Managing financial constraints: Undercapitalization and underwriting capacity in spanish fire insurance

MANAGING FINANCIAL CONSTRAINTS: UNDERCAPITALIZATION AND UNDERWRITING CAPACITY IN SPANISH FIRE INSURANCE Pablo Gutiérrez González Research Assistant at the Department of Economics and Economic History University of Seville Spain Lars-Fredrik Andersson Associate Professor at the Department of Geography and Economic History Umeå University Sweden ABSTRACT Reinsurance is a vital financial device for enhancing underwriting capacity, ceding risks and mitigating financial distress. By supplying financial resources and services, reinsurance can facilitate growth and expansion in the insurance business. Focusing on the insurance sector in the emerging Spanish economy and using a novel dataset on fire insurance companies, this paper examines the role of fire insurance in the capital formation, the importance of reinsurance as a vehicle for expanding the country’s domestic underwriting capacity, and how the capital import impacted on the balance of payment, from the introduction of the first comprehensive legislation regarding 2 insurance in 1908 to the outbreak of the Civil War in 1936. Considering the situation of undercapitalization, the singularities of the insurance market, and the changes in regulatory schemes, we find that foreign reinsurance became a key financial vehicle for increasing the underwriting capacity in Spain. We also show the struggle for an emerging market to find ways to keep balance of current accounts and raise capital when the financial infrastructure is underdeveloped. The diffusion of reinsurance networks from the core of industrial Western countries towards emerging economies was one of the mechanisms for financial modernization on a global scale. Technologies and devices developed in pioneering countries are expected to facilitate the performance of less developed economies. 1 Latecomers benefit from the diffusion of knowledge from advanced economies, generating higher return and growth in the real economy. From a financial perspective, the advancement and expansion of financial services and instruments in pioneering countries can facilitate the diffusion of capital and knowledge for the development of banking and insurance in developing countries. 2 Indeed, as shown by Pearson, continental Europe benefited from the insuring experience and new instruments that diffused from the United Kingdom in the nineteenth and early twentieth centuries. 3 Nevertheless, the lower levels of capital accumulation and the fragmentation of the market forced continental insurers to develop new devices beyond those complex pools of risk diversification, such as Lloyd’s of London, used by British 1 Gerschenkron, Economic backwardness and Krugman, ‘A model of innovation’. 2 Rousseau and Sylla, ‘Financial System, Growth and Globalization’. 3 Pearson, Insuring the industrial revolution. 3 offices. German and Swiss insurers benefited from new management focuses, and dealt with capital constraints by employing innovative financial devices. 4 From the mid-nineteenth century onwards, reinsurance was one of the growing global mechanisms of financial risk diversification. It became a key financial vehicle for enhancing underwriting capacity, ceding risks, and mitigating financial distress in Western economies. 5 In its capacity as a supplier of financial resources and risk services across countries, reinsurance early on became a global business. 6 During the phase of industrialization and later, the diffusion of global reinsurance networks met a growing need for direct insurers to reduce their exposure to unexpected losses following the growth of capital structures. 7 Moreover, by using reinsurance as an alternative to equity capital or accumulating resources, direct insurers could underwrite more risks without jeopardizing their financial solvency. Thus, reinsurance serves as an imperfect substitute for equity capital. 8 Indeed, reinsurance allows small companies and mutual societies to overcome their scarcity (or lack) of equity capital and to maximize their market share in terms of business performance. 9 As a consequence, reinsurance serves as an alternative to traditional sources of funding. Supplying an additional financial vehicle, the importance of reinsurance is expected to grow in undercapitalized economies and/or during periods of economic growth when the demand for funding is rising. 10 In this process, peripheral latecomers such as Spain seem to have been highly influenced by the diffusion of such 4 James et al., The value of risk, 9–10. 5 Pearson, ‘Development of reinsurance’, and Pearson, ‘Birth pains’. 6 Borscheid and Haueter (eds.), World Insurance, 45, and James et al., The value of risk, 68. 7 Idem, Golding, A history of reinsurance, and Gerathewohl, Reinsurance. 8 Mayers and Smith, ‘Evidence from reinsurance market’, Berger, Cummins and Tennyson, ‘Reinsurance’, Garven and Lamn-Tennant, ‘The demand for reinsurance’ and Powell and Somner, ‘Internal versus external capital’. 9 Prieto, El reaseguro: función económica, Gerathewohl, Reinsurance and Pearson, ‘Development of reinsurance’. 10 Pearson, Insuring the industrial revolution, Pearson, ‘United Kingdom’ and Kader et al., ‘Determinants of reinsurance’. 4 techniques. In barely 40 years of development, the Spanish insurance market managed to ride out financial rigidities and capital shortages in such a way that, as the First World War drew near, it was able to cover not only domestic risks but also foreign ones. 11 Focusing on the insurance sector in the emerging Spanish economy, this paper examines the determinants of foreign reinsurance demand in the Spanish fire insurance market from the introduction of the first comprehensive legislation in 1908 to the outbreak of the Civil War in 1936. As in other lines such as marine and life insurance, the strong presence of foreign companies and the supply of reinsurance from abroad became key elements in providing insurance coverage for the emerging Spanish economy. 12 However, unlike the marine insurance business, fire insurance showed a stable foreign presence over the whole period and was less affected by the shock of the First World War. 13 In addition, the weight of fire line in the insurance industry was increasingly higher than marine. Figure 1 shows the size of traditional lines and the total weight of insurance over GDP, from 1912 to 1934: along with the economic growth of the 1920s, fire insurance almost multiplied by five the size of marine market at the end of the decade, being one of the main components in the contribution of the insurance industry to the Spanish economy. In relation to life insurance, fire line was characterized by a more competitive and dynamic business environment. 14 Moreover, the risk structure of life business and the wide use of other actuarial devices as mortality tables made reinsurance less decisive for the management of the portfolio. In this scenario, fire business in Spain during this period experienced a fast and sustained 11 Tortella et al., Historia del seguro. 