Self-Fulfilling Prophecies of Failure: The Endogenous Balance Sheets of Distressed Companies

If only implicitly, discussions of corporate failure in law and economics inevitably call upon their sister discipline, accounting. In order to establish insolvency, law must ultimately compare the (accounting) value of assets with the value of liabilities. And, in order to establish whether the exit of a firm would add to efficiency, economics and finance must compare the (accounting) value of assets in their present use with their value in an alternative use. Yet the flourishing literatures on bankruptcy in those separate disciplines generally treat such accounting measures (of assets and liabilities) as straightforward, or else as outside their remit—in either case not warranting further discussion.

However, the major company failures of recent years have provided a reminder of the dangers of ignoring the accounting issues which arise with distressed businesses. Balance sheet manipulation was used to mask Enron's true indebtedness, and earnings management was being used to flatter the earnings being achieved by the company's assets in their current use.¹ Nor are such accounting responses to financial distress exceptional: Sweeney's (1994) econometric work, for example, shows that commonly ‘managers of firms approaching default respond with income-increasing changes’ (p. 307): faced with financial distress they responded with accounting changes—a process vividly narrated in the Australian case studies of Clarke et al. (1997). Moreover, it is not just the executives of distressed firms who exploit accounting to escape the disciplines of law and economics: in the Savings and Loan crisis of the 1980s, the regulatory authorities responded to the imminent bankruptcy of many institutions not with legal, economic or financial measures but, in effect, with changes in accounting rules which redefined insolvent businesses as solvent, and staved off their collapse.²

Such creative accounting in the face of financial pressures supports our general position that accounting matters in the process of company failure. But it is not the focus of this article, which instead investigates accounting problems which—even in the absence of creative accounting—modify in unexpected ways the conclusions of the law and economics literatures on inefficient bankruptcy. In particular, it shows that neglected problems of asset and liability measurement for distressed firms lead to economically inefficient decisions on exit, whether in a debtor-oriented legal regime such as the U.S. or the creditor-oriented legal arrangements of countries such as the U.K.

Section 1 outlines formally for the U.S. the place of balance sheet valuations in, first, the efficient exit condition in economics, and, second, the insolvency condition in law. Section 2 generalizes the framework to other regimes, showing that the balance sheet occupies a pivotal role in determining insolvency in jurisdictions outside the U.S., even where, prima facie, they accord lower or zero weight to the balance sheet in their bankruptcy codes (a contention supported empirically in the Appendix, with an analysis of some 70,000 sets of financial statements for U.K. listed companies). Section 3 then explores a disconcerting indeterminacy in the balance sheet valuations used to establish insolvency: it collates evidence from finance, economics, law and accounting to show that assets and liabilities are continually revalued (with the effect of eroding the residual, equity) as a company becomes financially more distressed. The probability that a firm will fail affects to an economically significant extent the valuations assigned to its assets and liabilities. At the same time, of course, the valuation of assets and liabilities determines how close the firm is to insolvency—in other words, its probability of failure! The conundrum then is that it becomes impossible to establish whether a firm is insolvent independently of the markets’ views of its survival prospects. Section 4 introduces very simple models to show some implications of this endogeneity in the balance sheet. First, under (debtor-oriented) U.S. Chapter 11, reallocation of control from an inferior incumbent management team is inhibited and delayed—agency costs are higher because of the endogenous balance sheet, and they are greater, the larger is the asset erosion analysed in this paper. Second, under the (creditor-oriented) U.K. receivership/administration code, unnecessary and costly bankruptcies are precipitated by the endogenous balance sheet; and the greater the asset erosion, the greater is the incentive to reach a Pareto-inferior outcome. Recent cases illustrate questionable countermanding of huge financial claims as a result of this endogeneity—triggered by false forecasts, rumour, or rulings by court or government.

1: COMPANY FAILURE IN LAW AND ECONOMICS, EXPRESSED THROUGH THE BALANCE SHEET

The economic criterion for company failure can be written as:

where:

A_pv is the present discounted value of net future cash flows generated by the company's assets if they are retained in their existing use, and

A_nrvis the net realizable value of the assets, if they are sold for an alternative use.

This is simply saying that economic efficiency would be enhanced if a company discontinued once its assets could yield higher returns in an alternative use.

One legal criterion for company failure might be written:

where A is the value of the firm's assets (we leave the valuation basis undefined at this stage) and L is the value of the firm's liabilities (debts).

