Risk management regulation and corporate demand for reinsurance in the Spanish autarky (1940-1952)

175 Received: 14 July 2020 – Fecha de recepción: 14 de julio de 2020 Accepted: 3 March 2021 – Fecha de aceptación: 3 de marzo de 2021 Revista de Historia Industrial – Industrial History Review Vol. XXXI, no. 84, 2022 - DOI: https://doi.org/10.1344/rhiihr.v31i84.31954 Risk management regulation and corporate demand for reinsurance in the Spanish autarky (1940-1952) • Pablo gutiéRRez gonzález Universidad de Sevilla https://orcid.org/0000-0002-0591-8254 • JeRònia Pons Pons Universidad de Sevilla https://orcid.org/0000-0003-0491-7038 Introduction Risk management is a vital issue for companies. Firms try to minimize their exposure to possible losses and thereby maintain stable cash flows. To achieve this goal, corporate risk management has relied on financial hedging strategies and the purchase of insurance (Altuntas et al., 2018, p. 211). However, although hedging has become widespread over the last three decades, insurance was the main risk diversification device before the financial revolution in the late twentieth century (Pearson, 2016; Capie, 2016). Thus, actuarial and financial theory has analyzed the determinants of the corporate demand for insurance. While empirical research has deepened understanding of firm-level elements boosting the demand for risk spreading, literature on financial theory has also underlined the weight of country-level determinants, that is, macroeconomic and, especially, regulatory features affecting the insurance decision (Mayers and Smith, 1982; Adiel, 1996; Adams et al., 2008). In the same way that insurance provides risk-diversification services to productive firms, reinsurance helps to mitigate the financial exposure of primary insurers by covering a share of their risk portfolios. Reinsurance enables insurance companies to design more homogeneous risk portfolios and reduce the volatility of losses (Blazenko, 1986; Adams, 1996, pp. 22-24). Moreover, by using reinsurance, insurance providers acquire the capacity to sell more policies and, therefore, maximize their presence in the market without jeopardizing their financial solvency. In this way, the use of reinsurance enhances the underwriting capacity of insurers without accumulating addi- 19182_RHI84_TRIPA.indb 175 15/3/22 8:28 Risk management regulation and corporate demand for reinsurance in the Spanish Autarky (1940-1952) 176 tional resources, which makes it an imperfect substitute for equity capital (Mayers and Smith, 1990, pp. 37-39; Garven and Lamn-Tenant, 2003, pp. 217-221; Powell and Somner, 2007, pp. 176-178). This financial nature turns reinsurance into a key device for those insurance markets suffering from capital constraints, which become dependent on external reinsurers to deal with risks arising in developing economies (Outreville, 1995; Gutiérrez and Andersson, 2018). As a channel for risk diversification, reinsurance is essentially an international business. Nevertheless, this global scope makes reinsurance extremely sensitive to regulations on cross-border financial transfers. Indeed, reinsurance has traditionally been resistant to state jurisdiction and, especially, to attempts to supervise the nature of traded risks (Pearson, 2012, pp. 74-75; Vec, 2016). In this way, literature has underlined the importance of the principle of confidentiality of the reinsurance treaty, since public information on risk cessions could raise doubts about the ceding company’s portfolio, as well as creating uncertainty for customers (Nebel, 2002, pp. 115-118). Nonetheless, the nature of the purchase of foreign reinsurance cessions as an outflow of currency constitutes one of the main drivers of public intervention in reinsurance: the control of access to international capital and foreign exchange markets (Gerathewohl, 1993, pp. 49-52). Hence, in countries suffering from financial constraints, regulation has traditionally had wide repercussions on the development of this industry, by restraining the free transfer of premiums. Focusing on the case of Spanish insurance, the goal of this paper is to examine the impact of regulation on reinsurance from the outbreak of the Second World War to the first steps towards liberalization in 1952. The international isolation of the country during this period and the chronic dependence of domestic insurers on foreign reinsurance turned the Spanish case into a conflict of opposing interests between the objectives of public regulations and the financial needs of private companies. The autarkic economic policy implemented by the Franco dictatorship sought tight control over cross-border currency transfers, while the country’s neutral condition placed Spanish companies among the main providers of marine insurance and, in turn, among the main customers of international reinsurance. In this context, we seek to evaluate the effects of these regulatory changes on the risk management strategies of insurance companies and, ultimately, to analyze the influence of reinsurance regulation on the capacity of the Spanish insurance industry to reduce its exposure to unexpected losses. To address these issues, we explore the links between reinsurance regulation, the financial leverage of the insurance industry and risk management strategies implemented by firms, in relation to the following questions: (1) How did regulatory changes modify the systemic role of reinsurance in the Spanish insurance industry? And (2), What costs arose from the new insur- 19182_RHI84_TRIPA.indb 176 15/3/22 8:28 Pablo Gutiérrez González, Jerònia Pons Pons 177 ance industry regulation, in terms of financial distress and actuarial performance? To deal with these questions, we use both institutional and corporate sources to build a novel dataset including figures on reinsurance transfers in Spain from 1940 to 1954 and firm-level statistics on insurance companies and their reinsurance practices. The remainder of the paper is organized as follows. The next section examines the long-term determinants of reinsurance demand in the Spanish insurance industry and the effects of regulatory changes introduced during the Second World War. The third section presents the data and the methodology for the empirical analysis. The fourth section discusses the results. Finally, we conclude the paper. Financial and regulatory constraints: The long-term determinants of reinsurance demand in Spain During the first decades of the twentieth century, the development of the Spanish economy triggered the need for risk management services, which resulted in the progress and growth of the insurance industry (Pons, 2003; 2007a). Indeed, Spanish insurers used complex devices such as reinsurance to connect isolated regions with national financial networks, as well as to expand their operations to other markets (Gutiérrez and Pons, 2017). However, domestic companies had to deal with major