Introduction

In the past decade, financial technology (FinTech) has become widespread. Due to the rise of FinTech, the financial industry faces a formidable challenge in the form of fierce competition with FinTech firms. FinTech now attempts to disrupt the global insurance industry. Traditional insurance companies must keep up with the FinTech trend and adopt novel tactics to coordinate with the growth of FinTech and InsurTech. As a result of the rapid growth of InsurTech since 2010, many industry professionals regard the digitisation of insurance as the next significant opportunity following Fintech. According to the report of World Economic Forum (2015), the development of FinTech will have a greater impact on the insurance industry than on other financial institutions (e.g. banking and the securities industry). Future disaggregation of the insurance value chain will profoundly affect the characteristics of the insurance industry.

According to Neale et al. (2020), the three primary roles of InsurTech companies in the insurance industry are disruptors, disintermediators, and enablers. InsurTech companies are technology firms founded on technological innovations (e.g. telematics and blockchain). To avoid disruption from InsurTech firms, insurance companies should invest in technology using external or internal InsurTech-oriented investments to increase firm efficiency and decrease operation costs. Several responses to the disruptive effect of InsurTech companies are offered. Concerning external investments, mergers and acquisitions (M&A), venture capital investments, and strategic alliances with InsurTech companies may be effective strategies. From the perspective of internal investments in technology, digitalisation, automation, and big data analytics (BDA), linked technological applications could be leveraged to boost the efficiency and synergy effects of the organisation. Under the influence of InsurTech, insurers must invest in external companies in the form of collaborative InsurTech ventures in addition to their internal InsurTech initiatives to achieve competitive advantages. Eling and Lehmann (2018) state that digital innovations impact most insurance industry activities, including the insurance value chain, product development and insurance policy, underwriting and claims, sales and distribution, pricing and asset–liability management, and risk management. Therefore, to raise a company's competitiveness, insurers must expand their technology hardware and software to facilitate the transformation of the value chain, process optimisation, and operational efficacy. It is preferable for insurers to develop a new business model that incorporates (IoT), BDA, and know-your-customer (KYC) in order to deliver complete financial services. Cappiello (2018) demonstrates how digitalization will profoundly alter the financial and insurance ecosystem, reshaping the competitive landscape and affecting all points along the insurance value chain. In addition, Grima et al. (2020) argue that blockchain technology can facilitate and expedite claims settlements and improve fraud control. They also propose that insurers better recognise how blockchain can be used across different insurance tasks to understand its benefits for employees, management, and customers.

In accordance with co-opetition theory, Hung and Luo (2016) propose that insurance companies can either invest in technology or collaborate with new InsurTech startups. On the one hand, insurance firms can invest in technology to implement digital financial reforms, such as telematics, blockchain, and digitalization. In addition, they can collaborate with InsurTech startups by investing in shares, M&A, and strategic partnerships. This paper analyses the technology investments of insurance companies rather than insurers' corporations with InsurTech startups to understand the impact of technology investments on firm performance.Footnote 1

Figure 1 depicts the total and average amount in dollars insurers spend on technology investments.Footnote 2 Overall, technology investments demonstrate the cycle phenomena. During the 2008 subprime mortgage crisis, insurers were forced to reduce technological spending until approximately 2010. After that, annual growth was consistent. It demonstrates that after 2010 insurance companies began to recognize the threat posed by FinTech and InsurTech and began investing in technology. As shown in Fig. 2, despite the rapid expansion of InsurTech, the average EDP_asset and EDP_inv of insurance companies decrease annually, indicating that insurers do not raise their InsurTech-oriented investments annually, where EDP_asset and EDP_inv are defined as electronic data processing equipment and software to total assets and total invested assets, respectively. This analysis reveals that EDP_asset and EDP_inv are dropping annually, which may be attributable to insurance companies' expanding size and investment position. Notably, Fig. 2 demonstrates that insurance companies' average EDP_EI and EDP_EP increase annually due to the cost or depreciation of EDP, demonstrating that insurance companies constantly expand their expenditures on technology, where EDP_EI and EDP_EP are defined as cost or depreciation of EDP to total expenses incurred and expenses paid, respectively. Even though insurers may utilise a variety of fixed-asset depreciation methods, the annual growth in cost or depreciation of EDP indicates that insurers consistently invest in technology. Despite the fact that insurers' investments in technology are declining annually (as depicted in Fig. 1), if insurers invest a certain percentage of their budgets in technology each year, the long-term cumulative effect of technology investments may enable insurance companies to generate long-term positive outcomes, such as the advantages of technology resource integration, long-term efficiency, and investment synergies. Moreover, the long-term firm performance of insurers would be enhanced.

