Why did shareholder liability disappear?

Limited liability is pervasive in modern financial systems. It can encourage investment, but it also has the potential to incentivize risk taking. Asymmetric payoffs between profits and losses can encourage financial firms to pursue more speculative projects, knowing that their shareholders will gain the full benefit of success, but a limited loss from failure. Its role in exacerbating risk taking may have played a role in the Global Financial Crisis of 2008.

Historically, banks in the United States were required to have double liability, meaning that shareholders faced additional costs beyond their initial investment if the bank became insolvent, and recent scholarship suggests that such banks were less likely to fail.

In the United Kingdom, most banks and insurance firms voluntarily chose to have extended liability, where shareholders could potentially be exposed to payouts which were much greater than the amount that they had invested. However, this is no longer the case. This raises the question: Why did shareholder liability disappear?

We address this by looking at the use of shareholder liability by British insurance companies over the long run. Until 1862, the only way for an insurance company in the UK to have limited liability was to have been incorporated via a Royal Charter or an Act of Parliament. Two famous examples of companies incorporated by an Act of Parliament are the London Assurance and Royal Exchange Assurance, which were a late addition to the Bubble Act of 1720. Because companies with tradeable shares were illegal under the Bubble Act, other insurance businesses were formed as unincorporated companies, which were clever legal workarounds utilising trust law. These insurance companies contracted in their corporate constitutions to create limited liability, but under the common law, these unincorporated companies were de jure and de facto unlimited. The implications of this for insurance companies was that shareholders could limit their liability inter se, but not to third parties such as policyholders. From a practical point of view, however, debates about the legality of limited liability clauses were immaterial because unincorporated insurance companies had such large amounts of uncalled capital, i.e., capital which could be called up from shareholders by directors at any time, or by policyholders and other creditors in the event of bankruptcy.

With the passage of the 1862 Companies Act, insurance companies could limit their liability by simply registering under this legislation, which most of them did. However, as can be seen from Table 1, they generally retained large amounts of uncalled capital. In 1880, 97% of insurance companies had this uncalled capital, but by 1930 this had fallen to 74%, and just 15% in 1965. By 1975 there were no UK insurance companies with shareholder liability. We also see that in 1880 the average shareholder faced over quintuple liability, meaning that for every £1 of equity that had been invested, the shareholders were liable for an additional £5.67. By 1930, this had fallen to close to double liability and by 1965 this ratio was close to zero.

Table 1: Shareholder liability of UK insurance companies

Source: Bogle, Campbell, Coyle and Turner (2024)

So, why did shareholder liability disappear? One possibility is that regulation and government-provided policyholder protection meant that shareholder liability was no longer required. However, there were no regulatory changes during or even immediately after the period when shareholder liability disappeared. We also find that nearly all companies removed their shareholder liability many decades before the passage of the Policyholders Protection Act in 1975.

The second possibility which we consider is that shareholder liability was removed because it was de facto limited. One plausible reason for this irrelevance is that there may have been nothing to prevent shares being sold to individuals who would have been unable to pay calls on them because insurance company shares were freely transferable. Thus, the removal of shareholder liability was simply acknowledging what was already a reality. Using detailed archival data on shareholder wealth for a large British insurance company which had quadruple liability, we find that shareholders had more than enough wealth to cover potential calls on their shares. Thus, shareholder liability was not de facto limited.

The third possibility is that the risks associated with extended liability led to a higher cost of capital, which incentivized companies to remove uncalled capital. Using a century of monthly share prices for insurance companies, we find that this was the case. Companies may have preferred to move away from this system, but they were reluctant to explicitly remove uncalled capital as it would have reduced the confidence of their customers that their policies would be honored.

However, the buffer provided by shareholder liability became less important as companies increased in size and scope, which they did through organic growth and mergers. Using hand-collected financial statement data, we find that size was a very important determinant of the level of shareholder liability that insurance companies possessed, and that increases in size were associated with decreases in shareholder liability.

Our findings ultimately suggest that shareholder liability disappeared in the British insurance sector because of the increasing size of insurance firms. The same could plausibly be argued for British banking, with shareholder liability disappearing after banks had merged and grown. This contrasts with the experience of the U.S. banking system, where the branching restrictions prevented U.S. banks from the growth necessary to remove extended liability. We would speculate that it was more likely the emergence of deposit insurance that led to the removal of shareholder liability in U.S. banks.

The complete paper is available for download here.