12 Contemporary authors noted this importance. For example, Cenamor, ‘El seguro sobre la vida’, xi, discussed the dependence of Spanish life insurers on foreign reinsurance as the only way to avoid to ‘perish without remedy’. 13 Pons, ‘A history of insurance’, 158–9. 14 Pons, ‘Multinational enterprises’. 5 growth affected by large and concentrated risk pools, which resulted in higher dependence on the supply of risk diversification services as reinsurance. Figure 1. Premiums earned in the main traditional lines of the Spanish insurance market, (million Pesetas, 1928-year price level) (left axis) and weight of total insurance premiums over GDP, in % (right axis), between 1912 and 1934. Note: Data on premiums earned by marine insurance companies are not available until 1920. Sources: Pons, ‘Estrategias de crecimiento’, Appendix 1 and Carreras and Tafunell (eds.), Estadísticas Históricas de España, 1339. The focus on the fire insurance business may be motivated by other reasons as well. As shown in previous studies, fire insurance played a vital role for emerging industrial economies by supplying both financial protection and indemnity of losses due to unexpected risks, and thereby mitigating the unanticipated risk of major financial losses associated with investment in physical capital. Also, by insuring one of the major collateral securities, fire insurance has been recognised for its importance in supporting an important part of the financial infrastructure of capital formation. 15 15 See e.g. Pearson, Insuring the industrial revolution. 0,5 0,75 1 1,25 1,5 0 20 40 60 80 100 120 1912 1913 1914 1915 1916 1917 1918 1919 1920 1921 1922 1923 1924 1925 1926 1927 1928 1929 1930 1931 1932 1933 1934 Life Fire Marine % Total Premiums / GDP 6 However, in meeting the demand from the emerging industrial economy, the Spanish fire insurance sector had to overcome a number of shortcomings. One of the major shortcomings was the shortage of capital in the domestic economy. Together with the singularities of the Spanish insurance market, and the changes in regulatory schemes, this paper argues that reinsurance became a vital financial vehicle for increasing the underwriting capacity in the undercapitalized Spanish insurance sector. This paper will address the issue, placing the fire insurance industry into a wider context, and examining the determinants of reinsurance in line with the following three questions; (i) How did the expansion of fire insurance relate to capital formation in Spain? (ii) How did the use of reinsurance affect the strong dependency on foreign capital imports and, in the end, the performance of the Spanish balance of payments? (iii) Was there a systemic preference for purchasing reinsurance beyond firm-specific conditions affecting Spanish companies? To identify such conditions, we employ a novel and extensive dataset including both domestic Spanish fire insurance companies and foreign offices, for which we have categorized firm-specific and systemic preferences such as nationality. The remainder of the paper is organized as follows. Section 2 traces the historical development of the Spanish financial and insurance sector from the late nineteenth century up to the late 1930s. Section 3 outlines the archival records, the data used for examining the role of reinsurance in the Spanish insurance sector, and descriptive statistics on fire insurance in Spain. Section 4 describes the empirical analysis and the key findings regarding the mechanisms underlying the role of reinsurance for Spanish insurers. Section 5 concludes the paper. 7 I The development of the Spanish insurance sector was slow in the nineteenth century. The demand for insurance was weak due to the unbalanced national growth trajectory, while political uncertainty and a restrictive legal framework presented obstacles for joint stock companies to supply insurance effectively. 16 These difficulties resulted in the proliferation of mutual societies along traditional lines as fire and life during the central decades of the century. 17 Most of the companies of this type were small local societies featured by the backwardness in managerial practices and a deep actuarial lag. 18 Nevertheless, in the turbulent context of the 1860s, a new legislation on stock companies was passed in 1869. 19 Together with the political stability of the restored monarchy in 1875, these changes improved the business environment in the latter part of the century. 20 Far from being a drastic structural change, investments in productive assets resulted in a steady trend of growth in the capital stock. 21 The increasing stock of buildings, machinery and transportation equipment triggered the demand for insurance coverage. Indeed, the growing needs of risk management devices made insurance a key element, supplying capital services to the manufacturing sector. 22 Figure 2 shows the aggregated development of the capital stock (1878–1936) and the growth in the fire insurance business (1908–1936). 23 Dwellings, other buildings and infrastructures made up the major components of the stock of capital during the 16 Frax and Matilla, ‘Los seguros en España’, 191–5, and Tortella et al., Historia del seguro, 88–90. 17 Pons, ‘International Influence’, 190. 18 Pons, ‘A history of insurance’, 146–8, 158. For more details about technical backwardness in the life branch, see Pons and Gutierrez, ‘Actuarial practices’, 11–6. 19 Frax and Matilla, ‘Los seguros en España’, 187–8, and Pons, ‘Multinational enterprises’, 90, have underlined this change as the starting milestone for the insurance business in Spain 20 Prados, El progreso económico, 230, points out the sustained growth before 1929, even though there were significant differences between the low rates until 1920 and the great expansion of the 1920s. 21 Prados and Rosés, ‘Capital accumulation’, 162, determined an annual rate of growth of 2.3 per cent from 1883 to 1920 and 3.5 per cent from 1920 to 1929. 22 A similar process for the Swedish case is described by Lilljegren and Andersson, ‘Variation in organization’, 79. According to Frax and Matilla, ‘Los seguros en España’, 193-4, number of companies in all lines increased from 32 in 1876 to 124 in 1908. 23 There are no aggregated data previous to 1908. 8 whole period. Moreover, the expansion since the First World War and the price trends made both dwellings and buildings maintain and increase their weight beyond machinery and transportation equipment values. Figure 2. Fixed-capital stock estimates by components (left axis), and fire insurance premiums (right axis), (million Pesetas, 1928-year price level) between 1876 and 1936. Sources: Prados and Rosés, ‘Capital accumulation’, Appendix Table A5, and Pons, ‘Estrategias de crecimiento’, Appendix 1. The intense capital formation went hand-in-hand with the expansion of the fire insurance industry in Spain. From 1912 to 1932, the size of the market tripled in real terms. 