Bank affiliation influence on life insurers’ performance before and after the financial crisis

IntroductionBoth in the US and in most European countries actuarial risk coverage is an exclusive domain of insurance companies. However, while the law forbids banks to perform this activity, it allows them to sell insurance products. Therefore, starting from the 1970s, European banks carved a role in the insurance business by distributing policies through their branch networks and creating the "bancassurance" phenomenon [Hoschka 1994].As documented in the following section, at least before the big financial crisis, European banks operating in the life insurance business through controlled insurance companies favoured the offer of "financial" policies - i.e. insurance products with a limited actuarial risk protection content and whose main objective is producing returns by investing premiums either in index linked bonds or in mutual funds. On the other hand, players that were not bank affiliated competed in the traditional area of demographic and financial risk insurance, mostly selling "with profit policies".The positive trend in financial markets' prices, lasting from the beginning of the 2000s until 2007, benefited bank affiliated insurers because it magnified the returns to holders of the products they marketed. New customers were enticed by the results coming from financial policies, enabling insurers to pocket part of considerable management fees [Fiordelisi & Ricci 2012]. On top of that, bank affiliated insurers could take advantage of lower distribution costs, due to the availability of their banking partners' distribution networks.The financial crisis changed all this, making products sold by non-bank affiliated insurers attractive again. Savers' increased risk aversion shifted demand towards safer investing, giving traditional insurers the chance to fight back banks controlled ones by offering products earning nice profits for them while being easy to sell by their agents.Has traditional insurers' performance benefited from this change, in spite of higher distribution costs? More in general, how relevant is being "bank affiliated" for succeeding in the life insurance business?In this article, we try to answer the above questions by analysing the performances of Italian life insurers from 2003 to 2013. We examine the impact that bank affiliation, product mix composition and distribution costs had on their economic results and the change that the big financial crisis determined on the influence that the above variables have on profitability.To our knowledge, this is the first systematic attempt at examining all the above factors' contribution to life insurers' performance. Moreover, we consid- ered their interaction with macroeconomic conditions, adding value and generalization extent to our analysis.In order to shed light on the above topic, Italy is an interesting case from various points of view. First, banks accession to the life insurance sector through the acquisition of ownership stakes in established players has been particularly swift due to the virtual absence of cooperative insurers. The "bancassurance" phenomenon had been developing for a number of years before 2000, allowing us to analyse it when the presence of banks in the life insurance market was full-fledged.Second, during the timeframe of our analysis, Italy never exited the great crisis that hit Europe in 2008. Except for 2010, from 2008 to 2013 Italian GDP growth was zero or negative. While this is unfortunate for obvious reasons, it means that both bank affiliated and independent insurers had to adapt to economic recession because it became soon clear that it was not a temporary occurrence. We could therefore observe how insurance companies reacted to it.Third, the Italian life insurance market is the fourth in Europe for volume of premiums written, making it meritorious of an analysis of its own.Limiting our analysis to one country allowed us to be sure that the same legal and accounting framework applies to the entire sample, avoiding data manipulation that could bias our results. …