obstacles in order to compete successfully: first, the fragmentation of the market, which was reflected in the high number of financially weak small companies and mutual societies; and second, the shortcomings of the financial system, which was barely capable of providing capital services and equity resources to such a complex business as insurance (Pons, 2002, pp. 13-15).1 As a result, the Spanish insurance industry, along with other sectors, experienced chronic difficulties to access financial resources in a period of market expansion. Furthermore, the new regulation on insurance passed in 1908 and emerging trends of economic nationalism stimulated the withdrawal of foreign companies from Spain, which left the market free for domestic offices (Pons, 2012, pp. 92-94). In this context, although insurance business grew steadily over the next four decades, this development suffered from deep financial imbalances. While the aggregate financial leverage of the Spanish insurance industry declined from the end of the First World War and during the 1920s and 1930s, 1. As noted by Martín Aceña (1987; 1993), Comín (2013) and Martínez Ruiz (2013), non-ad- hesion to the gold standard complicated the internationalization of Spanish finance, while the excessive weight of the Bank of Spain in the whole banking system and its subordination to fiscal needs hindered the development of a sound domestic capital market. 19182_RHI84_TRIPA.indb 177 15/3/22 8:28 Risk management regulation and corporate demand for reinsurance in the Spanish Autarky (1940-1952) 178 it remained at a high level compared with the figures obtained for other countries.2 Moreover, the financial leverage of Spanish insurers determined their capital structure and their risk management strategies: in order to overcome the chronic lack of capital funding, domestic offices showed a systematic preference for purchasing reinsurance. Indeed, by using reinsurance as an imperfect substitute for capital, Spanish insurance companies were able to meet the growing demand of the overall economy, but at the same time they became heavily dependent on foreign reinsurance networks throughout the whole period from 1900 to 1940 (Gutiérrez and Andersson, 2018). While this trend seems to have been a structural dynamic, the evolution of the international risk exchange networks and regulatory changes affecting reinsurance marked the evolution of the business during different stages. From the First World War onwards, restrictions on the circulation of reinsurance premiums were established, both in highly developed and peripheral countries, to avoid capital outflows and currency transfers (James et al., 2013, pp. 71-79). As a matter of fact, the conflict redirected the demand of freight and insurance services towards neutral countries.3 However, as reinsurance transfers were deemed to be a destabilizing element for the national balance of payments, national authorities implemented different measures with regard to foreign exchange control during the last years of the conflict. Moreover, the new regulations attempted to transform national reinsurance markets into monopolies ruled by national champions (James et al., 2013, pp. 171-179).4 Within this context, the international scope of reinsurance became seriously limited: as long as national regulatory agencies tried to force primary insurers to share information about their reinsurance cessions and reveal the conditions of contracts, the principle of confidentiality in reinsurance treaties would be threatened or broken. However, institutional restrictions were not the only problem for international reinsurers: monetary disturbances and inflationary trends during the 2. The premium to surplus (P/S) ratio, calculated as net premiums written to surplus (equity + reserves) is widely used as a proxy to measure the probability of bankruptcy and the financial leverage of insurance companies (see e.g. Mayers and Smith, 1982; Powell and Somner, 2007; or Cole and McCullough, 2006). Indeed, while Beckman and Tremelling (1972), p. 206, obtained P/S ratio values lower than 1 for the United States in the same decades and Kader et al. (2010), p. 276, got a mean value of 0.410 for Sweden, most of the figures in the Spanish case fluctuated between 1.5 and 2. Gutiérrez (2014), p. 48. 3. As noted by Larsson and Lönnborg (2014) and Petersson (2011) for Sweden, and by Gutiérrez (2014) and Gutiérrez and Pons (2017) for Spain, a drastic increase in the financial exposure of marine insurance companies triggered the demand for foreign reinsurance by domestic insurers in neutral countries, in order to alleviate their leverage. 4. Furthermore, when the constraints on international payments became widespread during the 1920s and, especially, the 1930s, several unsuccessful plans emerged in France, Belgium, Sweden, Greece and Spain, among other countries, to convert these private monopolies into state-owned ones, which alarmed the industry (Gutiérrez, 2014; Gutiérrez and Pons, 2017). 19182_RHI84_TRIPA.indb 178 15/3/22 8:28 Pablo Gutiérrez González, Jerònia Pons Pons 179 1920s and the disintegration of international economy in the 1930s posed additional difficulties to the development of the business during the interwar period (Capie, 2017). In addition, the social struggle and political instability during Spain’s Second Republic and the outbreak of the Civil War in 1936 contributed to increasing the isolation of the country. Besides the economic collapse during the years of conflict, the victory of General Franco with the support of the Axis powers left the Spanish economy in an unsteady position. Difficulties to access an almost blocked international equity market converged with the consequences of the war. As well as the physical destruction, the disturbances caused by three years of monetary dualism and the high exposure of the financial system posed additional obstacles to recovery. Besides these problems, the insurance industry had to confront the effects of the massive claims faced by companies due to the civil conflict, as well as problems to receive compensation for reinsured policies (Tortella et al., 2014, pp. 160-166). The outbreak of the Second World War added further complexity to the situation.5 Much as during the First World War, Spain, as a neutral country, was among the main providers of freight and insurance services for international trade (Martínez Ruiz, 2003, p. 122).6 The increase in the underwriting of new policies and the rise of loss ratios due to the riskiness of marine traffic jeopardized the financial equilibrium of Spanish insurers: as they had to face new liabilities in their balances, companies needed to raise their reserves or purchase reinsurance. The rigidities of the Spanish financial system, and the obstacles to accessing additional funding posed by the war, led these companies to resort massively to foreign reinsurance (Gutiérrez, 2017, pp. 172- 173). The outflow of capital in the form of reinsurance premiums, both in pesetas and in foreign currencies, alarmed Spanish monetary authorities at the Instituto Español de Moneda Extranjera (IEME). Pursuing the goal of identifying and controlling the cessions of premiums through reinsurance, an Official Committee on Marine Insurance was created in 1942. In parallel, a new regulation was passed to control reinsurance transfers from Spanish insurers to abroad.7 Within the new framework, insurers could only cede their policies to foreign reinsurers through an administrative authorization granted by the IEME. Moreover, a compulsory cession to the Committee on Marine Insurance of a 5. As noted by Egido León (2005), the Spanish government moved from neutrality to non-bel- ligerency in 1940 and back to neutrality in 1943, but these changes had no significant effects on the role of the Spanish merchant fleet, which mostly purchased insurance services provided by domestic companies. 