Collusive and Competitive Institutions in the Swedish Insurance Market

Artikelförfattare: Alexander Boksjö
Utgåva:
2, 1994
Språk: Engelska
Kategori:

139 NFT 2/1994 The Swedish insurance industry has under- gone astonishing development since the mid nineteenth century. From being a quite small sector in the economy it has grown to become one of the most important actors in the Swed- ish credit market, and one of the most domi- nant institutional owners in the stock, bond and real estate markets. A characteristic feature of this development has been the extensive co-operation between companies to stabilise an uncertain and fluc- tuating market. Both market forces and the Swedish legislature have influenced and giv- * The authors are Ph.D. candidates and engaged in a project that studies the structure and development of the Swedish insurance market. The project is led by Profes- sor Mats Larsson at Uppsala University, Department of Economic History, Financial History Unit. Alexander Boksjö Collusive and Competitive Institutions in the Swedish Insurance Market by M.A. Alexander Boksjö and M.A. Mikael Lönnborg-Andersson, Financial History Unit, Department of Economic History, Uppsala University* The present structure of the Swed- ish insurance industry is the re- sult of a long historical process, in which relational contracting between companies is of crucial importance. The policies adopt- ed by the firms towards each other — their market conduct — were early characterised by a mix of collaboration and compe- tition. On balance, this resulted in a market development, which in many ways is unique com- pared with other financial markets. The mutual interdependence of an oligopolistic market often results in agreements, which aim to cope with structural uncertainty; this is evident in the insurance business. Formal contracting was widespread, and many cartel agree- ments were signed, of differing spans and covering diverse areas of activity. Knowledge of insurance cartels is very incomplete. We neither know the persist- ence nor the frequency of cartel agreements. Another area where our knowledge is very random is with regard to which insurance branches the cartels have been founded and how this changed over time. Thus the purpose of this paper is to describe and analyse cartels of the Swedish insurance industry 1947—1980. Mikael Lönnborg- Andersson 140 en this co-operation many different forms. A definition of the insurance market would in- clude the insured, share owners, lenders and borrowers, who together with the government have created the environment in which the insurance companies function. The insurance companies make it possible for individuals to spread their risks over short- er periods, as well as over a lifetime. In other words the insurance industry is a risk-spread- ing institution, which both limits risks and increases safety.1 Thus, when insurance con- tracts relate to a long period, it is necessary that the companies survive to fulfil their as- signments. From this point of view it has been important for legislators to create an institu- tional form with the crucial goal of keeping the structure of the insurance industry intact. To this end, the private insurance supervisory service (the Swedish Private Insurance Su- pervisory) has sanctioned the formation of cartels and other forms of co-operation. How these cartels have functioned is very important, as is their effect on financial mar- kets, although we know very little about this. In which insurance branches the cartels have been founded and how this has changed over time is another area where our knowledge is very random. This paper aims to chart and analyse the cartels in the Swedish insurance industry between 1947—1980. In section 1 we will describe the origin of co-operation in the insurance market and stress different factors that have affected the struc- ture of the market. The discussion concen- trates on the period before 1947, and the intention is to give a short summary regarding what has characterised the insurance market. We also look at the influence of legislation on the co-operation practised and the effect of different trade associations on the market’s structure and behaviour. Section 2 examines different theories of mergers and describes why we have chosen to analyse cartels from this point of view. We then deal with the empirical study of the Swedish Private Insur- ance Supervisorys registry of cartels, in sec- tion 3. Finally, in section 4 we try to relate the theoretical framework with the empirical study, and to explain the motives for and barriers to co-operation and cartels in the Swedish insurance industry. 1.1 The origin of co-operation The first agreement for co-operation dates from 1866 between the Fire- and Life-Insur- ance Company Svea (founded in the same year) and Skandia, the first joint-stock insur- ance company in Sweden, which began busi- ness in 1855. Svea took the initiative in this agreement. The main purpose in constructing the new organisation was to develop day-to- day systems such as accounting routines and other operational devices, and to create a common policy to satisfy the total need of insurance in Sweden. It was an attempt to shut out other actors from the fire and life insur- ance market, although the formal agreement concerned only common fire insurance con- ditions.2 The incentive for this co-operation on the side of Svea was an endeavour to reduce their establishment costs and at the same time to construct an organisation that had already proved to be functional. For Skandia we may see this arrangement as a simple way of find- ing out more about their new competitor and in the long run also as an attempt to control the market. At the end of 1871 a third joint-stock com- pany, Nordstjernan Life Insurance Company, was set up. Nordstjernan developed contacts with Skandia to receive information about calculating premiums, tariffs, and other prac- 2) Framtiden livförsäkringsbolag, De förenade, Fram- tiden, Victoria — minnesskrift, Uppsala, 1972, p. 6 f. 1) Skogh, G., & Samuelsson, P., Splittring eller samman- hållning i svensk försäkring, Lund, 1985, 17 f. 141 tical aspects of the business. During this peri- od thought was given to the control of the market, and Skandia, Svea and Nordstjernan concluded a mutual agreement in 1877. This comprised additional premiums for journeys outside Europe, for some specific occupa- tional groups, and for some abnormal health hazards, and also common regulations for acquiring activity provisions and collecting provisions for medical fees. There was also an agreement regarding re-insurance between the companies.3 Through this support from already estab- lished firms, the new insurance companies acquired similar organisational structures. Because management and decision-making processes were organised in a comparable way, this facilitated a high degree of co- operation. Another important development was the rotation of senior management be- tween the companies, which created a degree of social networking in insurance. The inter- dependence and rotation of personnel be- tween the Swedish Private Insurance Super- visory, which was established in 1904, and the companies contributed to an even higher degree of co-operation.4 There was therefore close mutual relations, which influenced the structure of the insurance market. The recruit- ment of senior management and supervisors in the Inspectorate was, until mid 1950s, done exclusively from the insurance branch. These different factors contributed to making the Swedish insurance sector a closed system, with small possibilities for the uninitiated to enter or obtain experience, and so further strengthened the solidarity within the existing insurance industry. 1.2 Competition within the insurance market The first tendency towards competition can be seen in connection with the shift from fire to life insurance. When Skandia and Svea were founded, fire insurance was the most- profitable branch of business, but after some major fires with subsequent heavy disburse- ments the profits in the fire insurance branch decreased. At the same time life insurance had shown a satisfying yield: thus, the founding of companies that operated exclusively in the life insurance branch expanded.5 The share- capital required to start up activity was re- duced from the beginning of 1870, and the newly-founded companies directed their busi- ness towards more specified insurance branch- es. These two factors contributed to a more heterogeneous market structure. The life insurance company Thule was founded in 1872 and was the fourth joint- stock insurance company in Sweden. Thule increased the insured’s influence in the organ- isation, through the possibility of electing an accountant. The share-capital was fixed at a comparably low level and dividends to the shareholders were restricted. The surplus was instead to be refunded to the insured. Thule’s intention was to reduce costs and so to be in a position to offer lower premiums, and the insured’s influence was a means of winning market shares. Skandia, Svea and Nordstjern- an regarded these measures as unfair compe- tition and they reacted jointly against Thule. Hence, a strong antagonism arose between the established companies and Thule which divided the market into two spheres. From their beginning Thule had trouble reaching a satisfactory earning capacity. But gradually Thule’s ideas gained the public’s confidence, and newly-founded companies after 1872, and eventually also the established ones, ad- 3) Ibid., p. 10 f. 4) For a more detailed statement of the interactive rela- tion between the Supervisory and the insurance compa- nies, see Englund, K, Försäkring och fusion, Skandia, Skåne, Svea, Thule, Öresund 1855—1980, Stockholm, 1980, p. 54. 5) Framtiden livförsäkringsbolag, op. cit., p. 9. 