1Philip Cottrell, writing in 1980 about the finance of business in the long 19^th century, observed: ‘industrial firms had to look to intra-industry sources of finance or to themselves or acquaintances for accommodation. Basically, this means some form of trade credit, or the admission of a new partner, or a loan on mortgage, or profits generated by the concern. Unfortunately it is precisely for these areas of finance that little evidence is at present available and therefore the picture that can be established is partial and, with regard to profits, highly conjectural’.² This is still true today, largely because business historians on the whole eschew financial records, even though they survive in quantity and are frequently the only records that can be found.³ The American historian Dr Stephen Campbell bemoaned this state of affairs in his blog ‘How Bills of Exchange Worked⁴: Part of the reason is that the issue itself is rather arcane’. The accounting historian R. H. Parker urged historians ‘not to discard accounts but to learn more about them’.⁵ Unless historians follow his advice and make an effort to interpret the working of the money market they will fail to connect with current research and literature on the crucial questions of inter-firm finance and sources of working capital. There are a few exceptions, such as. T. J. Baldwin’s Cash flow and corporate finance in Victorian Britain.⁶

2Until the coming of limited liability in the mid-19^th century, the only surviving evidence for the way in which most businesses were managed are the ledgers, which included the partners’ private accounts.⁷ Their balances represented a firm’s capital on which interest was paid and partners and their families often provided additional loans to increase capital adequacy. Partners used their accounts to bind customers and suppliers through loans and investments in a web of inter-firm finance. The informal nature of such relationships makes it difficult to calculate the scale and complexity of such networks, but there is enough evidence to suggest that they were pervasive and persistent. Such linkages should not be confused with cartels that were widespread in the United Kingdom in the late 19^th century and bound by formal agreements, for example shipping conferences.⁸

3The ledgers of William Denny & Brothers of Dumbarton on the north bank of the river Clyde in the west of Scotland are held in Glasgow University Archives.⁹ Although the family had been shipbuilders for some time, the firm formally had its origins in 1844 when the brothers William, Alexander and Peter Denny formed a new partnership. The ledgers survive from 1845 until 1910.¹⁰ Peter Denny became in effect the sole partner in the enterprise from 1850 until his death in 1895. By investing forward in shipping companies, he built up a large business, specialising in small vessels for inland river navigation and long-haul vessels for routes to India and Australasia. Following the disruption in the Church of Scotland in 1843, Peter Denny had become committed to the Free Church of Scotland. This brought him into contact with the partners in the Glasgow shipping firm of Paddy Henderson & Co. with whom he collaborated in succouring the Free Church colony on the South Island of New Zealand. He began to build ships for the Hendersons and took shares in their ventures, notably the formation of the Albion Shipping Co. in 1864. The following year he again joined with the Hendersons in the establishment of the Irrawaddy Flotilla Co. to provide shipping services to the interior of Burma.¹¹ Such forward investment in the supply chain was a commonplace. Notable examples are the Belfast shipbuilders Harland and Wolff and Russell & Company of Port Glasgow.¹² Such investments not only brought business to the yards, but also income to the partners in the shape of directors’ fees and dividends. The financial consequences of such linkages are little explored, almost certainly because the evidence is hard to come by.¹³

4The detailed records of Peter Denny’s investments are to be found in 3,777 entries in his partner’s account in the firm’s ledgers. These were analysed by Michael Moss and Nicholas Morgan in 1989. The results were surprising. He earned only 60 percent of his income from the shipyard: the remainder was made up of ‘of dividends on other investments, rents from property, interest on loans made to individuals, capital gains on investments, and director’s fees’ (Figure 1).¹⁴

Figure 1. Peter Denny: Sources of income 1844‑1895

5The bulk of his income from William Denny& Brothers was reinvested in the firm to fund further expansion. He only started to make substantial investments in shipping in 1868 when he invested £12,000 in new ships under construction, £31,000 in shares in the Hendersons Albion Shipping Co. and £8,000 in the Irrawaddy Flotilla Company.¹⁵ Altogether during his lifetime he took shares in no less than 50 ships or shipping companies, totalling £227,000 which yielded him a return of just over £280,000 in interest and dividends, quite apart from the business it brought to the yard (Figure 2). Peter Denny also invested backwards in his suppliers, for example in the Consett Iron Company in the north east of England, the Steel Company of Scotland, the North British Railway Company and his largest investment of all, some £40,000 in the formation of the Rio Tinto Co.¹⁶

