I
The City of Glasgow Bank failure in October 1878 was an important event in the history of banking regulation and capitalism. The plight suffered by many of its 1,819 shareholders liable in an unlimited way for the bank’s losses led to a significant change in company legislation – the 1879 Companies Act – which laid the groundwork for banks’ widespread adoption, decades later, of pure limited liability applying to all shareholders.Footnote 1
Pure limited liability – as ‘limited liability’ is defined today – entails that when a bank makes losses, shareholders, including senior bank executives, can be asked to make up for those losses only up to the amount they initially invested in the bank, which contrasts with their potentially unlimited upside gains. If, in the end, this limited contribution amount is insufficient to cover total losses, some of the bank’s creditors, another bank, or indeed taxpayers may be asked to foot the bill. By incentivising the strict restriction of shareholder liability, the 1879 Act contributed to the advent of the modern era of British banking, one in which banks had no trouble attracting large numbers of shareholders now relieved of a heavy burden.
Yet this era was also one in which senior executives, to whom limited liability also applied, gradually lost a major stake in banks’ survival, with the costs of bank failure shifting progressively to society. Despite recent efforts to reverse that trend through better resolution practices (Philippon and Salord Reference PHILIPPON and SALORD2017), issues of moral hazard and a sense of injustice among taxpayers persist to this day (see Admati, Conti-Brown and Pfleiderer Reference ADMATI, CONTI-BROWN and PFLEIDERER2012; Cohan Reference COHAN2017; Greenwood et al. Reference GREENWOOD, HANSON, STEIN and SUNDERAM2017).Footnote 2
In this article, we question the foundations of this Act, which improved the majority of shareholders’ situation for good reason, but perpetuated the original problem of liability rules applying to all shareholders in the same way, despite significant asymmetries of power and information among them.
We do this in two steps. First, we revisit the history of the City of Glasgow Bank and, while we do not expose new facts about the causes of its failure, we draw a contrast between its directors, who had considerable information and control over the bank, and the other shareholders, who had little of either. We then analyse the public and parliamentary debates that followed, up to the passage of the 1879 Act, and assess the soundness of the arguments that were put forward in support of equal liability treatment of directors and other shareholders. We argue that those arguments had little regard for information and control allocation. An amendment was proposed in favour of retaining extended liability for bank directors, but was shelved, we argue, for insufficient reasons, thereby missing an opportunity for more thoughtful reform. Following Goodhart and Lastra (Reference GOODHART and LASTRA2020), we suggest that the resulting Act’s one-size-fits-all approach to liability was flawed. Examining the debates around the Act enables us to give further weight to the idea that an alternative framework is needed to meet concerns about both the ability to attract shareholders, by strictly limiting their liability, and keeping senior executives at greater risk, by adjusting their remuneration regime. This idea echoes emerging themes in the corporate governance literature (see, in particular, Macey and Miller (Reference MACEY and MILLER1992); Grossman (Reference GROSSMAN2001, Reference GROSSMAN2007); Kashyap, Rajan and Stein (Reference KASHYAP, RAJAN and STEIN2008); Edmans and Liu (Reference EDMANS and LIU2011); White (Reference WHITE2011); Conti-Brown (Reference CONTI-BROWN2012); Edmans et al. (Reference EDMANS, GABAIX, SADZIK and SANNIKOV2012); Grossman and Imai (Reference GROSSMAN and IMAI2013); Mitchener and Richardson (Reference MITCHENER and RICHARDSON2013); Bolton, Mehran and Shapiro (Reference BOLTON, MEHRAN and SHAPIRO2015); Haldane (Reference HALDANE2015); Acharya, Mehran and Sundaram (Reference ACHARYA, MEHRAN and SUNDARAM2016); Schwarcz (Reference SCHWARCZ2017); Salter, Veetil and White (Reference SALTER, VEETIL and WHITE2017); Huertas (Reference HUERTAS2018); Aldunate et al. (Reference ALDUNATE, JENTER, KORTEWEG and KOUDIJS2021); Koudijs, Salisbury and Sran (Reference KOUDIJS, SALISBURY and SRAN2021); Bogle et al. (Reference BOGLE, CAMPBELL, COYLE and TURNER2024); Acharya, Rajan and Shu (Reference ACHARYA, RAJAN and SHU2025)).Footnote 3
The plight of the great majority of the City of Glasgow Bank shareholders led to understandably critical reactions. The losses of the bank were so large (more than six times its capital) that many of the bank’s shareholders, under the rules of unlimited liability, became insolvent.Footnote 4 Many, especially in the Press, exaggerated the damage caused by portraying shareholders as men and women of modest means thrown into poverty overnight, when in fact most were relatively wealthy individuals, which enabled depositors to be repaid in full (Acheson and Turner Reference ACHESON and TURNER2008; Lee Reference LEE2012; Turner Reference TURNER2014). Nevertheless, public outrage also came from the fact that shareholders had been left out of some of the key decisions made by the executive directors over 1876–8. In the eyes of many observers, they could neither have had the adequate information nor an adequate amount of control over banking operations needed to be rightfully held responsible for the bank’s failure.Footnote 5 The public’s call for limiting shareholders’ liability was therefore justifiable.
But what about senior executives? The executive directors, for instance, were presumably those with both fuller information and necessary control powers to make banking decisions. Why didn’t the 1879 Act differentiate between their liability and that of the other shareholders? Senior managers’ liability was not a central part of the debate, which focused more on the obvious and popularly decried plight of the latter. This may have been, by itself, part of the problem.Footnote 6 But the question was asked by some. There was, at the time, an assumption that extensive forms of liability would attract ‘men of straw’, men of modest means, devoid of talent and with little capacity to make depositors good, providing further justification for limited liability. This assumption was based on erroneous grounds (Turner Reference TURNER2009, Reference TURNER2014). Nevertheless, it was widely held and was also applied to directors. This argument can still be heard today.