142 justed their regulations towards Thule’s mod- el.6 At the end of the nineteenth century a new form of insurance association developed, where the capital stock consisted of the mem- bers’ entry and yearly fees. These associa- tions eventually turned into mutual insurance companies, based on the principle that the insured also were the owners and that any surplus should be refunded to the insured. The first mutual life insurance company started in 1887, and by 1895 four more had been found- ed. Between 1895 and the end of the century this area of business rapidly increased, and eight mutual companies began their activities during this period. The mutual insurance com- panies established a certain degree of co- operation concerning similar insurance con- ditions and premiums. The new form of companies contrasted with the established joint-stock companies, who stated that the mutual companies were finan- cially weak and might thereby decrease the public’s confidence and so threaten the sur- vival of the whole market. This started the so- called ”big principle controversy”, the debate being mainly concentrated in the insurance journals. An effect of this schism was that co- operation between the joint-stock and mutual companies became restricted. The mutual companies tried to expand their business to a wider range of insured. Howev- er, neither the mutual nor the joint-stock com- panies were successful in spreading life assur- ance among the working class. Both forms of companies based their insurance fees on ad- vance payments, either for a year, six months or three months. This decreased the possibil- ity for any member of the working class to sign a life insurance contract, since they could not easily raise the amount of capital required in advance. Furthermore, the existing type of organisation could not absorb this potential market because it required a considerable number of collectors who could collect fees on a weekly basis and at a reasonable cost. To fill the need for insurance within a broader class of society, the people’s insur- ance companies were established. They worked with new groups of customers, with other methods and with different conditions of insurance than the established companies. Weekly fees were introduced, which meant that almost everyone could sign an insurance contract. The first people’s insurance compa- ny was Life-Insurance Institution Trygg — mutual, founded in 1899.7 Sweden’s Practical Life-Insurance Institu- tion Folket — founded in 1914 — was one of the most important companies in this sector. Folket was connected to the Consumers Co- operative Movement through agreements that Folket should invest parts of its assets in the Cooperation’s projects. Folket commenced an intensive propaganda campaign against the private insurance companies, especially on the issues of cartels and other measures of co-operation. The reason behind this was to increase competition and hence lower the insured’s premium. At the same time Folket officially took a stand against any form of co- operation.8 To sum up, there had been a strong antago- nism within the insurance sector against new actors. New organisations and different meth- ods had been strongly opposed by the estab- lished insurance companies. However, if the innovations had shown viability, they were often absorbed by the rest of the insurance companies. Here it is possible to see the mar- ket’s development as showing what we call a stair effect: when one company does some- thing that proves to be successful, the rest follow. 6) Ibid., p. 10 f. 7) Framtiden livförsäkringsbolag, op. cit., p. 32 f. 8) Larsson, M., Den reglerade marknaden — svenskt försäkringsväsende 1850—1980, Stockholm, 1991, p. 36 f. Framtiden livförsäkringsbolag, op. cit., p. 111 f. 143 1.3 The threat of nationalisation In the mid 1930s, as a result of the crash of the Kreuger empire, the demand for capital de- creased. In combination with a higher internal financing of the industry, both short and long term interest rates dropped.9 When the oppor- tunities for profitable investment contracted, the competition between insurance compa- nies and the commercial banks increased.10 High administration costs caused by the het- erogeneous structure of the insurance indus- try was a large problem, and this together with the lower profitability of the market, received much bad publicity. Another problem was the large number of cancellations of life assur- ance contracts. The main reason for the can- cellations was a large number of insurance agents of doubtful business acumen, for whom the insurance companies could not offer suf- ficient education. The agents’ lack of knowl- edge of what different insurance alternatives existed also increased the number of cancella- tions. Further, the generally costly structure with its many insurance agents was criticised by the public. The cancellations forced up the companies’ administration costs and the insured thus lost a part of the sum insured. Sections of the Social Democratic Party saw complete na- tionalisation as the only solution. Through such socialisation the government saw the possibility of achieving firmer control over the Swedish credit market, and thus facilitate the goals of fiscal and monetary policies. The threat of nationalisation brought about a stronger connection between the private insurance companies and the Swedish Private Insurance Supervisory. The Inspectorate’s chief inspector, O. A. Åkesson, initiated sev- eral actions aimed at tightening the co-opera- tion and reducing unnecessary transaction costs. The intention was to refinance through measures which decreased costs, and to in- crease extent of competition, the goal being to raise efficiency and so avoid nationalisation. In 1935 several branches of the industry start- ed up committees whose objective was to reduce the principal differences between the companies, for example, between joint-stock and mutual insurance companies, and to de- velop a wider range of co-operation. An im- portant aim was also to increase public confi- dence in the private insurance market and thereby eliminate the grounds for nationalisa- tion.11 1.4 The influence of legislation on co-operation Before the legislation of 1903 the insurance market was very unstable. Co-operation and agreements that restricted competition can, from this point of view, be seen as a substitute for a non-existent legislation. Further reason to introduce common regulations was to legit- imatise insurance activity and gain public confidence. The most important reason be- hind the mutual striving against general legis- lation was probably that the unregulated mar- ket ran the risk of becoming overcrowded. Co-operation and new regulations protected the established companies towards potential competitors, and thus reduced uncertainty in the market.12 The first insurance legislation in 1903 proved in fact a state intervention that opened up an even closer co-operation. It sought to bring uniformity among the companies concerning the mathematical basis for life insurance, i.e. on the calculation of premiums, profit, and9) Kock, K., Kreditmarknad och räntepolitik, 1924— 1958, Uppsala, 1961, p. 154 ff. 10) Eliasson, K., ”Försäkringsbolagens kapitalplaceringar under tidsperioden 1915—1985”, (Mimeo), Department of Economic History, University of Uppsala, 1992, p. 36 f. 11) Grip, G., Vill du frihet eller tvång? Svensk försäk- ringspolitik 1935—1945, Stockholm, 1987, p. 25 ff. 12) Larsson, M., op. cit., p. 49. 144 profit-sharing. The life insurance companies were required to furnish more detailed infor- mation on the calculation of their premiums, and the reserve of premiums and the regula- tions concerning repurchase of insurance. The companies had also to establish and put aside resources for an insurance fund. Hence, the 1903 legislation stabilised the insurance com- panies’ capital base and financial structure, which meant that regulations on premium and profit became alike among the insurance com- panies, and decreased as an instrument for competing. The Swedish Private Insurance Superviso- ry was also strengthened at the same time. The Insurance Supervisory gave its support to an even higher level of co-operation because of the stabilising effect it had on the market.13 The concession constraint for insurance com- panies, which became harder after the legisla- tion, reduced the number of new companies. The 1914 additions to the insurance legisla- tion of 1903 contained further regulations that decreased competition. The use of returned provisions among insurance agents had been a big problem in the acquisition activity. The policyholders received a part of the agent’s provision for policies written.These provi- sions were a competitive tool among the dif- ferent insurance agents, but were not sanc- tioned by the insurance companies. The legis- lation involved more strict guidelines for the agents, and the companies became more re- sponsible for the acquisition activities. This meant a standardisation, because the only approved way to refund surpluses to the in- sured was through premium divends or profit- sharing. The 1927 law of insurance agreements then increased the uniformity concerning policy- holders’ information duties and obligations accompanying any agreement. Through this regulation the insured received a better survey of the market, and insurance conditions, at the same time as the differences between the companies were reduced. The standardisation of the conditions decreased competition, but the Inspectorate thought it was necessary be- cause it created a financial security that better secured the insured’s rights. The further insurance legislation of 1948 sharpened the control of the establishment of new companies. The principle of a means test was added to the earlier concession examina- tion. According to this principle companies that wanted to enter the market had to fulfil a perceived need and offer a sound develop- ment to the insurance market. Even compa- nies that intended to diversify their activities had to apply for a new concession. Thus, the opportunity for the Swedish Private Insur- ance Supervisory to limit the number of new companies increased, as did the possibility of drawing up guidelines for the future develop- ment of the insurance market.14 A heterogeneous insurance market with many actors meant that the administration costs amounted to a considerable share of total costs. The Swedish Private Insurance Supervisory’s possibility to carry out a survey and to supervise each company was also con- straint. Another intention with the law of 1948 was to facilitate mergers and other meas- ures to increase the market’s degree of con- centration, and thereby reduce administration costs. It was also feasible to combine different insurance branches, except for life insurance, in one company which further increased the concentration ratio.15 Finally, the life insurance companies, in accordance with the 1948 law, had to follow 13) Folksams försäkringsutrednings betänkande, Stock- holm, 1962, p. 22 f. 14) At the same time the province— and district insur- ance companies were placed side by side with national companies, which further increased the Supervisory’s possibilities to adjust the structure of the insurance market. Larsson, M., op. cit., p. 60 f. 15) Ibid. 145 the principle of reasonable premiums.16 An effect was that the companies’ attentions were drawn to cutting overhead costs, through sim- plified administration routines, rationalisa- tion and co-operation with other companies. The Inspectorate’s intentions had instead been to increase competition between the compa- nies and in this way to stimulate rationalisa- tion and lower premiums. To sum up. On the legislation’s influence regarding co-operation, it basically stimulat- ed agreements involving the restriction of competition, thereby standardising the com- panies and making it difficult for new busi- nesses to enter the insurance market. It is also important to keep in mind the social value of insurance. Hence, the legislation had been specially attentive to the possibility of com- panies’ long-run survival and fulfilment of their contracts. The legislation in fact institu- tionalised a restriction of competition, to re- duce insecurity in the market. Together with the concession constraint, this policy has been successful in reshaping the structure of the insurance market. 