6Such investments across the supply chain made only a marginal contribution to the working capital of either William Denny & Brothers itself or his customers and suppliers. Much of the working capital came from bills of exchange. George Rae, writing in his well-known handbook, The Country Banker, in 1885 estimated that ‘the average of bills under discount in England may be safely put at 250 millions of pounds sterling’.¹⁷

Figure 2. Perter Denny: Expenditure, income and revenue from shipping investments

7This was a colossal sum; it is, however, much less than the figures given by Shizuya Nishimura in his study of the decline of the inland bill on the London market from 1855-1913. He reckoned the total of bills drawn in 1873 was £1,781 million which had fallen to £1,121 million by 1894 (Figure 3).¹⁸ The political economist Walter Bagehot observed in his essay of 1873 on Lombard Street – the financial heart of the City of London: ‘English trade is carried on upon borrowed capital to an extent of which few foreigners have an idea...’¹⁹

Figure 3. Bills in circulation in £ million, 1855‑1913

8Bills of exchange or promissory notes have a long history and were a simple way of securitizing debt. In the United Kingdom their use was codified in legislation in 1882, which included the definition that all students of banking learned by heart – ‘an unconditional order in writing addressed by one person to another, signed by the person or company giving it, requiring the person or company to whom it is addressed to pay on demand or at fixed or determinable future time a sum certain in money to, or to the order of, a specified person or company, or to bearer’.²⁰ In other words it is a binding promise to pay for a good or service at a future date. Bills were usually payable in three to six months, but could be renewed if a trade, for example the import or export of goods or commodities, had not been completed.²¹ The larger the denomination of a bill the longer the payment period tended to be.²² Bagehot in the same essay explained how by resorting to bill finance a merchant, and he could have added industrialists, could make a much better return than by relying solely on their own capital.²³ Bills and other credit instruments depended entirely on trust between the two partie: they could be endorsed by third parties who guaranteed to meet any payment if the acceptor failed to honour the obligation to pay. The novels of Anthony Trollope are full of references to the folly of endorsers. In Framley Parsonage, published in 1861, the Reverend Mark Robarts finds himself liable for £400-£500 through endorsing a bill unwittingly: ‘This was terrible to be borne. He had heard much lately which had frightened and scared him, but nothing so terrible as this; nothing which so stunned him, or conveyed to his mind so frightful a reality of misery and ruin’.²⁴ Such literature as exists tends to focus on such fraudulent use of bills, for example so-called ‘pig on bacon’ bills where no goods or services changed hands, but the bills were discounted by unsuspecting brokers.

9Providing they were not non-negotiable, bills could be sold to third parties, mostly through local banks which could rediscount them with London discount houses, and: ‘Because of its liquidity it is an attractive asset for banking and non-banking institutions to buy as a short-term investment’.²⁵ The way the system worked in the 19^th and early 20^th century was that commercial banks would receive bills from their clients. They could either hold them to maturity or if they wished re-discount them on the bill market through bill brokers (secondary banks) which would obtain guarantees from a merchant bank that would settle the debt in event of a default. The discount rate was and is still calculated as the principal less the discount interest rate times the number of months the bill has still to run. Discount interest was always above the monthly bank bill rate, but varied depending on a variety of factors, such as the credit-worthiness of the payee and the length of time the bill had left to run. The bills could then be sold by bill brokers to mostly London-based commercial banks. As time went on some bill brokers began to act in effect as secondary banks by holding bills to maturity. Most of the bills discounted were of large denomination. The Bank of England provided liquidity through an office called the Discount Window that gave ‘Discounts’ (exchanging through repurchase agreements at discount) on merchants’, brokers’ and other banks’ bills of exchange. This was in addition to ‘Advances’ also provided through the Discount Window.²⁶ This need not concern us here, but helps explain why it is so difficult to arrive at anything more than an estimate of the total value of bills in circulation at any one time.

