Investment strategies of bank-affiliated and independent venture capitalists: a focus on innovation in the fintech sector in the wake of the global financial crisis

Despite the significant increase in interest in fintech in the academic literature in recent years (Allen et al. 2021; Block et al. 2018; Bollaert et al. 2021; Farag and Johan 2021; Giudici et al. 2021; Goldstein et al. 2019; Ughetto et al. 2021), existing studies remain loosely connected (Goldstein et al. 2019; Kavuri and Milne 2019) and provide little insight into fintech VC. Current research on fintech VC is limited and examines either the macro drivers of fintech VC investment or the role of VCs in the emergence of the fintech market (Chemmanur et al. 2020; Cumming and Schwienbacher 2018; Kolokas et al. 2022). Some recent research emphasizes the interplay between fintechs and banks but without focusing on bank-affiliated VCs (Brandl and Hornuf 2020; Hornuf et al. 2021). This is surprising, given that VCs play an important role in fostering technological advances that are blurring industry boundaries and revolutionizing the financial industry with the introduction of new and alternative digital advisory and trading and payment systems (Farag and Johan 2021; Philippon 2016; Ughetto et al. 2021).

The study by Cumming and Schwienbacher (2018) analyzes the factors that influence the level of VC in the fintech industry. The study finds that differential regulatory enforcement between startups and large, established financial institutions after the GFC has driven VC investment in fintech. A related article is that of Haddad and Hornuf (2019), who analyze fintech startups in 55 countries and provide evidence that more fintech startups are created in more developed economies, and where VC funding is easily accessible. However, the emergence of fintech ventures is positively influenced by VC funding when there is a critical mass of fintech entrepreneurship in a country, such that VC and credit markets are substitutes, especially in countries characterized by more dynamic fintech entrepreneurship (Kolokas et al. 2022).

Two recent papers have also started to discuss how banks interact with fintechs. Hornuf et al. (2021) use hand-collected data on banks in Canada, France, Germany, and the UK to investigate the forms of collaboration between fintech startups and banks. They find that banks prefer forming alliances with startups with a well-defined digital strategy and/or employing a chief digital officer. Similarly, Brandl and Hornuf (2020) conducted a network analysis of banks and fintechs in Germany and concluded that most relationships between banks and fintechs are product-related collaborations.

The global economic downturn led to a significant reduction in VC activity due to the close relationship between the VC market and the general economic climate (Block and Sandner 2009; Mason 2009). The VC industry was significantly and adversely affected by the crisis due to several factors, including the lack of liquidity in the IPO and M&A markets, the decline in portfolio company valuations, and the lack of funding that affected many private VC investors as a result of the financial difficulties faced by insurance companies and large banks (Bertoni et al. 2019; Block and Sandner 2009).

The onset of the GFC created a new dynamic in the VC market, forcing VCs to reduce their activities (Mason 2009) and change their investment behavior (De Vries and Block 2011). Block and Sander (2009) compared average VC investment decisions in the US before and after the GFC. They found that after the crisis, VCs, on average, selected older companies and made later-stage investments in larger syndicates. In contrast, early-stage investments were less attractive to VCs, especially corporate VCs. During the liquidity supply shock, investing in core sectors became more attractive for VCs, with core startups receiving 9.4% more funding than non-core startups compared to normal times (Conti et al. 2019). VCs reduced their propensity to syndicate investments (De Vries and Block 2011) and IPOs became less common, largely replaced by acquisition exits (Cumming and Johan 2013).