Yet limiting senior managers’ liability in the same restricted way as the mass of shareholders could come with sizeable costs. First, it could introduce an incentive problem whereby senior executives with little to lose on the downside but unlimited upside potential might be tempted to take excessive risks. This moral hazard issue was recognised either explicitly or, more often, implicitly, and the remedy was found in regular audits. As long as a bank was regularly examined by external auditors who would check the truthfulness of published balance sheets, the argument went, depositors’ funds would be safe. With auditing already widespread, the 1879 legislation formalised a shift in responsibility for good banking practice from bank insiders to auditors and regulators, just as US legislation would do some 50 years later.Footnote 7 Second, limiting liability significantly increased the potential for injustice if it meant that senior executives lost little during taxpayer-funded bailouts and even, some might argue, bail-ins.Footnote 8
In this article, we first re-examine the well-known City of Glasgow Bank failure through the lens of information and control thanks to the material that has been preserved about it, including the verbatim account of the directors’ trial in 1879 (Section II). We show that, then as now, the size of the shareholder pool brought sizeable capital but precluded shareholders’ access to key information and reduced their agency in the bank. The bank’s shareholders could be considered ‘outsiders’, like most of today’s bank shareholders. In contrast, the executive directors all had information and control in a way that would qualify them as ‘insiders’.Footnote 9 In this sense, from a moral hazard point of view as well as from the point of view of justice, limiting most shareholders’ liability made sense, while limiting that of senior executives in the same way did not. This helps build the case for more tiered forms of liability.
We then analyse the parliamentary and public debates that followed the bank’s failure to understand and critique the arguments that gained traction and eventually led to the passage of the 1879 Companies Act facilitating the equal limitation of all shareholders’ liability, including that of executives (Section III). To this end we pore over pamphlets, newspaper articles, parliamentary speeches and parliamentary debates. Many of the arguments put forward on various sides had already been made in the decades preceding the crisis, and the move towards equal forms of limited liability had been a process long in the making. But this bank’s failure gave them extra weight and urgency.
The main argument against requiring more extensive liability for senior management was the same as that regarding other shareholders: that it would repel wealthy and skilled men from the profession. In fact, this argument was partly misconceived. Unlimited banks had no trouble attracting a large pool of wealthy and skilled individuals thanks to the greater trust they inspired, which widened business opportunities and gave a greater prospect of stability. As we will see, there are good reasons to believe that this would remain partly true today, especially with the relative rise of social responsibility as an integral objective of corporate agendas. Nevertheless, the issue of job attractiveness must be taken seriously in a hyper-competitive environment.
Finally, we set out the conditions for an efficient tiered liability regime, drawing on broader historical and theoretical considerations. First, even extensive liability should remain defined (such as double or triple). Second, it should be compensated financially. Third, a Court of Appeal should be set up for contentious cases. Fourth, safeguards should be put in place to reduce the risk of avoidance. We also discuss other current executive remuneration proposals and suggest that, from the point of view of incentives, payment in extended liability equity (rather than, say, bail-inable debt) would be optimal, with extended liability applying to the bank as a going concern.Footnote 10 While we refrain from drawing any firm conclusions about the exact shape of optimal total executive remuneration, we end by summarising how the City of Glasgow Bank failure and its parliamentary aftermath provide a strong case for liability reform. We realise that it would not be a panacea and argue for continuing regulatory efforts, supervision and lender-of-last resort operations.
II
Let us start with some basic facts. The City of Glasgow Bank was one of the largest Scottish banks, incorporated in 1839. It had 133 branches at the time of failure. All 1,819 share owners had unlimited liability, meaning that in case of losses they could be called for an unlimited multiple (initially equal across shareholders) of their subscribed capital (Official List of Shareholders 1878).Footnote 11 It had an original paid-up capital of £1 million, and its minimum share denomination was £100. From as early as 1873, but especially from 1876 onwards, the bank, under the leadership of Robert Stronach, started concentrating a great part of its assets in loans to individuals and companies in liquidation or close to it, instead of responding to the need to close those accounts and potentially close the bank itself early on before further losses were made.Footnote 12 In the end, the bank closed with £6.2 million in losses, and liquidators eventually called upon shareholders for 27.5 times the par value of their shares. Only 254 shareholders remained solvent after liquidation (Wallace Reference WALLACE1905; Couper Reference COUPER1879). The seven executive directors were tried and sentenced to several months in prison.
The bank’s corporate form was the most common one at the time in Britain, one which followed the original Scottish model. From early on, the Scots had realised that small partnerships, in which all partners had managing power and unlimited liability, had both an advantage and a drawback. The advantage was that, to borrow Goodhart and Lastra’s (Reference GOODHART and LASTRA2020) expression, liability ‘matched power’: partners had equally significant skin in the game and power to act, which was viewed as fair and efficient to all involved.Footnote 13 The problem was one related to size: partnerships could not expand significantly without spreading managing power over many heads and thus losing efficiency. In turn, size restrictions reduced access to capital, asset diversification and raised the probability of failure, as illustrated in the 1825 English banking crisis (Pressnell Reference PRESSNELL1956, p. 226).