24 As shown in Figure 2, fire insurance kept pace with the growth of the capital stock with the exception for the years during the First World War. This period of sustained growth coincided with a steady trend in internationalization. In the wake of the global expansion of financial business, several 24 Pons, ‘Diversificación y cartelización’, 570. 0 10 20 30 40 50 60 70 80 90 100 0 500 1000 1500 2000 2500 3000 3500 1876 1879 1882 1885 1888 1891 1894 1897 1900 1903 1906 1909 1912 1915 1918 1921 1924 1927 1930 1933 1936 Dwellings Other construction and infrastructures Transport equipment Machinery equipment Insurance premiums (Fire) 9 foreign and domestic insurance companies expanded into the emerging Spanish economy. Foreign offices coming from mature markets such as the United Kingdom and France supplied a substantial part of insurance coverage in Spain. 25 Before the First World War, in 1912, the fire insurance market was shared between 16 domestic companies and 29 foreign offices. The foreign business share was 37%. 26 The strong foreign presence resulted in the profuse adoption of innovative devices and practices from abroad: collusive practices in the way of the British Fire Offices Committee were employed around the Comité Español de Aseguradores de Incendios, which thus helped both domestic and foreign companies to set standard practices in the market. 27 In parallel, French and British societies introduced the use of reinsurance into the Spanish insurance industry: in a first step as a way to enter the market; after that, as a mean of diversifying their portfolios. 28 Together with this openness and the increasing size of the market, the economy concealed internal imbalances and systemic shortcomings that hindered the development of the domestic insurance industry and paved the way for the entrance of foreign offices. Certainly, despite the rates of growth experienced since 1875, the Spanish financial system remained underdeveloped at the turn of the century. 29 Based on the five basic characteristics of a successful financial system according to Rousseau and Sylla, Spain suffered from a number of structural shortcomings. 30 Indeed, with 25 Pons, ‘Estrategias de crecimiento’, 12, and ‘Influence of foreign companies’, Pearson, Insuring the industrial revolution, Alborn, Regulated lives and Wilkins, ‘Multinational enterprise’. 26 Pons, ‘Influence of foreign companies’, 51. Foreign insurers were registered as branch offices of their mother company. They did not declare equity capital, and their only financial support against liabilities in Spain were their accumulated reserves, although head offices were expected to back their operations. 27 Pons, ‘Las empresas extranjeras’, 200. Pons, ‘Diversificación y cartelización’, 584–5, also describes the Comité Central de Compañías de Seguros contra Incendios as a precedent in the nineteenth century. 28 Pearson, ‘Las compañías de seguros extranjeras’, 116-8. 29 Malo de Molina and Martín-Aceña, ‘Introduction’, 4–5, and Pons, ‘The main reforms’, 73–6. 30 Rousseau and Sylla, ‘Financial System, Growth and Globalization’, 374–7. These features are presented as follows: (1) public finance, including its debt management, is run prudently; (2) a stable currency; (3) private commercial banks involved in international and/or domestic financial business; (4) monetary authority that can act to stabilize domestic finances and manage international financial relations, and (5) a well-functioning securities market. 16 theory framework, managers would have had to decide between three approaches in order to reduce or keep stable this probability of ruin: (1) improve the performance of the portfolio, (2) accumulate capital, or (3) purchase reinsurance. 54 In the first case, managers could increase the net surplus of the portfolio either by raising the price of insurance or by increasing low-variance risk contracts. Such a strategy, however, seems less likely in a competitive market such as the Spanish one in the period of study. It also seems less likely that companies were able to increase the number of low-variance risk contracts, which even in the case of avoiding reducing prices would be difficult since, according to Deelstra and Plantin, such risks are the most inelastic part of the market. In the second option, increasing capital could be a difficult task for Spanish insurance companies during this period. Though the financial system had progressed since 1900, a lack of diversification or specialization and business instability remained, as did the problems of the banking industry, which continued to be the main component of the financial system. 55 These elements obstructed the capitalization of the Spanish insurance industry. Indeed, immediately before the outbreak of the Spanish Civil War, the insurance industry showed shortages of capital that resulted in extraordinarily high levels of leverage. 56 Faced with an undercapitalized financial market and a high level of liabilities, reinsurance seem to have been more attractive to domestic insurers than foreign companies on the expanding Spanish fire insurance market. In order to determine whether such a preference was predominant among Spanish insurance companies or the outcome of the Spanish insurance market in general, the following analysis will focus on the determinants of reinsurance. The 54 Doherty and Tinic, ‘Reinsurance under equilibrium’, Plantin and Rochet, When insurers go bust and Deelstra and Plantin, Risk theory and reinsurance, 27–8. 55 Malo de Molina and Martín-Aceña, The Spanish Financial System. 56 The P/S ratio (premiums earned over surplus) is widely accepted as a valid indicator of leverage. The sample of fire insurers used in Kader et al., ‘Determinants of reinsurance’ in the case of Sweden between 1919 and 1939, shows mean values for the P/S ratio of 0.410. Using official aggregate data from the Dirección General de Seguros for 1942, Gutiérrez, El control de divisas, 48, displays a value for Spain of 1.741. 17 analysis aims at identifying whether there was a systemic preference for purchasing reinsurance based on the country of origin (nationality), or whether reinsurance purchase was the outcome of firm-specific conditions. II There are no complete published or edited official statistical series on Spanish reinsurance until 1942. The available data for the insurance industry as a whole are the annual reports in the Boletín Oficial de Seguros, Boletín Oficial de la Inspección Mercantil y Boletín Oficial de Seguros y Ahorro, which include the premiums underwritten from 1911 to 1935. Edited information on the premiums reinsured, reserves, claims experience, liquidity indicators, and other financial data are only available for a few years. 57 To overcome this lack of statistics, we have collected the financial balances sent by companies to the regulation agency after 1908 and currently located at the Archivo General de la Administración (AGA). 