BVCs have been compared with IVCs regarding their selection dynamics, monitoring practices, and value-added activities (see Table 16 for an illustration of key differences). In terms of investment selection practices, it has been recognised that financial institutions that extend their knowledge to VC investments create opportunities to build future relationships with their clients. The aim is to establish a connection with a company at the VC stage with the expectation of generating potential clients for their underwriting and lending activities in the future (Andrieu 2013; Croce et al. 2015; Hellmann 2002; Hellmann et al. 2008). As such, BVCs tend to pay more attention to credit scoring variables (based on financial statement analysis) than to assessing a company's innovation potential (Croce et al. 2015). They are under less pressure than IVCs to exit early and generate abnormal returns (Manigart et al. 2002) because they can obtain additional sources of funding from the bank and do not need to raise funds from limited partners (Andrieu and Groh 2012; Croce et al. 2015). Given the generally risk-averse nature of parent banks, BVCs prefer to invest in companies that can be easily monitored (e.g., local) and from which it is easy to gain insights into the local market (Bertoni et al. 2015). They also seem to prefer to invest in more diversified and later-stage portfolios, with a lower probability of default, to reduce overall risk (Croce et al. 2015; Cumming et al. 2008; Dimov and Gedajlovic 2010; Yoshikawa et al. 2004) compared to IVCs. In contrast, IVCs actively seek high-growth innovative companies to maximise returns by betting on a particular innovative product's technological feasibility and market viability. The primary objective for IVCs is to achieve successful portfolio exits or abnormal returns as they manage funds from external investors (Dimov and Gedajlovic 2010; Gompers and Lerner 2001). In terms of monitoring practices and value-adding activities, BVCs tend to be less actively involved in the management and monitoring of their portfolio companies than IVCs, which can ultimately affect the overall performance of the firm (Baum and Silverman 2004; Hellmann and Puri 2002). BVCs are generally not involved in the day-to-day operations of the companies in which they invest and, like other captive investors, have limited decision-making autonomy. This can hinder their ability to help companies achieve a successful exit (Croce et al. 2015; Cumming et al. 2008; Yoshikawa et al. 2004).

We draw on neo-institutional theory, in particular the concept of mimetic isomorphism, to explain how BVCs respond to the GFC and align with the selection strategies of IVCs in fintech investments. Mimetic isomorphism involves the achievement of uniformity with other organisations through imitation. Institutional theorists argue that the tendency for organisations to copy each other stems from their efforts to gain or enhance legitimacy in the field (Barreto and Baden‐Fuller 2006; Deephouse 1999; DiMaggio and Powell 1983; Oliver 1991, 1997; Pedersen and Dobbin 2006) and to improve their performance (Oliver 1991; Sirmon and Hitt 2009). Mimetic isomorphism is also a simple and effective response to uncertainty (DiMaggio and Powell 1983, p. 151). Indeed, one of the pressures driving the mimicry of business practices is the turbulence in the organisational and environmental environment in which organisations operate (DiMaggio and Powell 1983). In a stable and predictable environment, organisations look to their peer group and benchmark themselves against the group’s recognised standards. In times of crisis, the boundaries and membership of reference groups can become blurred, affecting the dynamics of mimetic isomorphism (Davis et al. 1994). The emergence of a crisis can destabilise an organisation, undermine resource flows, and reshape the distinctive patterns and processes that characterise it, possibly redirecting its search for new points of reference. Thus, organisational efforts to emulate other organizations are intensified in times of heightened uncertainty because organizations believe that doing so reduces the costs associated with decision-making (Cyert and March 1992) and improves their performance (Oliver 1991). Additionally, emulating other organisations in turbulent times may stimulate the exploration of new opportunities that were previously out of reach or little considered (Paruchuri and Ingram 2012), and improve the ability to make complex decisions through peer observation and iteration (Lévesque et al. 2009).

Drawing on neo-institutional theory and the concept of mimetic isomorphism, we theorize that the GFC exerted mimetic isomorphic pressure on BVCs to align with the selection strategies pursued by IVCs in fintech investments. This alignment is related to the increased role that the incorporation of technological advances plays in the business of banks.