The Scots had thus devised an alternative corporate form which allowed for bank owners to be as numerous as necessary, thanks to a transfer of controlling power to a small set of managing partners. This separation of ownership (by numerous shareholders) and control (by senior executives) was a significant innovation as it allowed firms to raise unprecedented amounts of capital and expand. The numerous ‘outsider’ shareholders were now simply ‘sleeping partners’, while control was retained in the hands of a few managing directors (Turner Reference TURNER2009; Acheson, Hickson and Turner Reference ACHESON, HICKSON and TURNER2010). In the process however, the original advantage of small partnerships was lost, in our view. By retaining unlimited liability, sleeping partners risked their personal wealth despite having lost much control in, and information about, the bank’s doings. Power and information on the one hand, and liability on the other, had become disconnected.
In spite of this disconnect, this partnership model was adopted in England with the passage of the Irish and English Copartnership Acts of 1825 and 1826, as the advantages relative to small partnerships were more obvious than the downsides. Not only could banks now expand, but they could also convince uninsured depositors that their money was safe, with so many unlimited shareholders (Pressnell Reference PRESSNELL1956).Footnote 14 There might have been a danger that unlimited shareholders sold their shares to individuals of low wealth, especially with rising insolvency risk, but managing boards had the power to vet share transfers, and the law ensured that owners remained liable after a transfer for three years anyway (one year after 1862) (Turner Reference TURNER2009; Button et al. Reference BUTTON, KNOTT, MACMANUS and WILLISON2015). The ability to attract depositors was a major reason why, despite limited liability being allowed for banks from 1858 onwards, by 1875 most British banks had still chosen unlimited liability (Anderson and Cottrell Reference ANDERSON and COTTRELL1974).Footnote 15
But the disconnect between power/information and liability remained. This is nowhere more apparent than at the City of Glasgow Bank. All unlimited shareholders except for the shareholding directors were kept out of key decisions made by the bank. Their power mainly consisted in the ability to vote for the appointment of directors at the Annual General Meeting (n.a. 1859). Most of them were from relatively wealthy backgrounds: lawyers, medical doctors, farm owners and merchants (Acheson and Turner Reference ACHESON and TURNER2008). They had bought or inherited their shares with confidence and believed that the appointment of what they thought of as respectable directors was enough to keep the bank on the right path. This was true even of the large shareholders, such as Robert Brown, a Glaswegian merchant who owned £5,000 in shares, or James McGowan, a manufacturer with £5,500. The only information they had access to was that provided at the Annual General Meeting, especially in the form of the annual published balance sheet prepared by the directors (Wallace Reference WALLACE1905, pp. 370–408). The latter was provided as part of the ‘Report of the Directors’, which initially also included a very succinct verbal summary of how the bank had fared over the previous year, but from 1864 onwards even did away with that (n.a. 1864).
On the other hand, the seven ‘executive’ directors were under the same liability regime as the other shareholders, but benefited from a very different position. By virtue of their original contract, they had the right, if not the duty, to examine the bank’s books at any time.Footnote 16 They could also make important investment decisions without waiting to consult the other shareholders at the annual meeting. (There were also four non-executive directors who did not have the same rights and duties, formally or informally defined, and in consequence were not tried).Footnote 17
Some might argue that the shareholders would have known about the bad loan decisions had the executive directors not broken the law. Indeed, all seven of them were convicted of issuing balance sheets ‘knowing them to be false’ at the annual meetings in 1876, 1877 and 1878. Two of them, the managing (or chief) director Robert Stronach and Lewis Potter, were found guilty of actively falsifying the balance sheets and incurred a longer prison sentence.Footnote 18 The balance sheets were falsified in obvious ways, for instance by subtracting loan items from the loan categories and adding them to safe security categories. However, one can doubt whether more truthful balance sheets would have revealed all the management issues at the bank. Even today there are more subtle, legal ways of overstating asset values to shareholders. In general, it may be said that most shareholders within joint-stock companies have been and continue to be ‘outsiders’ by design. Already in 1856, Walter Bagehot, who later became The Economist’s editor, pointed out that ‘very few shareholders can know anything of the management of the concern … only very few persons can be really aware how things are going on’ (Bagehot [1856] Reference BAGEHOT1968, vol. IV, pp. 308, 312).
In a sense, therefore, one can understand the injustice felt by many shareholders in the face of the outsized calls for capital during liquidation. Depositors, whom everyone readily accepted did not have the information necessary nor the power to monitor their bank, were due to be paid back in full. But did the vast majority of shareholders have much more information and control than depositors? Following the City of Glasgow Bank failure, there was a widespread realisation that they did not, and that even if the law had been upheld and truthful balance sheets published, they might not have had all the information or control necessary to significantly alter the course of the bank’s operations. A shareholder signing ‘One Who Will Be Properly Ruined’ wrote to the Glasgow Herald on 12 November 1878: ‘We, the miserably ruined shareholders of the City of Glasgow Bank … feel … that surely we will not be allowed to be entirely beggared, through the acts of a coterie of men, whose doings, had we but known of, would have been stopped long ago [emphasis added].’ As far as Quebec the press frequently referred to the unlucky shareholder ‘whose management he has little or no control over’ (Quebec Daily Telegraph, 30 Sept. 1879).Footnote 19
Many, on the other hand, perceived as fair that the executive directors were not only prosecuted, but also made to lose significant amounts through unlimited liability. The seven directors of the City of Glasgow Bank had been contractually required to hold a minimum of nine shares for a total of £900 each.Footnote 20 This was designed to reduce moral hazard by ensuring that they, too, had skin in the game. The evidence shows that each of the seven directors did hold at least £900 in stock at the time of the trial (n.a. 1879, pp. 3–4). Selling this stock would have meant losing the directorship. Even in the case where they would have preferred to step down as director, they would have remained liable for their shares after one year of doing so, which provided a disincentive for selling.