1.5 Trade associations Another important factor in developing co- operation in the insurance market was the several trade associations that were estab- lished. The first association dealing with is- sues regarding common interests in the insur- ance market was the Swedish Fire Tariff As- sociation, founded in 1873 by Skandia and Svea. Hence, non-life insurance was the first activity with a formal organisation. The In- surers Association (Sjöassuradörernas Fören- ing) provided for the interests of underwriting companies and was set up in 1893. When the Association for Liability Insurance, founded in 1937, joined with the Swedish Fire Tariff Association in 1945, the name changed to the Swedish Tariff Association, which was active in the whole non-life insurance branch of business.17 The Insurance Society was founded in 1875, and in 1919 Swedish was added to its name. The aim of the Swedish Insurance Society was to disseminate and increase the public’s knowledge of the insurance industry. The organisation also acted as a discussion forum for insurance people. The Insurance Society encouraged the companies to seek a higher degree of co-operation, and was responsible for some common insurance education. Be- fore the 1903 insurance legislation the Socie- ty strove to achieve a governmental regula- tion for the insurance industry.18 The Swedish Life Insurance Companies Directors Association was established in 1906 (in 1936 its name was changed to the Swedish Life Insurance Companies Association). Its aim was a tighter collaboration between the life insurance companies, and it consisted of an arbitration board to handle and solve prob- lems between the affiliated companies. Another trade association was the National Association of Swedish Insurance Compa- nies, whose goal was to debate questions of common interest in the insurance market. Almost all the companies joined the associa- tion, except for those connected with Con- sumers Cooperative Movement. Within the association an information service provided information to the public on insurance issues. In 1946 this was complemented by the publi- cation of the Insurance Journal.19 16) The principle of a reasonable premium means that the insured’s final cost, (premium minus refund), should be in proportion to the standard of the insurance contract. SOU 1986:8, Soliditet och skälighet i försäkringsverk- samheten, delbetänkande av försäkringsverksamhets- kommittén, Stockholm. 17) Englund, K., op. cit., p. 65. 18) Sweden was the first country in Scandinavia with an Insurance Society. Similar organisations were founded in Denmark in 1883, in Norway in 1900 and in 1911 in Finland. Grenholm, Å., Svenska Försäkringsföreningen 1875—1935 — minnesskrift, Stockholm, 1935, p. 9 ff. 19) Framtiden livförsäkringsbolag, op. cit., p. 21 ff. 146 The task of the Swedish Tariff Association was to determine premium and insurance con- ditions for the associated organisations which normally were active in the same branch of business. The associations functioned as in- dependent investigation authorities and with an advisory service for the related companies. Agreements of the Swedish Tariff Associa- tions divided specified insurance areas among the companies. There also existed agreements that constrained the companies’ area of activ- ity and prohibited certain forms of insurance, because they were inappropriate to the in- sured. The associations even regulated the acquisition activities of the companies. An example was that the associations did not allow companies to recruit employers that already had an agreement of compulsory in- surance with another company.20 One conclusion regarding the activities of the trade associations, especially the tariff associations, is that they constrained compe- tition. Therefore, they could be, and also have been, classified as cartel agreements. 2.1 Merger theories — a survey In this section we will outline different theo- ries of mergers as a starting point for creating a model to interpret cartels. Co-operation can be said to show various degrees of stability where the merger is the most solid form. That is, there is a scale going from mergers to verbal agreement or even tacit understand- ings. Cartels will be somewhere in between these extremes and include verbal agreements. Thus we have chosen to start with the firmer type of cooperation and work down the line towards the more loose institutional cartel. There are, however, a few problems con- nected with theories of mergers. They result from studies mainly concentrating on the phe- nomena of industrial mergers, and hence their validity for the service sector is questionable. But as we study cartels this should not prove an insurmountable obstacle. Besides, some premises for business can be held to be funda- mental, irrespective of the character of the business. The difficulty of surveying the various merger theories represents another problem, as the impression is formed that every study has its own theory of the motivations and aims behind mergers. In the fields of economics it is hard to find any literature that presents more contradictory hypotheses and empirical re- sults than that dealing with the causes of mergers. Nevertheless we will try to system- ize the theories into different groups, accord- ing to their analyses of the causes of mergers. There are two main types of theories. One assumes that mergers are mostly acts of irra- tional behaviour.21 This theory is deeply root- ed in the literature of business economics and may provide an explanation for less profitable or non-profitable mergers. Especially in times of boom a sort of mass psychosis breaks out when companies merge as never before. This phenomenon could be the cause of what is known as ”waves of mergers” in economic history. The other main kind of merger theories assumes that mergers are undertaken because of well-founded decisions and rational behav- iour. We have chosen to sort these theories — there are many of them — into four groups, to do with real profitability, financial profitabil- ity, risk distribution and reduction and growth maximisation. Theories of real profitability emphasise that mergers enable increases in productive effi- ciency, because of economies of scale or consolidated market positions. Economies of scale can also involve administration and marketing, as well as distribution and re- search. The market result may be a new com- pany formation that holds a monopoly or 20) Folksams försäkringsutrednings betänkande, p. 55. 21) Alarik, B., Fusioner — Drivkrafter beslut samord- ning, Göteborgs universitet, 1982. 147 oligopoly position. In the literature the mo- tive to obtain a monopoly position is seen as a strong force behind mergers. There are also theories that argue that mergers also contrib- ute to lower information and transaction costs. A third school tries to explain why mergers are preferable to individual stock buying, by referring to tax issues and the politics of trade and industry.22 Theories of financial profitability start from the assumption that changing stock market conditions are crucial to the occurrence of mergers, while the real profitability in com- modity markets and in companies are not paid attention or are regarded as given.23 These theories are based on the relationship between stock prices and mergers, and are not appro- priate for building a model to interpret cartels. The third group, theories of risk distribu- tion and reduction, hold that the forces behind mergers are that companies may obtain a wider product range and/or new geographical markets, and thus are able to spread their risks. A merger can also make the move from mature to more expansive markets easier. However, these theories cannot explain why a merger is preferred to a diversified portfolio. We meet the same problem with theories of growth maximisation. Alone they cannot ex- plain mergers and have more of an ad hoc nature. They can be connected to the concept of managerial capitalism, which argues that the separation between management and own- ership often leads to a situation where the former tend aim for the maximum growth of sales, production and physical capital rather than maximise the rate of return on the share capital. Why is that? Managers’ salaries, sta- tus and other benefits relate to the size of the company and not the rate of return on the companies shares. Theories of growth max- imisation may be useful in explaining some diversified mergers, if a merger does not result in an explicit profitability. The relation- ship between growth maximisation and merg- ers is that companies that are run with a growth maximising goal are assumed to have a greater propensity for take-overs than com- panies with profit maximising goals. 