Thus the U.S. bankruptcy code of 1978 states: ‘ “Insolvent” means, in relation to an entity, that . . . the sum of such entity's debts is greater than all of such entity's property, at a fair valuation’.³

Figure 1 enables us to categorize the firms which fall foul of one or more of these inequalities, as well as the families of critiques presented in the law/economics/finance literatures of present bankruptcy procedures. To help develop the argument, failing firms are shown as two intersecting sets. On the one hand, members of set x are those firms which are liable to failure because they have violated the legal insolvency condition—assets fall short of liabilities. On the other, set z comprises those firms which economics would single out for failure because their assets could be better employed elsewhere (A_pv < A_nrv). Sets x and z could be disjoint, given that the two failure criteria do not coincide. The important feature of the figure, for the literatures in law, economics and finance, is that sets x and z are likely to differ.

Subset p has been analysed, for example, by Jackson (1982) and Webb (1991): on legal grounds they are liable to fail (A < L: they are insolvent); but on economic grounds the assets would best remain employed in the present business: the business is worth more as a going concern than if it is liquidated. But Jackson and Webb have analysed cases where, under a creditor-oriented legal system (such as the U.K.'s), the incentive system may induce destructive creditor races, and economically undesirable exit.

Subset y, the intersection of x and z, represents the fortunate cases where economists and lawyers agree the firm should fail: the assets would generate greater returns in an alternative use, and the legal system provides the powers to achieve this.

Finally, subset q comprises the firms which, on economic grounds should be disbanded (A_pv < A_nrv), but are not compelled to by the legal system. Such companies are the focus of Jensen's (1986) analysis of cash-rich firms free to squander stockholders’ money because of agency problems; and of his prescription of raising their leverage, thus compelling them to return some cash flow to wealth-holders every year in the form of interest payments: unless their performance improves they will ultimately be driven into the insolvency net, with the prospect of law intervening to part executives from their under-performing assets. A related case analysed by Stiglitz (1972) (see also White, 1992) has managers continuing in business under the protection of Chapter 11, gambling with stockholders’ money, even though liquidation is warranted on economic grounds.

So the precise relationship between debts, the disposal value of assets, and their value in their present use, is crucial in triggering bankruptcy, and in determining whether firms which should fail, on economic grounds, are actually singled out for failure by the legal system.

2: GENERALIZING THE INSOLVENCY CONDITION TO JURISDICTIONS OUTSIDE THE U.S.

Figure 1 and the discussion in section 1 incorporate a balance sheet test of insolvency: following the U.S. model, insolvency has been defined as negative equity—assets fall short of liabilities. But this is not the test incorporated in the law of some other jurisdictions. For example, at the opposite end of the spectrum is Australia, under whose law the only test of insolvency is (Corporations Law (Aust), s. 95A):

This is a cash flow, rather than a balance sheet test. It asks not whether liabilities exceed assets, as in the U.S., but whether cash flows are sufficient to meet the payments which the company is required to make. U.K. law⁴ appears to be a hybrid of U.S. and Australian rules: it includes a cash flow test, one component of which is (Insolvency Act 1986, s. 123): ‘(1) {c} . . . the company is unable to pay its debts as they fall due’. But it also includes a balance sheet test: ‘(2) A company is also deemed unable to pay its debts if it is proved to the satisfaction of the court that the value of the company's assets is less than the amount of its liabilities’. However, Sealy (1999) explains that closer analysis of actual U.K. cases shows that: ‘in practice virtually all orders are made on the basis of the presumption that follows the failure to comply with a statutory demand. This is essentially “cash flow insolvency”.’

Does this mean that section 1's discussion of balance sheet insolvency is irrelevant in these other jurisdictions? We argue not. Closer analysis of the dynamic interaction of ‘flow’ and ‘stock’ financial statements during financial distress suggests that the balance sheet is still crucial. Suppose an Australian company (in that cash flow bankruptcy regime) has encountered a cash flow difficulty and does not have cash in the bank with which to pay a ‘debt due and payable’. What happens next? Managers (and equity-holders) have strong incentives to avoid the legal process of failure, since it entails bankruptcy costs (elaborated, e.g., in Altman, 1984; Andrade and Kaplan, 1998; and section 4 below), and would typically deprive equity-holders of their wealth, and managers of their control and their income. They can therefore be expected to seek cash with which to meet the debt. They could do this by selling off any non-cash assets which are not required for operations (surplus inventories, real estate or equipment): DeAngelo et al. (2002) chart this process in their clinical study of L.A. Gear—the company funded no less than $180 million of cash losses through devices such as selling excess inventories to deep discount outlets. Or, if there were no such surplus assets, they could borrow against the assets they wished to continue operating. For example, in the case of real estate they could construct a mortgage or a sale and lease back arrangement; in the case of inventories they could follow common practice and use these as collateral for short-term bank borrowing; and in the case of receivables, they could adopt factoring to ‘turn them into cash’ ahead of the due payment day.⁵