58 These include the main accounting sheets of the companies: statements of financial positions and profit and loss accounts. From these documents we have collected the premiums earned, claims paid, reinsured premiums, cash and liquid assets in bank accounts, equity capital declared, and ongoing-risks reserves. Between 1908 and 1936, there were 117 companies authorized to underwrite fire insurance policies in Spain. However, only a third of them operated throughout the period without interruption. New entrances and exits from the fire branch as well as liquidation processes determined a highly changing scenario in which the total number 57 Comisaría General de Seguros, Memoria sobre las entidades de seguros que operan en España, 1911, 1912, 1913, 1915. 58 AGA (1) 26 - Cajas I1 to I-141 Top. 13/31 and AGA (1) 26 - Cajas 11/02131 to 11/02180. The obligation to report their operations was established within the legal framework of the 1908 Inspection and Registry Act. Although there are companies reporting after 1908, the standard model of report was not defined until 1912. 18 of companies registered for each year varied between 46 in 1912 and 79 in 1934. Besides this volatility, there is a great deal of missing data due to different reasons: many offices sent incomplete information on their operations; some diversified companies reported mixed data from various branches; and, finally, insurers authorized to underwrite fire business that were retreating from the branch, or even not selling policies, showed abnormally low values for main indicators. Consequently, those entries showing a large amount of missing and/or extreme data have been excluded from the dataset. Our full sample includes 62 insurers distributed along the period 1908–1934 and fluctuating from 41 in 1915 to 33 companies in 1931. The years before 1910 and after 1932 show a clear lack of data from fire insurers, with fewer than 15 companies reporting their operations, so these have been analysed separately. As a result, compared to the data from Pons, our dataset includes between 80% (in 1912) and 45% (in 1931) of total companies authorized for each year. 59 However, in terms of market share, the sample represents no less than 85% of total premiums earned (as we have excluded most local companies and small mutual societies with less than 1% of market share). 60 Regarding the nationality of the companies, the sample includes 27 Spanish, 15 British, ten French, three Swiss, three German companies, and one each from the United States, Italy, Denmark, and Portugal. 61 Among the Spanish companies, we include only three mutual societies; despite the number of this kind of society in the Spanish fire branch, most of them operated within a local scope and underwrote only minor amounts of business. 59 Pons, ‘Estrategias de crecimiento’, Appendix 3. 60 For example, Pons, ‘Estrategias de crecimiento’, estimates total fire insurance premiums earned in Spain in 1930 at 76.5 million Pesetas, while our sample reaches 73.4 million. 61 As shown in the first section, the actual nationality of companies was a controversial issue in this period. This sample includes three companies with a significant presence of foreign capital: La Unión y el Fénix Español, Hispania and Numancia. 19 Table 2 includes the main descriptive statistics (mean and standard deviation) for the key variables built from the data collected. Our dependent variable, reinsurance (REIN), is expressed as the percentage of reinsured premiums over the total amount of premiums earned. With regard to our independent variables, the previous literature has stated a positive relation between leverage and the probability of ruin, which, in turn, connects it positively with reinsurance. 62 Though the premium to surplus ratio (P/S) has frequently been used as an indicator of leverage in the literature, it shows several limitations in our sample. 63 Several Spanish stock companies showed differences up to 70% between equity capitals declared and effectively deposited. Mutual societies add further heterogeneity, since they have no equity capital. Foreign offices did not declare equity capital in Spain because, according to insurance regulations, they received financial support from their headquarters. 64 Searching for a homogeneous indicator, we seek a variable expressing the available resources allocated by the company to face the liabilities resulting from policies in force. Therefore, we have chosen the ongoing-risks reserves as a variable common to all kinds of insurers and one that is only subjected to managerial decisions, unlike other statistics strictly marked by legal obligations. The Registry Act of 1908 established the minimum amount of ongoing risk reserves as one third of written premiums, and the lack of control until the end of each exercise let many companies operate during most of the year under this limit, while others decided to form larger reserves, always depending on the office strategy. 65 62 Mayers and Smith, ‘Corporate demand for insurance’ and Mayers and Smith, ‘Evidence from reinsurance market’. 63 P/S ratio is defined as net premiums written over total surplus. Idem, Cole and McCullough, ‘A reexamination of the corporate demand’ and Kader et al., ‘Determinants of reinsurance’ used the sum of equity and reserves as surplus value. 64 After several controversies on the issue, the Royal Order of 1st of April of 1913 obliged foreign insurers to declare as equity capital the amount corresponding to their head offices. Actually, these sums were not deposited in Spain nor they were subjected to Spanish legislation. 65 See Regulation on the implementation of the Registry Act of 1908, of February 2nd of 1912, in Gaceta de Madrid 49, 18/02/1912. The firms in our sample show that the average leverage was too high to meet these legal requirements: ongoing risk reserves made up only 25.4% on the written premiums. 20 Consequently, our variable acts as an indicator of the solvency risk of the company and as a proxy for partial leverage (LEV), since it is confronting the total business written by the insurer to the amount of resources that the managers are willing to tie up in order to face eventual liabilities. 66 LEV is calculated as the percentage of total premiums earned over ongoing-risks reserves. Given that reinsurance serves to reduce the exposure to unexpected claims, previous authors have considered loss ratio to be closely and positively related to reinsurance purchases. 67 In line with this, we use loss ratio (LOSS), expressed as the percentage of total claims paid over premiums earned. Plantin argues that insurers with fewer liquid assets available are expected to purchase more reinsurance in order to overcome their cash-resources constraints. 68 In our research, liquidity (LIQ) is equal to the percentage of total liquid assets (cash and bank accounts) over current liabilities, expressed as ongoing reserves. According to the previous literature, size is expected to be negatively related to reinsurance, since small companies are more sensitive to losses. 69 We express size (SIZE) as the natural logarithm of total business underwritten. Finally, older companies are expected to have a wider knowledge of the market and to be more efficient. 70 In the case of insurance, this experience would appear as a lower loss volatility in the risk portfolio; thus, age would be negatively related to reinsurance. In this research, age (AGE) is expressed as the total years of operation of the company on the Spanish insurance market. 