Fig. 1
figure 1

The trend of technology investments ($)

Fig. 2
figure 2

The trend of technology investments (%)

In the insurance industry, investments in FinTech and InsurTech are well known. However, these InsurTech-oriented investments are not readily visible until insurers voluntarily disclose their InsurTech-oriented investments. Even if insurers declare their FinTech- or InsurTech-oriented investments to the market, it remains challenging to predict how and how much they will invest. Neale et al. (2020) investigate the factors influencing insurance companies' technology investments. They used the EDP expense allocation to underwriting and claims adjustment processes to examine the relationship between insurance production and capital investments in insurance technology. Therefore, this paper primarily adheres to the measurement rationale of Neale et al. (2020) and uses the EDP on the asset and expense sides to measure insurers' technology investments.,Footnote 3Footnote 4

In addition, a survey of insurers’ InsurTech-oriented investmentsFootnote 5 revealed that EDP expenses are relatively high among insurers with many InsurTech-oriented engagements and investments. Whether insurance companies are partners in strategic alliances or invest in InsurTech startups, these investments or engagements will inevitably increase the investments in related hardware and software equipment due to the need to connect and match the core technology or system of InsurTech startups. This indicates that the associated cooperation costs will increase when the insurance company collaborates with InsurTech startups. These findings demonstrate that an insurance company's EDP expenses strongly correlate with its overall InsurTech-oriented engagements and investments. Thus, the untabulated survey results also indicate that the EDP measure utilized in this study is a reliable proxy for measuring insurers' overall InsurTech-oriented engagements and investments.

Previous research has focused on the relationships between firm performance and various firm-specific characteristics, such as corporate governance, CEO compensation, risk-taking, organisational form, business diversification, and corporate social responsibility (e.g. Simsek 2007; Core et al. 1999; Kang and Shivdasani 1995; Sun et al. 2013; Fields et al. 2012; Lai and Limpaphayom 2003; Elango et al. 2008; McGuire et al. 1988; Cummins and Weiss 2000). Since technology investments bolster a company's fundamental competitiveness, it is essential to comprehend how technology investments contribute to achieving business objectives. The non-financial sector has produced a substantial amount of literature examining the effects of technology investments on corporate performance (e.g. Kauffman and Weil 1989, 2000; Rai et al. 1997; Mahmood and Mann 1993). Overall, the literature provides contradictory findings on the effect of technology investments on company performance and productivity. literature also revealed that the effectiveness of technology investments might emerge after multiple investment periods (Brynjolfsson 1993; Devaraj and Kohli 2003; Mahmood and Mann 2000; Lee et al. 2016). The time lag or cumulative effect explains why the relationship between technology investments and short-term corporate performance is negative or null. Therefore, insurers' investments in overall InsurTech-oriented investments may result in improved long-term performance, as the benefits of such cumulative InsurTech-oriented investments are recognised over time.

Harris and Katz (1991) propose that insurers with superior performance have a greater ratio of IT costs to overall operating costs and a lower ratio of IT costs to premium income. According to Francalanci and Galal (1998) and Cummins and Xie (2013), a rise in IT expenditures or technology investment correlates favorably with productivity growth and company efficiency. Although Harris and Katz (1991) drew a perceptive conclusion regarding the relationship between technology investments and business performance, they did not consider InsurTech-oriented investments under the InsurTech wave. This study differentiates from Harris and Katz (1991) by incorporating InsurTech-oriented investments.Footnote 6 Che et al. (2021) suggest that insurers' underwriting and overall performance will be enhanced if they adopt usage-based insurance (UBI) for their auto line of business early. The findings indicate that insurers' investments in InsurTech improve their business success.

In the insurance literature, relatively few researchers explore the association between technology investments and firm performance. To fill the gap in the literature, the purpose of this study is to examine how insurers' overall InsurTech-oriented investments affect their performance in the short-term and whether insurers' cumulative InsurTech-oriented investments under the wave of InsurTech increase their firm's performance in the long term, utilising the special and unique information set of technology investments provided by the NAIC annual statement.

This study utilises unbalanced panel data for U.S. property-liability insurers from 2006 to 2019 to assess the performance of firms under the influence of InsurTech. The Two-Stage regression analysisFootnote 7 reveals that the effects of InsurTech-oriented investments are predominantly adverse to short-term firm performance, suggesting that insurers who invest in InsurTech-oriented technology tend to perform poorly due to immediate expenditure increases. In addition, the results of Heckman's two-stage regression model for the EDP dummy measure indicate that insurers' InsurTech-oriented investments have zero net present value and are worth exactly what they cost (Dos Santos et al. 1993; Mahmood and Mann 2000; Chae et al. 2014). Importantly and intriguingly, insurers' cumulative InsurTech-oriented investments are significantly positive for overall firm performance, indicating that insurers' InsurTech-oriented investments will lead to better long-term performance because the benefits of such cumulative investments will be realised in the years to come (Campbell 2012; Che et al. 2021).