And yet, a basic argument against extending senior executives’ liability, which was not formulated at the time, would go something like this. Given how much the City of Glasgow Bank’s directors had to lose financially, judicially, and therefore reputationally, the fact that they acted illegally and in ways that seriously compromised the bank’s survival suggests that no amount of liability would have prevented them from straying. This line of reasoning lessens the argument for extended liability based on moral hazard for senior management (but not the argument based on social justice).
Granted, in many cases the threat of future losses will not prevent excessive risk-taking due to a form of irrational behaviour often referred to as ‘gambling-for-resurrection’ that can make bank executives invest in unprofitable projects against their own interest (Rochet Reference ROCHET2009; see also Massenot and Baghestaniany Reference MASSENOT and BAGHESTANIANY2015).Footnote 21 However, from a theoretical point of view there is no reason why, on average, rational senior managers should not respond to incentives such as greater skin in the game. For this reason, as well as for social justice considerations, the case for extending senior executives’ liability remains strong.
In addition, there may be specific reasons why the threat of financial liability was reduced for some of the City of Glasgow Bank’s directors, which may have played some role in their behaviour. Four of the seven directors had become heavily indebted to the bank by the time of the trial. Although their personal wealth is unknown at this point, this fact may reveal that, at least in the couple of years preceding failure, they owned little and did not have much to lose, which may have increased their incentives to gamble for resurrection.Footnote 22 It also appears that at least three of them held larger amounts of stock earlier on but sold in the years preceding the failure, leading them to own only the minimum required amount.Footnote 23 This may be a sign that the set minimum was too low. Finally, two of the directors had also been directors or partners in enterprises that were connected to some of the firms that were deeply in trouble and to whom the bank was continuing to lend. This may have disincentivised them from resolving to close the relevant accounts.Footnote 24 Importantly, given the meagre information that the other shareholders were given at the annual AGM, it is very unlikely that they would have known anything about the directors’ financial situation. Their power to monitor it was very low.
In Sections III and IV, we come back to these deficiencies and suggest that specific safeguards need to be put in place to prevent this kind of situation from arising.
III
The failure of City of Glasgow Bank led not only to public outrage, but also to a relatively well-structured debate in the press, across pamphlets and within Parliament about the proper extent of shareholder liability. This debate eventually led to the passage of the 1879 Companies Act on 13 August 1879.Footnote 25 This Act did not impose pure limited liability. In many ways, it was not so different from the 1862 Act, which had already given a choice to banks between unlimited and limited liability. What was new in this 1879 Act was the added option of ‘reserve liability’. This addition was important in that, as a halfway house between unlimited and limited liability, it led to better acceptance of the concept of limited forms of liability, thereby facilitating the move away from unlimited liability.
This legislation had positives, and its impact should not be overstated. As stated earlier, there were very good reasons for encouraging the limitation of most shareholders’ liability. Moreover, in terms of depositor protection banks’ new ability to choose between these three options did make a difference, but not a huge one at the time. Usually, when subscribing to a share, a shareholder ‘paid up’ an initial portion of the price while the rest could be called in at directors’ discretion, for example to expand the business, or in case of losses. Many limited liability banks, to maintain depositors’ trust in the absence of an extensive government backstop mechanism, in fact asked for greater amounts of paid-up capital to ensure a reasonable buffer.Footnote 26 Under ‘reserve liability’, the only difference was that the uncalled portion could only be used in case of failure (so was kept in escrow, so to speak), and banks could ask for a defined multiple (double or triple) of this uncalled portion to be kept in reserve.Footnote 27 As a matter of fact, nearly all of these banks, whatever regime they chose, kept their depositors reasonably well protected and ensured a significant degree of skin in the game for executives, at least up to the 1890s (Turner Reference TURNER2014, pp. 126–9).Footnote 28
The problem with this legislation, however, was, in our view, mainly threefold. First and foremost, like many pieces of legislation before and after it, its one-size-fits-all approach failed to differentiate between different types of shareholders with different levels of involvement in the bank. This is an issue regardless of the level of liability chosen: at extended levels of liability, executives’ liability might be just right but that of the other shareholders might be too high; at low liability levels (for instance, pure limited liability), the liability of shareholders might be right but not that of executives.
Second, by refraining from imposing a set level of liability, and instead giving a choice between different levels, the law made it more likely that, with the rise of central bank interventions and deposit insurance in the twentieth century, banks would choose the most limited (pure) form, thus leading to the case where executives’ liability would be too low from the point of view of moral hazard and fairness. This is because, as the cost of failure was being gradually externalised to other banks and society as a whole, banks worried less and less about maintaining high reserves of capital. While moral hazard and fairness would still call for greater liability among executives, if not to help the bank survive, at least to reduce costs to society, they became secondary concerns to banks who bore only a small portion of the cost of failure. As a result, the choice of pure limited liability became more likely, even as the need to hold executives to account remained from a societal perspective.
Third, even a significant level of liability (say double or triple) would not hold executives to account if the total amount due was meaningless relative to their remuneration and/or to the size of the bank’s liabilities. In 1885, around 70 per cent of aggregate deposits were covered by shareholder capital, including any uncalled capital, reserve liability, paid-up capital and liquid reserves. By 1921 this ratio had fallen to 15 per cent. By 1950 it had fallen yet further (Turner Reference TURNER2014, p. 128). One of the primary reasons behind this decline was the rise in deposits due to war inflation, which was not matched by concurrent capital increases. Of course, deposit insurance does reduce the need for capital backing. But double or triple liability will not effectively reduce moral hazard and enhance fairness if this liability applies only to tiny amounts. This means that good legislation should require that executives hold a significant stake not only in absolute terms but also relative to their remuneration and to the bank’s liabilities (and that their liability endures for a time after transfer; see Section IV for more detail).Footnote 29
While the latter two issues were not explicitly discussed, the one-size-fits-all problem was mentioned in places. There were in fact proposals for differentiating between shareholder types, in particular requiring more extensive forms of liability for senior management specifically. However, these proposals were little discussed, and in one case, rejected offhand. In what follows we explain why and offer a preliminary critique of those reasons.