66 As shown by Pons, ‘Las entidades aseguradoras’, 345–6, these reserves should be formed with cash resources and low yield securities as Spanish State bonds, public securities from foreign states with rates of return lower than 4%, corporate debt securities and bonds traded in Spanish stock market but only if they maintained the 75% of their nominal value, and mortgage bonds, which could not exceed 25% of the overall reserves value. 67 Mayers and Smith, ‘Evidence from reinsurance market’ and Lamn-Tennant and Starks, ‘Stock versus mutual’. 68 Plantin, ‘Does reinsurance need reinsurers?’. 69 Adams, ‘The reinsurance decision’ and Adiel, ‘Reinsurance’. 70 Jovanovic, ‘Selection’. 21 Table 2. Fire insurance companies in Spain (1908-1934). Descriptive statistics for firmspecific variables by nationality and organizational form. Mean and (standard deviation). Sign Full Sample Spanish British French Other All Mutual REIN . 30.21 41.68 82.87 16.26 25.77 41.52 (24.72) (24.03) (6.21) (16.31) (18.47) (32.66) LEV + 392.77 442.34 650.23 350.53 358.62 414.12 (202.12) (249.88) (121.29) (131.74) (158.41) (227.32) LOSS + 35.62 31.36 28.64 44.65 34.52 27.50 (25.26) (18.61) (27.38) (32.95) (20.12) (22.75) LIQ - 114.57 121.35 64.35 120.85 52.84 185.32 (332.40) (134.50) (34.81) (493.28) (72.86) (466.20) SIZE - 12.30 12.71 12.16 12.17 12.41 11.26 (1.47) (1.71) (0.36) (1.00) (1.42) (1.14) AGE - 30.07 25.56 20.58 36.65 33.16 18.89 (20.03) (18.18) (8.31) (24.10) (14.80) (14.98) No. of firms 62 27 3 15 10 10 Note: REIN = Premiums reinsured / Premiums earned, in %; LEV = Premiums earned / Ongoing-risks reserves, in %; LOSS = Claims / Premiums earned, in %; LIQ = Cash + Bank accounts / Ongoing-risks reserves, in %; SIZE = Natural logarithm of premiums earned; AGE = nº of years since the first registration in fire branch in Spain. Source: AGA (1) 26 - Cajas I-1 to I-141 Top. 13/31 and AGA (1) 26 - Cajas 11/02131 to 11/02180; Comisaría General de Seguros, Memoria sobre las entidades de seguros que operan en España, Years 1911, 1912, 1913, 1915. To examine whether reinsurance purchases go beyond that of firm-specific characteristics, we have employed nationality as a proxy identifying systematic preference. We classify our sample according to nationality (Spanish, British, French, and Other), given that we suppose a greater inclination to purchase reinsurance for domestic companies facing more of the structural shortcomings of the Spanish financial system. Given that foreign companies were integrated into their home country’s capital markets, we expect a lower need for reinsurance to overcome financial system shortcomings. 22 To identify a preference of reinsurance based on organizational form, we have separated mutual societies from joint-stock companies, since previous literature states that organizational form has a deep influence on the performance of the risk portfolio. 71 According to this, we expect mutual societies to demand more reinsurance than do stock companies, as they suffered from capital constraints derived from their lack of equity capital. Finally, for country-level control variables we consider two macroeconomic indicators that are expected to affect reinsurance demand. First, we use the GDP real growth (GRTH) from Carreras and Tafunell, Estadísticas históricas, since economic growth is expected to increase insurance covering demand, which, in turn, would positively affect reinsurance services. 72 Second, we use the Peseta/French Franc (FF) exchange rate in order to measure the monetary effect on the demand for foreign reinsurance. Thus, we consider that the lower/higher rate of exchange would improve/worsen the position of Spanish insurers in order to purchase foreign reinsurance, which would result in higher/lower reinsurance ratios. Between official exchange rates, we use the French Franc not only because of the close links with the French market but also for its better performance compared to the British Pound, with respect to other macro indicators such as inflation and legal interest rates. As shown in Table 2, the mean reinsured portion of risks in the Spanish fire insurance industry reached almost 30% of the business. However, looking at the standard deviation, there were substantial differences within the sector. Organizing the data according to nationality it can be seen that, as we expected, Spanish companies purchased more reinsurance than did British or French ones. Certainly, these offices 71 Prieto, El reaseguro: función económica, Pearson, ‘Development of reinsurance’, Adams, ‘The reinsurance decision’ and Kader et al., ‘Determinants of reinsurance’. 72 We use GDP data at market prices and the main price index, both by Prados de la Escosura, in Carreras and Tafunell, Estadísticas históricas, 1291; 1331. 23 would have access to the financial support of their mother companies, which in turn had direct contact with more capitalized financial markets. Leverage shows a similar trend, with Spanish companies reaching higher values than British and French ones, and mutual societies surpassing the rest. Companies from other countries are rather heterogeneous and include only a few companies of each origin. Danish, US, Italian, and Portuguese companies showed stable values next to the full sample mean levels for reinsurance and leverage. However, German companies presented much higher values: reinsurance ratios oscillated from 60 to 95% after the outbreak of the First World War and during the monetary distortions of the 1920s, while leverage in this period in the case of Nord Deutsche was no lower than 600. Finally, Swiss companies displayed an interesting asymmetry: while leverage levels remain between 200 and 300 for all companies – that is, lower than the rest of the sample – reinsurance ratios varied from the approximately 55% of La Baloise to the 8% of La Federale. This could be explained as a Swiss anomaly, since these companies had direct access to a highly efficient capital market and, at the same time, can be expected to have been in close contact with one of the most advanced reinsurance markets in the world. In this scenario, the decision to reinsure would not arise from a financial necessity, but from a managerial decision related to other elements. For example, the low rates of retention of La Baloise could be explained with the tight links the company had with the reinsurance industry: certainly, the office was very close to the Basler Handelbank, which was one of the main shareholders and founding member of Swiss Re. 73 As a consequence of this high heterogeneity, we include these companies in the full sample, but not as a separate analysis group. 73 James et al., The value of risk, 247. 