The primary contributions of this study are as follows. First, in contrast to Harris and Katz (1991), this study adopts a sample of insurers to analyze the relationship between InsurTech-oriented investments and firm performance during the wave of InsurTech. Even though Che et al. (2021) have specifically discussed the relationship between insurers' firm performance and the UBI program in auto line, to the best of my knowledge, this study is the first to investigate how insurers' overall InsurTech-oriented investments influence their financial performance in terms of integration firm-level data of InsurTech-oriented investments. Hence, this paper fills the literature gap on firm performance analysis by focusing on insurers' InsurTech-oriented investments. In addition, the findings of this paper are complementary and consistent with those of Che et al. (2021), indicating that insurance companies' long-term performance is dependent on technology investments. The effect of technology investments on insurance companies' performance should be considered.

The second feature of this study is that it utilizes a unique and specialised dataset (i.e. EDP) to assess insurers' InsurTech-oriented investments. EDP presents insurers' investments or expenses in electric data processing, equipment, and information software, including hardware and software InsurTech-oriented investments. Technology investments in this study refer to the sum of investments related to digital transformation and technology innovation. These capital investments or expenditures (expenses) relate to investments in InsurTech, such as AI, BDA, HRM, CRM, IoT, and blockchain. Thus, during the wave of InsurTech, insurers update hardware and software devices through digital transformation and collaborate or partner with InsurTech startups to improve the efficiency and cost-effectiveness of their value chain, including claim settlement, underwriting, customer service, marketing, and pricing. Based on the measurement rationale of Neale et al. (2020) and the survey of insurer’s overall InsurTech-oriented investments, an insurers’ EDP investments and expenses are highly correlated with its overall InsurTech-oriented investments, implying that EDP is a good proxy for measuring insurers' InsurTech-oriented spending during study periods.

Third, the research indicates that insurers' InsurTech-oriented investments significantly negatively impact their short-term performance, supporting the yearly expenditure and spending rise argument. Intriguingly, the evidence indicates that the cumulative effect of insurers' InsurTech-oriented investments strongly correlates with their long-term business performance. Time lag and cumulative effects are offered to support this conclusion. Importantly, insurers consistently participate in InsurTech-focused initiatives that enable firms to benefit from synergy and integration efficiency. The results emphasise the policy implications of InsurTech development and InsurTech-oriented investments.

The remaining sections of this paper are structured as follows. "Research hypothesis" section describes the hypotheses. In the "Methodology and variables" section, variables and methodology are discussed. a description of the data, the empirical results, the robustness check, and a discussion"Data and empirical results" section, provides Final remarks are provided in the. "Conclusion remarks" section.

Research hypothesis

Much research on the implications of technology investments on corporate performance has emerged. Kauffman and Weil (1989) examine 13 empirical studies and discuss the challenges researchers face in identifying robust techniques and gaining insight into creating IT business value. According to Mahmood and Mann (1993, 2000), correlations between IT investments and organizational performance and productivity may not necessarily indicate causation. They argue that empirical evidence contradicts the positive association between IT investments and organizational performance and productivity; they refer to this as “the big lie of the information age.” Revathi (2020) suggests that digitising and updating a company's core systems is costly and complicated. Consequently, most insurer transformation initiatives focus on enhancing client engagement platforms.

Furthermore, Brynjolfsson (1993) contends that the absence of positive evidence on how IT investments influence organizational performance and productivity is due to mismeasurement of outputs and inputs, lags in learning and adjustment, redistribution and dissipation of profits, and mismanagement of IT. According to Rai et al.’s (1997) survey, hardware, software, and telecom expenditures negatively correlate with company organizational productivity. Nonetheless, they conclude that IT investments contribute positively to firms' business performance and labor productivity. Cao and Li (2017) demonstrate that investments in information hardware infrastructure are favorably associated with company performance, whereas investments in information software systems negatively influence firm performance.

Harris and Katz (1991) examine the correlation between company success and the intensity of information technology investments in the home office operations of 14 systems technology leaders in the life insurance industry. Better performing insurers have a higher ratio of IT costs as a percentage of overall operating costs and a lower ratio of IT costs as a percentage of premium income than insurers with poor performance. Francalanci and Galal (1998) contend that an increase in IT expenditures is connected with a change in labor composition or a reduction in the proportion of professionals. In addition, Cummins and Xie (2013) suggest that insurers with greater investments in technology are significantly and positively associated with efficiency and productivity enhancements. Recent research by Che et al. (2021) demonstrates that early UBI adopters have superior underwriting performance in gaining market share. Additionally, UBI schemes improve insurers' overperformance by 1% for return on assets (ROA) and 3% for return on equity (ROE).

