Simple is simply not enough—features versus labels of complex financial securities

The high complexity of some financial products like Asset-Backed Securities (ABS) has led to severe problems such as adverse selection and moral hazard, and ultimately to the financial crisis. While tightened regulatory requirements addressing security design features seem beneficial in mitigating these issues, investors might be inclined to rely on aggregated, easily accessible labels signaling a high quality. Empirically, existing studies have predominantly focused on analyzing the consequences of individual design features, such as risk retention (Ashcraft et al., 2019), specific categories of design features, such as complexity (Ghent et al., 2019), or quality labels, such as AAA ratings (Chen et al., 2020; Mählmann, 2012). However, a joint consideration, which relevance design features on the one hand and quality labels on the other hand have for the pricing of complex financial securities, is missing. Thus, it remains uncertain to what extent investors prioritize detailed, albeit time- and cost-intensive, information for evaluating the quality of financial products. Alternatively, it is unclear to what extent they lean towards easily accessible information, leading prices to be driven more by the presence of quality labels rather than detailed features. Similarly, concrete evidence on how the performance of complex financial products can be explained by features versus labels is lacking. In light of these considerations, our primary objective is to investigate the impact of design features and quality labels on both the pricing and performance of complex financial products. In light of these considerations, our primary objective is to shed light on the following research question: How do design features versus quality labels affect the pricing and the performance of complex financial products?

To address this research question, we leverage the regulatory framework known as the “European Union’s Securitization Regulation” (EUSR) framework, enacted in January 2019. This framework aims to reduce asymmetric information of securitizations and restore trust in the EU securitization market. The EUSR framework introduced, first, tightened minimum requirements regarding the features of the security design, where “features” refer to attributes of financial securities, encompassing criteria and procedures related to the securitization structure and involved parties. Second, it introduced the optional STS label (“Simple, Transparent, and Standardized”), which indicates higher standards regarding the securitizations’ simplicity, transparency, and standardization, to facilitate investors’ risk assessment (EC, 2015). Within the context of the EUSR framework, our investigation focuses on understanding how the design features and labels of complex financial securities impact both (I) the pricing of these securities and (II) the performance of securitized loans. As a crucial case study, we analyze residential mortgage-backed securities (RMBS), representing a significant example of highly intricate financial instruments. A typical RMBS deal comprises around 20,000 underlying mortgage loans, with investor prospectuses explaining the deal’s structure in documents ranging from 250 to 450 pages. Thus, even if RMBS deals address institutional investors, the immense complexity makes a proper risk assessment a very difficult and expensive task.

Our analyses are grounded in a dataset comprising over 31 million quarterly loan-level observations from 186 RMBS deals. Additionally, we utilize hand-collected data encompassing 48 features of the security design, derived directly from the regulatory text. Our key findings reveal that the STS label has a substantial impact on tranche prices, especially for AAA tranches. Conversely, we observe no positive impact resulting from improvements in the securitization features. Furthermore, our analysis furnishes evidence indicating that the pricing cannot be justified solely by the performance of underlying loans. The implications of these results extend to regulators, investors, and originators. Regulators contemplating the adoption of concepts like STS (e.g., in the US) can leverage these findings to assess and formulate effective regulatory policies. We find that introducing the regulatory STS label induces the problem of investors relying too much on the quality signal again—similar to AAA-ratings for Asset-Backed Securities (ABS) before the subprime crisis (Coval et al., 2009a, b; Mählmann, 2012)—instead of analyzing financial products’ riskiness, which is particularly pronounced for risk-averse investors of AAA tranches. This finding starkly contradicts the regulatory authorities' intention to facilitate the risk assessment for investors, as the STS label induces a reliance on quality signals rather than a thorough analysis of financial products' riskiness. It is crucial to emphasize that investors should prioritize analyzing the design that shapes the incentive structure within deals, rather than relying on simplistic quality signals such as the STS label. However, originators currently mainly benefit from receiving the STS label and not from complying with better design features because the label induces lower financing costs.

To be more specific, based on various regression models—e.g., two-way fixed effects, logistic, and instrumental variables regressions—we find that particularly AAA investors seem to focus simply on the STS design label, whereas the actual performance of STS deals is not improved. Although the EUSR improves loan performance, the spreads of deals containing obligatory features increase, indicating inappropriate pricing by investors. STS-labeled deals benefit from substantially reduced spreads; we provide evidence that receiving the STS label—and not the superior security design of STS deals—explains this spread reduction. We show that even complying with all features is irrelevant for pricing, but only the label assignment has a spread reductional effect. Supporting this result, we find a within-tranche spreads’ decline of 13% at the time of receiving the label—although the information about the design features has already been available in the investor prospectus. This corresponds to a value of the STS label of € 2.3 million per annum for an average tranche.