3.1 An empirical study of the Swedish Private Insurance Supervisory’s registry of cartels In section four we will examine the insurance cartels, based on an empirical study of the Swedish Private Insurance Supervisory’s reg- istry of cartels. The intention is, first, to chart in which branch of the insurance industry the cartels were formed. The register includes all cartel arrangements made by the insurance companies after 1947, but some earlier cartels are also included if they were still active after 1947. There have been 90 cartels established in the insurance market. Half of them were still valid in 1980, with the others cancelled during the period. It is important to point out that the Swedish Private Insurance Supervi- sory’s registry of cartels began in 1947. Of course cartels had established before this year, but before 1947, in a legal sense, they were not regarded as cartels. In fact, as pointed out above, the first formal agreement constrain- ing competition, the Swedish Fire Tariff As- sociation, was established in 1873. Insurance cartels were conducted in two different ways, either as agreements or as recommendations. An agreement was a con- tract between at least two companies and was based on mutual interest. On the other hand, a recommendation consisted of a formal direc- tive to members of an association. Thus, the 22) Statens Offentliga Utredningar 1988:38, bilaga 7, Ägande och Inflytande i svenskt näringsliv, Allmänna Förlaget, Stockholm, pp. 17—22. 23) For futher information see, Gort, M., (1969) ”An Economic Disturbance Theory of Mergers”, The Quar- terly Journal of Economics, 83:4, pp. 639—42, see also Lewellen, Wilbur G., (1971), ”A Pure Financial Ration- ale for the Conglomerate Merger”, The Journal of Fi- nance, vol 26, pp. 522—5. 148 recommendations had tighter arrangements than the agreements. Of the 90 cartels, 73 can be defined as agreements. A rule of thumb is that the recommendations mostly referred to cartels in the life insurance branch of busi- ness, the recommendations amounting to some three fifths of all life insurance cartels. 3.1 Insurance cartels over time We here discuss when the cartels were found- ed and the division of cartels between life and non-life-insurance companies. To give the broadest view we have chosen to account for all cartels, even established before 1947. One- third of the cartels (28) pertained to life insur- ance, and the rest (62) to non-life insurance, with one cartel in 1948 including both life and non-life insurance.24 Cartels increased in number during the 1930s, reaching a peak at the end of the 1940s and the beginning of the 1950s. Non-life insurance companies dominated the estab- lishment of cartels, except in the years 1951— 55 and 1976—80. The life insurance compa- nies’ share of cartels was insignificant until 1950. The threat of nationalisation had en- couraged a looser and more informal co- operation, and made cartel agreements un- necessary. The life insurance branch had also been characterised by co-operation on a meso level, for example among joint-stock compa- nies on the one hand and mutual companies on the other. The life insurance branch had been very rigid, because of the high entry barriers caused by the restricted concessions to new life insurance companies. This gave the co- operation a long-range stability. During the 1940s the life insurance compa- nies were responsible for one-third of the cartels, but the number rose quickly and by the 1950s they accounted for three fifths of the total. Half of all the cartels made by the life insurance branch were in fact concentrated in the 1950s. As Figure 1 shows the number of cartels was especially large between 1941— 55. The non-life share of cartels decreased steadily from 1941 to 1960. The decline was principally due to the group of cancelled agree- ments, while valid cartels had a more stable development. The life insurance branch had a completely different pattern of cartels during the same period. The number of agreements rose grad- ually during the 1940s and from 1950 strong- ly expanded, decreasing finally at the end of the decade. Except for the years 1946—50, both valid and cancelled cartels within the life insurance branch followed this pattern. One reason for this distribution of cartels between life and non-life insurance compa- nies could be institutional. The life insurance branch probably had less reason to seek com- mon regulations between the companies, be- cause legislation had already accounted for common rules for this branch of business. The non-life business had a wider frame to work in, but coincidentally this increased the need for controlling the market. While the business environment changed slowly in the life insur- ance field, the non-life insurance cartels quick- ly became out of date, and a modified or thoroughly new cartel was required. This was probably the main reason for the large number of cancelled cartels in the non-life insurance market. 3.2 The division by sector The second approach is based on a more detailed level of analysis and concerns the different sectors in the non-life and life-insur- ance branches of business. The discussion here also deals with valid and cancelled car- tels. The non-life insurance business has been divided into six sectors, fire, traffic, under- 24) Financial Supervisory Authority, (FI), Insurance Supervisory’s archive, Registry of cartels (RC), Contract on constraint competition division on valid and can- celled, D 1 BA, D 1 BB, D1 BC and D 1 BD. RC F 57. 149 Figure 1 Number of cartels divided into five year periods, valid and cancelled cartels, 1873—1980 Year Quantity -1925 1926-1930 1931-1935 1936-1940 1941-1945 1946-1950 1951-1955 1956-1960 1961-1965 1966-1970 1971-1975 1976-1980 0 2 4 6 8 10 12 14 16 0000000000000000000000000000000000000000000000 0000000000000000000000000000000000000000000000 0000000000000000000000000000000000000000000000 0000000000000000000000000000000000000000000000 0000000000000000000000000000000000000000000000 0000000000000000000000000000000000000000000000 0000000000000000000000000000000000000000000000 0000000000000000000000000000000000000000000000 0000000000000000000000000000000000000000000000 0000000000000000000000000000000000000000000000 0000000000000000000000000000000000000000000000 0000000000000000000000000000000000000000000000 0000000000000000000000000000000000000000000000 0000000000000000000000000000000000000000000000 0000000000000000000000000000000000000000000000 0000000000000000000000000000000000000000000000 0000000000000000000000000000000000000000000000 0000000000000000000000000000000000000000000000 0000000000000000000000000000000000000000000000 0000000000000000000000000000000000000000000000 0000000000000000000000000000000000000000000000 0000000000000000000000000000000000000000000000 0000000000000000000000000000000000000000000000 0000000000000000000000000000000000000000000000 0000000000000000000000000000000000000000000000 0000000000000000000000000000000000000000000000 0000000000000000000000000000000000000000000000 0000000000000000000000000000000000000000000000 0000000000000000000000000000000000000000000000 0000000000000000000000000000000000000000000000 0000000000000000000000000000000000000000000000 00000000000000000000000000000000000000000000000 00000000000000000000000000000000000000000000000 00000000000000000000000000000000000000000000000 00000000000000000000000000000000000000000000000 00000000000000000000000000000000000000000000000 00000000000000000000000000000000000000000000000 00000000000000000000000000000000000000000000000 00000000000000000000000000000000000000000000000 00000000000000000000000000000000000000000000000 00000000000000000000000000000000000000000000000 00000000000000000000000000000000000000000000000 00000000000000000000000000000000000000000000000 00000000000000000000000000000000000000000000000 00000000000000000000000000000000000000000000000 00000000000000000000000000000000000000000000000 00000000000000000000000000000000000000000000000 00000000000000000000000000000000000000000000000 00000000000000000000000000000000000000000000000 0000000000000000000000000000000000000000000000 0000000000000000000000000000000000000000000000 0000000000000000000000000000000000000000000000 0000000000000000000000000000000000000000000000 0000000000000000000000000000000000000000000000 0000000000000000000000000000000000000000000000 0000000000000000000000000000000000000000000000 00000000000000000000000000000000000000000000000 00000000000000000000000000000000000000000000000 00000000000000000000000000000000000000000000000 00000000000000000000000000000000000000000000000 00000000000000000000000000000000000000000000000 00000000000000000000000000000000000000000000000 00000000000000000000000000000000000000000000000 00000000000000000000000000000000000000000000000 00000000000000000000000000000000000000000000000 00000000000000000000000000000000000000000000000 00000000000000000000000000000000000000000000000 00000000000000000000000000000000000000000000000 00000000000000000000000000000000000000000000000 0000000000000000000000000000000000000000000000 0000000000000000000000000000000000000000000000 0000000000000000000000000000000000000000000000 0000000000000000000000000000000000000000000000 0000000000000000000000000000000000000000000000 0000000000000000000000000000000000000000000000 0000000000000000000000000000000000000000000000 0000000000000000000000000000000000000000000000 0000000000000000000000000000000000000000000000 0000000000000000000000000000000000000000000000 0000000000000000000000000000000000000000000000 0000000000000000000000000000000000000000000000 0000000000000000000000000000000000000000000000 0000000000000000000000000000000000000000000000 0000000000000000000000000000000000000000000000 0000000000000000000000000000000000000000000000 0000000000000000000000000000000000000000000000 0000000000000000000000000000000000000000000000 00000000000000000000000000000000000000000000000 00000000000000000000000000000000000000000000000 00000000000000000000000000000000000000000000000 00000000000000000000000000000000000000000000000 00000000000000000000000000000000000000000000000 00000000000000000000000000000000000000000000000 00000000000000000000000000000000000000000000000 0000000000000000000000000000000000000000000000 0000000000000000000000000000000000000000000000 0000000000000000000000000000000000000000000000 0000000000000000000000000000000000000000000000 0000000000000000000000000000000000000000000000 0000000000000000000000000000000000000000000000 