From the perspective of capital investments or capital expenditure, enterprises that significantly expand capital investments achieve negative benchmark-adjusted returns in the coming years (Titman et al. 2004; Cordis and Kirby 2017). The literature implies that capital investment or capital expenditure is negatively associated with a firm's stock performance due to increased spending and ineffective short-term integration (Anderson and Garcia-Feijoo 2006; Chae et al. 2014). In addition, Cooper et al. (2008) demonstrate a robust negative association between asset growth and abnormal returns. In contrast, the literature suggests a positive correlation between capital expenditures and business performance. Taipi and Ballkoci (2017) indicate that capital expenditures in the Albanian construction industry are statistically significant and favorably connected with the firm's overall performance. Jiang et al. (2006) discover, using manufacturing businesses listed on the Taiwan Stock Exchange, that the link between capital expenditures and future corporate profitability is significantly positive, even after controlling for current corporate earnings.

In conclusion, the existing research provides contradictory findings on the effect of technology investments on organizational performance and productivity. Therefore, the purpose of this study is to investigate the following hypotheses:

Hypothesis 1A

Insurers invest in InsurTech-oriented investments, their spending increases, and their short-term performance suffers.

Hypothesis 1B

Insurers that invest in InsurTech-oriented investments tend to increase spending and perform worse in the short term than insurers that do not invest.

Hypothesis 2A

Insurers that invest in InsurTech-oriented investments typically gain efficiency and competitive advantages, resulting in improved short-term performance.

Hypothesis 2B

Insurers that invest in InsurTech-oriented investments have competitive advantages and are more efficient in the short term than insurers that do not invest.

Another body of research contends that correlations between IT investments and a company's overall performance do not necessarily suggest causality. Dos Santos et al. (1993) explore the effect of an IT investment announcement on a company's market value. According to their findings, IT investments have zero net present value (NPV) and are worth exactly what they cost. Mahmood and Mann (2000) and Chae et al. (2014) also contend that the causality relationship may not exist, showing no meaningful relationship between IT competency and company performance. According to Forman and Gron (2011), vertical integration has no effect on the use of electronic communication tools by an insurer's sales force. In addition, vertical integration does not affect the adoption of non-distribution-related internet technologies. Neirotti and Paolucci (2007) demonstrate that there is no association between competitive advantages and IT spending levels or the type of IT investments that enabled overall productivity growth in the industry. Overall, it is anticipated that IT expenditure will have no effect on firm performance, and the hypothesis is as follows:

Hypothesis 3

Insurers invest in InsurTech-oriented investments that have zero NPV  and are worth exactly what they cost.

The literature indicates that poor short-term firm performance is due to the time lag effect of IT investments. The effectiveness of technology investments may become apparent over multiple investment periods. Brynjolfsson (1993) and Devaraj and Kohli (2003) contend that lags in learning and adjustment in IT investments are one explanation for the lack of positive evidence in the short term. Moreover, Mahmood and Mann (2000) demonstrate that the benefits of IT investments can only be recognized over extended periods. The positive association between IT and organizational performance and productivity only exists in succeeding periods, suggesting that the cumulative effect of IT investments drives improved firm performance. In addition, Lee et al. (2016) demonstrate that the IT conundrum arises in specific investment categories due to the time required for full realisation. In conclusion, the literature suggests that the time lag or cumulative effect is one reason IT investments are negative or zero for short-term firm performance. Most of these difficulties appear to have been effectively addressed in recent years due to ongoing technological improvements and improved digital tool utilisation. While technologies such as robotic process automation (RPA), artificial intelligence (AI), block chain, and advanced analytics act as promoters to increase the significance of insurance, insurers strive diligently to build a more streamlined and connected insurance system. Therefore, this study anticipates that the advantages of cumulative IT expenditure will be realized over time. The hypothesis is as follows:

Hypothesis 4

The cumulative advantages of insurers' InsurTech efforts will be realized over the long term; therefore, insurers' InsurTech-oriented investments will improve their long-term performance.

Methodology and variables

Methodology

The following functional form examines how insurers' InsurTech-oriented investments impact their firm's performance during the InsurTech wave. Specifically, this study estimates the following specification:

$${{\text{Firm}_\text{Performance}}}_{it} = \, f\left( {{\text{EDP}}_{it} ,{\text{ CV}}_{it - 1} } \right) + e_{it} ,$$
(1)

where EDPi,t indicates the insurer i's technology investments in year t.Footnote 8 Firm_Performancei,t is an indicator of the insurer i's firm's performance in year t. CVi,t-1 represents the control variablesFootnote 9 and year dummies. Lastly, ei,t indicates the model disturbance that follows i.i.d. ND(0, σ2). Notably, prior research suggested the existence of reverse causation, endogeneity concerns, and sample selection bias between technology investments and firm success (e.g. Maiga and Jacobs 2011; Che et al. 2021). Consequently, two-stage regressions are employed to address endogeneity issues and sample selection bias.Footnote 10