6. Indeed, the size of the marine insurance market, in terms of premiums underwritten, rose from almost 100 million pesetas in 1940 to more than 300 million in 1943. See Pons (2002), p. 115, and Memoria de la Dirección General de Banca, Bolsa e Inversiones (1951). 7. Decree of 13 March 1942, in the Official State Gazette (BOE) no. 87, 27 March 1942. 19182_RHI84_TRIPA.indb 179 15/3/22 8:28 Risk management regulation and corporate demand for reinsurance in the Spanish Autarky (1940-1952) 180 share of all marine insurance policies written in Spain was introduced. Consequently, one of the main elements of the reinsurance treaty was broken with these measures: the principle of confidentiality. While the initiative tried to prevent the free movement of foreign currencies, the new regulation introduced an anomalous element into corporate risk management strategies: companies seeking risk diversification had to share their portfolio information with the regulatory agency. This situation provoked strong protests from private offices but, far from causing a conflict between the Franco dictatorship and the industry, an agreement was reached: the Committee on Marine Insurance would retrocede a share of those risks taken from the market that did not generate currency outflows.8 When the Second World War was coming to an end in Europe, Spanish monetary authorities designed a new framework to deal with the postwar scenario. While the disturbances affecting the marine insurance market were vanishing, the isolation of the Spanish economy increased due to the dictatorship’s links with the Axis powers. Moreover, the slow reconstruction of the productive sectors and the government’s obstinate policy of pursuing autarky affected the demand for insurance services.9 In parallel to this trend, the relaxing of requirements to establish new companies in legislation passed in March 1944 led to an increasing fragmentation of the market.10 In this situation, the key motives for direct intervention in marine insurance lost weight in the face of a wider problem: the increasing lack of foreign currencies and the weakness of the peseta in the new global financial and monetary context.11 Consequently, the Committee on Marine Insurance was remodeled as an Official Committee on Reinsurance. This new agency retained a decreasing role as a reinsurer for marine risks, but its main function was redefined as supervisor of all reinsurance operations abroad involving Spanish risks. It was to operate in all branches of the market by taking a 1 percent share of every policy underwritten in Spain by domestic or foreign companies and reinsured 8. Furthermore, as noted by Gutierrez (2014), pp. 34-35, this capture and reallocation of marine insurance was implemented with a clear inclination towards domestic marine insurers. Indeed, while foreign companies had to cede a share of their policies to the Committee, the retrocession strategies implemented by the public agency took only domestic companies into consideration. In line with the trend of economic nationalism initiated three decades before, these measures forced a shift of market share from foreign offices to Spanish insurers. 9. See Prados (2017) and Maluquer de Motes (2016) for a detailed analysis of the evolution of the economic indicators during this period. García Ruiz and Caruana (2009), Pons (2010) and Gutiérrez (2017) note the stagnation of written premiums in all lines. 10. In fact, as noted by Pons (2015), the dictatorship boosted the proliferation of mutual societies formed by businessmen and employers. This process, along with the constitution of new companies with low financial resources, resulted in a drop of 30 percent in the average premiums earned by companies between 1942 and 1951 (Pons 2010, p. 65). 11. This was one of the main points cited in the arguments behind the new legislation. See Decree of 8 July 1945, in BOE no. 197, 16 July 1945. 19182_RHI84_TRIPA.indb 180 15/3/22 8:28 Pablo Gutiérrez González, Jerònia Pons Pons 181 abroad.12 Moreover, companies ceding premiums abroad were required to send the authorities all the information of the reinsurance contract, which could only be signed once approved by the Committee. Private companies such as La Unión y el Fénix alleged that this new mechanism implied the end of the confidentiality of reinsurance operations, which would impose severe difficulties on accessing international risk exchange networks.13 Furthermore, the criteria used by the Committee to permit the transaction were quite restrictive: all operations sending reinsurance abroad had to be compensated by an equal amount of reinsurance premiums coming from abroad, remunerated in foreign currencies. TABLE 1 ▪ P/S ratio in life and non-life branches in Spain (1942-1951) 1942 1943 1944 1945 1946 1947 1948 1949 1950 1951 Life 0.210 0.211 0.221 0.217 0.201 0.205 0.193 0.190 0.189 0.186 Non-life 1.847 1.797 2.025 2.363 2.368 2.210 2.076 2.043 1.893 1.853 Source: Memoria de la Dirección General de Banca, Bolsa e Inversiones, 1951 After the new regulation was passed in July 1945, insurance companies abruptly reduced their reinsurance operations abroad. Committee officials noted this trend, when the amount of premiums ceded to foreign reinsurers declined by 70 percent in the second half of 1945.14 Meanwhile, the amount of domestic risks reinsured by Spanish companies rose from 10.3 percent of life premiums underwritten in Spain in 1944 to 18.3 percent in 1946, while in the case of non-life branches the figure rose from 32 percent to 59 percent in the same period (Gutiérrez, 2017, p. 226). In parallel, the performance of solvency indicators worsened. Indeed, as shown in Table 1, insurance companies in non-life branches experienced a significant increase in their financial leverage from 1943 to 1946.15 This trend coincided with the regulatory obsta- 12. The compulsory cession of 1 per cent of all risks reinsured abroad was actually introduced by a previous regulation passed in September 1944 but, as explained by officers of the Committee, it was not fully implemented until the transformation of the institution. See Decree of 29 September 1944, in BOE no. 293, 19 October 1944. 