0000000000000000000000000000000000000000000000 0000000000000000000000000000000000000000000000 0000000000000000000000000000000000000000000000 0000000000000000000000000000000000000000000000 0000000000000000000000000000000000000000000000 0000000000000000000000000000000000000000000000 0000000000000000000000000000000000000000000000 0000000000000000000000000000000000000000000000 0000000000000000000000000000000000000000000000 0000000000000000000000000000000000000000000000 0000000000000000000000000000000000000000000000 0000000000000000000000000000000000000000000000 0000000000000000000000000000000000000000000000 0000000000000000000000000000000000000000000000 0000000000000000000000000000000000000000000000 0000000000000000000000000000000000000000000000 0000000000000000000000000000000000000000000000 0000000000000000000000000000000000000000000000 0000000000000000000000000000000000000000000000 0000000000000000000000000000000000000000000000 0000000000000000000000000000000000000000000000 0000000000000000000000000000000000000000000000 0000000000000000000000000000000000000000000000 0000000000000000000000000000000000000000000000 0000000000000000000000000000000000000000000000 0000000000000000000000000000000000000000000000 0000000000000000000000000000000000000000000000 0000000000000000000000000000000000000000000000 0000000000000000000000000000000000000000000000 0000000000000000000000000000000000000000000000 0000000000000000000000000000000000000000000000 0000000000000000000000000000000000000000000000 0000000000000000000000000000000000000000000000 0000000000000000000000000000000000000000000000 0000000000000000000000000000000000000000000000 0000000000000000000000000000000000000000000000 0000000000000000000000000000000000000000000000 0000000000000000000000000000000000000000000000 0000000000000000000000000000000000000000000000 0000000000000000000000000000000000000000000000 0000000000000000000000000000000000000000000000 0000000000000000000000000000000000000000000000 0000000000000000000000000000000000000000000000 0000000000000000000000000000000000000000000000 0000000000000000000000000000000000000000000000 0000000000000000000000000000000000000000000000 0000000000000000000000000000000000000000000000 0000000000000000000000000000000000000000000000 0000000000000000000000000000000000000000000000 0000000000000000000000000000000000000000000000 0000000000000000000000000000000000000000000000 0000000000000000000000000000000000000000000000 0000000000000000000000000000000000000000000000 0000000000000000000000000000000000000000000000 0000000000000000000000000000000000000000000000 0000000000000000000000000000000000000000000000 0000000000000000000000000000000000000000000000 0000000000000000000000000000000000000000000000 0000000000000000000000000000000000000000000000 0000000000000000000000000000000000000000000000 0000000000000000000000000000000000000000000000 Both Non-L. V. Non-L. C. Life V. Life C. 000000000000000000000000000000000000 000000000000000000000000000000000000 000000000000000000000000000000000000 000000000000000000000000000000000000 Note: V = Valid cartels; C = Cancelled cartels; Both = cartel F 57 of 1948, where all the insurance companies were involved. Source: The Insurance Supervisory’s registry of cartels. writing, accident, cattle and health. The larg- est sector was fire, followed by the traffic and accident sectors. Life insurance proved to be the most dom- inant sector with regard to valid cartels, while the cancelled agreements were particularly to be found within fire insurance. Health was the only sector that did not take part in any can- celled cartels. Within the fire, accident and traffic sectors, half of the agreements became cancelled. These three sectors had an abso- lutely dominating position regarding cancelled cartels, principally for two reasons. First, to- gether they comprised a major part of the total Swedish insurance industry, and were repre- sented in a large number of all cartels and consequently also in the cancelled agreements. Secondly, these sectors were characterised by a fluctuating market pattern of business (ex- cept for life insurance), and new insurance needs developed quickly. Hence, the need to modify the nature of the cartels was high, while it was easier to set up a new cartel, rather than change a current one. If we examine the cartels both over time and by sector, it is possible to see when different sectors have most resorted to cartels. Fire insurance proved to be the dominating sector to 1934, and during the years 1941—45 and 1963—68. But in the last period, 1969—80, this sector was not represented at all. During the first decades of the twentieth century the formation of cartels in the fire sector in- creased as a consequence of the percieved need to regulate this expanding industry. For instance, the fire insurance companies stand- 150 Figure 2 The division of cartels by sectors, valid and cancelled, 1873—1980 ardised the regulations for installing electric- ity. The companies also co-operated in creat- ing unified insurance conditions and to broad- en the preventive measure planning. The fire insurance sector expanded through the developing of new insurance needs and sub-branches. One example was through com- bined insurance (especially seen in home in- surance, involving fire, burglary, water main and liability insurance) that started up at the end of the 1930s. The increase of cartels within the fire sector in 1941—45 was owing to World War II and involved cartels that regulated the risks connected with the war. Traffic insurance had its peak during the two first periods, to 1934 and between 1935— 40. This was mainly caused by the establish- ment of compulsory motor car insurance in 1929. Thirty-three traffic insurance compa- nies were founded between 1925 and 1939 (in 1929 alone 28 companies were established), who required standardised regulations in the sector. First, the market was quite new and the need for common regulations was high. And secondly, an unregulated market could easily acquire a bad reputation and so lose the public confidence. At the end of the 1960s the number of cartels increased, probably because of the car industry’s expansion and the spread of cars to wider spectra of Swedish society. The underwriting insurance sector had a certain number of cartels until the mid 1940s, but then vanished until 1950, and only re- vived towards the end of the period under investigation. The development of the under- writing sector varied with international de- velopments. During World War I this activity expanded, due to the intensive submarine Sector Quantity Life Fire Traffic Accident Cattle Health Underwriting 0 2 4 6 8 10 12 14 16 18 20 Cancelled Valid Source: The Insurance Supervisory’s registry of cartels. 151 war. The inter-war period saw a decreased turnover, caused by the decrease in foreign trade, competition from foreign insurance companies, and the stable structure of the Swedish merchant navy.25 The underwriting companies probably sought higher co-ordina- tion to match the decreasing need of insur- ance. But World War II constituted another peak period for turnover and cartels in the underwriting branch of business. The accident insurance market expanded markedly until 1915. But after the legislation of 1916, concerning compulsory insurance for accidents at work, the market for accident insurance decreased.26 This sector then stag- nated until 1928, when 12 companies started to co-operate on common premiums and in- surance conditions. At the end of the 1920s and in the early 1930s activity then expanded, due to the establishment of new insurance fields, such as polio and school accident in- surance, which had not been included in the compulsory insurance.27 Such expansion was correlated to the estab- lishment of new cartels in the accident insur- ance sector. The development of new fields demanded a common ground for company activity, and from 1935 to 1945 six new cartels were formed. Thus, the peak periods of cartel formation within this sector was at the end of the 1930s and the beginning of the 1940s. The first part of the period under investiga- tion was dominated by fire and traffic insur- ance, while the 1950s were dominated by the life insurance sector. Fire, traffic and under- writing insurance (and, to a certain extent, life insurance) were in the later period the most important sectors in cartel formation. The main reason for the decrease in fire insurance cartels was probably that the combined insur- ance firms took market shares. In the 1950s the legislative and tax reforms stimulated a higher degree of cartelisation in the life insurance branch. The tax reform of 1953, for example, contributed to a consider- able speculation in capital insurance, the life insurance companies were forced to establish a temporary lending obstacle to counteract this tendency.28Another example was the agreement of 1951 that forced the life insur- ance companies to become more restrictive in granting combined capital and pension insur- ance, the aim being to steer the insured to a new pension insurance.29 During the 1960s and 1970s new market conditions and new, more specific branches and sectors of the insurance business increased the number of cartels in the traffic and under- writing areas. Fire insurance proved to be the most dom- inant sector in cartel creation. There are dif- ferent explanations for this. First, the laws that controlled the non-life insurance compa- nies were not formulated strictly enough to disallow cartels. Cartels became a substitute for weak legislation. Secondly, fire insurance encompassed such a large market share, with so many of different products, that there was a great need for standardisation. 3.3 The aims of cartels The purpose of a cartel means what, official- ly, in every cartel, was stated as the reason for forming the cartel. Study of the Swedish Pri- vate Insurance Supervisory’s registry of car- tels reveals, according to us, six different aims in forming a cartel. They are, the geographical division of markets, the division of markets by product, the standardisation of premiums and conditions, the standardisation of terms, 28) The Insurance Supervisory’s registry of cartels, no. F 65. 29) Ibid., no. F 51. 25) Larsson, M., op. cit., p. 46. 26) Städernas försäkringsbolag minnesskrift, 1953, Stockholm, p. 96 f. 27) Städernas försäkringsbolag, op cit., p. 101 f. Larsson, M., op. cit., p. 47. 