This study estimates the following model:

$${\text{Firm}_\text{Performance}}_{it} = \, f \left( {{\text{EDP}}_{it} ,{\text{EDP}_{\text{Stock}}}_{it} ,{\text{ CV}}_{it - 1} } \right) + e_{it} ,$$
(2)

where the EDP_stockit is configured to record and evaluate the cumulative long-term effect. Similarly, two-stage GMM and G2SLS regression models are used to address endogeneity problems. This research adheres to the essential principle established by Kyock (1954) and Mirzaei et al. (2016).Footnote 11 The cumulative EDP for InsurTech-oriented investments is derived from the cumulative EDP at time t from time t-4 and is calculated as follows:

$${{\text{EDP}_\text{stock}}}_{it} = {\text{ EDP}}_{it} + \, r \times {\text{EDP}}_{it - 1} + \, r^{2} \times {\text{EDP}}_{it - 2} + \, r^{3} \times {\text{EDP}}_{it - 3} + \, r^{4} \times {\text{EDP}}_{it - 4} ,$$
(3)

where r is set to 1.1,Footnote 12 indicating the cumulative effect of firms' InsurTech-oriented investments. An r of larger than one suggests that cumulative InsurTech-oriented investments will create an annual multiplier effect.

Variables description

Measuring the performance of a company is crucial to this paper. The dependent variable, firm performance, is the risk-adjusted return on asset (RAROA) and risk-adjusted return on capital (RAROC), which are calculated as an insurer's ROA (ROC) divided by its standard deviation of ROA (ROC) over the past 5 years, where the standard deviation is calculated on a moving average basis (Elango et al. 2008).Footnote 13

It is difficult to determine how many technology resources an insurer invests in InsurTech and FinTech. Data on EDP and software included on the assets sheet of the NAIC annual report is the primary proxy for an insurer's InsurTech-oriented investments during the wave of InsurTech. The EDP data might be described as an insurer's technological input on software systems and innovative hardware infrastructures, such as InsurTech, BDA, e-commerce, and IoT, to raise operation efficiency and firm performance. Insurers update hardware and software devices through digital transformation and collaborate or partner with InsurTech startups to improve the efficiency and cost-effectiveness of their value chain, including claim settlement, underwriting, customer service, marketing, and pricing.

When an insurance company invests in InsurTech-oriented investments, the application of the platform or equipment will incur costs and depreciation. In addition, there may be technology-related costs associated with strategic alliances or partnerships with InsurTech startups. Therefore, EDP expenses are used to measure insurers' InsurTech-oriented investments (e.g. Neale et al. 2020).

This study identifies the top 20 EDP expense insurance companies over the past five years and then searches and compares the top 20 EDP expense insurers for InsurTech-oriented investments to demonstrate that high EDP expense insurers have relatively high InsurTech-oriented investments.Footnote 14 The untabulated survey results showed that 16 of the top 20 EDP expense insurers (80%) have InsurTech-oriented engagements and investments, indicating high EDP expenses for insurers with relatively high InsurTech-oriented investments and cooperation with InsurTech startups (e.g. partnerships and strategic alliances). Insurance companies, for instance, invest in InsurTech, which consists primarily of digital front-end platforms, BDA, IoT, blockchain, AI, and digital ecosystems. Moreover, most InsurTech-oriented investments are utilized for product management, marketing, sales and distribution, underwriting and policy services, and claims management.

Neale et al. (2020) investigate the factors influencing insurance companies' technology investments. They used the EDP expense allocation to underwriting (UW) and claims adjustment processes to examine the relationship between insurance production and capital investments in insurance technology. Therefore, this paper primarily adheres to the measurement rationale of Neale et al. (2020) and uses the EDP on the expense side to measure insurers' technology investments. In addition, the survey supports that EDP expenses are relatively high among insurers with many InsurTech-oriented engagements and investments. Whether insurance companies are partners in strategic alliances or invest in InsurTech startups, these investments or engagements will inevitably increase the investments in related hardware and software equipment due to the need to connect and match the core technology or system of InsurTech startups. This indicates that the associated cooperation costs will increase when an insurance company collaborates with InsurTech startups. These findings demonstrate that an insurance company's EDP expenses strongly correlate with its overall InsurTech-oriented engagements and investments. Thus, the untabulated survey results indicate that the EDP measure utilized in this study is a reliable proxy for measuring insurers' overall InsurTech-oriented engagements and investments.

Notably, the EDP measure in this study may contain measurement errors. Possible reasons include (1) data integrity and accuracy and (2) limitations on investment information that insurers may choose not to disclose because technological inputs may constitute a business secret. For measurement errors, this study addresses the issue of endogeneity to mitigate the effects of omitted variables, error-in-variables, and simultaneous causality. Moreover, this study improves estimation efficiency by controlling fixed and random effects. It is believed that this study could mitigate the influence of omitted variables, error-in-variables, and simultaneous causality. Therefore, despite the fact that the EDP measure in this study may contain measurement errors, it still has credibility and contributes to the literature.