It is widely accepted that BVCs, by their nature, tend to pay more attention to credit scoring variables (based on financial statement analysis) than to assessing a company’s innovation potential (Croce et al. 2015). They are under less pressure than IVCs to exit early and generate abnormal returns (Manigart et al. 2002) because they can obtain additional sources of funding from the bank and do not need to raise funds from limited partners (Andrieu and Groh 2012; Croce et al. 2015). Given the generally risk-averse nature of parent banks, BVCs prefer to invest in companies that can be easily monitored (e.g., local) and from which it is easy to gain insights from the local market (Bertoni et al. 2015). They also seem to prefer to invest in more diversified and later-stage portfolios to reduce overall risk (Cumming et al. 2008; Dimov and Gedajlovic 2010; Yoshikawa et al. 2004) compared to IVCs. As such, BVCs are more similar to banks than VCs in their selection activity, favoring firms with a lower ex-ante probability of default (based on financial data) and potentially disregarding innovation potential. This description of BVCs fits well with the pre-GFC period when technology integration was not yet central to discussions on bank stability. The notion that technological advances could improve monitoring and screening, thereby increasing resilience in times of crisis, was not prevalent during this period. The stability of the environment in those years favoured internal benchmarking practices within the group of BVCs. This meant that BVCs were less active than IVCs in investing in start-ups introducing disruptive technologies, thus favouring the selection of firms that appeared more promising in terms of financial strength rather than innovation potential compared to IVCs. In other words, BVCs were more tolerant of firms with lower innovation intensity (and innovation quality) and potentially more subject to innovation failure than IVCs.

The destabilisation caused by the GFC meant that BVCs faced external pressures that imposed difficult imperatives on how to bring technological advances to the core banking business in order to remain competitive, with little guidance on how to do so. In fact, while the banking industry has been increasingly confronted with new disruptive technologies and has been able to integrate most of them into the digital back-end, it has been much slower to internalize the digital servitization of front-end services and financial products (Navaretti et al. 2017). The GFC accelerated the overwhelming impact of technology on the financial sector, making banks aware of the importance of incorporating technology-driven configurations into their businesses. The GFC reinforced the notion that technological advancements can significantly alter how information is processed, affecting both lending and underwriting activities. We argue that the turmoil caused by the GFC exerted mimetic isomorphic pressure on BVCs, which changed their investment practices towards those of IVCs, with a tendency to select more innovative fintech firms.

The unfolding of the GFC led BVCs to pursue a strategy of imitating IVCs as a viable solution to reduce uncertainty (Ashworth et al. 2009; DiMaggio and Powell 1983; Hu et al. 2007) and bring technological advances to the core banking business, which has traditionally been characterised by low levels of innovation (Beck et al. 2016). As the fintech and innovation environment was rapidly evolving, the GFC pushed banks to reorganise their activities by improving the digital services offered to their customers in order to be more resilient in a future crisis. Banks became keen on integrating big data by leveraging fintech technologies to improve the efficiency of processing hard information and overcome the time-consuming and labour-intensive processing of soft information (Balyuk et al. 2020). In the case of fintech BVC investment, the GFC acted as a catalyst for mimetic isomorphism (DiMaggio and Powell 1983). Mimicking the IVC selection strategy towards more innovative ventures meant that BVCs could respond to the uncertainty of the changed economic environment and explore newly available technological alternatives to legitimise their parent banks’ current business practices (DiMaggio and Powell 1983; Hannan and Carroll 1992; Meyer and Rowan 1977). Our overarching hypothesis is, therefore, as follows:

We obtained the data used in this paper from Crunchbase, a unique database that captures worldwide information about VC investments in high-technology ventures. For each portfolio venture, Crunchbase reports information on the date of each investment and the investing VC firms. The database provides additional details on the ventures, including the foundation year, the country of operation, a description of the industry in which the venture operates, the number of financing rounds received, the amount of money raised in each financing round, and the type of funding received. The dataset also reports information on investors that can broadly be classified as individuals and financial organizations (e.g., VC and private equity firms). The present analysis is based on data obtained from Crunchbase in November 2019.

The initial dataset contains information about 440,810 investments related to 120,427 ventures. We proceeded through several steps in the selection of the sample.

First, we only focused on fintech ventures. Following Haddad and Hornuf (2019), we resorted to the following procedure to identify fintech ventures:

Second, we focused on first-round investments (i.e., those in which a given venture receives financing for the first time) since follow-on investment decisions are qualitatively different from initial investment decisions (Podolny 2001). This filter reduced our dataset to 26,169 first investments, referring to 12,087 fintech ventures.