152 Figure 3 Cartels divided by sector and over time, valid and cancelled, 1873—1980 Year Quantity -1934 1935-1940 1941-1945 1946-1950 1951-1954 1955-1962 1963-1968 1969-1980 0 2 4 6 8 10 12 14 00000000000000000000000000000000000000000000000000000000000000000 00000000000000000000000000000000000000000000000000000000000000000 00000000000000000000000000000000000000000000000000000000000000000 00000000000000000000000000000000000000000000000000000000000000000 00000000000000000000000000000000000000000000000000000000000000000 00000000000000000000000000000000000000000000000000000000000000000 00000000000000000000000000000000000000000000000000000000000000000 00000000000000000000000000000000000000000000000000000000000000000 00000000000000000000000000000000000000000000000000000000000000000 00000000000000000000000000000000000000000000000000000000000000000 00000000000000000000000000000000000000000000000000000000000000000 00000000000000000000000000000000000000000000000000000000000000000 00000000000000000000000000000000000000000000000000000000000000000 00000000000000000000000000000000000000000000000000000000000000000 00000000000000000000000000000000000000000000000000000000000000000 00000000000000000000000000000000000000000000000000000000000000000 00000000000000000000000000000000000000000000000000000000000000000 00000000000000000000000000000000000000000000000000000000000000000 00000000000000000000000000000000000000000000000000000000000000000 00000000000000000000000000000000000000000000000000000000000000000 00000000000000000000000000000000000000000000000000000000000000000 00000000000000000000000000000000000000000000000000000000000000000 00000000000000000000000000000000000000000000000000000000000000000 00000000000000000000000000000000000000000000000000000000000000000 00000000000000000000000000000000000000000000000000000000000000000 00000000000000000000000000000000000000000000000000000000000000000 Health Cattle Accident Underwriting Traffic Fire Life 0000000000000000000000000000000000 0000000000000000000000000000000000 0000000000000000000000000000000000 Source: The Insurance Supervisory’s registry of cartels. the standardisation of acquisition activity, and other forms of distribution risk. In the geographical division of markets (GDM) the companies divided the market spatially. The companies virtually claimed their territories. This reason for forming a cartel was the least frequent, accounting for only three per cent of all agreements (see Figure 4). The division of markets by product (DMP) implies that the members of a cartel divided the right to sell certain products among themselves. In this case the companies gained a right to sell without the competition of other members.30 The division of markets by prod- uct accounts for a somewhat larger share of cartels than GDM, but was still not consider- able, altogether one-tenth of all agreements. The standardisation of premiums and con- ditions (SPC) concerns companies that had agreed upon common pricing and duplicate conditions for certain products. This purpose for forming cartels was the most dominant and constituted almost half of all agreements. Standardisation of terms (ST) dealt with fields that did not directly affect the price. This concerned additional regulations, definitions of risks, reactions to legislative changes and the making of common institutional and or- ganisational regulations. Standardisation of acquisition activity (SAA) handled the re- cruitment of the companies’ customers. The insurance companies had a number of times- fixed commissions and other remunerations for the sellers. Accounting for seven per cent, 30) A good example is cartel agreement no. F 39 where the health insurance company, Eir, received all short- term health insurances, and the life companies got the long-term. 153 SAA did not make up a large share of the cartel agreements. Other distribution of risk (ODR) comprised, first, agreements that created or directed in- surance pools. When insurance companies undertook insurance projects affiliated to large risks, or projects that would in case of com- pensation being due be a great burden to them, the companies joined together in a risk-dis- tributing cartel. Here were to be found agree- ments like the Swedish nuclear insurance pool, but also secondly non-pool cartels such as consortium agreements providing finan- cial back-up to introduce new products.31 Other distribution of risk comprised more than one-tenth of all cartel agreements. Figure 4 indicates that the standardisation of premiums and conditions predominated in the first half of the twentieth century. The 1920s saw the highest establishment rate of insurance companies.32 This period should therefore also have been the phase in which the strongest competition existed. If the insur- ance companies wanted to develop a sound business they had to avoid competing over pricing. Some representatives of the Social Figure 4 Cartels divided according to aim of the agreements, 1873—1980 Aim Quantity GDM DMP SPC ST SAA ODR 0 5 10 15 20 25 30 35 40 Valid Cancelled Note: GMD = geographical division of markets; DMP = The division of markets by product; SPC = standardisation of premiums and conditions; ST = standardisation of terms; SAA = standard- isation of acquisition activity; and ODR = other distribution of risk. Source: The Insurance Supervisory’s registry of cartels. 31) The Insurance Supervisory’s registry of cartels, nos. F 58, F 59 and F 83. 32) Larsson, M., op. cit, p. 32. 154 Democratic Party argued that the large extent of insurance with varying conditions and pre- miums made the market difficult to survey. Hence, there was a political interest in stand- ardising premiums and conditions. At the beginning of the 1950s there was a shift in predominance between the SPC and the standardisation of terms, most likely brought about by legislative changes. The new insurance law of 1948 brought about more stringent establishment regulations. This implied a diminishing competition between older and more newly established businesses, which thereby decreased the need for a stand- ardisation of premiums and conditions. Dur- ing this decade the Swedish insurance compa- nies had become considerable financial insti- tutions, which attracted the attention of the state in several ways. Their investment poli- cies became an object of very stringent regu- lation by the government, as well as by the companies themselves. The National Associ- ation of Swedish Insurance Companies rec- ommended that their members should avoid investment policies that, as a result of compe- tition, would depress interest rates.33 The government considered the company assets as perfect capital funds that could swallow up a good part of treasury bond issues. These treasury bonds above all mobilised capital which financed the so called ”Miljonprojek- tet”, that sought to construct one million fam- ily apartments during the 1960s. Other distribution of risk cartels were main- ly concentrated in the latter part of the 1930s and 1970s. During the 1930s ODR-cartels concerned accident insurance, and they were carried out as pools. In the 1970s the aim of distributing risk included more of a building financial back-up by consortium.34 In the periods 1955—1962 and 1969—1980 geo- graphical division of market and division of market by product were both a little more significant compared to other periods. A pos- sible explanation might be the changing mar- kets of the 1950s which saw an increased efficiency in pricing and a higher concentra- tion of companies by mergers. A more oli- gopolistic market — the result of this devel- opment — competed with other means than premiums. A typical feature was that compa- nies competed for market shares.35 Standard- isation of acquisition activity was equally spread over the whole period, which meant a continuous overhaul of the conditions for sellers and other agents. 3.4 Number of members per agreement The structure of the Swedish insurance mar- ket changed during the 1950s and 1960s due to a widespread wave of mergers.36 The growth of the market and the 1948 legislation stimu- lated concentration by promoting and facili- tating the acquisition of insurance enterpris- es. Between 1940 and 1980 the number of national insurance companies decreased by more than 60 per cent, from 144 to 55.37 This amalgamation of insurance companies im- plied a need for financially strong insurance companies. At the same time the insurance companies moved their field of business more towards combining insurance, so reducing the market for specialised insurance. The in- surance market became increasingly homo- geneous.38 The developing number of cartels displayed large changes over time, and was clearly relat- ed to the diminution in the number of insur- ance companies. Until 1937, agreements with over 35 members (group V) and 11—19 (group 33) The Insurance Supervisory’s registry of cartels, no. F 34. 34) The Insurance Supervisory’s registry of cartels, nos. F 72, F 83 and F 84. 35) Skogh, G., & Samuelsson,P., op. cit., p. 49. 36) Englund, K., op. cit., p. 25 ff. 37) SOS, Enskilda försäkringsanstalter, 1940 and 1980. 