Based on the measurement rationale of Neale et al. (2020) and the untabulated survey of insurers’ InsurTech-oriented investments, this study concludes that using the EDP as a proxy for insurers' technology investments or overall InsurTech-oriented investments is reasonable. Although this paper lacks direct evidence to prove that insurers’ EDP investments are the same as those of FinTech- and InsurTech-oriented investments, this paper demonstrates that insurers’ EDP investments are highly correlated to the firm’s FinTech- and InsurTech-oriented investments. Moreover, certain technology investments may constitute business secrets for insurance companies, and insurers may be unwilling to disclose the specifics of their EDP investments in FinTech and InsurTech. Realistically, it is not easy to know the complete details of each company's technology investment or integration of InsurTech-oriented investments. For instance, Che et al. (2021) can only discuss the UBI issue with UBI engagement samples separately. Therefore, the EDP measure, the best measurement according to Neale et al. (2020), is currently used as a proxy for an insurer's technology investments or overall InsurTech-oriented investments. Future research on InsurTech-oriented investments can utilize the EDP measure as a surrogate.

This study employs EDP_asset and EDP_inv ratios as surrogates for an insurer's InsurTech-oriented investments during the wave of InsurTech. Both EDP_asset and EDP_inv ratios are calculated by dividing the EDP by the total admitted and invested assets, respectively. In addition, this study uses the EDP_dummy variable to determine whether insurers invest in EDP based on EDP usage. It equals 1 if the EDP is greater than 0 and 0 if it equals 0. In addition, in accordance with Neale et al. (2020), the cost or depreciation of EDP to total expenses incurred (EDP_EI) and the cost or depreciation of EDP to total expenses paid (EDP_EP) are utilized, where EDP expenses include three components: underwriting, claims adjustment, and investments. The measurements of EDP_stock include EDP_assstock, the cumulative effect of EDP_asset at time t from time t − 4; EDP_invstock, the cumulative effect of EDP_inv at time t from time t − 4; EDP_eistock, the cumulative effect of EDP_EI at time t from time t − 4; and EDP_epstock, the cumulative effect of EDP_EP at time t from time t − 4.

This study draws the company-level control variables from earlier research on the determinants of firm performance (e.g. Pottier and Sommer 1999; Lai and Limpaphayom 2003; Wang et al. 2007; Elango et al. 2008; Ma and Elango 2008). Accordingly, following the firm performance equation setting of these studies, several firm-specific variables, such as firm size, commercial lines, business, and geographic Herfindahl indices, group insurer, organizational form, marketing channel, underwriting margin, invested assets, leverage, premium ratio, liquidity, and the standard deviation of the loss ratio across all lines, are chosen as the control variables for the regression analysis. According to the research, most of these firm-specific characteristics link to insurers' firm performance. This article summarises the definitions and expectations for these variables in Table 1.

Table 1 Variable definitions and summary statistics

Data and empirical results

Data

This study obtained annual and unbalanced panel data on the U.S. property-liability insurance industry from the NAIC and A.M. Best's key rating guide for 2006–2019. Initially, there were 3315 total insurers. This study removes data that do not meet sample requirements for regression analysis, such as missing values, negative assets, negative surplus and net premium written, and irrational values (or illogical values). The final dataset contains 1781 insurers and 19,253 firm-year observations after these filtering processes. In addition, this study winsorises all variables between the 1st and 99th percentiles, except for dummy variables, to reduce the effect of extreme values.

Descriptive statistics for all variables at the firm level are provided in Table 1. In the sample, the mean (median) firm performance (i.e., RAROA and RAROC) is 1.7304 and 1.6821 (1.3064 and 1.2511), respectively. In addition, the average ratios of EDP to total assets and invested assets for insurers are 0.0638% and 0.0804%, respectively. It proposes that investments in InsurTech are pretty low in the property–liability insurance industry. In addition, approximately 38.23% of insurers participate in InsurTech. It suggests that most insurers pay less attention to the future effects of the wave of InsurTech.Footnote 15 Regarding the control variables, the results for the majority of variables indicate that the sample utilized in this study is an appropriate selection (Pottier and Sommer 1999; Lai and Limpaphayom 2003; Wang et al. 2007; Elango et al. 2008; Ma and Elango 2008).