What is the limit on such borrowing? For a credit-worthy company operating in efficient capital markets, the limit will be set by its potential collateral. In other words, the question will be: does it have more assets than liabilities? Cash flow insolvency will only matter, therefore, if it is accompanied by balance sheet insolvency: A < L. Otherwise the rational manager will solve the problem of cash flow insolvency with some form of re-financing package.⁶

This rather dry point can be illustrated with the most famous statement of cash flow insolvency to be found in English literature: ‘Annual income twenty pounds, annual expenditure nineteen, nineteen, six, result happiness. Annual income twenty pounds, annual expenditure twenty pounds, ought six, result misery.’

Our modification of Mr Micawber's position in David Copperfield lacks Dickens’ elegance, but shows how taking account of the balance sheet, alongside the cash flow statement, can transform a firm's position:

The Appendix provides some corroborating evidence for listed U.K. companies. The experience of the listed company population is consistent with the proposition that balance sheet insolvency is sufficient (though not necessary) for failure.

3: THE ENDOGENOUS BALANCE SHEET: ASSET EROSION AS FINANCIAL DISTRESS INCREASES

The earlier discussion having argued that, ultimately, whatever the jurisdiction or the wording of the regulations, a balance sheet test is required to establish insolvency, this section now shows that the balance sheet itself depends crucially on whether the company is deemed near to insolvency or insolvent. The probability of the company failing, and the market's perception of that probability, are likely to affect the valuation of both assets and liabilities. This creates a disconcerting indeterminacy: the markets look to the balance sheet to establish whether a firm is insolvent, yet it is not possible to establish whether a company is insolvent without first knowing the markets’ view of its survival prospects.

Table 1 traces changes in the balance sheet in various phases of financial distress, to suggest at which points in the failure process assets might be eroded (or, equivalently, liabilities increased). An asterisk (*) marks each combination of phase and balance sheet component where the erosion effect of the endogenous balance sheet is to be expected. The balance sheet headings are highly aggregated, for simplicity. And the phase headings are intended only to be suggestive: they indicate successive steps in financial decline, but these will take a different form in different jurisdictions. Phase 1, early distress when a firm finds difficulty in paying bills as they fall due, is common to all jurisdictions, as is phase 5, liquidation. But some of the intervening phases differ significantly across jurisdictions (see section 4): the U.K. has no Chapter 11 (phase 3); and the appointment of a trustee in the U.S. is not formally equivalent to the appointment of a receiver or, more recently, an administrator in the U.K. (phase 4)—they are bundled together here just to suggest the legal state of play fairly late in the process of failure.

The first of the asset categories, cash, is depleted from phase 3 onwards by ‘direct bankruptcy costs’—court fees, professional fees, etc. associated with administering a bankruptcy under court supervision. This is the most familiar and best documented asset erosion in the literature: Altman (1984) estimates typical orders of magnitude as 6 per cent of firm assets. The figure rises, the more protracted and complicated the bankruptcy: case studies reveal figures such as Wickes’$250 million (Altman, 1984), Eastern Airlines’$114 million (Weiss and Wruck, 1998), and around $1 billion for Enron and for WorldCom/MCI.⁷

Altman (1984) also introduced a second reason for cash to be eroded, but this from the very earliest phases of distress, and accumulating further if the firm moves through to later phases: indirect bankruptcy costs. These are the profits lost because of the damaging effect upon the firm's product market arising from news of financial distress. Doubts about the firm's survival can influence firms’ customers in numerous ways. For example, automobile purchasers fear for the availability of spare parts, the value of their warranties,⁸ and the residual value of a product if a brand has discontinued. Thus, in the U.S., Chrysler lost sales representing no less than 2 per cent of the whole national car market because potential buyers feared the company would go bankrupt (Altman, 1984, p. 1071); and in the U.K., Leyland DAF saw its market share fall from 19 per cent to just 6 per cent in three months after the appointment of a receiver (phase 4 in Table 1) (Franks et al., 1996). Airline passengers develop fears about the airworthiness of aircraft operated by a distressed airline (Altman, 1984). And the clients of insurers fear that their policies will become worthless: the clients of First Executive dramatically increased the surrender rates of life insurance policies as doubts were published about the firm's survival prospects (DeAngelo et al., 1994). In the latter case, the adverse effects began eroding the firm's assets in phase 1—once adverse publicity surfaced. But the erosion may increase sharply as the process moves on to a new phase. This became a bargaining factor at Eastern Airlines: the disgruntled creditors’ committee was deterred from exerting public pressure on the trustee to shut down by the knowledge that this would in the meantime depress sales, further increase losses and deplete their eventual payout (Weiss and Wruck, 1998). The importance of these factors will obviously vary a good deal from product to product: typical orders of magnitude suggested by Altman's (1984) study were 10 per cent of assets—more than his estimates of direct bankruptcy costs.