24 Regarding the Spanish, British, and French companies, the figures indicate a positive correlation between reinsurance and leverage, despite a high dispersion of values. With regard to loss ratio indicators, one can see that British and French fire insurance companies had high figures compared to the Spanish ones (and the other nationalities). Figures on volatility show that British insurers especially faced great variations in loss ratio. A close examination shows that most jumps in the loss ratio are attributed to two kinds of extreme values: those produced by abnormally low figures for premiums earned due to entering or exiting the market, as in the case of Legal and General and General Accident and Fire; but also six entrances corresponding to catastrophic losses during 1909, 1912 and 1914. Liquidity indicators show the managerial practices of foreign insurers that would keep minimum resources in the country in order to maximize the returns of their investment in the host market, while Spanish companies displayed a somewhat greater preference for maintaining liquid assets. However, both groups are highly heterogeneous, as indicated by standard deviation values. With regard to size, it shows homogeneous values for all nationalities, while in the case of age there are large internal deviations inside groups and differences between nationalities in favour of foreign companies, which could balance the levels of market knowledge for different countries of origin. Finally, as a cross variable, data on mutual societies have been separated from the Spanish sample. As shown in the table, these societies are the most prone to use reinsurance, but are also those with higher levels of leverage and lower levels of liquidity. In parallel, their loss ratios are below the mean of the market and they are among the youngest companies operating in the fire branch. 25 III To empirically examine the determinants of reinsurance, we have employed a panel data approach. Since one of the key variables (nationality) is time-invariant, we start with a basic random-effects model, using the following regression equation: REINit = α + ∑ β Xit + β Yt + β Ut + β Vt + εit where REINit is the ratio of reinsured premiums of company i in year t; Xit represents the nationality of the company (NATSP for Spanish; NATUK for British; NATFR for French); Yt refers to organizational form (OF); Ut is the vector of macro variables including exchange rate (EXCH) and GDP growth in real terms (GRTH); and Vt is the vector of firm-specific variables including LEV, LOSS, LIQ, SIZE and AGE. To identify the pairwise relation across the variables employed in our analysis, we build a correlation coefficient matrix including nationality, organizational form, country-level data, and firm-specific indicators so as to identify the sign, intensity, and statistical significance of the pairwise relations between variables. As shown in Table 3, nationality and organizational form have a significant correlation with reinsurance: Spanish companies are expected to reinsure more than British ones, while mutual societies are more prone to purchase reinsurance than are stock companies. With regard to macroeconomic variables, GDP growth has an insignificant impact on reinsurance, while exchange rate shows a significant negative correlation, as hypothesized. Firm-specific variables perform in diverse ways. Leverage is significantly and positively correlated with reinsurance, as predicted. Loss ratio shows a weak but statistically significant inverse correlation with reinsurance, which is contradictory to the risk aversion framework. As we expected, age is significantly and inversely correlated with reinsurance, while size and, especially liquidity, have less impact. Regarding correlations between independent variables, nationality indicators show the 32 very weak. With regard to size, it comprises a statistically significant impact on reinsurance (p-value < 0.001): as stated in the previous literature, small companies purchased more reinsurance so as to increase their presence on the market without accumulating new capital resources. Furthermore, they are more sensitive to loss volatility since their ratio of liabilities over assets is higher. 79 In addition, the characteristic of this variable as a “catchall” for the variations in the rest of the firmspecific variables contributes to its high significance (see Appendix 4). Finally, the other firm-specific variables show no significance in our model, as they are affected by the aforementioned multicollinearity problems. In relation to the role of organizational form, the Hausman test was not statistically significant (p-value > 0.010), indicating that a random-effects model was more appropriate. Table 5 displays the second set of estimators applied over a sample restricted to Spanish companies. The performance of organizational form shows a strong and positive impact in the use of reinsurance (p-values < 0.005). This is consistent with the previous literature: mutual societies purchased more reinsurance than did stock companies, since they suffered from extra capital constraints due to their lack of equity capital. 80 In the case of Spanish mutual societies, this tendency to reinsure has been noted as a growth strategy in individual cases like Mutua General de Seguros. 81 According to the results, this was likely a common practice in property insurance, in such a way that mutual societies collected high commissions and kept their position in the local markets, where they operated through an intense use of reinsurance. 79 Mayers and Smith, ‘Evidence from reinsurance market’, Adams, ‘The reinsurance decision’, Adiel, ‘Reinsurance’ and Dickson and Waters, ‘Reinsurance and ruin’. 80 Pearson, ‘Development of reinsurance’, noted the use of reinsurance as a way for small stock companies and mutual societies to increase their market share, while Adams et al., ‘Mutuality’, Adams et al., ‘Competing models’, Kader et al, ‘Determinants of reinsurance’ and Lilljegren and Andersson, ‘Variation in organizational form’ found that Swedish mutual societies were less prone to reinsure since they were relatively less exposed to unexpectedly high losses. 81 Gutiérrez and Pons, ‘Risk management’. 33 Table 5. Reinsurance determinants in Spanish fire insurance companies (1908-1934). Random effect linear regression with organizational form, exchange rate and firmspecific variables. (1) (2) ORG 41.378a (12.205) 35.302a (13.108) FALSE 14.100c (7.262) 12.702c (7.404) EXCH -15.578a -15.880a (2.840) (2.996) LEV 0.016b (0.008) LOSS 0.068 (0.046) LIQ -0.009 (0.008) SIZE -3.407b -4.985a (1.501) (1.702) Const. 