Empirical results

Using the two-stage GMM regression model, Tables 2 and 3 construct four baseline equation specifications (based on EDP_asset, EDP_inv, EDP_IE, and EDP_EP measures) to examine the influence of InsurTech-oriented investments on firm performance. In Column (1) and Column (4) of Table 2, the first-stage regression results for EDP_asset and EDP_inv, respectively, are displayed. The findings indicate that the bulk of exogenous and instrumental variables substantially correlate with firm performance. In addition, Kleibergen–Paap rank LM statistics (for the under-identification test; 40.69*** and 43.20***), Cragg–Donald Wald F statistic (for weak identification test; 19.21** and 19.36**), and Hansen J statistics (for overidentification test; 0.71 and 2.37) of Table 2 demonstrate that the IVs are adequate. In columns (2) and (5) of Table 2, the link between RAROA, EDP_asset, and EDP_inv is analysed as the baseline analysis. In accordance with Hypothesis 1A, the results indicate that the coefficients of EDP_asset (− 71.2148***) and EDP_inv (− 53.0868***) are negative and statistically significant at the 1% level. The baseline models showing the association between RAROC and EDP_asset and EDP_inv are displayed in Columns 3 and 6 of Table 1. EDP_asset (− 69.0769***) and EDP_inv (− 51.4266***) metrics are also consistent with Hypothesis 1A. Overall, the empirical findings support Hypothesis 1A and the counterargument that insurers who invest in InsurTech-oriented projects tend to increase spending, resulting in poor short-term performance. In addition, the predicted coefficients of firm-level control factors are consistent with prior studies. Firms have superior overall performance when they are a single entity, have an independent marketing channel, have higher invested assets, a higher leverage ratio, or have greater liquidity. In contrast, their overall performance is lower when they have more commercial lines, a greater business and geographic concentration, a larger combined ratio, or a higher ratio of net premium written to gross premium. In Table 3, Columns (1) through Column (6) re-estimate the baseline regressions using alternative EDP metrics (i.e., EDP_EI and EDP_EP). However, consistent with Hypothesis 2A, EDP_EI (164.1347*** and 152.4319*** for RAROA and RAROC) and EDP_EP (188.9446*** and 175.1387*** for RAROA and RAROC) are positively and significantly connected to the performance of enterprises. According to the intriguing findings, insurers with superior productivity improvement and business success had a higher ratio of IT costs or technology investments (Harris and Katz 1991; Francalanci and Galal 1998; Cummins and Xie 2013).

Table 2 The results of the two-step GMM regression model (EDP_asset and EDP_inv measures)
Table 3 The results of the two-step GMM regression model (EDP_EI and EDP_EP measures)

The estimation model with fixed and random effects is implemented in Table 4 to mitigate the problems of omitted variables, simultaneous causality, and variable errors. The results of a two-stage regression with firm and year fixed effects are shown in Panel A of Table 4. Overall, Table 4 reveals that all EDP measures are negatively correlated with the performance of enterprises. In addition, in Panel B of Table 4, the regression findings adjusted for firm and year random effects indicate that EDP_asset and EDP_inv continue to be adversely connected with firm performance, whereas EDP_EI and EDP_EP are insignificantly and positively connected to the performance of enterprises. In sum, the empirical findings presented in Table 4 are overall compatible with Hypothesis 1A.

Table 4 The results of the two-stage regression model with fixed and random effects

Since the literature suggests that sample selection bias of a firm's InsurTech-oriented investments may exist, Table 5 presents the findings of the Heckman two-stage regression model with multiple settings of firm and year fixed and random effects for the EDP_dummy measure. The negative significance of the coefficients of the Inverse Mills Ratio (IMR) from Columns (2) through (7) indicates the existence of sample selection bias and a detrimental impact on the performance of enterprises. However, the coefficient of EDP_dummy in Column (5) is considered negative, supporting Hypothesis 1A. The insignificantly negative correlations in Columns (2), (3), (4), (6), and (7) confirm Hypothesis 3 and suggest that insurers' InsurTech-oriented investments have zero net present value and are worth precisely what they cost (Dos Santos et al. 1993; Mahmood and Mann 2000; Chae et al. 2014). Overall, the empirical results of Table 5 demonstrate that the fixed and random effects can alleviate the difficulties of omitted variables, simultaneous causality, and variable errors.

Table 5 The results of the Heckman two-stage regression model (EDP_dummy measure)

The literature suggests that time lag or cumulative effects cause the short-term relationship between InsurTech-oriented investments and firm performance is negative or zero. To examine the effect on overall firm performance, Panels A, B, and C of Table 6 re-estimate the baseline regression with the long-term cumulative effect of InsurTech-oriented investments for controlling firm and year fixed or random effects. Consistent with Tables 2 and 3, the EDP-related metrics of Table 6 indicate a strong negative relationship between InsurTech-oriented investments and firm performance. Importantly and intriguingly, the EDP_stock linked indicators demonstrate that cumulative InsurTech-oriented investments positively impact the firm's overall performance, which is consistent with Hypothesis 4. When insurers invest in InsurTech-oriented investments, their long-term performance will improve because the benefits of such cumulative investments are realised over time. Consistent with the finding of Che et al. (2021), insurers who join the UBI program early will have a greater return on assets and capital in the coming years.