Third, we restricted the analysis to fintech ventures financed by VCs, thus excluding individual investors. More in detail, we included in our sample BVC investors and, for comparison purposes, IVC investors. Thus, we only considered fintech ventures that received their first financing round from an IVC or a BVC, thus excluding rounds syndicated by the two types of VC investors. This filter is motivated by the aim of comparing the selection strategies of these two VC types in terms of the innovation level of invested ventures. The presence of co-invested ventures in our sample would influence and bias our results. We obtained 7175 first investments in fintech ventures, 225 of which (3.14%) were financed by BVC investors, and the remaining 6950 (96.86%) were financed by IVC investors.

Finally, we excluded observations for which we had missing data for the control variables used in our econometric model. The final sample consists of 6711 fintech ventures invested by either IVCs (6509 ventures, 96.99% of the sample) or BVCs (202 ventures, corresponding to 3.01% of the sample).

Table 2 reports the distribution of our samples in terms of foundation year, country of operation, fintech services category, and investment year. Both total distribution and distribution by typology of VC investors are reported.

Looking at the total sample, almost 11.43% relates to investments made in the years before 2008 (i.e., the starting year of the GFC), while the remaining 88.57% refers to investments made in the years following the GFC. This is driven by the fact that the Crunchbase coverage has increased over the years. As to the geographical distribution, 51.26% of sample ventures are located in the U.S., while most of the remaining ventures are located in Europe (24.51%), with 618 U.K. ventures (9.21%) and the remaining 1027 from countries belonging to the European Union (15.30%). Sample ventures provide different fintech services, even if the largest majority (34.72%) operate in the General Fintech Services sector. Payment (18.01%) and Financing (15.78%) represent other significant categories of fintech services providers in our sample. Significant differences exist between companies backed by BVCs and those backed by IVCs based on χ² tests comparing their foundation year, location, and fintech services. This implies the importance of adding these variables as controls in our selection model, which will be detailed in Sect. 5. The same reasoning holds for the amount received by VCs in the first round of financing, with BVC-backed companies receiving significantly higher financing than IVC-backed ones. As a result, we will also incorporate the received amount in our control variables.

We measured innovation by resorting to the number of patents and forward patent citations as a proxy of patent quality. In the fintech sector, patent filings have increased steadily over the years, although the quality of these patents has often been reported to be low (Lerner et al. 2015). The innovative potential of a venture is often signaled to investors by means of the patents owned (Lerner 2002). Besides being an exclusive right that grants legal protection to an invention from imitation by competitors, a patent can be an important devise to signal that the firm has innovation capabilities and a specific technological positioning (Chang 2012; Hoenig and Henkel 2015) and that an invention is worth being protected (Caviggioli et al. 2017; Caviggioli and Ughetto 2016).

Patents are a reliable indicator of a company's innovation endeavors for two main reasons. Firstly, applying for a patent is expensive and requires significant resources, implying that only highly valuable innovations are likely to be patented. Secondly, patenting activity is typically associated with product innovation rather than process innovation (Teece 1988). However, the effectiveness of patents in protecting technological innovations is contingent upon the product's characteristics, which varies across different sectors. For instance, certain innovations that form the basis of digital products, such as computer code, may not be effectively protected by patents. Other digital products that are novel or distinct may be instead effectively protected by patents (Boudreau et al. 2022).

Research has shown that startups filing patents have a greater likelihood of receiving VC financing (Hoenen et al. 2014; Mann and Sager 2007; Munari and Toschi 2015) and that patents are associated with favorable firm valuations by VCs (Baum and Silverman 2004; Hsu and Ziedonis 2013). For VCs, the fact that startups own patents indicates the presence of collaborative R&D, inventiveness, and technological dynamicity of the firm.