‘Justice has been served’
First, an arguably important feature of the popular psychology of the time was that, while the executive directors had certainly been the focus of attention during the trial itself, with detailed and lengthy commentary in the national and local presses, now that the trial was over and the sentence given, eyes turned to the fate of the mass of shareholders who were struggling to make good their commitments. Regarding the directors, popular feeling was that the possibility of moral hazard and injustice had been dealt with, that justice had been served – quite literally so, in court – whereas something remained to be done for the other shareholders. Their plight took centre stage, and whatever happened to directors remained secondary. This is evident in Chancellor Stafford Northcote’s speech on 21 April 1879, which announces the drafting of a Banking and Joint Stock Companies Bill.Footnote 30 There may have been an expectation that future bank failures would see the directors tried in court in the same way, with possibly prison sentences, so that a system of tiered liability would have been unnecessary from an economic efficiency and justice point of view. In addition, as legislation needed to be passed quickly for political reasons, it was simpler to treat all individuals equally.Footnote 31
‘Reserve liability and audits will tackle depositor protection and moral hazard’
Despite its defects, there was a belief that reserve liability, as opposed to pure limited liability, would ensure effective depositor protection, as transpires from Mr Rathbone MP’s defence of the bill (Hansard, HC Debate 12 August 1879, col. 838).Footnote 32 But there was also an admission on the part of some economic commentators, such as Alexander Wilson and several parliamentarians, that the increasingly widespread adoption of more limited forms of liability, including reserve liability, might lead to excessive legal risk-taking on the part of executives (Wilson Reference WILSON1879, p. 129; Hansard, HC Debate 12 August 1879, col. 838). However, in their view this could easily be solved by way of an amendment to the bill requiring the external professional audit of balance sheets each year.Footnote 33 Alexander Wilson, for example, put great hope in it: ‘the mere prospect of a periodical bank audit would practically do more to check abuses of this nature than any legal restrictions which could be devised’ (Wilson Reference WILSON1879, p. 135; see also Evans and Quigley Reference EVANS and QUIGLEY1995; and Willison Reference WILLISON2018).Footnote 34 This was not a new idea (Brown Reference BROWN1905; Soll Reference SOLL2014; Willison Reference WILLISON2018), and the amendment was eventually adopted.Footnote 35
Some MPs did express scepticism at the idea that audits would efficiently deal with moral hazard. Already in his 1856 article entitled ‘Sound banking’, Walter Bagehot had warned that ‘if the auditor knows anything of the business of the bank, it must be by being connected with the management, and then his audit ceases to be independent’ (Bagehot [1856] Reference BAGEHOT1968, vol. IV, p. 308).Footnote 36 And as the bill was being discussed at committee stage in Parliament, Mr Bristowe MP likewise insisted that ‘it was quite impossible that the auditors could dive into all the accounts of the money in the hands of the bank’ (Hansard, HC Debate 12 August 1879, col. 859). Such scepticism is still widely shared today (see, for instance, Kanagaretnam, Krishnan and Lobo Reference KANAGARETNAM, KRISHNAN and LOBO2010).
Among the most sceptical was William Mitchell, a Scottish solicitor to the Supreme Court, who wrote a 170-page pamphlet arguing for amendments to the banking bill (Mitchell Reference MITCHELL1879). Instead of relying on audits, he proposed to extend bank executives’ liability in a way that would make it greater than general shareholders’ liability. To the best of our knowledge, this was one of the first proposals for a tiered form of liability, and it was phrased as follows:
There does not seem to be any undue hardship to Directors in expecting them to undertake unlimited liability while that of the ordinary shareholders is limited. The Directors alone have the opportunity of satisfying themselves as to the soundness of the Bank, and by undertaking unlimited liability in connection with it they will evince their own confidence and at the same time increase that of shareholders and the public. (Mitchell Reference MITCHELL1879, p. 130)
A similar suggestion was made by an anonymous commentator in 1859, arguing that ‘surely it is most proper that the risk of all undertakings should be on those who have the means of knowing, and the power to regulate the amount of that risk’ (n.a. 1859, p. 40).
The proposal for some form of ‘two-tier’ liability was also made in Parliament. Joseph McKenna MP drew a contrast between shareholders’ access to information and directors’, describing the latter as the ‘only class of proprietors of a bank who could not be deceived’. This for him justified an amendment to the bill requiring that directors retain unlimited liability, up to one year after leaving office.Footnote 37 Yet the amendment was swiftly rejected by Chancellor Northcote himself, who immediately replied that ‘the objection to the Amendment was that it would discourage men who would make desirable Directors from incurring the responsibility’ (Hansard, HC Debate 12 August 1879, col. 863). It is to this argument that we now turn.
‘We will attract wealthier and more talented directors’
There was a fear that imposing greater liability on senior management would attract men of low wealth, and inadequate talent and skill. This fear had its roots in a much more general argument about all shareholders: keep liability unlimited, and bank shareholders will predominantly end up being ‘men of straw’ with neither the skill to understand bank balance sheets nor the wealth to make depositors good. This argument was assumed to apply to senior executives as well.
In a world without deposit insurance, the possibility that all shareholders might lack wealth could be a very significant issue. On the face of it, it seemed logical that an individual who had more to lose would be reluctant to take on unlimited liability, so that adverse selection would occur and only individuals of low wealth would end up owning the company. In case of insolvency, even an unlimited call might not bring much. This could have devastating consequences for the depositing public, reduce confidence in the banking system, with potentially self-fulfilling effects at times of panic.