Table 6 The results of the cumulative effect of the two-stage regression model with fixed and random effects

Robustness check: market response

Understanding the market's reaction to companies' technology investments is crucial.Footnote 16 This paper further uses U.S.-listed to confirm the market's response to firm performance regarding technology investments by using listed insurers on the stock exchange markets (e.g. AMEX, NASDAQ, and NYSE) with Standard Industrial Classification Codes (SIC) of 6311 from 2006 to 2019.Footnote 17 This study compares insurers’ company names to identify all listed property and liability insurers in the NAIC dataset and finally confirms the existence of approximately 151 listed property and liability insurers. After removing missing and unreasonable values, 138 listed companies and 1233 firm-year observations are finally used. The study adopts market-to-book ratio (M/B)Footnote 18 and annual stock return (Return) to proxy the firm’s market reaction to the firm’s financial performance, and the testing equation is set as follows:

$$M/B_{it} \,{\text{or Return}}_{it} = \, f\left( {{\text{EDP}}_{it} ,{\text{ CV}}_{it - 1} } \right) + e_{it} ,$$
(4)
$$M/B_{it} \,{\text{or Return}}_{it} = \, f\left( {{\text{EDP}}_{it} ,{{ \text{EDP}_\text{Stock}}}_{it} ,{\text{ CV}}_{it - 1} } \right) + e_{it} .$$
(5)

The empirical results are shown in Table 7. Overall, when listed insurers are utilized, Panel A of Table 7 shows that EDP and firm financial performance (M/B and Return) are marginally negative. In addition, Panel B of Table 7 shows that the cumulative EDP results are consistent with the main findings of this article, however, with a lower significance level. Moreover, based on the findings of Panel B of Table 7, it is hypothesised that specific cumulative benefits of technology investments may manifest after three to five years of EDP investments, supporting the argument of Campbell (2012).

Table 7 The results of the two-step GMM Regression model with the analysis of the market response

Discussion

The evidence presented in this paper indicates that insurers' InsurTech-oriented investments decrease the firm's overall performance in the short term, but bring long-term benefits due to the cumulative effect of InsurTech-oriented investments. Why has a negative influence appeared quickly? In general, technology investments can be categorised as capital expenditures. Thus, when firms engage in capital investments, their annual expenditures and expenses would increase, resulting in negative short-term performance (Titman et al. 2004; Cordis and Kirby 2017). Inadequate investments, a lack of synergy, or integration of technology investments may also contribute to poor overall performance (Anderson and Garcia-Feijoo 2006; Chae et al. 2014). In addition, a company may plan to invest in technology or renew technology investments annually; thus, the issue of underinvestment arises, resulting in poor short-term performance.

Notably, firms with cumulative technology investments generate beneficial long-term performance. If firms' technology investments are viewed as capital expenditures, the yearly accumulation of capital expenditures will result in favorable long-term performance. This study provides the evidence necessary to confirm this argument. In addition, the long-term accumulation of technology investments has produced synergistic benefits, such as increased operational production or sales efficiency, accurate pricing, and less adverse selection in underwriting and moral hazard in claims settlement (Che et al. 2021). Moreover, Campbell (2012) indicates that the cumulative benefits of technology investments may manifest after three to five years. This study concludes that technology investments by insurers have a cumulative effect over time.

Conclusion

An essential question is whether insurers' investments in technology boost their competitive advantages. The evidence of this study indicates that insurers' InsurTech-oriented investments have a significant negative impact on their short-term performance, which is consistent with the yearly expenditure and spending increase argument. Intriguingly, the evidence also demonstrates that the cumulative effect of insurers' InsurTech-oriented investments is significantly and favorably associated with their long-term performance, suggesting that the arguments of time lag and cumulative effects are valid.

Under the wave of InsurTech, insurers must acknowledge that firms are increasing their InsurTech-oriented investments and that IT or digitalization help them achieve synergy and become more efficient in addressing the crucial challenge new InsurTech startups pose. The findings of this study are highly significant and exceptional. This research fills a gap in the literature concerning the performance of the insurance industry. Insurers, investors, and regulators will find this study to be a valuable resource, as well. Insurers and regulators must pay close attention, particularly when making decisions about InsurTech-oriented investments and InsurTech development.

It is worth noting that although this paper cannot directly demonstrate that EDP investments are directly related to FinTech- or InsurTech-oriented investments, it is evident from the discussion of Neale et al. (2020) Consequently, using EDP as a proxy for insurers' technology investments may be the most suitable metric currently. It would be preferable if insurers' actual investments in FinTech and InsurTech could be precisely identified for further study. Moreover, the definition of EDP investments cannot adequately describe the pertinent investment activities associated with share investments (e.g. joint venture, M&A, and strategic alliance). Similarly, the topic of share investments merits further discussion.