13. Historical Archives of the Bank of Spain (ABE), Departamento Extranjero, IEME, Comité Oficial de Reaseguros, Correspondencia y Actas del Comité, c.85, f. 6. 14. ABE, Departamento Extranjero, IEME, Comité Oficial de Reaseguros, Correspondencia y Actas del Comité, c.84, f.12. While a minimum share of this fall could be caused by the concealment of operations, the trend was also acknowledged by private companies in their internal reports. See the Report to the Meeting of the Board of Directors of Compañía General de Reaseguros, 21 January 1946. 15. Note the difference between life and non-life branches. As shown by Gerathewohl (1993), among others, the nature of the risks covered make life insurance less leveraged, since companies in the branch accumulate mathematical reserves in accordance with the mortality tables used. More- 19182_RHI84_TRIPA.indb 181 15/3/22 8:28 Risk management regulation and corporate demand for reinsurance in the Spanish Autarky (1940-1952) 182 cles imposed by the IEME on the use of foreign reinsurance, which in turn hindered the development of risk management strategies. In response to this situation, Spanish offices began to create alternate risk transfer mechanisms, which helped alleviate financial distress as from 1947. In fact, following the theoretical framework of ruin theory, two choices are available to reduce financial leverage in insurance companies: increase equity and reserves or purchase reinsurance. While the first option was difficult to achieve within an almost blocked Spanish financial system, the second was complicated by the legal requirements to purchase foreign reinsurance. Moreover, private companies were reluctant to share their policy information with public officers.16 Indeed, since insurers serve as managers of actuarial information, their main goal was to retain their knowledge of the market, so domestic companies had to avoid control by the Committee and search for alternate options within the country. Even after the first attempts to control reinsurance flows in 1942, several primary insurers constituted subsidiary companies devoted to providing reinsurance services (Caruana, 2017). While there were 24 Spanish pure reinsurers operating in the period 1943-1953, 11 of them were directly owned or affiliated to direct insurance companies. Moreover, by taking the sample of the 75 most important companies accepting reinsurance in Spain in 1946 (both pure reinsurers and primary insurers), we found that more than 21 percent of the members of these companies’ boards of directors were common directors (that is, shared by two or more companies); furthermore, when looking at chief executive officers employed by these insurers, more than 50 percent (38) were affiliated to other firms at the same time (Gutiérrez, 2017, pp. 247-250).17 over, the relatively small size of the risks reduces the volatility of losses, which in turn affects financial leverage. 16. ABE, Departamento Extranjero, IEME, Comité Oficial de Reaseguros, Correspondencia y Actas del Comité, c.64, f. 1-12. 17. Although the concept of interlocking directorates described by Mizruchi (1996) has been challenged, Lilljegren (2019) has shown the impact of these corporate networks in Swedish insurance, while Rubio and Garrués (2016) have successfully used them to explain the structure of Spanish business in sectors such as energy, banking and manufacturing. In a wider context, Fohlin (2006) underlines the relevance of interlocking directorates involving banking and industry firms for German industrialization. 19182_RHI84_TRIPA.indb 182 15/3/22 8:28 Pablo Gutiérrez González, Jerònia Pons Pons 183 TABLE 2 ▪ Primary insurers and pure reinsurers in the Spanish market, ranked according to the aggregate premiums reinsured in the main lines (life, fire, marine and accident) (1943-1951) Companya Year of foundation Market share (%) Owner/Head office 1Mediterránea de Reaseguros 1942 10.74 Compañía Hispano-Americana 2Compañía General de Reaseguros 1942 6.41 Mutua General de Seguros 3La Equitativa Hispano Americana 1944 6.28 La Equitativa (Fundación Rosillo) 4La Unión y el Fénix Español 1864 5.98 – 5Cía. Española de Reaseguros 1941 5.52 La Unión y el Fénix Español 6La Equitativa Reaseguros 1928 5.13 La Equitativa Nacional 7Nacional de Reaseguros* 1942 4.53 – 8Reaseguradora Española 1940 4.08 Bilbao 9Compañía Hispano Americana 1924 4.01 – 10 Continental 1943 2.67 Occidente a Pure reinsurers in italics. * Constituted in Seville in 1939, and then moved to Madrid. Source: Anuario Financiero y de Sociedades Anónimas, 1941-1952 and Revista del Sindicato Vertical del Seguro, 1943-1953 As shown in Table 2, interlocking directorates and property links helped shape a national reinsurance market characterized by a high level of concentration. Two of the top ten reinsurers in the main lines were primary insurers authorized to reinsure Spanish risks, while the rest were specialist companies. Among the latter, only La Equitativa Reaseguros had been constituted before the outbreak of the Second World War and the subsequent changes in the international reinsurance market. In addition, the last column shows, where applicable, the primary insurer’s links to the company, usually as its owner. Thus it seems clear that the national champions in Spanish insurance, which were La Unión y el Fénix Español for the fire line, La Equitativa (Fundación Rosillo) for the life branch, Mutua General de Seguros for industrial accident business and Compañía Hispano-Americana for marine insurance, built a network of reinsurers around them during these years of isolation. Figure 1 shows the aggregated data of the transformation caused by these legal and management changes: the introduction of the first waves of control in 1942 led to a decline in cessions abroad through reinsurance and, in turn, to a slow growth of the exchange of risks among Spanish reinsurers. Moreover, the amount of premiums accepted by Spanish reinsurers increased almost fourfold between 1942 and 1951. As a result, the chronic trend of capital out- 19182_RHI84_TRIPA.indb 183 15/3/22 8:28 Risk management regulation and corporate demand for reinsurance in the Spanish Autarky (1940-1952) 190 Regarding the rest of the variables, we focus on the features of dependent reinsurers and primary insurers inside corporate groups (DEP and HQ). Indeed, the positive and statistically significant relationship between these variables and the Reinsurance ratio, in the case of HQ companies, and with Leverage and Loss ratio, in both cases, could indicate the following. First, companies acting as the head of a corporate group would face higher levels of leverage and loss ratio which, in turn, would