She has also done postdoctoral research on the economics of plastic waste management infrastructure, and her research continues to focus on the the intersection of sustainable urban infrastructure and climate finance.
Aneil Tripathy is a PhD Candidate in economic anthropology at Brandeis University, where his research focuses on the growth of climate finance and the green bond market. Aneil has been a visiting researcher at Cass Business School, University College London, and the Pentland Centre at Lancaster University. He has worked in climate finance for five years as a researcher, executive associate, and academic research consultant at the Climate Bonds Initiative, and as a consultant for the Clean Energy States Alliance. Aneil has training in environmental economics, ethnographic research, and systems thinking.
Almeida, Miguel; Caroline Harrison, and Pietro Sette. 2019. “Green Bond European Investor Survey.” Climate Bonds Initiative with analysis support from Henley Business School. Available from https://www.climatebonds.net/files/files/GB_Investor_Survey-final.pdf.
Baker, M., D. Bergstresser, G. Serafeim, and J. Wurgler. 2018. “Financing the Response to Climate Change: The Pricing and Ownership of U.S. Green Bonds.” Brookings Institution. Available from https://doi.org/10.3386/w25194 or https://www.nber.org/papers/w25194.
Bachelet, M. J., L. Becchetti, and S. Manfredonia. 2019. The Green Bonds Premium Puzzle: The Role of Issuer Characteristics and Third-Party Veriﬁcation. Sustainability 11(4): 1–22. ISSN 20711050. Available from https://doi.org/10.3390/su11041098 or https://www.mdpi.com/2071-1050/11/4/1098.
Basar, S., and A. Krebbers. 2019. “Green Bonds May Have Halo Effect.” Available from https://www.marketsmedia.com/green-bonds-may-have-halo-effect/.
Beunza, Daniel, Iain Hardie, and Donald MacKenzie. 2006. “A Price Is a Social Thing: Towards a Material Sociology of Arbitrage.” Organization Studies 27(5): 721–745.
Buhr, Bob, and Ulrich Volz, Charles Donovan, Gerhard Kling, Yuen Lo, Victor Murinde, and Natalie Pullin. 2018. Climate Change and the Cost of Capital in Developing Countries. London and Geneva: Imperial College Business School, London; SOAS University of London; UN Environment. Available from http://unepinquiry.org/wp-content/uploads/2018/07/Climate_Change_and_the_Cost_of_Capital_in_Developing_Countries.pdf.
Çalışkan, Koray. 2010. Market Threads: How Cotton Farmers and Traders Create a Global Commodity. Princeton: Princeton University Press.
Callon, M. 2016. “Revisiting Marketization: from Interface-Markets to Market-Agencements.” Consumption Markets & Culture, 19(1): 17–37. Available from https://doi.org/10.1080/10253866.2015.1067002.
Climate Bonds Initiative (CBI). 2019. “Agenda Outline—Climate Bonds Conference.” Available from https://cbi19.climatebonds.net/files/files/CBI19-Agenda-18-Feb.pdf.
Climate Bonds Initiative (CBI). 2020. “US Munis mar20.” Climate Bonds Initiative.
Ehlers, Torsten, and Frank Packer. 2017. “Green Bond Finance and Certification.” BIS Quarterly Review (September). Available from https://www.bis.org/publ/qtrpdf/r_qt1709h.htm.
Ferry, Elizabeth. 2016. “Gold Prices as Material-Social Actors: The Case of the London Gold Fix. The Extractive Industries Society 3:82–85.
Flammer, Caroline. 2019. Green Bonds: Effectiveness and Implications for Public Policy (No. w25950). National Bureau of Economic Research. Available from https://www.nber.org/papers/w25950.
Guyer, Jane. 2009. “Composites, Fictions, and Risk: Toward an Ethnography of Price.” In Market and Society: The Great Transformation Today, edited by C. Hann and K. Hart, 203–220. Cambridge, UK: Cambridge University Press.
Hachenberg, Britta, and Dirk Schiereck. 2018. “Are Green Bonds Priced Differently from Conventional Bonds?” Journal of Asset Management 19(6): 371–83. Available from https://doi.org/10.1057/s41260-018-0088-5.
Harrison, Caroline. 2017a. “Green Bond Pricing in the Primary Market: Q4 2016 Snapshot.” London: Climate Bonds Initiative. Available from https://www.climatebonds.net/files/files/March17_CBI_Briefing_Primary_Market.pdf.
Harrison, Caroline. 2017b. “Green Bond Pricing in the Primary Market: Jan 2016–March 2017.” London: Climate Bonds Initiative and IFC. Available from https://www.climatebonds.net/files/files/CBI-Green-Bond-Pricing-Q2-2017.pdf.
Harrison, Caroline. 2018. “Green Bond Pricing in the Primary Market: Q3 2017, July–September 2017.” London: Climate Bonds Initiative. Available from https://www.climatebonds.net/files/reports/greenium_q3-final-20180219.pdf.
Kapraun, Julia, and Christopher Scheins. 2019. “(In)-Credibly Green: Which Bonds Trade at a Green Bond Premium?” Proceedings of Paris December 2019 Finance Meeting EUROFIDAI – ESSEC. Available from SSRN: https://ssrn.com/abstract=3347337 or http://dx.doi.org/10.2139/ssrn.3347337.
Karpf, Andreas, and Antoine Mandel. 2018. “The Changing Value of the ‘Green’ Label on the US Municipal Bond Market.” Nature Climate Change 8:161–165. Available from https://doi.org/10.1038/s41558-017-0062-0.
Laborda, Ricardo, and Jose Olmo. 2014. “Investor Sentiment and Bond Risk Premia.” Journal of Financial Markets 18:206–233.
Larcker, David F., and Edward Watts. 2019. “Where’s the Greenium?” SSRN Electronic Journal. Available from https://doi.org/10.2139/ssrn.3333847.
Mackenzie, Donald. 2006. An Engine, Not a Camera: How Financial Models Shape Markets. Cambridge: The MIT Press.
Muniesa, Fabian. 2007. Market technologies and the pragmatics of prices.” Economy and Society 36(3): 377-395. Available from https://doi.org/10.1080/03085140701428340.
Partridge, Candace, and Francesca Romana Medda. 2020. The evolution of pricing performance of green municipal bonds. Journal of Sustainable Finance & Investment, 10(1): 44–64. Available from https://doi.org/10.1080/20430795.2019.1661187.
Preclaw, Ryan, and Anthony Bakshi. 2015. “The Cost of Being Green.” Barclays.
Wulandari, Febi, Dorothea Schäfer, Andreas Stephan, and Chen Sun. 2018. “The Impact of Liquidity Risk on the Yield Spread of Green Bonds.” Finance Research Letters 27 (December 2018): 53–59. Available from https://doi.org/10.1016/j.frl.2018.02.025.
Zaloom, Caitlin. 2009. “How to Read the Future: The Yield Curve, Affect, and Financial Prediction.” Public Culture 21(2): 245–68.
Zerbib, Olivier David. 2018. “The Effect of Pro-Environmental Preferences on Bond Prices: Evidence from Green Bonds.” Journal of Banking & Finance 98. Elsevier B.V.: 39–60. Available from https://doi.org/10.1016/j.jbankfin.2018.10.012.