Cash may be depleted for a third reason at the liquidation stage.⁹ In some jurisdictions, including the U.K., liquidation crystallizes a liability which had previously only been contingent, and therefore left off the statutory balance sheet: that for non-discretionary redundancy payments. These will deplete the firm's assets by more, the greater its reliance on labour inputs, and the longer its workforce has been employed by the firm. In one of the U.K. cases analysed by Armour and Deakin (2003), an automobile producer with a large labour force whose members typically had long records of service, the potential liability summed to some hundreds of millions of pounds.

Moving down the balance sheet in Table 1, once the firm has passed to the later stages of distress, and is unlikely to survive, the cash realized from receivables is likely to fall. The increased uncertainty about the distressed firm's future as a trading partner reduces the incentive for the debtor to maintain good business relations: the sanction of cutting off supplies has less force (no force at all by the liquidation phase). Debtors can make complaints, real or contrived, in an effort to delay or avoid payment. Armour and Frisby (2001) provide telling interview evidence of this response during U.K. receiverships, as one interviewee explained: ‘What started out as a million pounds of book debts becomes a million pounds of disputes and aggravation’ (p. 95). The recovery rate of receivables is clearly likely to fall as the firm moves through successive phases of distress, most sharply when longer term trading relationships are broken as control shifts from the firm's directors to a trustee, receiver or liquidator.

The next asset category in Table 1's balance sheet, Inventories, will be threatened by some of the same forces as depleted cash. As loss of customer confidence in the brand of the distressed firm erodes profit margins, the firm and its auditors will have to assess whether the recoverable value of inventories has fallen below their cost; and following GAAP, inventories will have to be marked down where realizable value has fallen below cost. The impact of this will again vary greatly across industry, being particularly powerful where brand differentiation is strong and where the production and distribution process requires large inventory holdings.

In the case of real estate, plant, etc., the next balance sheet category in Table 1, there are again reasons why valuations will fall in successive phases of distress, but the mechanisms are rather different from those for inventories. Before a firm becomes distressed, its real estate, plant, etc. will typically have been recorded in the balance sheet at cost.¹⁰ Implicitly it will be assumed that the firm as a going concern would recover these costs (via depreciation provisions and residual values) in the course of future business. Once the going concern assumption can no longer be maintained, however, the directors and auditors must consider realizable value. And this is likely to be smaller, in relation to cost, the less fungible (or ‘redeployable’ in Williamson's, 1996, terms) the firm's assets are to use in other activities. In a case study of the specialist assets of a machine tool manufacturer, Holland (1990) cites discounts of 50 to 70 per cent on a distress sale. Shleifer and Vishny (1992) discuss the common case where the problems are compounded because the distressed firm's asset sale is precipitated by an industry- or economy-wide exogenous shock, of the sort analysed by Goudie and Meeks (1991, 1998): the most likely potential buyers—members of the same industry—‘themselves are hurting’ and ‘are unlikely to be able to raise funds to buy distressed firms’ assets’ (Shleifer and Vishny, 1992, p. 1344). Pulvino (1998) provides some ingenious estimates of the impact of financial distress on equipment values for the U.S. airline industry. When the industry was depressed he concluded that: ‘[distressed] airlines with above-industry-median leverage ratios and below-industry-median current ratios . . . sell aircraft at a 14% discount to the average market price’. When aircraft were sold outside the industry, particularly big discounts were applied: during market recessions financial institutions paid a discount of 30 per cent to average market price. And this was recognized by professional valuers:

Again, Shleifer and Vishny (1992) cite rule of thumb discounts of 15 to 25 per cent for a rapid sale compared with the proceeds of an orderly sale over several months, even with relatively fungible real estate assets.¹²Brown (2000) provides evidence in the same vein for U.S. commercial real estate in the late 1980s and early 1990s. And Andrade and Kaplan (1998) report that, of their sample of 31 financially distressed businesses (all of whom had positive operating income and eventually survived, but suffered distress because of high leverage) twenty-two reported ‘desperation asset sales’, and ten of these with ‘proceeds less than expected’ (p. 1475).