result in a higher inclination to reinsure. Second, dependent reinsurers would have to face higher levels of financial leverage and loss ratios, since they had to assume the risks ceded by their mother companies instead of designing a sound risk portfolio. As a result, dependent reinsurers would act as the alleviators of the financial distress of their owners. However, we do not find any statistically significant evidence of a higher inclination of these dependent companies to retain risks (that is, an aversion to purchase reinsurance). Finally, we find that the country-level variables show quite different results: while the exchange rate with the dollar shows a significant and negative correlation (as expected) with reinsurance, the GDP growth rate shows no relationship with reinsurance practices, which could be attributed to the restrictive regulatory environment of the market. Corporate responses to regulatory change: Reinsurance determinants in an isolated economy To address the impact of regulatory changes on the financial balance and actuarial performance of the Spanish insurance industry, we utilize a panel data regression analysis, testing for the determinants of reinsurance through the period of study. To examine these interactions, we seek to estimate the following equation: REINit = α + ∑ β Xit + β Yi + β Ut + εit where REINit is the rate of reinsurance of company i in year t; Xit is the vector of firm-specific variables including Size, Loss ratio, Leverage, Liquidity and Profitability of company i in year t; Yi is the vector of firm-specific variables defining the role of company i, namely: Pure reinsurer (PURE), dependent reinsurer (DEP) or head of a corporate group (HQ). Finally, to control for country-level variations, we insert the vector Ut including the exchange rate of the peseta with the US Dollar. Since a relevant set of our explanatory variables (the company’s condition of affiliated or non-affiliated and its role within a corporate group) are time 19182_RHI84_TRIPA.indb 190 15/3/22 8:28 Pablo Gutiérrez González, Jerònia Pons Pons 191 invariant, we run the specification tests in order to identify if a random-ef- fects model may be used instead of a fixed-effects one. A Hausman test was conducted to assess unobserved heterogeneity. As the test was not statistically significant (p-value > 0.10), we assume that unobserved influences among firm-specific variables are unimportant and a random-effects model is appropriate. To deal with the problems of multicollinearity among firm-specific variables, we use Size as an instrumental variable capturing most of the variations of the set (see Appendix A1). In parallel, as we try to explain the impact of regulation on reinsurance practices, we run different models for the whole period 1944-1953. Then we introduce the period 1944-1945, in which the restrictions on international reinsurance flows were not in force (we have not included the 1952-1953 period of re-liberalization, because the low number of observations did not allow statistically significant results to be obtained from the model). Finally, we examine reinsurance determinants during the period of tight regulation and control on foreign reinsurance operations, that is: the years 1946-1951. Table 5 displays the determinants of reinsurance demand, including firm-specific indicators and the role of the company as explanatory variables. As shown in the table, the random-effects models confirm the negative impact of the exchange rate of the peseta on reinsurance purchases. At the same time, the coefficients underline the relevance of corporate networks in the design of reinsurance strategies. Certainly, while the condition of being a specialist reinsurer (PURE) shows a diverse effect on the inclination to cede risks, companies acting as the head office of a corporate group (HQ) were systematically more prone to reinsure. Moreover, this preference would be especially strong during the period in which restrictions to accessing international reinsurance networks were in force. At the same time, those dependent or subsidiary reinsurers within corporate groups also showed a particular behavior in their risk management strategies. In this case, subsidiary companies owned by the heads of corporate groups were significantly less prone to reinsure, that is: they were systematically oriented to retain the risks accepted instead of ceding them. Furthermore, comparing the coefficients for the different models, we see that this systematic preference to retain risks would be significantly higher during the years 1946-1951, when the controls on reinsurance flows were tighter (see Equations 5 and 6). In this way, and beyond firm-specific conditions, the role of the company accounted for around 20 percent of purchases of reinsurance services by head offices; in the case of subsidiary companies, the condition of dependent reinsurer determines between 11 and 21 percent of reinsurance decisions. The contrast with the fixed-effect aggregated term by corporate role confirms this impact: Appendix A2 shows 19% lower reinsurance ratio for dependent companies compared to the rest of the sample. 19182_RHI84_TRIPA.indb 191 15/3/22 8:28 Risk management regulation and corporate demand for reinsurance in the Spanish Autarky (1940-1952) 192 In parallel, the explanatory power of the firm-specific financial indicators remains insignificant when we take the whole set of variables. Indeed, only Size shows explanatory power for reinsurance, since it seems to serve as a catchall variable for the rest of the firm-specific indicators. As expected, Size shows a substantial and positive effect on reinsurance purchases during the TABLE 5 ▪ Determinants of reinsurance in the Spanish insurance industry. Randomeffects linear regression models (1944-1953) 1944–53 1944–45 1946–51 (1) (2) (3) (4) (5) (6) PURE 4.748 5.774 –1.292 –5.855 9.569** 11.292** (3.621) (4.183) (5.210) (6.194) (5.005) (5.096) HQ 7.465*** 7.718*** 0.674 0.351 10.651*** 11.351** (2.912) (3.269) (4.596) (4.999) (3.774) (3.798) DEP –16.305*** –16.036*** –12.504** –11.008*–21.519*** –22.221*** (3.747) (3.994) (5.983) (6.565) (5.149) (5.170) EXCH_US –0.117*–0.119*–0.016** 0.571 –0.133** –0.179 (0.066) (0.072) (0.002) (0.390) (0.098) (0.119) SIZE 1.073*0.072 2.187** 9.345*** –1.142 –5.030 (0.615) (4.085) (1.083) (3.453) (0.760) (4.626) LEV –0.003 –0.132 –0.015 – (0.082) –(0.356) –(0.082) – LOSS –0.066** ––0.006 ––0.022 – (0.029) –(0.041) –(0.044) – LIQ 0.001 ––0.001 –0.002 – (0.002) –(0.003) –(0.002) – PROFIT –0.001 –0.352** ––0.001 – (0.002) –(0.139) –0.002 – Const. 33.052*** 54.306 4.434 –99.371 50.481*** 106.415 9.633 (59.091) (17.841) (52.293) (12.059) (67.342) R2 within 0.022 0.004 0.075 0.038 0.034 0.017 between 0.113 0.115 0.114 0.061 0.145 0.120 overall 0.097 0.083 0.088 0.056 0.137 0.085 N. of obs. 816 245 478 p-values in parentheses. *** p<0.01, **p<0.05, *p<0.1 Note: Models (2), (4), and (6) use SIZE as an instrumental variable. Source: Boletín Oficial de Seguros y Ahorro, 1945-1951; Revista del Sindicato Nacional del Seguro, 1943-1954; Anuario Español de Seguros, 1946, 1948, 1952-1954; Anuario Financiero y de Sociedades Anónimas de España, 1945, 1949, 1954. 