We identified all patents associated with sample ventures by resorting to the proprietary patent search database Derwent Innovation provided by Clarivate. We collected information on patents and patent quality for each venture over time. The innovation measures are based on the patent application year (i.e., the year in which the patent application is filed) since this is closer to the time of the actual innovation (Griliches et al. 1987). The patent count is measured as the number of patent applications filed (and subsequently granted) by the venture in a given year. To capture the importance (technical merit) of each patent, we constructed two measures of quality based on forward-citation counts. The technological worth of an innovation is often estimated in scientific literature by analyzing forward patent citations, which are indicative of the patent's value and technical significance. The use of forward citations as indicators of patent value was first introduced by Trajtenberg in 1990. Since then, numerous empirical studies have validated the usefulness of forward citations in assessing the value of patents (e.g., Harhoff et al. 2003; Jaffe and De Rassenfosse 2017; Reitzig 2003; Sapsalis et al. 2006). Following Hall et al. (2001, 2005), the citation truncation problem was corrected using citation-lag distribution. Consistent with the literature, we used two variables: (1) forward 4-year citation, defined as the number of forward citations within four years of filing for all patents filed in a given year, and (2) citation per patent– defined as the average number of forward citations received by each patent. Natural log transformation was used to counter the right-skewness of the variables.

As shown in Table 3, out of 6711 ventures (total sample), 1016 resulted in having at least one patent (15.14% of the total sample), both before and/or after the investment round. More in detail, patenting ventures comprise 991 IVC-backed ventures (15.23% of the IVC-backed sample) and 25 BVC-backed ventures (12.38% of the BVC-backed sample).

We observe 3441 patents filed by these 1016 patenting ventures, 72 of which are associated with BVC-backed ventures (2.09%), while the remaining 3369 refer to IVC-backed ventures (97.91%). Out of the total 3441 patents, 537 (15.61%) are filed before the entry of the VC in the venture’s equity capital, 325 (9.44%) are filed in the same year of the receipt of VC financing, while the remaining 2579 (74.95%) relate to patents filed after the entry of the VC in the venture’s equity capital. The availability of data about patents and patent quality in the post-investment period allows us to provide some evidence of the impact of VC investments on the innovation rate of invested ventures, as shown in Sect. 6.

Our final sample is thus composed of 6711 observations relative to investment years (325 of which are also patenting years) plus 3116 observations relative to additional patenting years before or after the investment year, for a total of 9827 observations.

Secondly, we complemented our data with information about the innovation level of the invested ventures by resorting to different proxies deriving from text mining and semantic network analysis.

First, we considered a measure of innovation based on semantic text analysis performed on the business description of ventures invested by BVCs and IVCs (i.e., considering the products and services they offer and the other prominent attributes emerging from their business description). For this analysis, we combined methods and tools of text mining and network analysis (Wasserman and Faust 1994). In particular, we used the SBS BI app (Fronzetti Colladon and Grippa 2020) to pre-process the textual data by removing stopwords (i.e., words that add little meaning to each description, such as the terms “and” or “the), punctuation, links, and special characters, transforming all the text into lowercase, and removing word affixes. This last procedure is known as stemming (Porter 2006); it makes sure that related terms (such as “mystery” and “mysterious”) map to the same word stem.

After preprocessing, we transformed the corpus into semantic networks, where each node represents a word/concept, and links among words are weighted based on their co-occurrence in the text (Diesner 2013). This procedure is preliminary to identifying the main topics emerging from the different venture descriptions, which we carried out following a network topic modeling approach. This choice is aligned with past research that showed the advantage and reliability of adapting methods from community detection in networks (Lancichinetti et al. 2015).⁷ Specifically, we identified the main topics by finding the most meaningful clusters in the semantic network using the Louvain algorithm (Blondel et al. 2008). In addition, we used the formula proposed by Fronzetti Colladon and Grippa (2020) to identify the most representative words for each topic and their importance scores. In particular, we valued more terms with strong links within their cluster and weak external connections. This process led to the identification of four main topics, described in Table 4. At the same time, we obtained information to carry out a reverse fit and evaluate how much each venture description was related to the different topics.