The EU’s introduction of the STS label was intended to facilitate investors’ risk assessment instead of directly relying on the STS label as a quality signal of the underlying exposures. However, our results reveal that investment decisions are not driven by the deal’s security design features but mainly by the design label. To identify the relevant features of the security design, we collect the requirements concerning the design features stated throughout the regulatory text. These features comprise the EUSR features, which are obligatory for all EU securitizations issued after 2018, such as risk retention and the prohibition of cherry picking, and additional STS features for obtaining the STS label. While complying with the STS features is optional, an STS deal must fulfill all these additional features, e.g., the true sale of the assets, providing historical performance data, and hedging interest rate and currency risk. We assess the deals’ compliance with each EUSR and STS feature—totaling 48—before and after the regulation’s effective date.

Since the regulation recently entered into force, literature regarding the EUSR and the STS concept is scarce. We contribute to the more general literature on security design, quality labels of financial securities, and the optimal design of regulations. First, regarding the design of financial securities, the theoretical literature suggests that poor security design and subsequent informational differences between securitization parties can induce exploiting informational advantages (Leland and Pyle, 1977; Boot and Thakor, 1993; DeMarzo and Duffie, 1999; DeMarzo, 2005). The literature on the design of financial securities focuses particularly on risk retention, which is one of the required EUSR features, and complexity. Retention harmonizes originators’ and investors’ interest and induces increased loan and tranche performance as well as lowers credit spreads (DeMarzo, 2005; Hartman-Glaser et al., 2012; Chemla and Hennessy, 2014; Guo and Wu, 2014; Begley and Purnanandam, 2017; Hartman-Glaser, 2017; Vanasco, 2017; Hébert, 2018; Ashcraft et al., 2019; Flynn et al., 2020; Agarwal et al., 2021; Hibbeln and Osterkamp, 2024). Furthermore, originators use complexity to obfuscate the securitizations’ quality, as more complex securitizations have a lower quality which is not reflected in ABS pricing (Furfine, 2014; Ghent et al., 2019; Griffin et al., 2014). Thus, it seems promising that “simplicity” features are required in the STS framework. More generally, studies addressing the market’s complexity for funds and retail financial products provide evidence for strategic complexity, leading investors to misprice these financial products, which, in turn, incentivizes financial engineers to exploit their informational advantage (Stoimenov and Wilkens, 2005; Carlin et al., 2009; Carlin and Manso, 2010; Henderson and Pearson, 2011; Gennaioli et al., 2012; Sato, 2014; Ammann et al., 2017; Celerier and Vallee, 2017; DeHaan et al., 2021). We contribute to the literature by evaluating the effect of a feature score based on 48 features induced by the EUSR/STS framework instead of focusing on one individual design feature, and we show that investors seem to ignore these detailed design features when pricing complex financial securities but rather focus on the easily accessible label.

A second strand of literature focuses on quality labels of financial securities. Particularly for complex financial securities, investors have an incentive to focus on aggregated information and rely on labels as a quality signal to reduce asymmetric information. The investors’ reliance on credit ratings of securitizations before the financial crisis—particularly for “AAA” products—exemplifies the overdependence on such quality measures. Studies suggest that investors outsource parts of their risk analysis to rating agencies and shirk on their due diligence when investing in tranches from complex securities. A stronger reliance on labels for more complex securities implies that labels are more valuable if investors face complexity (Adelino, 2009; Griffin and Tang, 2012; Mählmann, 2012; Chen et al., 2020). For investment funds, investors consider quality measures and react disproportionally positively to funds labeled as high-quality while punishing funds whose quality label decreases below the top one-third category, with the effect seeming less pronounced for institutional investors. Therefore, their reaction, measured by flows into the funds, is driven only by changes in the quality measure and not by the underlying fund performance (Blake and Morey, 2000; DelGuercio and Tkac, 2008; Ammann et al., 2018). These findings are supported by studies on funds as well as structured and retail financial products, which add that investor groups tend to neglect certain risks and are less attentive to details (Arnold et al., 2021; Ferman, 2015; Gennaioli et al., 2012). We contribute to this field of literature by showing that complex financial securities pricing is mostly influenced by the label and not by the incentive structure established by the features. Institutional investors react positively to the quality label by demanding lower spreads while punishing non-labelled deals with significantly higher spreads, regardless of security design and loan performance. These findings reveal that introducing a quality label can entice investors to focus on a label and neglect a proper risk analysis despite the availability of extensive information on complex financial securities. This also highlights the danger that investors might rely too heavily on easily accessible information, similar to “AAA-ratings” for securitizations before the financial crisis.