This argument – that the mass of shareholders might be adversely selected – had been made before. One of the first proponents of the idea was William Clay, a parliamentarian, who as early as the 1830s had argued that unlimited liability attracted men of low wealth, intelligence and respectability (Willison Reference WILLISON2018). The Bank of England had voiced similar concerns.Footnote 38 But its most vocal supporter was Walter Bagehot. He brought up the argument in several essays between 1856 and 1862. In his Saturday Review article, ‘Unfettered banking’, he expressed worry that ‘instead of allowing persons of real wealth to become shareholders in banks, we practically confine their foundation and management to comparatively needy and adventurous men’ (Bagehot [1856] Reference BAGEHOT1968, vol. IV, p. 311).Footnote 39
The failure in 1857 of the Western Bank of Scotland, a large unlimited liability joint-stock bank, may have led to fears that wealthy shareholders of other unlimited concerns would quickly sell their stakes in realisation of the concrete risks they were taking. This in turn would reduce the aggregate wealth available to shore up depositors in Britain. A piece of legislation in 1858 allowed banks to register as limited liability companies just like other joint-stock ones. Some banks, such as Overend and Gurney, did re-incorporate as limited liability concerns.Footnote 40 And yet, most in the banking community were against the spread of limited liability as they worried the limited form would play a bigger role in reducing the amount of capital backing the banking system and scare depositors away (Turner Reference TURNER2009).Footnote 41
After the City of Glasgow Bank failure, however, the argument gained further traction. Chancellor Northcote feared that
gentlemen of large means, who are just those whom the public would desire to see occupying positions as shareholders in great institutions, will be unwilling to put themselves in that position, and will withdraw; and so this unlimited liability will lead to the substitution as shareholders of a very inferior class of persons whose liability, although unlimited in name, will not be worth nearly so much to the public and the creditors of the concern as might be the limited liability of a superior class of persons. (Hansard, HC Debate 21 April 1879, col. 792)
The argument was also expounded at length by Alexander Wilson.Footnote 42 It also occurs in the Bankers’ Magazine.Footnote 43
The argument applied to bank directors as well. Wilson worried that retaining unlimited liability for directors would ‘undoubtedly produce deterioration in the quality of directorates’ (Wilson Reference WILSON1879, p. 129). Mr Heygate MP believed that
everyone connected with banking business knew there was no greater difficulty than that of finding a good set of men to undertake the management of a joint-stock bank; and if the principle of the Bill was adopted, so as to become general, he thought it would tend to the advantage of the public in securing both a better class of shareholders and a better class of directors. (Hansard, HC Debate 21 April 1879, col. 804)
As we saw, it was the chief reason advanced by the Chancellor of the Exchequer when Joseph McKenna proposed his amendment.Footnote 44
Yet the argument lacks evidence. Bank managers kept a close eye on shareholders’ identity, and thus their wealth. They had, in most cases, the right to vet share sales, which gave them significant power to deny entry to a prospective shareholder of insufficient means. In addition, since 1862 a clause in the law ensured that unlimited liability would be kept for a year in case of share transfer, which created a disincentive to sell even as a banks’ prospects diminished. Carr and Matthewson (Reference CARR and MATHEWSON1988), Acheson and Turner (Reference ACHESON and TURNER2008) and Hickson and Turner (Reference HICKSON and TURNER2003) have convincingly shown that thanks to this vetting process, most bank shareholders were quite wealthy (see also Lee Reference LEE2012). At the time of their appointment, most of the directors of the City of Glasgow Bank were also likely to have been wealthy men meeting contemporary criteria for ‘respectability’ (with one exception).Footnote 45 Their incentives to check each other’s credentials were strong under unlimited liability. It remains true, however, that as the years went by following their appointment, these incentives were weakened by other factors. This calls for specific safeguards to be put in place (see Section IV below).
Importantly, this evidence suggests that the pool of prospective shareholders was usually large. An unlimited bank inspired trust, which in turn brought in more investors and depositors.Footnote 46 This might raise the probability of profitable business opportunities. Unsurprisingly, the potential for growth and stability was an attractive prospect to shareholders and directors alike. The responses to the questionnaires sent by the Royal Mercantile Laws Commission in 1854 show that, although lawyers, academics and MPs were in favour of limiting liability for banks, those leaning against a change in the law were more likely to have been be bank executives themselves (Willison Reference WILLISON2018; see also Bryer Reference BRYER1997).Footnote 47 Walter Bagehot himself only believed that adverse selection might become more common in future, and admitted that this belief was not based on observed facts. On the contrary, as he openly conceded in 1862:
Experience shows that under a system of unlimited liability a good board of directors may … be readily obtained. The success of the great London joint stock banks, which were all founded when limited liability was not permitted by law, is on this point conclusive … they could not have been steadily and regularly successful, with scarcely a check or drawback during so many years, if their management had not been admirable or excellent. (Bagehot [1862] Reference BAGEHOT1968, vol. IV, p. 395)Footnote 48
It seems that there was no shortage of talent and skill in London at the time, despite most other industrial and trade concerns being incorporated under limited liability (see also Goodhart and Lastra Reference GOODHART and LASTRA2020; and Bair et al. Reference BAIR, CHANG, GOODHART, GOODMAN, NOVICK and SANDOR2023).