88.492a 100.606a (20.003) (22.382) R2 within 0.136 0.172 between 0.394 0.425 overall 0.195 0.232 a, b, c Statistically significant at the 1%, 5% and 10% level. Note: ORG = 1 for mutual, 0 for stock company; FALSE =1 if false Spanish company, 0 othervise; EXCH = Official exchange rate of Peseta / 100 French Francs (FF); GRTH = % rate of GDP growth in real terms; LEV = Premiums earned / Ongoing-risks reserves, in %; LOSS = Claims / Premiums earned, in %; LIQ = Cash + Bank accounts / Ongoing-risks reserves, in %; SIZE = Natural logarithm of premiums earned. Source: AGA (1) 26 - Cajas I-1 to I-141 Top. 13/31 and AGA (1) 26 - Cajas 11/02131 to 11/02180; Comisaría General de Seguros, Memoria sobre las entidades de seguros que operan en España, Years 1911, 1912, 1913, 1915. To control for the potential effect of foreign governed Spanish companies, we included a dummy for ‘false Spanish companies’. The impact of foreign governance on reinsurance was statistically significant and positive in both models. However, the performance of the three false Spanish companies included in our sample requires further explanation. Numancia was included in a group of German-owned capital with others offices such as Plus Ultra, which operated in the transport branch. In the second case, Hispania was founded in 1902 by investors from Catalonia in order to sell workaccident insurance policies. In 1911, it began operations in the fire line but, three years 34 later, it was bought by the Swiss Zurich. Attempting to specialise the company in the work-accident line, the new managers assumed the policies in force, and in 1916 left the fire branch. As a result, most of the business in 1914, 1915 and 1916 was reinsured to Zurich. 82 Finally, La Unión y el Fénix Español was constituted with a 58.22% of nonSpanish capital, though it was declining over the following decades. 83 In parallel though, the company maintained a special agency in Paris whose direct business was devoted to the Spanish market. Additionally, the financial sources of the company were based in Spain, showing close relations with the Banco de España and, during the 1920s, with the Banco Hipotecario de España. In this way, the special reinsurance performance could be related not to its foreign shareholders, but to its role as the gateway to foreign reinsurance for a substantial share of Spanish insurers. 84 Supporting the results obtained in the model presented in Table 4, exchange rate holds a significant and negative correlation with reinsurance purchases. As regards firm-specific variables, size is again the only one with a statistically significant impact on reinsurance (at the 5% level in the complete model). In contrast, leverage, loss ratio, and liquidity show no significance within the model. Taking into account the results shown in both models, reinsurance seems to be a widely used device by Spanish companies in order to enhance their underwriting capacity. In fact, controlling for firm-specific variables and paying attention to variable coefficients (Table 4), this indicates that Spanish nationality holds a strong preference for reinsurance. In view of the differences derived from nationality regarding access to capital resources, this strong systematic preference could be considered to be due to the 82 Pons, El grupo Zurich. 83 Tortella et al., La Historia del seguro, 121. 84 Gutiérrez and Pons, ‘Risk management’, 6–9, describes this role and connects it to the international scope of the company. 35 contribution of foreign reinsurance as a funding provider to enhance the underwriting capacity on the undercapitalized Spanish insurance market. Furthermore, the financial role of reinsurance as a key element in the development of the Spanish insurance market connects with other dimensions noted in the previous literature and related to market structuring and internationalization management. It seems clear that, as stated in the previous literature, reinsurance served as the main channel for small local companies and mutual societies to participate in a dense network managed by big companies, such as La Unión y el Fénix Español, capable of covering local risks by using their interregional and even international underwriting capacity. 85 Certainly, reinsurance played a major role in the globalization of Spanish risks, but also in the strategy of internationalizing Spanish offices such as La Unión y el Fénix Español. Following the procedure of British companies like Phoenix Assurance, analysed by Trebilcock and Kobrak, this company used reinsurance so as to break into new markets such as France, Italy, Belgium and the United States. 86 IV Knowledge and technology transfer benefitted insurance providers in continental Europe. The development of new devices like reinsurance allowed them to deal with larger risks and with financial constraints. Certainly, as an imperfect substitute for capital, reinsurance enhanced underwriting capacity and acted as a funding provider for insurers. Simultaneously, reinsurers built wide networks of risk exchange which, in turn, contributed to modernizing peripheral latecomers. 85 Gutiérrez and Pons, ‘Risk management’ 86 Idem. For the case of Phoenix, see Trebilcock, Phoenix Assurance, Kobrak, ‘USA: International attraction’. 36 In the Spanish case, the insurance industry developed quickly together with the rest of the economy from 1870 to the outbreak of the Civil War in 1936. Foreign investment and new management focuses contributed to the growth of the industry. However, the scarcity of capital resources and the undercapitalization issue remained the main shortcomings on the Spanish insurance market throughout the period. Systemic imbalances and rigidities, as well as a lack of diversification in the financial system, defined an insurance industry affected by capital constraints and high levels of leverage. According to the empirical results, reinsurance played a key role in avoiding these shortcomings as an enhancer of the underwriting capacity of domestic insurers and, in parallel, as a device for dealing with undercapitalization issues. As expected, nationality and organizational form operated as high-impact factors defining reinsurance strategies. Certainly, when controlling for firm-specific variables, we note that the performance of Spanish insurers relied heavily on their access to the reinsurance market, as they were significantly more prone to reinsure than were British and French offices. In contrast to foreign companies, and taking into account the differences related to nationality regarding the access to capital markets, reinsurance served as a major source of capital funding for domestic insurers. These findings are consistent with previous studies dealing with the organization of reinsurance flows in Spain and the links between local markets and global networks. In this sense, our study contributes by empirically testing the impact of reinsurance on undercapitalized and developing economies, and by offering a quantitative measure of this influence on the modernization and development of the insurance market. Our analysis further suggests that around 20% of the underwriting capacity on the undercapitalized Spanish insurance market was supplied by foreign reinsurance 37 networks. We show that the provision of foreign capital was a necessity to keep pace with the demand arising from the process of capital formation; fire insurance meet a critical need in both improving financial protection (collateral) and mitigating indemnity of losses due to unexpected risks. However, the capital import may have exacerbated the balance of payment difficulties unless the productivity growth following the reinforced capital formation was strong enough to balance by greater commodity export. REFERENCES Adams, M.B., ‘The reinsurance decision in life insurance firms: an empirical test of the risk-bearing hypothesis’, Accounting and Finance, 36 (1996), pp. 15-30. Adams, M.B., Andersson, L.F., Jia, J.Y. and Lindmark, M., ‘Mutuality as a control for information asymmetry: a historical analysis of the claims experience of mutual and stock fire insurance companies in Sweden 1889-1939’, Business History, 53 (2011), pp. 1074-91. Adams, M.B., Andersson, L.F., Lindmark, M. and Veprauskaite, E., ‘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 (2012), pp. 990-1014. Adiel, R., ‘Reinsurance and the management of regulatory ratios and taxes in the property-casualty insurance industry’, Journal of Accounting and Economics, 22 (1996), pp. 207-40. Alborn, T., Regulated lives: life insurance and British society, 1800-1914 (Toronto, 2009). 