10One of the advantages of financing working capital through bills of exchange is that they were held off balance sheet as a contingent liability and not as a charge on any asset.²⁷ This would be important where bills were being used to finance an asset, such as a ship or a railway, that might cost more than the value of the fixed assets (plant and machinery) used in its construction. When the bill broker Overend & Gurney failed spectacularly in 1866 the largest loans on its books were to shipowners and shipbuilders. It had advanced the Atlantic Royal Mail Steam Packet Company £839,345; the Millwall Ironworks, a London shipbuilder, £422,565; the East London Shipping Company £397,653; Thomas Howard of the Rotherhithe Foundry, another London shipbuilder, £331765; and the Greek and Oriental Steam Navigation Company £144,144. Since these loans were in the shape of bills at discount and held off balance sheet, they were not secured and in the case of the Millwall Ironworks and Thomas Howard valueless and those to African Royal Mail Steam Packet Co. were written down to £160,000, to the East India and London Shipping Company to £25,000 and the Greek and Oriental Steam Navigation Company to £7,000.²⁸ A collapse of this magnitude sent shock waves through the economy and brought down many other businesses. There are clearly parallels with the 2008 banking crisis that led the Bank of England to revisit the failure of Overend, Gurney & Co., attributed to ‘poor lending practices and a lack of management oversight’. Overend’s failure caused financial panic in the City of London and the Bank of England had to ‘draw greatly on its own reserves to support the bills market’.²⁹ What was different was that bills, although they function as derivatives, had far fewer associated risks than the special purpose vehicles (SPVs) in 2008 held off balance sheet to securitize mortgages in bundles in so-called collateral debt obligations (CDOs).³⁰

11The problem confronting historians in exploring working capital and liquidity is finding the evidence. Since bills were held off balance sheet, they are usually recorded in separate volumes from the main financial series in bill books. These record bills receivable and bills payable, in other words bills that the firm had accepted in payment of a good or service and those that had been accepted by third parties for goods or services supplied. They are easily overlooked by archivists unaware of their significance and the sometimes large transactions they contain.³¹ Although any well managed business must have kept records of bills running either held or discounted, such records rarely survive. Entries in the ledgers simply record payments and receipts, from which it is impossible to calculate the scale of contingent liability. In most cases there is no record apart from the bill books of the use of bills to fund a trade, unless, as in the case of shipbuilding, it is specified in a contract. Contracts, even for expensive goods such as ships, do not often survive. Those for William Denny & Brothers do.

12Contracts could be very simple, just a letter offering to build a ship at a certain price, as for example ‘We have much pleasure in offering to build for you a Screw Steamer of moulded dimensions 390 feet by 40 feet by 28 feet for the sum of £44,000’. This contract with Paddy Henderson & Co. was later changed to a joint-venture at a reduced price of £43,000. Dennys wrote – ‘a cheaper or finer money making [ship] we have never built. It is ticklish work with everything rising round us, but we shall “hedge”. The ship is “now going on”’. It was later sold to the British India Steam Navigation Company at a handsome profit of £6,000.³² Joint-ventures and hedging were other ways of increasing working capital, particularly in times of crisis. Very little has been written about the futures market. In the commodity trades in its modern form it depended on the introduction of the telegraph and in the United Kingdom the London brokers Lewis & Peat are credited with its development.³³

13One of the drawbacks of simple contracts is that they could lead to confusion if circumstances changed, such as the failure or takeover of the purchaser.³⁴ Because ships took a relatively long time to construct, they were paid for in instalments at various stages in construction and after delivery. The contract for ship number 233 of 3,500 tons at a price of £78,900 for the Peninsular & Orient Steam Navigation Company with Dennys in 1879 specified that from the outset the vessel should be the property of the company and that it would be paid for in four instalments paid in bills at three months after the work had been certified by an inspector. The last instalment was to be paid after ‘the ship (together with machinery and equipment) has been completed and handed over’.³⁵ More often payment was a mixture of cash payments and bills. For example, ship number 236 which was priced at £4,300 was to be paid for £1,500 in cash at the time of the launch and the balance by six months ‘acceptances [bills] when completed (£1,000 of the latter to be renewed for another 6 mos’.³⁶ Renewing bills and delayed payments were common and were sometimes accompanied by taking security in the shape of mortgages over the completed vessel.