Over the past 80 years or so, the rise of government backstops has reduced stability worries somewhat, which has meant that, while fairness had always been a secondary concern for banks, reducing moral hazard also increasingly became a secondary objective, as it would make a smaller difference in terms of nurturing trust in the bank, and thus increasing business opportunities. This could potentially reduce the pool of prospective managers attracted to a job with extended liability attached to it.Footnote 49
It is therefore still possible that, especially in today’s hyper-competitive corporate world, extended senior management liability in some sectors or countries might deter prospective candidates relative to other sectors and nations. But solutions can be found to keep the job attractive without jeopardising the initial objectives. In what follows we discuss possible practical solutions to this issue.
IV
This last section steps away from the City of Glasgow Bank episode and draws on broader historical and theoretical considerations to set out some practical conditions necessary for an efficient tiered liability regime. In addition to these essential conditions, it also discusses current proposals for executive remuneration and suggests payment in extended liability equity as a potentially attractive compromise between various forms.
As early as the twelfth century, some of the earliest legal corporate forms had tiered liability. For instance, whenever a mill was established in southwestern France, a contract was drawn between the miller, with responsibility both for management and losses, and the individuals who brought in capital, and gained from the enterprise but would not be responsible for shortfalls (Cartier and Charlier Reference CARTIER and CHARLIER2012).
Later, the well-known economic success of medieval Venice had some of its roots in the efficient legal organisation of maritime commerce, which was partly based on tiered liability. There, unlimited liability partnerships, compagnia, coexisted with commenda – limited liability joint-stock companies in which managers held more power (governo) and information and were subject to unlimited liability. Those were essentially tiered liability companies (Carmona Reference CARMONA1964). Almost every trading ship was a commenda, in which limited liability investors were keen to invest while the ship’s captain remained solely responsible for the enterprise’s (and his own) survival while at sea.
In the commenda, as the managing partners ran greater risks than the shareholders, they were commonly allowed special remuneration. Compensation practices differed across companies. Executives could, for instance, have an especially large amount of capital invested, so that any gains would be very significant. Or they could, given a specific level of capital, gain more relative to this capital than would be the case in a limited company. The latter case seems to have been more common (Carmona Reference CARMONA1964). The commenda corporate form gradually became preponderant in such city-states as Florence (after 1408), Lucca (after 1554) and Livorno, encompassing all kinds of trades, including banking.
More recently, there is evidence that France has also sought to compensate managing partners for their increased liability relative to the other shareholders. France still has around 300 sociétés en commandite, and this number has remained stable since 2000, including a few among the largest French businesses (Cartier and Charlier Reference CARTIER and CHARLIER2012).Footnote 50 In sociétés en commandite, as in medieval Venice and Renaissance Tuscany, executives have unlimited liability while the mass of shareholders is only liable in a limited way. To compensate the former for their extra liability, sociétés en commandite give them not extra remuneration, but extra power. Executives usually cannot be ousted by shareholders except in extreme circumstances, and even then, it might be difficult.Footnote 51 This quasi-autocratic framework may continue to attract candidates to the job, but is facing an increasing backlash among shareholders, as the recent abandonment of the corporate form by Lagardère Group, an entertainment and travel retail giant, has shown.Footnote 52 If it is not already the case, it is likely that this form will soon seem archaic.
In the end, it appears that it would be best to go back and think of compensation as Renaissance merchants from Lucca envisioned it, that is, to think of optimal ways to attract candidates to senior management roles without resorting to extreme empowerment.
A basic way of ensuring a consistent, highly skilled labour supply at executive level might first be to define liability clearly. This is the first condition of an efficient tiered liability regime. Indeed, even Walter Bagehot, sometimes portrayed as supporting pure limited liability, in fact insisted simply on executive directors’ liability being defined rather than unlimited, even if that meant up to five times the original amount invested. In his mind it was important that senior managers should take risks in the enterprise: ‘It is necessary that directors should have a large risk in the undertaking’ (Bagehot [1862] Reference BAGEHOT1968, vol. IV, p. 397).Footnote 53 But there would be no shortage of good directors, provided their liability was well defined: ‘they would be ready, perhaps, to go to five times that amount. They are anxious there should be a limit, but they would not care comparatively where the limit was placed’ (Bagehot [1856] Reference BAGEHOT1968, vol. IV, p. 310).Footnote 54
Within this framework, one could envisage various systems of tiered liability, which would be applied as soon as losses occur. In one of them, liability would be pre-defined at a specific, single multiple of accumulated revenue for all ‘insiders’ (for instance, double), whereas the ‘outsider’ shareholders would retain pure limited liability. Deciding what constitutes an ‘insider’ would have its challenges, but the advantage of such a two-tier system would be that, once defined, it would be relatively simple and intelligible to all. Goodhart and Lastra (Reference GOODHART and LASTRA2020) suggest that ‘insiders’ include all of the board of directors, including the externals, as well as any employee earning remuneration in excess of 50 per cent of that of the CEO.Footnote 55 As some employees might avoid liability by adjusting their remuneration, the regulatory authority should also be able to designate ‘insiders’, including retrospectively.Footnote 56
Another system might be one which also distinguishes between ‘insiders’ and ‘outsiders’, but where liability would be tiered among the ‘insiders’ themselves. As CEOs have much more information and power than anyone else, they could have triple liability, while board members and chief officers would have double liability and anyone else within the ‘insider’ category would have less. Large shareholders might also be included in the ‘insider’ category and might face tiered liability according to the amount of their holdings.