19182_RHI84_TRIPA.indb 192 15/3/22 8:28 Pablo Gutiérrez González, Jerònia Pons Pons 193 period 1944-45; however, when we examine its performance after the change in the regulatory environment, it seems to lose its explanatory power for reinsurance practices. All the above would be in accordance with the hypothesis of a network of captives serving as preferential reinsurers for the head of the corporate groups. Therefore, regulatory restrictions on accessing foreign reinsurance would lead primary insurers to constitute their own channels of risk diversification. In this scheme, dependent or subsidiary reinsurers would have to assume and retain those risks that could not be spread abroad because of barriers to purchasing foreign reinsurance. Moreover, this could be explained as a systemic function in which dependent reinsurers accepted the risks of their head offices as a way to alleviate the financial and actuarial distress of their owners. Furthermore, we could affirm that, when leading primary insurers such as La Unión y el Fénix Español, La Equitativa, Mutua General de Seguros or Bilbao, constituted dependent companies as corporate reinsurers (such as Compañía Española de Reaseguros-CERSA, Minerva, CRESA, La Constancia, La Equitativa Hispanoamericana or Guipúzcoa Reaseguros) they created the densest network of captive reinsurance companies in Europe, as an alternative to sharing their risks and their portfolio information with the regulatory agencies.24 The link with these captive reinsurers was either established through a direct ownership relationship or by means of the personal proprietorship of company managers. In this way, these reinsurers were exploited not only as a mechanism to obtain reinsurance services, but also to earn additional profits through commissions on ceded reinsurance, on top of the surpluses generated by the reinsurers. However, the trends of liberalization initiated in 1952 alleviated the tight control of the IEME over the foreign exchange market. In this new context, interest in these subsidiary reinsurers was modified: the opening up of the international reinsurance industry made it easier to access wider risk diversification networks, while national dependent reinsurers appeared to be financially exhausted due to the managerial strategies implemented by their owners (Pons, 2015, pp. 211-212; Gutiérrez, 2017, p. 284). Consequently, the interest of primary insurers in the reinsurance business became insignificant, and they returned to the international risk exchange networks. We have confirmed this process in several cases. The managers of the market leader, La Unión y el Fénix, planned the creation of a dependent reinsurer in a meeting of the board of directors on February 23, 1940. Concerns about possible changes to the regulation of the foreign exchange market and 24. Although the concept of captive reinsurer is applied in a strict sense to the management trend initiated in the 1950s, Gerathewohl (1993) identified several examples of dependent reinsurers since the 1850s. 19182_RHI84_TRIPA.indb 193 15/3/22 8:28 Risk management regulation and corporate demand for reinsurance in the Spanish Autarky (1940-1952) 194 their effects on the company’s reinsurance business in the French market led the board to anticipate events and to constitute a pure reinsurer devoted to providing the mother office with risk spreading services. The drafting of statutes and procedures to register the new company were accelerated in the following months, and CERSA started its operations in 1941.25 Corporate control of subsidiary reinsurers became a key issue for Spanish insurers in the 1940s. When the group La Equitativa (Fundación Rosillo) had to face the family conflict between its founders in 1944, access to reinsurance services was one of the main concerns of the managers, since the corporate reinsurers, La Equitativa Reaseguros, split from the group.26 Indeed, after the separation of the group, Miguel Rosillo received the life insurance company La Equitativa, Fernando received the property-casualty insurer La Equitativa Riesgos Diversos, and Fermín took the reinsurance company. Miguel and Fernando kept their companies acting as a group and, trying to rebuild the family corporation, they founded the reinsurer La Equitativa Hispano Americana, with a share capital of 5 million pesetas. This company was intended to provide the group’s insurers with reinsurance services. The president of the company was Miguel Rosillo, while the vice-president was Fernando (Pons, 2007b, pp. 360-361). Mutual societies were also involved in these practices. The case of Mutua General de Seguros is especially relevant because of the multiple and divergent interests involved in the creation of a subsidiary reinsurer. In fact, one of its subsidiaries, CRESA, was created in 1942 and registered as a reinsurer in 1943. In 1944, it was acquired by the members of Mutua General’s board of directors. Although during its first years it concentrated on reinsuring a large part of Mutua General’s fire business, in 1947 it was registered as a primary insurer in order to initiate a line of business based on coinsurance. In fact, coinsurance treaties in non-life branches were signed with Mutua General: with a 10 percent share for CRESA and 90 percent for Mutua, the company alleviated its demand for reinsurance. Following this unusual strategy performed with CRESA, in 1948 Mutua General’s management board decided to purchase a relevant share of Compañía General de Reaseguros (REASE). In this case, Mutua’s main goal was to gain direct access to reinsurance services for its own portfolio, but also to make profit on the business of reinsurance itself, since mutual societies were not allowed to accept reinsurance premiums. Subsequent to the acquisition of a share of REASE’s capital, the mutual society enjoyed beneficial treatment which can be confirmed in the company’s reinsurance treaties. For ex- 25. Book of Minutes of the Board of Administration of La Unión y El Fénix, 27 January 1940, 17 May 1940 and 28 January 1941. 26. See Pons (2007b) for a concise history of the group and its evolution until 1960. 