14Remarkably the bill book of William Denny & Bothers bill book contains loose sheets detailing shipbuilding bills that were running monthly from 1903 to 1911 (Figures 4 and 6). These confirm at a glance Bagehot’s assertion that firms could greatly expand their business by using bill finance rather than simply relying on their own capital. The volume of bills running was normally around £400,000, but occasionally as much as £800,000 when at the time the capital of the partnership was just £275,000.³⁷ (Figure 4). These were colossal sums with high gearing and huge associated risks if any of the acceptors failed.

Figure 4. Shipbuilding bills running monthly, drawn by William Denny & Bros (the drawer) on ship owning customers (drawees)³⁸

15The short ledger line shows the total of bills receivable monthly from the ledger, which only records payments when bills came back on to the balance sheet, illustrating the importance of bill books in any attempt to calculate working capital. Dennys did not discount all the bills they drew, preferring to hold some to maturity. There is a separate list of bills discounted, showing the discount rate and whether they made a profit or a loss on the transaction (Figure 6). Dennys made a profit on most transactions as already mentioned discount rate was always in excess of the monthly bank bill rate and sometimes well above bank rate. In 1905 for example they made a profit on every bill they discounted. They used well known London discount houses, such as Alexander & Co., United Discount through its manager J.R. Macdonald³⁹ and the London branch of the Commercial Bank of Scotland. Sometimes bills were discounted by shipping companies linked through shareholding to Dennys.⁴⁰ To be able to hold bills to maturity must suggest that a company had good control of its cash flow and had a keen eye to the margin between the bill and bank rates.

16Inland bills for paying for goods and services sourced in the United Kingdom were abandoned, according to both W. T. C. King and Nishimura, in favour of running accounts, now more commonly referred to as circulating credits, that were agreed between the parties and provided a month’s credit with late payment resulting in the loss of any discount. Lines of credit with commercial banks provided the necessary liquidity in much the same way as the Discount Window. These specified the credit ceiling and the rate of interest which like the monthly bill rate was usually less than bank rate. Running accounts in effect internalised bill finance, doing away with the need for discounting and made cash flow much easier to manage. King claimed that bank amalgamations and the extension of branch banking so strengthened balance sheets that ‘it became a matter of indifference to the banks whether they financed their customers by discounting bills or by granting loans or advances, and to the customers the flexibility of the overdraft system had definite attractions’.⁴¹ As Nishimura pointed out, King’s argument repeated ‘in almost identical words’ R. G. Hawtrey’s in his study A Century of Bank Rate.⁴² Nishimura was right to draw attention to the fact that neither King nor Hawtrey’s argument was based on any detailed analysis and that the decline in inland bills predated the amalgamation movement.⁴³ Nishimura with equal lack of evidence claimed that improvements in communications meant that it was no longer necessary for firms to hold large stocks and that they could deal directly with suppliers rather than through brokers and middlemen, ‘thereby gaining the rebate on cash payments’.⁴⁴ In his review of Nishimura’s book, Derek Aldcroft was surprised ‘that the author has very little to say either about the changing pattern of inventory investment or the role which communications played in the process’.⁴⁵ Unless runs of balance sheets survive, it is impossible to prove or disprove Nishimura’s assertion. There is only one balance sheet for the period in the Denny records for 1892, which shows stocks of £37,644, of which, as might be expected, iron comprised £4,279, wood £18,902 and stores £11,163 made up the bulk. These figures suggest that Dennys were already managing their inventory effectively.⁴⁶

17The disadvantage for the banks as George Rae cautioned is that an overdrawn account, unlike a bill, is not negotiable and Hawtrey to support his argument added ‘the early legal doctrine of assumpsit suitable to the creation of the debt itself’, which has become ‘the basis of almost all transaction on all markets, and is in effect, the assumption that a mere offer and acceptance of a price on the commodity market creates at that price a debt’. The concept emerged in the sixteenth century and was finally enshrined and codified in the Sale of Goods Act of 1893.⁴⁷ Hawtrey was so preoccupied in arguing for what he described as ‘debit money’ that he paid little attention to the underlying security and more significantly the statutory landscape.⁴⁸ The marked decline in the use of inland bills identified by Nishimura coincides almost precisely with the Sale of Goods Act of 1893 (Figure 3).⁴⁹ This landmark legislation is considered to be an outstanding example of codifying legislation, as the distinguished judge Lord Denning pointed out in his judgements in the Mihalis Angelos case in 1971:

‘Sir Frederick Pollock divided the terms of a contract into two categories; Conditions and Warranties. The difference between them was this: If the promisor broke a condition in any respect, however slight, it gave the other party a right to be quit of his future obligations and to sue for damages: unless he by his conduct waived the condition, in which case he was bound to perform his future obligations but could sue for the damage he suffered. If the promisor broke a warranty in any respect, however serious, the other party was not quit of his future obligations. He had to perform them. His only remedy was to sue for damages.
This division was adopted by Sir Mackenzie Chalmers when he drafted the Sale of Goods Act, and by Parliament when it passed it’.⁵⁰

18The Act established in clauses 39 to 41 that if the buyer was unable to pay then the seller had a lien on the goods even if they had been delivered. The protection offered by the Act effectively did away with the security that bills afforded in the domestic market and the associated cost of stamp duty. Clause 8 on price was left open, ‘it could either be fixed by contract, or may be left to be fixed in a manner thereby agreed…’, and clause 28 on payment made no mention of rebates, just stating ‘delivery of the goods and payment of the price are concurrent conditions’. As Chalmers later explained: ‘It is always easier to amend an Act than to alter common law. Legislation, too, is cheaper than litigation. Moreover, in mercantile matters, the certainty of the rule is often or more importance than the substance of the rule’.⁵¹

19Short term secured bank loans were nothing new. A case in the Scottish Court of Session in 1863 between John Reddie and John W. Williamson explained how overdrafts or lines of credit were secured: ‘Three parties granted to a bank a cash credit bond in the usual style for £400 on an account to be operated by one of their number’. The whole sum was drawn by 1844 and thereafter the account remained overdrawn until the death of the holder of the bill of credit in 1850, when £562 was due to the bank. The point at issue was whether the cautioners were liable for the whole sum, including the unsecured advance of £162, or just the £400 stipulated in the bond.⁵² Technically lines of credit were then different and still are from overdrafts which were and are not secured, and the Bankers Magazine condemned in 1882 as ‘always objectionable’ if they represented a large proportion of liabilities.⁵³ Rae warned repeatedly of such over-reliance by banks on overdrafts for profits.⁵⁴ He explained: ‘Overdrafts, then, and especially those of a permanent character are deficient in the first requisite of a banker’s assets, namely – ready convertibility into money in case of need’.⁵⁵ In such circumstances bankers could easily fall into the trap of committing the cardinal sin of lending long and borrowing short in the money market to cover their positions, as happened in 2008. In Scotland bonds ‘formed to a large extent the security taken by Scottish banks’.⁵⁶ Philip Cottrell’s analysis of overdrafts granted by the London bankers Glyns ‘reveal[ed] a bewildering spectrum of maturities and amounts.⁵⁷ Although there could be difficulty in realising such securities in a downturn,⁵⁸ the attraction of switching to lines of credit for domestic transactions was, as King pointed out, their flexibility. Rae advised bankers only to accept sureties that were marketable and to be particularly wary of personal guarantees’.⁵⁹ J.&G. Thomson, the owners of the Clydebank Shipyard down river from Glasgow, later better known as John Brown’s, used mortgages over ships under construction for which they held six-month bills as security for their loans from the Union Bank of Scotland.⁶⁰

20King and Nishimura’s assertion that the switch to lines of credit cut out brokers and middlemen ‘thereby gaining the rebate on cash payments’ remains unexplored. If brokers and middlemen could offer better prices than those at the factory gate business would continue and the fact that they often described themselves as ‘stockholders’ suggests they bore a proportion of the costs of financing inventories and in some cases had to process material before delivery. In any event the commodity trades were largely in the hands of brokers, who traded in huge volumes and developed the expertise in the futures market. The most plausible explanation is that local producers of materials, such as coal, iron and steel, and brokers for goods, such as raw cotton, rubber and wood, all switched to running accounts or circulating credit in the domestic market after the passing of the Sale of Goods Act that offered greater security, while financing their overseas trades where there was no such protection in bills. Dennys were sourcing the bulk of their iron directly from Mounsey & Co. at Bedlington in Northumberland far away in the north east of England as early as 1855, but also buying through local stockholders.⁶¹