Next, extra remuneration during good times for those bearing greater liability in bad should be considered. The logic behind this would be twofold. First, it might maintain job attractiveness in a world where other sectors would retain pure limited liability. Second, greater liability comes with greater involvement, and greater involvement could be rewarded in good times. Walter Bagehot was clear that remuneration should be significant, if only because otherwise the threat of multiple liability would be meaningless.Footnote 57 Greater remuneration in good times would not defeat the purpose of greater liability, as liability itself would always be calculated as a function of accumulated revenue in good times (see below).Footnote 58 It was also supported by Mitchell when arguing for tiered liability (Mitchell Reference MITCHELL1879, p. 131).Footnote 59
Third, allowance should be made for cases where extended senior management liability would seem unfair in light of general circumstances. For instance, if an ‘insider’ gains awareness of management practices that need changing, but does not succeed in convincing others, they could send a letter to the regulator making their views known (these would then need to be examined and evaluated by the latter).Footnote 60 One might also consider the constitution of a Court of Appeal to deal with cases where banks make losses due to highly unusual but significant events originating outside the banking system, such as natural disasters.
Some might add that an initial failure of a bank may lead quickly to a systemic crisis in which there is a generalised withdrawal of deposits from banks, both good and bad. But, in practice, the normal progression of such a crisis involves depositors and other creditors first looking around for the next weakest bank (to the initial failing bank). Recent examples include Signature Bank after Silicon Valley Bank in 2023 and Bradford and Bingley after Northern Rock in 2007–8. By the time concerns turn into a more generalised panic, the central bank is virtually bound to step in, in order to protect the financial system as a whole. For these other banks liability rules then become less relevant.
Finally, some safeguards need to be designed and put in place to ensure that the incentives to reduce moral hazard are not jeopardized by ways of escaping liability or becoming immune to its effects. As mentioned earlier, some of the City of Glasgow Bank executive directors had initially held large stakes and gradually reduced them to the bare minimum required by the time the failure occurred. This reinforces the importance of calculating liability as a function of accumulated bank liabilities and senior executive revenue in good times. Some of the directors were also heavily indebted to the bank, and one of them even borrowed from the bank to buy his stake. This calls for restrictions on borrowing, and for some monitoring of directors’ financial situation.Footnote 61 Finally, some of the directors’ firms had connections to some of the firms the bank was lending to. These conflicts of interest change the balance of incentives and should be strictly regulated. It is also likely that tiered liability would need to be applied to nonbanks.
One of the closest competing proposals to ours, that executives be partly paid in bail-inable debt (sometimes referred to as contingent capital), takes its inspiration from the seminal paper by Edmans and Liu (Reference EDMANS and LIU2011), who suggest that company executives be remunerated through a mix of limited liability equity (to incentivise profit-making and align with shareholders’ motives) and debt (to align incentives with those of creditors and other depositors).Footnote 62 The advantage of such a proposal is that it would, like extended liability equity, reduce incentives for risk-taking. At the same time, the right mixture of equity and debt would have to be chosen, the maturity of the debt predetermined, and bail-in would have to be triggered by events that would need to be pre-specified. For these reasons such proposals are relatively complex and subject to a high degree of arbitrariness. In addition, executives’ incentives would only be affected by the prospect of an extreme tail event, as bail-in would not apply to a bank as a going concern. This would preserve some bias towards risk-taking, which would only disappear if they were affected by the prospect of losses irrespective of bail-in.
Other authors, such as Acharya, Mehran and Sundaram (Reference ACHARYA, MEHRAN and SUNDARAM2016), have suggested that bankers be remunerated partly in deferred (escrowed) cash or bonuses, whereby write-downs could occur while the bank is still a going concern but undergoing stress.Footnote 63 In addition, under a subset of such proposals directors would incur first loss, before common equity (as well as being subordinated to general creditors), so that they would in effect count towards the bank’s Total Loss Absorbing Capacity (TLAC) and Minimum Requirement for Own Funds and Eligible Liabilities (MREL) requirements during the deferral period.
Making sure that losses can be absorbed while the bank is a going concern is essential from the point of view of incentives. Remuneration in extended liability equity fulfils this criterion and thus, like the latter proposal, reduces any potential complexity and arbitrariness. It also has the additional benefit of maintaining the attractiveness of senior executive roles through immediately cashed-in profits during good times.Footnote 64
Much work remains to be done in the precise calibration of an optimal tiered liability system. Tiered liability would not be a panacea – we have insisted on past examples where individuals with large stakes in banks still acted recklessly, possibly due to ‘gambling for resurrection’ behaviour, including within the City of Glasgow Bank itself (Rochet Reference ROCHET2009). Effective, judgement-based supervision, for instance, also plays a critical role in assessing risk culture within a bank’s leadership, and would help detect this kind of behaviour (Enria Reference ENRIA2025). And while greater liability would reduce the need for some forms of financial regulation, capital and liquidity requirements would remain essential to limit systemic risks arising from panic and contagion effects. Finally, central bank intervention would be required once the crisis spreads.
V
The failure of the City of Glasgow Bank in 1878 revealed how unfair it was for ‘outside’ shareholders, with little, or no, access to information or power over the decisions of their bank to suffer unlimited liability. That such shareholders should have limited liability was then, and remains, generally accepted. But that then raises the question of what should be the liability of ‘insider’ shareholders, a question that was, indeed, considered at the time. The main argument that was levied against some form of multiple, or tiered, liability was that it ‘would discourage men who would make desirable Directors from incurring the responsibility’.
In this article we argue that such concerns were not properly based on evidence and were much overstated. There remains a strong case for some form of additional tiered liability for ‘insiders’ on grounds both of social justice and for the limitation of moral hazard. Moreover, the lack of any such grading has led to a developing thicket of regulatory and supervisory controls over banks that may make financial intermediation less efficient.
While we argue for a positive reconsideration of tiered liability for bank executives, there are many outstanding questions, such as the definition of (various rungs of) insiders, the scale of such extra liability, and whether there should be the ability to appeal against penalty when failure was due to unpreventable external forces. So, the adoption of tiered liability would be a complex matter. Nevertheless, we think that it could, and should, be attempted.
Open access