38 Bartolomé Rodríguez, I., Martín-Aceña, P. and Martínez Ruiz, E., ‘A Spanish tale: the Great Depression versus the Great Recession’. Communication presented at the XI Spanish Economic History Association Congress, Madrid, 4 September 2014. (Consulted on 10/10/2014). http://www.aehe.net/xicongreso/pdfs/sp_bartolome_martinez.pdf Berger, L.A., Cummins, J.D. and Tennyson, S., ‘Reinsurance and the liability insurance crisis’, Journal of Risk and Uncertainty, 5 (1992), pp. 253-72. Borscheid, P. and Haueter, N.V., eds., World insurance. The evolution of a global risk network (Oxford, 2012). Canals i Vilaró, S., Cuestiones económicas: el Comité Oficial de Seguros y el reaseguro obligatorio del Estado (Barcelona, 1923). Carreras, A. and Tafunell, X., dirs., Estadísticas históricas de España (v. III) (Madrid, 2005). Cenamor Val, H., El seguro sobre la vida en España en 1934 (Madrid, 1935). Cole, R.C. and McCullough, K.A., ‘A reexamination of the corporate demand for reinsurance’, The Journal of Risk and Insurance, 73 (2006), pp. 169-92. Comín Comín, F., ‘Las crisis de la deuda: el largo camino desde los impagos a la gestión responsable’, in P. Martín-Aceña, E. Martínez Ruiz and M.A. Pons Brías, coords., Las crisis financieras en la España contemporánea, 1850-2012 (Barcelona, 2013), pp. 197-240. Cuevas Casaña, J., ‘Las crisis bursátiles, 1850-2000. De la burbuja ferroviaria a la tecnológica’, in P. Martín-Aceña, E. Martínez Ruiz and M.A. Pons Brías, Las crisis financieras en la España contemporánea, 1850-2012 (Barcelona, 2013), pp. 159-96. Deelstra, G. and Plantin, E., Risk theory and reinsurance (London, 2014). 39 Dickson, D.C.M. and Waters, H.R., ‘Reinsurance and ruin’, Insurance: Mathematics and Economics, 19 (1996), pp. 61-80. Doherty, N.A. and Tinic, S.N., ‘Reinsurance under conditions of capital market equilibrium: a note’, The Journal of Finance, 36 (1981), pp. 949-53. European Commission, International Monetary Fund, Organisation for Economic Cooperation and Development, United Nations and World Bank, System of National Accounts 2008 (New York, 2009). Frax Rosales, E. and Matilla Quiza, M.J., ‘Los seguros en España: 1830-1934’, Revista de Historia Económica, 14 (1996), pp. 183-203. Gadea, M. D. and Sabaté, M., ‘The European periphery in the Era of the Gold Standard: the case of the Spanish Peseta and the Pound Sterling from 1883 to 1931’, Open Economies Review, 15 (2004), pp. 63-85. Garven, J.R. and Lamn-Tennant, J., ‘The demand for reinsurance: theory and empirical tests’, Assurances, 71 (2003), pp. 217-38. Gerathewohl, K., Reinsurance: principles and practice (London, 1985). Gerschenkron, A., Economic backwardness in historical perspective (Cambridge, 1962). Golding, C.E., A history of reinsurance (London, 1927). Gutiérrez González, P., El control de divisas durante el primer franquismo. La intervención del reaseguro (1940-1952) (Madrid, 2014). Gutiérrez González, P. and Pons Pons, J., ‘Risk management and reinsurance strategies in the Spanish insurance market (1880-1940)’, Business History, (2016). http://dx.doi.org/10.1080/00076791.2016.1187136 James, H., Borscheid, P., Gugerli, D. and Straumann, T., The value of risk. Swiss Re and the history of reinsurance (Oxford, 2013). 40 Jovanovic, B., ‘Selection and evolution of industry’, Econometrica, 50 (1982), pp. 64970. Kader, H.A., Adams, M. Andersson, L.F. and Lindmark, M., ‘The determinants of reinsurance in the Swedish property fire insurance market during the interwar years, 1919-39’, Business History, 52 (2010), pp. 268-84. Kobrak, C., ‘USA: The international attraction of the US insurance market’, in P. Borscheid and N.V. Haueter, (eds.), World Insurance. The Evolution of a Global Risk Network (Oxford, 2012), pp. 274-310. Krugman, P., ‘A model of innovation, technology transfer and the world distribution of income’, Journal of Political Economy, 87 (1979), pp. 253-66. Lamn-Tennant, J. and Starks, L., ‘Stock versus mutual ownership structures: the risk implications’, Journal of Business, 66 (1993), pp. 29-46. Lilljegren, J. and Andersson, L.F., ‘Variation in organizational form across lines of property insurance: Sweden, 1913-1939’, Financial History Review, 21 (2014), pp. 77-101. Malo de Molina, J.L. and Martín-Aceña, P., eds., The Spanish financial system. Growth and development since 1900 (Basingstoke, 2011). Martín-Aceña, P., ‘Desarrollo y modernización del sistema financiero, 1844-1935’, in N. Sánchez Albornoz, ed., La modernización económica de España (Madrid, 1987), pp. 121-46. Martín-Aceña, P., ‘Spain during the classical Gold Standard years, 1880-1914’, in M. Bordo and F. Capie, eds., Monetary regimes in transition (Cambridge, 1993), pp. 135-72. 41 Martín-Aceña, P., ‘The Spanish banking system from 1900 to 1975’, in J. L. Malo de Molina and P. Martín-Aceña, eds., The Spanish financial system. Growth and development since 1900 (Basingstoke, 2011), pp. 99-144. Martín-Aceña, P., ‘Crisis bancarias. Nada nuevo bajo el sol’, in P. Martín-Aceña, E. Martínez Ruiz and M.A. Pons Brías, coords., Las crisis financieras en la España contemporánea, 1850-2012 (Barcelona, 2013), pp. 53-114. Martínez Ruiz, E., ‘Papel mojado. Crisis inflacionarias’, in F. Comín Comín and M. Hernández Benítez coords., Crisis económicas en España. 1300-2012: lecciones de la historia (Madrid, 2013), pp. 203-26. Mayers, D. and Smith, C.W., ‘On the corporate demand for insurance’, Journal of Business, 55 (1982), pp. 281-96. Mayers, D. and Smith, C.W., ‘On the corporate demand for insurance: evidence from the reinsurance market’, Journal of Business, 63 (1990), pp. 19-40. Pearson, R., ‘The development of reinsurance markets in Europe during the nineteenth century’, Journal of European Economic History, 24 (1995), pp. 557-72. Pearson, R., ‘The birth pains of a global reinsurer. Swiss Re of Zürich, 1864-79’, Financial History Review, 8 (2001), pp. 27-47. Pearson, R., Insuring the Industrial Revolution: fire insurance in Great Britain, 17001850 (London, 2004). Pearson, R., ‘The growth, organization, and diffusion of the British insurance industry’, in Instituto de Ciencias del Seguro, ed., Encuentro internacional sobre la historia del seguro (Madrid, 2010), pp. 13-42. Pearson, R., ‘Las compañías de seguros en España antes de 1914’, in J. Pons Pons and M. A. Pons Brías, eds., Investigaciones históricas sobre el seguro español (Madrid, 2010), pp. 101-30.