21Although Dennys account with the Commercial Bank of Scotland was occasionally overdrawn, it was not until 1879 that they began heavily to depend on bank borrowing (Figure 5). This is well before the marked decline in inland bills in the early 1890s and at a time when Dennys were still using inland bills to provide working capital.

Figure 5. William Denny & Co., balances with the Commercial Bank of Scotland 1878‑1881

22This is exactly the period that Michael Collins in his analysis of the balance sheets of twelve bankers identified as a turning point: For the 12 banks during the 1870s the already persistent decline in the bills/advances ratio – from about 100 per cent to 60 per cent, 1868‑74 – was the most impressive feature. The ratio then levelled off, at about 50-60 per cent, for the remainder of the decade’. Like King and Nishimura he attributed this change to improved liquidity, but endorsed again with no evidence Nishimura’s claims about improved communications reducing ‘the necessity of businessmen to maintain large stocks of materials’.⁶² The evidence at least from Dennys would suggest that the switch to overdrafts and the decline in the use of inland bills was unrelated to the management of inventories and rebates on cash sales.

23Running accounts have a long history. It is impossible to find out much about such credits from a business’s financial records. It may be referred to in correspondence, but is much more likely to survive in the evidence in legal disputes and bankruptcy proceedings. In 1899 there was a dispute between the Spanish Government and the Clydebank Shipbuilding & Engineering Company, formerly J. & G. Thomson, about the late delivery of destroyers that were to be deployed in the war with the United States. The builders attributed part of the blame on ‘the unprecedented difficulty in obtaining the materials, especially the steel which was to be used in the construction of the vessels.

Figure 6. Denny’s discount rate as against Bank Rate and Monthly bill rate

24The Steel Company of Scotland admitted that there had been very heavy demand, but denied they had been slow in delivering orders. The reasons for the delay are of no interest here, but what is of interest is the terms of the contract. The Steel Company of Scotland explained that they had a ‘running account’ with Clydebank Shipbuilding & Engineering Company and had offered the following terms for the steelwork: ‘Free on trucks Clydebank. Cash less 5% discount on 10^th of month following date of invoices’, which had been accepted.⁶³ Even though Dennys held shares in the Steel Company of Scotland, they were not able to exact larger rebates than the standard 5 per cent for prompt payment.⁶⁴ However, not all contracts with the Steel Company of Scotland attracted rebates. The contract in 1883 to supply 12,000 tons of steel to Tancred, Arrol & Co. for the construction of the celebrated Forth Bridge to the north of Edinburgh was priced advantageously on the assumption that all the steel would be supplied by the Steel Company of Scotland. There is no mention of any discount for prompt payment.⁶⁵ This confirms that rebates were not automatic and, as the Sale of Goods Act made clear, the price was concurrent.

25It is only by case studies such as this that the use of inter-firm credit can be explored. It is painstaking and time-consuming research that depends on the survival of records. Fundamentally it depends on asking the essential question – how much did it cost? And where did the money come from? The cost can usually be found out, but unravelling the source of finance can be difficult, particularly when it was structured using a variety of facilities, cross-shareholdings, bill finance, mortgages and overdrafts. Multiple funding streams demanded effective financial management if success was to be guaranteed and failure avoided. By overlooking financial management and concentrating on other aspects of an enterprise, historians fail to address these two essential questions that speak directly to the events in 2008. There is nothing intrinsically wrong in securitizing debt and holding it off balance sheet as this paper has shown, provided the underlying security is good and there are remedies for failure to repay as provided in the Sales of Goods Act. The problem in 2008 was that long term mortgage debt was packaged and securitized and there was no straightforward mechanism for recovery when the value of the underlying security fell. The result was catastrophic and systemic failure of the financial system on an unprecedented scale.