Our semantic network analysis of the venture descriptions led to identifying four significant topics, described in Table 4. As shown in the table, Topic 1 is the most relevant and concerns the most innovative traits of the analyzed ventures. It includes information about the advanced technologies they use and the systems/devices they developed to offer innovative products and services—for example, related to cyber security, data protection, machine learning, or business intelligence and analytics. Topics 2 and 3 are less relevant in the corpus. The former concerns trading and payment systems offered by some ventures, including online platforms used for e-commerce or other financial transactions, such as money lending. Topic 3 is focused on people, both employees and customers. In this case, descriptions relate to how ventures can satisfy customers’ needs, improve their loyalty, or provide solutions to support teamwork and improved decision-making. Topic 4 is the second most relevant topic, and it is related to financial services and products, software, and management systems designed for banks and financial institutions. Among all these topics, only the first one is related to the innovative traits of venture activities. We thus resorted to Topic 1 as a measure of innovation in ventures invested by BVCs and IVCs investors.⁸ Through the SBS BI app, we matched the words used in the company descriptions with those most representative of each topic. In this way, we could measure the weight of Topic 1 in the text and assign to all ventures a metric of innovation (Innovation topic). Moreover, we also defined a high-innovative dummy (Top innovator) indicating ventures for which the value assigned to the Innovation topic exceeded the 90th percentile of the distribution. See Table 5 for a complete description of the variables used in this study.

Despite the growing interest in fintech and some preliminary evidence on fintech VC (Chemmanur et al. 2020; Cumming and Schwienbacher 2018; Kolokas et al. 2022), little is known about the relative attention that different VC types devote to the degree of innovation of target ventures in their selection processes and how this is affected by the GFC. The literature has shown that different governance configurations may affect the risk attitude, investment preferences, and expected returns of VC investors (Bertoni et al. 2015; Croce et al. 2015), in a way that can vary with changing market conditions (Bertoni et al. 2019).

Given the highly innovative nature of the fintech sector, our main focus was examining how BVCs compare to IVCs in their selection dynamics of innovative ventures before and after the GFC. Our study offers several contributions to the stream of research on entrepreneurial finance and crisis response. On the theoretical forefront, we extend the applicability of the theory of mimetic isomorphism to the equity finance sector and add insights by disentangling the pre- and post-crisis periods. An important extension of previous studies on VC financing is the development of the basic argument that different VC types can uniform to each other in their investment practices following an exogenous shock. By establishing a link between the notion of mimetic isomorphism, the GFC and the investment selection strategies of VCs, the comparison between the investment strategies of BVCs and IVCs becomes a novel angle to study. Similarly, we inform neo-institutional theory more generally by proposing a novel application context beyond the conventional discourse within international business management (e.g. to explain firms' decisions on offshore outsourcing or entry strategies into foreign markets).

Our results indicated that, in the pre-crisis period, BVCs did not seem to pay exclusive attention to the innovation potential of target firms and selected less innovative firms (in terms of innovation potential, number of patents, and patent quality) compared to IVCs. This attitude is explained by the tendency of BVCs to be less subject to pressures determined by time-oriented performance, as it happens with IVCs (Andrieu 2013; Croce et al. 2015; Hellmann 2002; Hellmann et al. 2008). Accordingly, BVCs endorse a longer-term investment horizon that makes them more tolerant of innovation failures or quality. We then developed theoretically grounded arguments based on the concept of mimetic isomorphism from neo-institutional theory. The GFC exerted mimetic isomorphic pressure on BVCs which changed their investment practices towards those of IVCs, with a tendency to select more innovative fintech firms. This change in attitude is linked to the fact that after the crisis, banks went under greater pressure to endorse a digital transformation and enhance their digital capabilities. Banks became aware: i) that advances in fintech allow the implementation of faster, more customer-centric, and user-friendly operations, which ultimately increase the overall efficiency and transparency of banking services (Barrett et al. 2015; Brandl and Hornuf 2020; Chemmanur et al. 2020; Sangwan et al. 2019); ii) of the need to rapidly adapt their financial services offerings to evolving demand in order to remain competitive and to face the entry of new fintech players, which are increasingly taking over functions that were traditionally within their domain (Hornuf et al. 2021; Jakšic and Marinc 2019; Sangwan et al. 2019). Mimicking the IVC selection strategy towards more innovative firms was seen as a viable solution to reduce uncertainty and nurture banks’ core business banks with technology-driven solutions.