The final category of fixed assets, intangibles, can also be expected to decline, ceteris paribus, as financial distress intensifies. This is because a number of the components of intangibles will be valued on the basis of discounted abnormal future returns attributable, for example, to brands. Impairment tests specifically provide for such assets to be written down as the estimates of future cash flows are revised downwards. Just as net current cash flows will be eroded because of the indirect bankruptcy costs identified and quantified in the literature (see above), so, pari passu, discounted future cash flows embodied in certain intangibles will decline too.¹³ Thus an impairment review by Marconi in 2002 led to £3677 million of goodwill disappearing from its balance sheet, equivalent to 48 per cent of its total assets at the beginning of the year.

As the probability of failure grows, the value of some liabilities may begin to change adversely too. Wilner (2000) analyses the effect of financial distress (corresponding to phase 1 in Table 1) on firms’ use of trade credit. He reports that common terms of sale are 2/10 net 30: the full purchase price is payable within thirty days, but a 2 per cent discount is obtained if payment occurs within ten days of sale. As the implicit annual interest rate for taking advantage of the supplier's credit from the eleventh day to the thirtieth is 44.6 per cent, only a distressed firm with interrupted access to ordinary lines of credit will make use of this source; and the consequence is a higher effective value for trade credit: residual equity is further diminished.

Some comparable developments occur with longer term liabilities as a firm becomes distressed. One mechanism is that, when the markets perceive distress, credit ratings will fall, debt covenants may be triggered, and the terms on which the business can renew loans will deteriorate: interest liabilities will rise for a given sum of credit. Neal and Rolph (1999) give the example of Daiwa Bank to illustrate this process. Some modern credit instruments embed such a process in a contract, so that, for example, an existing liability increases automatically (the interest rate rises) if the firm fails to meet a specified target within its financial statements. Lawton (1999) provides a general discussion. Asquith et al. (2005) analyse this development: the debt to EBITDA ratio is the most commonly used performance measure; and in an example they describe (BWAY Corporation) the interest rate spread over LIBOR could be reduced to 50 basis points or raised to 225 at the end of each quarter, depending on the debt to EBITDA ratio for the preceding four quarters. This would produce a similar result to a conventional loan covenant which enabled the lender to call in a loan if the borrower's financial statements failed to meet some predetermined target (i.e., in the early phase of financial distress, before default), and the borrower then had to contract new borrowings from a bargaining position weakened by its diminished credit rating.¹⁴ Both the performance pricing contract and the traditional loan contract are, of course, designed to shift onto the borrower some of the costs of rising credit risk; so inevitably, in terms of Table 1, they erode the borrower's net assets, one way or another.¹⁵

The final liability category in Table 1's balance sheet is the finance lease. Suppose the firm has assets held on finance lease, and default on the terms of the lease because of cash flow problems empowers the lessor to repossess the asset. On default, the asset disappears from the balance sheet and so does the liability to the lessor. This obviously may have damaging consequences for the firm's ability to continue operating. But equity is also likely to suffer directly too, because the liability will generally be less than the value of the asset (reflecting the lessor's safety margin); and this part of equity will pass automatically to the lessor.¹⁶

Thus the row totals in Table 1 show the aggregate erosion of particular classes of asset (or, correspondingly, increase in liability) over all phases of distress and failure. And the column totals show the aggregate depletion of net assets in each phase of distress approaching failure. Section 4 now shows how the asset erosion in a particular phase impinges on the exit decision.

4: IMPLICATIONS OF NET ASSET EROSION FOR THE EFFICIENT RE-ALLOCATION OF CORPORATE CONTROL UNDER BANKRUPTCY

5: CONCLUSION

Section 1 showed formally the relation between the insolvency condition in law and the efficient exit condition for a firm in economics. The two cannot be relied upon to coincide, as the modern literature in law and economics has made clear. This formal presentation focuses on valuations of assets and liabilities in the balance sheet—a treatment which corresponds with the institutional framework in the U.S. In some other jurisdictions, however, such as the Australian and British, bankruptcy law is couched in terms not of balance sheet insolvency (or, in the British case, not just of balance sheet insolvency) but of cash flow insolvency.