38) Larsson,M., op. cit., p. 4 f. 155 Figure 5 Cartel agreements divided by their aims and over time, 1873—1980 Year Quantity -1934 1935-1940 1941-1945 1946-1950 1951-1954 1955-1962 1963-1968 1969-1980 0 2 4 6 8 10 12 14 00000000000000000000000000000000000000000000000000000000000000000000000 00000000000000000000000000000000000000000000000000000000000000000000000 00000000000000000000000000000000000000000000000000000000000000000000000 00000000000000000000000000000000000000000000000000000000000000000000000 00000000000000000000000000000000000000000000000000000000000000000000000 00000000000000000000000000000000000000000000000000000000000000000000000 00000000000000000000000000000000000000000000000000000000000000000000000 00000000000000000000000000000000000000000000000000000000000000000000000 000000000000000000000000000000000000000000000000000000000000000000000000 000000000000000000000000000000000000000000000000000000000000000000000000 000000000000000000000000000000000000000000000000000000000000000000000000 000000000000000000000000000000000000000000000000000000000000000000000000 000000000000000000000000000000000000000000000000000000000000000000000000 000000000000000000000000000000000000000000000000000000000000000000000000 000000000000000000000000000000000000000000000000000000000000000000000000 000000000000000000000000000000000000000000000000000000000000000000000000 000000000000000000000000000000000000000000000000000000000000000000000000 000000000000000000000000000000000000000000000000000000000000000000000000 000000000000000000000000000000000000000000000000000000000000000000000000 000000000000000000000000000000000000000000000000000000000000000000000000 000000000000000000000000000000000000000000000000000000000000000000000000 000000000000000000000000000000000000000000000000000000000000000000000000 000000000000000000000000000000000000000000000000000000000000000000000000 000000000000000000000000000000000000000000000000000000000000000000000000 000000000000000000000000000000000000000000000000000000000000000000000000 000000000000000000000000000000000000000000000000000000000000000000000000 000000000000000000000000000000000000000000000000000000000000000000000000 000000000000000000000000000000000000000000000000000000000000000000000000 000000000000000000000000000000000000000000000000000000000000000000000000 000000000000000000000000000000000000000000000000000000000000000000000000 000000000000000000000000000000000000000000000000000000000000000000000000 000000000000000000000000000000000000000000000000000000000000000000000000 000000000000000000000000000000000000000000000000000000000000000000000000 000000000000000000000000000000000000000000000000000000000000000000000000 000000000000000000000000000000000000000000000000000000000000000000000000 000000000000000000000000000000000000000000000000000000000000000000000000 000000000000000000000000000000000000000000000000000000000000000000000000 000000000000000000000000000000000000000000000000000000000000000000000000 000000000000000000000000000000000000000000000000000000000000000000000000 000000000000000000000000000000000000000000000000000000000000000000000000 000000000000000000000000000000000000000000000000000000000000000000000000 000000000000000000000000000000000000000000000000000000000000000000000000 GDM DMP SPC ST SAA ODR 00000000000000000000000000000000000 00000000000000000000000000000000000 00000000000000000000000000000000000 00000000000000000000000000000000000 Note: GMD = geographical division of markets; DMP = The division of markets by product; SPC = standardisation of premiums and conditions; ST = standardisation of terms; SAA = standard- isation of acquisition activity; and ODR = other distribution of risks Source: The Insurance Supervisory’s registry of cartels. III) predominated, see Figure 6. A hetero- geneous insurance market with a large number of active companies explained a high number of members. Regarding the agreements that dealt with life insurance we see the companies associated with the Swedish Life Insurance Companies Association, composed of 17 com- panies.39 The growing number of agreements with 11—19 members after 1950 coincided with an increased number of agreements within life insurance, see Figures 1 and 6. As a result of the concentration of the insurance industry, after 1965, the cartel agreements included fewer and fewer members. As company con- centration increased, the tariff associations were left with little real influence. The major insurance combines had a key role in forming cartels. Skandia were involved in almost 40 per cent of all agreements, whereas Göta participated in one-third.40 In addition, RKA as well as Trygg and Hansa each took part in one-fifth of the cartels. Among the Consumer Cooperative insurance companies there had always been a strongly negative official attitude towards agreements aimed at 39) The Insurance Supervisory’s registry of cartels. 40) The Skandia combine included from 1963 Svea, Skåne, Thule and Öresund. 156 restricting competition.41 It is thus rather sur- prising to find that Folksam (previously Folket and Samarbete) participated in 26 different competition-restricting agreements, nearly 30 per cent of the total number of cartels. It is also documented in the registry of cartels that Folksam not formally, but implicitly, fol- lowed the different general pattern of other insurance cartels. This fact makes the Con- sumer Co-operative insurance companies the companies that most frequently directed their activity according to different principles of cartel agreements. Even foreign insurance companies that had been active in the Swedish insurance market participated in forming cartels. There was a particular tendency that they participated in cancelled agreements: two fifths of these agree- ments included foreign companies, whereas for the corresponding part of valid agree- ments it was 25 per cent. Altogether this means that foreign insurance companies took part in one-third of all cartel agreements. The foreign companies were especially involved in traffic insurance. Zürich, Winterthur, Na- tional and Motor Union participated together 41) Framtiden minnesskrift op. cit., p. 112 f. Note: Group I = 2-5 members; group II = 6 -10; group III = 11-19; group IV = 20-35; and group V = over 35 members. Source: The Insurance Supervisory’s registry of cartels. Figure 6 Cartel agreements divided by the number of members over time, 1873—1980 Year Quantity -1937 1938-1945 1946-1953 1954-1965 1966-1980 0 2 4 6 8 10 12 14 16 18 20 000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000 000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000 000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000 000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000 000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000 000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000 000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000 000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000 000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000 000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000 000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000 000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000 000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000 000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000 000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000 000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000 000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000 000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000 000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000 000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000 000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000 000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000 000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000 000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000 000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000 000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000 000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000 000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000 000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000 000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000 000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000 000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000 000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000 000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000 000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000 000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000 000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000 000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000 000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000 000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000 000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000 000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000 000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000 000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000 000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000 000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000 000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000 000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000 000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000 000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000 000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000 000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000 000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000 000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000 000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000 000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000 000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000 000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000 000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000 000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000 000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000 000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000 000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000 000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000 000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000 000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000 000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000 000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000 000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000 000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000 000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000 000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000 000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000 000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000 000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000 000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000 000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000 000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000 000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000 000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000 000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000 000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000 000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000 000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000 000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000 000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000 000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000 000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000 000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000 000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000 000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000 000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000 000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000 000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000 000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000 000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000 000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000 000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000 000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000 000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000 000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000 I II III IV V 0000000000000000000000000000000000 0000000000000000000000000000000000 0000000000000000000000000000000000 157 in three such agreements. The two first-men- tioned also took part in The Nordic Pool for Air Traffic Insurance (Nordiska poolen för luftfartsförsäkring), which seems to have been an interest group or some sort of network, between the foreign companies. It is not sur- prising that there was more solidarity be- tween two foreign companies operating on comparable terms than between a foreign and a Swedish company. 