Commercial banking, insurance and economic growth in Sweden between 1830 and 1998

Abstract

We examine empirically the dynamic historical relation between commercial bank lending, insurance and economic (income) growth in Sweden using time-series data from 1830 to 1998 and performing tests for Granger causality. Because of the non-stationary nature of the time series examined the procedure of Toda and Yamamoto (1995) is used. Our results, which have accounted for possible regime changes due to different exchange rate mechanisms over time, indicate that insurance has Granger-caused economic growth and bank lending. Therefore, we conclude that insurance is an important prerequisite for stimulating economic growth and that this could have important implications for contemporary developing economies

Keywords:

• banking
• insurance
• economic growth
• history
• Sweden

Acknowledgements

The authors thank Anna Andersson, Kurt Brännäs, Steve Diacon, Ola Grytten, Philip Hardwick, Kul Luintel, Philip Molyneux, Chris O'Brien, Erik Sørensen, Damian Ward and Hong Zou for their comments and assistance. The paper also benefited from the suggestions of commentators at seminars held at the Department of Economics, Umeå University, Sweden, the School of Business and Economics, Swansea University, UK, the Cardiff Business School, Cardiff University, UK and the Wharton School, University of Pennsylvania, USA. In addition, the financial support of the Handelsbanken-Wallander and Sveriges Riksbank Research Foundations is greatly appreciated. However, the normal disclaimer applies.

Notes

Gårdlund (1947) suggests that the early 1830s mark the approximate beginning of Sweden's modern industrialisation period. The 1830s also witnessed greater monetary stability (e.g. through the re-establishment of a silver standard in 1834 that had been abandoned in 1809) and the formation of the first commercial banks. In line with international developments at the time, Sweden changed from a silver standard to a gold standard in 1874 (Ögren 2006).

Financial intermediation is essentially concerned with the transfer of resources from savers (surplus units) to users of funds (deficit units) in the economy (Boot and Thakor 1997a, 693-4). Skogh (1991, 66) contends that the differences between banks and insurance companies as financial intermediaries lie in their respective specialisations in risks and customers. For example, banks have traditionally specialised in personal and commercial financial (e.g. credit) risks and insurance companies have specialised in personal and commercial accident (e.g. liability) risks.

This attribute does not necessarily mean that our results do not have wider appeal. For example, our results could be generalised to the financial services sector of countries that have experienced a similar economic history and institutional legal-regulatory framework to Sweden – for example, other Scandinavian countries.

Sweden has long-established national accounting and economic data archives (e.g. see Lindahl, Dahlgren, and Kock 1937). Levine (1999, 22) also reports that institutional information (e.g. accounting) systems can also be important prerequisites for the development of financial intermediation in an economy. The fact that Sweden has, by international standards, good economic data archives could to some degree contribute to the development of the domestic financial infrastructure and promote national economic growth.

Aspects of Sweden's welfare state (e.g. the public pensions system and work-place accident compensation) date back to the early twentieth century but it became particularly omnipresent in the domestic economy after the Second World War (e.g. see Magnusson 1997).

However, some local fire and marine insurance companies in Sweden can trace their origins back to the eighteenth century. For example, the present day Stockholm Fire Insurance Office (Brandförsäkringskontor) was founded in 1746 (see Lindmark, Andersson, and Adams 2006).

Ögren (2006, 88–9) reports that the right of Enskilda banks to issue notes was a key factor in Sweden's economic development in the nineteenth century as it ‘solved the problem of creating the necessary conditions for a financial system in a poor, under-monetized, and geographically dispersed country’. This helped to enhance liquidity and financial deepening in the economy. Ögren (2006, 89) also observes that the Enskilda banks were important in the development of deposit banking in Sweden.

Additionally, other European countries, such as Germany and the UK, have experienced a growth of national social insurance since 1945 without affecting the growth of private insurance provision.

In Sweden the Wallenberg family have for much of the twentieth century played an important role in the banking sector (more recently through their controlling interest in the universal Swedish bank, SEB). In fact, it is primarily through their banking activities that the Wallenberg family have maintained ‘pyramidal’ control of other sectors of the domestic economy. Morck, Wolfenzon, and Yeung (2005, 688–9) report that in many economies outside of the UK and US such arrangements give wealthy (oligarchal) families (and other concentrated ownership interests) in banks significant influence on the direction and pace of national economic growth; they also help to bind banks into long-term financing relationships with companies. The membership of the borrowing firm and bank to the same ‘pyramidal’ group does, however, help to mitigate market informational asymmetry problems in the raising of finance and facilitate long-term investment planning.

The underinvestment problem arises because under corporate limited liability rules shareholders have a default option that allows them to choose not to reinvest in fixed assets damaged as a result of/by severe mishap (e.g. due to fire). This is because the economic benefits of reinvestment could be greater for other claimants – for example, debtholders that hold fixed claims over the assets as collateral for loans granted. Insurance resolves this issue by indemnifying shareholders against the cost of asset reinstatement (see MacMinn 1987).

The supply (and price) of insurance in an economy in any particular period is likely to be affected by the so-called ‘underwriting cycle’ that functions in the insurance market, which is itself highly dependent on domestic and international macroeconomic factors (e.g. interest rates) and the supply of reinsurance in world markets (Doherty and Garven 1995). When the cycle contracts, premium rates rise and underwriting capacity falls thereby increasing demand among insurers for reinsurance. However, lack of data precluded us from testing for the effects of underwriting cycles on the supply of insurance in the Swedish market over the 169 years covered by this study.

The Swedish National Statistics Office derives annual population estimates from published census data (carried out every 10 years or so). This source of data extends back to the 1930s when the first annual national income figures for Sweden covering the period 1861–1930 were derived and first published in Lindahl, Dahlgren, and Kock (1937).

We aggregate non-life and life annual insurance premiums as risk transfer, indemnification and financial intermediation are functions common to both life and non-life insurance (see also Ward and Zurbruegg 2000, 490). We also assume that in the long term the public demand for life and non-life insurance will exhibit similar income elasticity and that both sectors of the insurance market emerged more or less concomitantly in Sweden. Some support for these maintained assumptions can be found in Hägg (1998) who emphasises the enthusiasm of the growing (and increasingly affluent) Swedish middle class of the late eighteenth/early nineteenth centuries for life and pensions (widows and orphans funds) as well as property insurance protection. Furthermore, aggregating non-life and life insurance activity can be justified as both non-life and life insurers perform an important intermediary in the domestic economy as institutional investors. Unfortunately, analysis of the effect of price elasticity by line of insurance (e.g. motor, property, liability and so on) could not be carried out in the current study due to the absence of publicly available data.

In 1905 the union between Sweden and Norway was dissolved; prior to this date, however, Norway and Sweden had separate governments and parliamentary systems, with their own laws and regulatory structures. Therefore, our data set excludes entirely Norwegian data prior to 1905.

Insurance penetration is sometimes viewed in the literature as reflecting domestic demand (consumption), but in equilibrium it could equally reflect the supply (sales) of insurance in an economy. Therefore, insurance penetration can be viewed as a measure of insurance activity reflecting both the demand and supply in an economy.

Initial tests using dummy variables to examine the possible effects of regulatory changes on the demand and supply of insurance in Sweden (e.g. after the Insurance Regulations of 1903 and 1948) were found not to be statistically significant and are thus excluded from the full analysis in the interests of producing more parsimonious models.

Also, though not statistically significant, test 3 of Table 3 shows that the relation between ECON and INS is just outside the 10% critical region, suggesting that economic activity could nevertheless be a driver of insurance demand over the period of analysis.