19182_RHI84_TRIPA.indb 194 15/3/22 8:28 Pablo Gutiérrez González, Jerònia Pons Pons 195 ample, the agreement signed in Barcelona on November 9, 1953, established a surplus reinsurance treaty under which Mutua would cede all its surpluses to REASE. One of the major features of the agreement was that Mutua retained the right to choose the potential reinsurers to which REASE could retrocede reinsured risks. In parallel, Mutua would receive 25 percent of the profits generated by these retrocessions. As a result of these onerous conditions, Mutua’s relationship with its subsidiary remained completely asymmetrical: although the reinsurer served as an extraordinary source of profits for the mutual society, Mutua imposed spurious interests on REASE’s management, which led to a decline in its financial position. With these practices, Mutua achieved several objectives: first, it overcame the institutional barriers to accessing foreign reinsurance introduced under the autarky; second, as a mutual society, it gained indirect access in quite beneficial conditions to a line of insurance, the reinsurance business, from which it was legally excluded. Nevertheless, the new institutional trend of liberalization in the 1950s and the financial distress faced by REASE forced the decision to get rid of the reinsurer. In 1955, Mutua’s board of directors decided to cancel the agreements with the reinsurer and sign new reinsurance treaties with the British Royal Insurance. Finally, the public regulator of the insurance business, the Dirección General de Banca, Bolsa e Inversiones, ordered the dissolution of REASE in 1959, given that its balance losses exceeded 50 percent of share capital.27 Conclusions As with other fields of the financial industry, insurance and risk management services are extremely sensitive to regulatory changes. Restrictions on accessing international risk exchange networks cause profound changes to both the actuarial and financial performance of the insurance industry and the management strategies implemented by firms. These modifications, which affect the capacity of the overall economy to manage the risks arising from economic growth, become structural in those countries dependent on global insurance and reinsurance networks. In the case of Spain, the convergence of domestic shortcomings with global events at the beginning of the 1940s triggered the transformation of the insurance industry. The changes to the global marine insurance market from the outbreak of the Second World War, Spain’s non-belligerency and the 27. Archivo Mutua General de Seguros, REASE files. “Documento convenio entre Compañía General de Reaseguros, S.A. y sus acreedores”. 12 November 1959, and Carta del Ministerio de Hacienda, Junta Consultivo de Seguros a la Compañía General de Reaseguros, S.A., 27 January 1959. 19182_RHI84_TRIPA.indb 195 15/3/22 8:28 Risk management regulation and corporate demand for reinsurance in the Spanish Autarky (1940-1952) 196 financial shortcomings of Spanish insurers after the Civil War boosted the demand for foreign reinsurance by domestic companies. Nevertheless, this trend, which had been a structural feature of the Spanish insurance industry, was interrupted by the interferences of the Franco dictatorship and its autarkic orientation. Indeed, regulatory restrictions on the access to foreign currencies and the breakdown of the principle of confidentiality of reinsurance treaties posed additional costs to accessing foreign reinsurance networks. Although Spanish insurers had relied heavily on the international reinsurance industry to manage their growth for decades, when access to those financial services was blocked by regulatory restrictions, they had to seek alternate tools to implement their risk management strategies. In this scenario, primary insurers created subsidiary and dependent reinsurers devoted to providing risk diversification services to the main corporate groups. Moreover, these dependent reinsurers actually served as a substitute for international risk exchange networks. According to the empirical results, the flows of risk exchange within corporate groups played a key role in maintaining the underwriting capacity of Spanish insurers and served as a device to deal with unexpected losses. As hypothesized, the role of the company as a group leader or as a dependent reinsurer showed high explanatory power for reinsurance strategies. We certainly find that the design of risk management strategies relied heavily on membership of these corporate groups and the role assumed therein. As we have illustrated in specific cases, among the leaders of the market, both stock companies and mutual societies used these strategies. Indeed, corporate groups headed by La Unión y el Fénix, La Equitativa (Fundación Rosillo) and Mutua General de Seguros resorted to dependent reinsurers to overcome the blockade against foreign reinsurance. Furthermore, while group leaders used these circumstances to obtain extraordinary profits from a closed market, the practices imposed on those dependent reinsurers made them financially weak and extremely leveraged. Consequently, when the regulatory restrictions to reinsure abroad were lifted, foreign reinsurers recaptured the Spanish market. From a macroeconomic viewpoint, our analysis shows that this dynamic entailed a temporary breakdown of the structural effects of reinsurance in Spain’s international position. While the use of foreign reinsurance for the first third of the twentieth century had exacerbated the deficits of the balance of payments due to importing services, regulatory restrictions forced the Spanish insurance industry to work without foreign financial assistance. Moreover, by working as an imperfect substitute for capital, the provision of reinsurance services by dependent companies helped Spanish insurers to deal with the problems deriving from the post-Civil War recession and the subsequent stagnation during a period of international isolation. 19182_RHI84_TRIPA.indb 196 15/3/22 8:28 Pablo Gutiérrez González, Jerònia Pons Pons 197 References adaMs , M. (1996), “The reinsurance decision in life insurance firms: An empirical test of the risk-bearing hypothesis”, Accounting and Finance 36 (1), 15-30. adaMs , M., HaRdwiCk, P . & zou, H . (2008), “Reinsurance and corporate taxation in the United Kingdom life insurance industry”, Journal of Banking and Finance 32 (1), 101- 115. adaMs, M ., andeRsson, l .F ., lindMaRk, M . & vePRauskaite, e . (2012), “Competing models of organizational form: Risk management strategies and underwriting performance in the Swedish fire insurance market between 1903 and 1939”, Journal of Economic History 72 (4), 990-1014. adiel , R. (1996), “Reinsurance and the management of regulatory ratios and taxes in the property-casualty insurance industry”, Journal of Accounting and Economics 22, 207- 240. altuntas, M ., gaRven, J . & RauCH, J . (2018), “On the corporate demand for insurance: evidence from the global reinsurance market”, Risk Management and Insurance Review 21 (2), 211-242. beCkMan, R .w . & tReMelling, R .n . (1972), “The relationship between net premium written and policyholder surplus”, Proceedings of the Casualty Actuarial Society 59 (111-2), 203-220. blazenko , G. (1986), “The Economics of Reinsurance”, Journal of Risk and Insurance 53 (2), 258-277. CaPie , F. (2016), “The monetary and financial environment, 1860-2013”, in