4.1 The analysis of cartels — a model Here we aim to relate merger theories with the different ways of structuring cartels, partly by using information from section 2 and 3 above. Mergers have both positive and negative ef- fects for companies. The positive effects may include rising profits, a larger control of a market, or an expansion into new ones. The negative effects for a company could be a diminishing freedom of action or the need to dismiss people. A compromise means to achieve positive profitability without under- taking a merger and without expensive inter- nal expansion could be to form a cartel. Car- tels should not be interpreted as a substitute for mergers, in that the latter may be a result of cartels that did not function properly with respect to their aims. Considering the gener- ally increasing degree of concentration in the Swedish industrial industry, it would seem more likely that mergers occurred in a later phase. It will be convenient to discuss the four groups of rational theories of mergers, and systematically analyse their relevance, from a cartel perspective. We see a clear connection between theories of real profitability and car- tels. Generally, cartels within the Swedish insurance industry were associated with effi- ciency gains and lowered information and transaction costs. Some cartels expressly act- ed to create institutions for exchanging infor- mation.42 And clearly others aimed at strength- ening the market positions of the members, as a large and strong cartel could operate in a way similar to a monopoly company. It was also possible to achieve economies of scale by forming a certain type of cartel called a trust, where one company takes care of the business administration of all the members. Factors of a political institutional nature had a great influence on the propensity to build cartels, especially the threat of nationalisation during the 1930s when the frequency of cartel forma- tion increased noticeably compared to the previous period. There are, as was previously touched upon, problems in connecting theories of financial profitability and cartels, as the former directly relate to oscillations in stock prices and the propensity to mergers. We can, however, dis- cern clear links between theories of risk distri- bution and reduction, and cartels. We refer principally to pools, which basically aimed to distribute risks. Some cartels did not even operate on the basis of a pool but still strive to distribute risks. There is empirical evidence of cartel agreements that resulted in a joint reinsurance system, or even a sort of consor- tium agreement, which with a joint base of indemnity provided a back-up for a new prod- uct.43 A cartel agreement may even interact in accordance with theories of growth maximi- sation. Maximising growth must not be equat- ed with the acquisition of stocks but rather be viewed as a higher concentration of capital that allows a greater power to keep potential new companies out of the market. In this sense growth maximisation resembles a consolidat- ed market position and hence may link with theories of real profitability. Naturally, all mergers amount to a convey- ance of ownership, which usually consists of 42) The Insurance Supervisory’s registry of cartels, no. F 28a. 43) Ibid., nos. F 70 and F 83. 158 shares. Cartels on the other hand do not in- clude any transfer of shares but rather of confidence. Thus shares have to be trans- formed into something else, which means that those merger theories that are strictly based on the valuation of shares not can be applied to cartels. Theories that can apply to cartels are theories of real profitability, of the distri- bution and reduction of risk, and to some extent theories of growth maximisation. As the latter may correspond to theories of real profitability we have two groups of theories that we can use in building a model to interpret cartels: theories of real profitability, and the- ories of the distribution and reduction of risk. By studying the Swedish Private Insurance Supervisory’s registry of cartels we found six different ways cartels were created in the Swedish insurance industry: the geographical division of markets; the division of markets by product; standardisation of premiums and conditions; standardisation of terms; stand- ardisation of acquisition activity; and other distribution of risk. The Swedish insurance industry established cartels either by division, standardisation or distribution of risks. These ways of establishing cartels consti- tute the first of six different levels of abstrac- tion in our model of explaining cartels. The other five levels are, in turn, practical meas- ures, general measures, general aims, theoret- ical approach and, finally, the level of funda- mental assumption. Figure 7 shows that the six different ways of defining cartels group into four spheres of practical measures. Both the geographical division of markets and divi- sion by product are a division of markets, or what we call a concession of territory, where- as the three items of standardisation split in the way they affect pricing. Standardisation of premiums and conditions rules pricing in a more direct manner than standardisation of terms and acquisition activity. Nevertheless, the latter affect the price by changing the administrative costs. Other distribution of risk has its practical measures in making consorti- um agreements or forming pools. Companies generally measure the conces- sion of territory in securing their market shares, which represents the next level of abstraction. Direct and indirect pricing come together in what we call RCP-advantages. This is short hand for the advantages in revenue, cost and price that companies may achieve by forming a cartel. Similarly, the aim of consortium agreements and the forming of pools general- ly can be measured in the distribution of risk. The securing of market share and RCP-ad- vantages have in common the general aim of achieving an improved market position. Com- panies may not necessarily want to form a cartel in order to improve their market posi- tions. The advantage of stability is another general aim of forming cartels, especially for insurance companies where some undertak- ings are associated with large risks. Now we can link the general aims of form- ing cartels in the Swedish insurance industry with the two theoretical approaches based upon the theories of mergers: theories of real profitability and theories of distribution and reduction of risk. These theories build on the fundamental assumption that mergers are per- formed on grounds of rational behaviour and to seek the gains of co-ordination. The aims of the cartels are that companies in one way or another set their minds on getting something positive or advantageous out of the co-ordina- tion. If a company participates in a cartel because of pressure, this implies a somewhat less negative position on the risk of being forced out of the market. If the option is, being forced out of the market or joining a cartel, we assume that a company obviously chooses the latter. We assume that all cartels imply gains of co-ordination and offer a way of co-opera- tion that every company should be eager to join. According to what has been said above we should have had nothing but cartels in the 159 Swedish insurance industry. There are, how- ever, what we call contract barriers that may counteract the propensity towards cartels. The different types of organisation existing in companies or personal and social networks can restrict the propensity to build cartels. Other barriers could be the increased trans- action costs or establishment costs, or the possibility that a company wants to keep a market for itself. In this paper we have touched upon and discussed things that constitute con- tract barriers: a) Personal and social networks imply more informal co-operation than is the case with cartels. b) The big principal controversy on the meso level between stock companies and mutual companies contributes to difficulties in co- operation. c) The distrust and suspicion that character- ised older-established companies towards the newly-established results in evident difficul- ties of co-operation. d) The possibility of merging provides the opportunity for companies to go directly from informal co-operation to mergers. e) Legislation is a counteracting force as in many respects it meticulously directs types of insurance and the business compass of insur- ance companies. The policy of the Swedish Private Insurance Supervisory was also an institutional means that worked for more for- mal regulations, and thereby decreasing the need for cartels in the Swedish insurance market. Figure 7 The relation between cartels and the theories of mergers Note: 1. level of abstraction = performing; 2. = practical measures; 3. = general measures; 4. = general aims; 5. = theoretical approach; and 6.= fundamental assumption.