A third strand of literature addresses the broad topic of the optimal design of regulations for financial products. The literature on securitization highlights the potential adverse effects of introducing regulations and reveals that market forces might be more beneficial in achieving welfare gains (Keys et al., 2009; Martin and Parigi, 2013). Focusing on the benefits of transparency as a regulatory tool, studies have revealed that mandating enhanced disclosure might lead unsophisticated investors to be disadvantaged (Pagano and Volpin, 2012; Balakrishnan et al., 2021). However, empirical studies of regulatory transparency initiatives show that the reduction of asymmetric information improves loan and pool performance (Ertan et al., 2017; Klein et al., 2020). Chen et al. (2020) emphasized the significance of stronger risk management by investors. Studies focusing on consumer protection measures reveal that while, generally, policies mandating transparency, standardization, and competition are beneficial, transparency measures might discourage private learning (Gabaix and Laibson, 2006; Carlin and Manso, 2010; Bertrand and Morse, 2011; Inderst and Ottaviani, 2012; Agarwal et al., 2015, 2020). Literature on the general effect of regulations on financial products reveals that transparency measures induce decreased informativeness of security prices. Mandating investors to strengthen their risk management can mitigate the issue of bounded rationality (Banerjee et al., 2017; Schwarcz, 2014). Furthermore, costly monitoring incentivizes small investors to free ride on monitoring efforts of large shareholders (Maug, 1998; Stoughton and Zechner, 1998). Research on consumer protection and consumer behavior includes consumers’ cognitive limitations and biases, which can reduce the effectiveness of regulatory transparency initiatives. Furthermore, a lack of consumer trust, in general, can cause consumers to avoid specific financial products (Campbell, 2006, 2016; Lusardi and Mitchell, 2007; Guiso et al., 2008; Choi et al., 2009; Christelis et al., 2010; Bertrand and Morse, 2011). We contribute to this literature field by providing evidence that introducing extensive transparency measures, combined with a publicly available quality measure, can disincentivize consumers and institutional investors to conduct a risk analysis and, moreover, foster investors’ reliance on pricing decisions made by other investors. This can lead institutional investors to misprice complex financial securities and thereby reduce the effectiveness of regulation.

In order to illuminate the influence of design features versus quality labels on the pricing and performance of complex financial products, we formulate several testable hypotheses by leveraging the European securitization regulation. The aggregate effect of the EUSR mandates enhanced design features, anticipated to diminish asymmetric information. Consequently, investors are expected to reward this improvement with reduced risk premiums across all tranches subject to the regulation. Furthermore, investors perceive the easily accessible STS label as a reliable quality signal. Consequently, it is expected that the credit spread of STS tranches will decrease. Thus, we formulate hypothesis H1a&b:

However, for EUSR tranches, which on average have improved design features but no STS label as a quality signal, two opposing effects might influence the spreads. While, on one hand, we anticipate a decrease in the spreads of EUSR tranches if the reduction of asymmetric information is predominant (albeit to a lesser extent than for STS, given the modest nature of the features' improvement), on the other hand, investors might interpret the decision not to issue an STS deal as a negative signal. Given the uncertainty ex-ante about which effect prevails, we formulate hypothesis H2a&b to address the opposing influences.

In addition, asymmetric information is of different importance to different investors. Tranches with a high credit rating, especially AAA tranches, exhibit relatively stable credit spreads over time and are less responsive to new information compared to equity or mezzanine tranches. As a result, investors in AAA tranches, such as credit institutions and insurance companies, are inclined to allocate less effort toward reducing asymmetric information through extensive risk analysis. They are more likely to rely on simpler measures, characterized by low information processing costs, such as a quality label. This leads to hypothesis H3:

The pricing of STS and EUSR tranches could not only be influenced by the information derived from the features and the label but also by different capital requirements due to the EUSR, which are particularly relevant for credit institutions and insurance companies; these financial institutions account for approximately one third of investments into EU securitizations. Nevertheless, our expectation is that the spread-reducing effect of the STS label is only partially attributed to reduced capital requirements. Accordingly, we test hypothesis H4:

Next, we focus on the performance of securitized loans by investigating whether the improved features of complex financial securities and their labels affect loan performance. Ex-ante, the aggregate effect of the EU securitization regulation on loan performance is unclear. The EU asserts in the preamble of the regulation that its primary objective is not to diminish pool risk but rather to address risks stemming from suboptimal security designs. However, the mandated enhancements in security design are intended to alleviate asymmetric information by aligning the incentives of originators and investors. This alignment is expected to prompt more rigorous screening and monitoring by originators, leading to superior loan performance. Consequently, we propose the following hypothesis:

If an enhanced performance is observed for such deals, it is reasonable to conclude that this improvement is primarily attributed to enhanced design features. However, if investors demand higher spreads for deals lacking the STS quality label (see H2b), it may suggest that they hold private information about lower loan quality in EUSR deals. Therefore, the additional risk premiums for EUSR tranches might serve as compensation for the potential inferior loan quality. If this holds true, the performance of EUSR loans should be significantly worse than STS loans and pre-EUSR loans. Nonetheless, we anticipate that enhanced design features will positively influence performance, regardless of whether the deal is classified as an STS or EUSR deal. Consequently, we posit that the label does not exert a significant effect on performance beyond capturing the improvements in features:

We implement several alternative specifications and robustness checks to underline the reliance of our results. In the context of our pricing analyses (chapter 4), we check for their robustness regarding alternative estimators and run the regression of Tables 5 and 7 by random effects instead of two-way fixed effects models. To check for functional form misspecifications, we make sure that our results hold for tranche ratings and sovereign ratings in non-linear functional forms; furthermore, we test whether the log transformation of the dependent variable Spreads has an impact on the results. Additionally, instead of trading day fixed effects, we check if our results hold if we control for macroeconomic factors such as the ABSPP when applying quarter or year fixed effects.

In the context of our loan level analyses (chapter 5), we check if our results are robust to changes in the estimation methods by implementing a linear probability model instead of logit for the regressions of Tables 8 and 9. Furthermore, in the IV setting, we utilize IV probit instead of 2SLS. In all of our analyses we check if our results hold if we do not winsorize the control variables or exclude outliers, run our analyses on subsamples (e.g. only post-EUSR period) and implement combinations of the abovementioned specifications and robustness checks. In all, our results do not change substantially and remain robust. The corresponding results are available upon request.

Our study aims to investigate the influence of features compared to labels on both the pricing and performance of complex financial securities. We leverage the European securitization regulation, a framework that on the one hand enhances requirements for the features of complex financial securities and on the other hand introduces the STS label for securitizations meeting supplementary optional criteria. The EUSR framework was designed with the purpose of easing the risk assessment process for investors and revitalizing the securitization market.

Our findings indicate that a substantial share of investors, particularly risk-averse AAA investors, tend to lean on the readily available information offered by the label, neglecting to consider the enhanced features when pricing complex financial securities. Contrary to pricing decisions, our analysis reveals that the actual performance is primarily driven by enhanced design features rather than the presence of the label. Our observations show a reduction in spreads for tranches in STS-labeled deals and an increase in spreads for tranches in non-STS deals, as compared to tranches issued prior to the introduction of EUSR. Interestingly, this observation stands in contrast to the reality that pre-EUSR deals exhibit inferior features in security design and inferior ex-post performance concerning non-performing loans and defaults. In summary, our results strongly suggest that investors tend to overlook a thorough risk assessment of complex securities. The misinterpretation lies in perceiving the issuance of deals with STS labels as a signal of superior performance, while deals without STS labels are erroneously considered signals of inferior performance.

Our findings carry significant implications, especially for regulators, investors, and originators in the realm of complex financial securities. For regulators, our results indicate that introducing extensive regulation can effectively mitigate asymmetric information in securitization markets. Nevertheless, a regulation incorporating both enhanced security design features and quality labels may lead to investors overly relying on the label rather than conducting a thorough risk assessment. This echoes the pitfalls of overreliance on AAA ratings prior to the financial crisis. Investors, based on our findings, should give precedence to the analysis of security design, as it significantly shapes the incentives of originators and, consequently, influences the resulting performance. However, it is imperative for originators to acknowledge that, at present, it is not solely the features of a proper security design but also obtaining the STS label that is crucial for signaling high asset quality. This results in significant spread reductions and lower financing costs. As an illustrative example, spreads, on average, decrease by 13% within a tranche after obtaining the STS label, equating to a reduction of the risk premium of € 2.3 million per annum for an average tranche. In conclusion, our results underscore that even well-intentioned regulatory measures aimed at enhancing simplicity, transparency, and standardization fall short of guaranteeing a thorough risk assessment by investors and mitigating asymmetric information in highly complex financial securities. Instead, measures fostering a harmonization of interests between originators and investors, such as the minimum risk retention, seem more promising. This ensures that even in the presence of information asymmetries, there are no significant disincentives. Furthermore, rather than solely emphasizing the STS label as a means to facilitate an appropriate risk assessment, regulators could compel financial institutions to actively conduct a proper risk assessment if they want to benefit from reduced capital requirements for investments in complex financial securities.