Some possible limitations of our study and areas for future research merit consideration. Firstly, the study focuses on the fintech sector, which, while appropriate given its innovative nature, may limit the generalizability of our findings to other industries. Future research could explore whether similar patterns hold in other high-tech sectors or traditional industries. Secondly, the study relied on patents and a novel metric incorporating text mining and semantic network analysis to capture the innovation potential of fintech startups. While these metrics provide valuable insights, they may not capture the full innovation potential of startups within the rapidly evolving fintech landscape. Future research could explore alternative metrics, or a combination of indicators, to offer a more comprehensive understanding of innovation in this context. Furthermore, our innovative methodology could be employed to analyze various textual documents pertaining to the company, extending beyond its description. This includes scrutinizing project documents, examining patent content, or delving into the knowledge exchanged among employees through email communications. For example, researchers could consider patent citations or use text mining to derive measures of patent impact (Arts et al. 2021; Kelly et al. 2018). This would enable a classification of the innovative capabilities of fintech startups based on the characteristics of their patent outputs.

Additionally, the study predominantly focused on the GFC, leaving room for further investigation into the dynamics of VC selection strategies during other economic downturns or periods of financial instability—such as the COVID-19 pandemic, which represented another significant exogenous shock to the global economy. Moreover, our research highlights the shift in the behavior of BVCs towards more innovative ventures post-crisis. However, the specific mechanisms and internal changes within BVCs that facilitated this transformation remain unexplored. Future studies could delve into the organizational structures, decision-making processes, and internal adaptations within BVCs that enable a shift towards greater innovation focus. New methods, such as experimental approaches (e.g., Lohrke et al. 2010), may allow researchers to delve deeper into these issues, as has been seen in other types of venture financing—see, for example, Ademi et al. (2023) in the corporate venture capital (CVC) context, or Block et al. (2019) in VCs, BAs, and family offices.

Lastly, while our study distinguishes between BVCs and IVCs, it does not fully account for the potential heterogeneity within these categories. For instance, BVCs affiliated with different types of banks (e.g., commercial vs. investment banks) or IVCs with varying strategic focuses might exhibit different investment behaviors. Future research could delve into more granular categorizations of VCs and consider additional types, such as CVCs and family offices.

Our study has important policy and managerial implications for fintech entrepreneurs looking for financing and BVCs investing in the fintech industry. First, it seems that a good innovation potential (number of patents and quality) is important when searching for BVC financing after the GFC, as it is for IVCs. Thus, besides solid financial statements, fintech ventures must also be aware that BVCs are concerned about innovation. This is conducive to the idea that the potential opportunities for fintech entrepreneurs supported by BVCs are broad, as with IVCs. Besides receiving capital resources and value-added services to fuel innovation, BVC-backed fintech ventures may have the possibility of accessing banks’ broader customer base, receiving better credit conditions from banks afterward, getting access to banks’ advice on how to cope with regulation, extensive social relations, industrial hub contacts. From an investor’s point of view, our results offer insights into how the shift in the investment strategy of BVCs after the crisis can potentially affect their portfolios’ risk/return profile. Moreover, a broader question is how the banking industry will be reshaped as banks increasingly absorb technological breakthroughs of fintechs through their VC arms.

Exploring the patterns and attitudes to innovation of BVCs compared to IVCs in the light of the exogenous shock caused by the GFC in fintech is relevant and timely because it presents both a challenge and an opportunity for policy: a challenge in terms of moving beyond the conventional wisdom that VC types have different investment strategies that fail to adapt to the dictates of a changing economic environment and an opportunity in terms of transforming the VC'ecosystem' into a more vibrant and competitive environment. Since fintech startups are increasingly growing in number, there is a pressing need for policymakers to develop regulatory initiatives that can enhance the development and application of digital technologies in the financial sector. A deeper understanding of the conditions under which VCs select innovation in the fintech sector and how BVCs are paving the way is important for policymakers who want to enhance the funding opportunities for fintech startups, promoting the fintech industry and, at the same time, setting the boundaries so that innovations can be compliant with regulations without losing their enabling technological power.