Nevertheless, section 2 shows that ultimately the balance sheet is pivotal in these regimes too: a priori, the incentive scheme facing managers combines with the dynamics of corporate finance to mean that cash flow insolvency is not binding unless it is accompanied by balance sheet insolvency.

Section 3 then reports a series of case material from several disciplines on valuing different assets and liabilities at different phases of corporate financial distress. This shows the fragility of the balance sheet numbers. Assets and liabilities are continually re-valued (almost always with the effect of eroding equity) as the firm moves to more severe states of financial distress. The processes include the absorption of cash by ‘direct’ and ‘indirect’ bankruptcy costs, the erosion of receivables as bargaining power vis-à-vis the debtor diminishes, declining inventory values, problems in liquidating long-term assets which in Williamson's (1996) terms are not readily ‘redeployable’, the impairment of intangibles, the crystallization of previously contingent liabilities, and the swelling of liabilities as perceived credit risk increases.

The probability that the firm will fail has a serious effect on the valuations assigned to its assets and liabilities. But, of course, section 1 showed that managers, stockholders and creditors look to the valuations of assets and liabilities to determine how seriously distressed a firm is—its probability of failure.

Section 4 traces some implications of this disconcerting interrelationship (valuations influence the probability of failure, yet the probability of failure influences the valuations). Two simple models are introduced. One is set in a U.S. (debtor-oriented) institutional regime. And we show that the endogenous balance sheet raises agency costs, because it inhibits the reallocation of control from inferior management under Chapter 11. Under the (creditor-oriented) regime of the U.K., on the other hand, the same balance sheet problem has the opposite effect: we show that it provokes unnecessary costly bankruptcy.

The more significant are the valuation problems we outline, the more wasteful they make the reallocation of control and claims in the U.S., and the more likely they make the economically inefficient reallocation of control and claims in the U.K.

Four recent cases illustrate the problem in practice: questionable countermanding of huge financial claims followed false forecasts of insolvency, rumours of insolvency, and court or government ordered (counterfactual) assumptions of insolvency.

6: SOME IMPLICATIONS FOR PRACTITIONERS

APPENDIX
BALANCE SHEET INSOLVENCY IN THE U.K. LISTED COMPANY SECTOR

Table 3 explores the role of balance sheet insolvency in practice—again from London, but a century and a half later than Dickens. It addresses the question: is balance sheet insolvency sufficient for company failure? And it asks this question through an analysis of the balance sheets of a population of large companies. This analysis identifies any company which survives despite balance sheet insolvency. If many companies survive despite balance sheet insolvency, this will challenge our contention that this criterion warrants the priority we have accorded it.

Balance sheet insolvency (BSI) is defined as occurring where:

And we test for this inequality in the annual published accounts of every non-financial company listed on the U.K. Stock Exchange in the period 1948–90 and fulfilling certain minimum size conditions.²²

Table 3 reports that 5,088 companies qualified for inclusion in the test, though not all of these were in the population for the full forty-two years. These companies had 71,111 qualifying annual financial statements.

How often did these companies survive despite BSI? Of the 5,088 companies, just twenty-nine exhibited BSI at any point in their lives in the dataset—0.6 per cent of the total. This corresponds to just 0.1 per cent of the available company-years. Very few companies indeed within this population remained in business with BSI long enough to publish their annual financial statements.

Moreover, of those twenty-nine companies which did not at first sight comply with the proposition that BSI is sufficient for failure, all but two did subsequently die—with a lag. Four went into receivership, nine were taken over and fourteen died of other causes. The remaining two, which defied our sufficient condition, turned out to be special cases in that, although they enjoyed an independent listing on the U.K. Stock Exchange, they were recognizably not financially independent entities: for example, one was Vauxhall Motors, whose finances were intimately connected with General Motors in the U.S.

These results are, then, consistent with the proposition that, although the U.K. system is on a first reading of legal statute and custom driven by cash flow insolvency, in practice it does not tolerate balance sheet insolvency for this population of companies. If cash flow insolvency had been the only effective criterion for failure, so that BSI was of no significance in the process, many more cases of BSI might have been expected; and the discussion of balance sheet movements in this article would have been much less significant.

We are not, of course, arguing that BSI is necessary for failure, and therefore could represent, by itself, a reliable predictor of failure. There are various other sufficient conditions for liquidation which are not necessarily associated with BSI—for example, if the present value of the company's assets is damaged by an exogenous development such as an oil shock, or the net realizable value of its assets in an alternative use rises because of a government planning decision, so that exit becomes efficient.