Haueter, N.V. & Jones, G. (eds.), Managing risk in reinsurance. From city fires to global warming, Oxford University Press, Oxford, 92-117. CaRReRas, a . & taFunell, X . (eds.) (2005), Estadísticas históricas de España. Siglos XIX-XX, Fundación BBVA, Madrid. CaRuana de las Cagigas, l . (2017), “The development of Spanish reinsurance from 1940 to the present day”, in Straus, A. & Caruana de las Cagigas, L. (eds.), Highlights on Reinsurance History, Peter Lang, Bruxelles, 135-152. CavalieRi, e . (2014), España y el FMI: la integración de la economía española en el Sistema Monetario Internacional, 1943-1959, Banco de España, Madrid. Cole, C .R . & MCCullougH, k .a . (2006), “A Reexamination of the Corporate Demand for Insurance”, The Journal of Risk and Insurance 73 (1), 169-192. CoMín CoMín, F . (2013), “Las crisis de la deuda: el largo camino recorrido desde los impagos a la gestión responsable”, in Martín-Aceña, P., Martínez Ruiz, E. & Pons Brías, M.A. (eds.), Las crisis financieras en la España contemporánea, 1850-2012, Crítica, Barcelona, 197-240. deelstRa, g . & Plantin, g . (2014), Risk theory and reinsurance, London, Routledge. egido león, a . (2005), “Franco y la Segunda Guerra Mundial. Una neutralidad comprometida”, Ayer 57 (1), 103-124. 19182_RHI84_TRIPA.indb 197 15/3/22 8:28 Risk management regulation and corporate demand for reinsurance in the Spanish Autarky (1940-1952) 198 FoHlin, C . (2006), Finance capitalism and Germany’s rise of industrial power, Cambridge University Press, New York. gaRCía Ruiz, J .l . & CaRuana de las Cagigas, l . (2009), “La internacionalización del seguro español en el siglo XX”, Revista de Historia Industrial 41 (3), 17-48. gaRven, J .R . & laMM-tennant, J . (2003), “The demand for reinsurance: theory and empirical tests”, Insurance and Risk Management 7 (3), 217-237. geRatHewoHl, k . (1993), Reaseguro: Teoría y Práctica, Reaseguros Gil y Carvajal, Madrid. gutiéRRez gonzález, P . (2014), El control de divisas durante el primer franquismo. La intervención del reaseguro (1940-1952), Banco de España, Madrid. — (2017), “El mercado reasegurador en España (1870-1952). Estrategias empresariales, prácticas financieras e iniciativas de control público”, Doctoral thesis, University of Seville. gutiéRRez gonzález, P . & Pons Pons, J . (2017), “Risk management and reinsurance strategies in the Spanish insurance market (1880-1940)”, Business History 59 (2), 292-310. gutiéRRez gonzález, P . & andeRsson, l .F . (2018), “Managing financial constraints. Undercapitalization and underwriting capacity in Spanish fire insurance”, The Economic History Review 71 (2), 567-592. HaRold, J ., boRsCHeid, P ., gugeRli, d . & stRautMann, t . (2013), The value of risk. Swiss Re and the history of reinsurance, Oxford University Press, Oxford. kadeR, H .a ., adaMs, M ., andeRsson, l .F . & lindMaRk, M . (2010), “The determinants of reinsurance in the Swedish property fire insurance market during the interwar years, 1919-39”, Business History 52 (2), 268-284. laRsson, M . & lonnboRg, M . (2014), SCOR Sweden RE. 100 years of Swedish (re)insurance industry, Dialogos, Stockholm. lilJegRen, J . (2019), “Networks that organised competition. Corporate resources sharing between Swedish property underwriters 1875-1950”, Doctoral thesis, University of Umea. MaluqueR de Motes, J . (2016), España en la economía mundial, Instituto de Estudios Económicos, Madrid. MaRtín-aCeña, P . (1987), “Desarrollo y modernización del sistema financiero, 1844- 1935”, in Sánchez Albornoz, N. (ed.), La modernización económica de España, Alianza Editorial, Madrid, 121-146. — (1993), “Spain during the classical Gold Standard years, 1880-1914”, in Bordo, M. & Capie, F., Monetary regimes in transition, Cambridge University Press, Cambridge, 135-172. MaRtínez Ruiz, e . (2003), El sector exterior durante la autarquía. Una reconstrucción de las balanzas de pagos de España (1940-1958), Banco de España, Madrid. — (2013), “Papel mojado. Crisis inflacionarias”, in Comín Comín, F., Hernández Benítez, M. (eds.), Crisis económicas en España. 1300-2012: lecciones de la historia, Alianza Editorial, Madrid, 203-226. 19182_RHI84_TRIPA.indb 198 15/3/22 8:28 Pablo Gutiérrez González, Jerònia Pons Pons 199 MayeRs, d . & sMitH, C .w . (1982), “On the corporate demand for insurance”, Journal of Business 55, 281-296. — (1990), “On the corporate demand for insurance: evidence from the reinsurance market”, Journal of Business 63, 19-40. MizRuCHi, M .s . (1996), “What do interlocks do? An analysis, critique and assessment of research on interlocking directorates”, Annual Review of Sociology 22, 271-298. nebel, R . (2002), “The case for liberal reinsurance regulation”, The Geneva Papers on Risk and Insurance 27 (1), 113-121. outReville , J.F. (1995), “Reinsurance in developing countries”, Journal of Reinsurance 2 (3), 42-51. PeaRson , R. (1995), “The development of reinsurance markets in Europe during the nineteenth century”,Journal of European Economic History 24 (3), 557-572. — (2012), “United Kingdom: Pioneering insurance internationally”, in Borscheid, P. & Haueter, N.V. (eds.), World Insurance. The evolution of a global risk network, Oxford University Press, Oxford, 67-97. — (2016), “The evolution of the industry structure”, in Haueter, N.V. & Jones, G. (eds.), Managing risk in reinsurance. From city fires to global warming, Oxford University Press, Oxford, 70-91. PeteRsson , G.J. (2011), “Insurance and cartels through wars and depressions. Swedish marine insurance and reinsurance between the World Wars”, Doctoral thesis, University of Umea. Plantin , G. (2006), “Does reinsurance need reinsurers?”, Journal of Risk and Insurance 73, 153-168. Pons Pons, J . (2002), “Las estrategias de crecimiento de las compañías de seguros en España 1900-1940”, Working Paper, Fundación Empresa Pública. Programa de Historia Económica, 2002/1. — (2003), “Diversificación y Cartelización en el seguro español 1914-1935”, Revista de Historia Económica 21 (3), 67-92. — (2007a), “The influence of foreign companies in the organization of the Spanish Insurance market: diversification and cartelisation 1880-1939”, in Borscheid, P. & Pearson, R. (eds.), Internationalisation and Globalisation of the Insurance Industry in the 19th and 20th Centuries, Philipps-University and Swiss Re Corporate History, Zurich, 49-65. — (2007b), Mutua General de Seguros (1907-2007). Cien años de historia del seguro español, Unpublished document. — (2010), “The difficulties of Spanish Insurance companies to modernise during the Franco years. The mechanisation of administrative tasks (1950-1970)”, in Pearson, R. (ed.), The development of international Insurance, Pickering and Chatto, London, 63-101. — (2012), “Spain: International influence on the domestic insurance market”, in Borscheid, P. & Haueter, N.V. (eds.), World Insurance. The evolution of a global risk network, Oxford University Press, Oxford, 189-212. 19182_RHI84_TRIPA.indb 199 15/3/22 8:28