Anatomy of a Financial Fraud A Forensic Examination of HealthSouth By Leonard

Anatomy of a Financial Fraud

A Forensic Examination of HealthSouth

By Leonard G. Weld, Peter M. Bergevin, and Lorraine Magrath

A forensic audit conducted by Pricewater- houseCoopers concluded that HealthSouth Corporation’s cumulative earnings were overstated by anywhere from $3.8 billion to $4.6 billion, according to a January 2004 report issued by the scandal-ridden health-care concern. HealthSouth acknowledged that the forensic audit discovered at least another $1.3 billion dollars in suspect financial reporting in addition to the previously estimated $2.5 billion. The scandal’s postmortem report found additional fraud of $500 million, and identified at least $800 million of improper accounting for reserves, executive bonuses, and related-party transactions. This billion-dollar-plus admission failed to garner financial media headlines, further evidence of the public’s inurement to financial reporting scandals.

Contemporary corporate fraud in the United States has affected market values, decimated private 401(k) plans, and devalued public pension funds. Tyco, Dynegy, WorldCom, and others joined Enron’s fraudulent accounting ranks in 2002. Birmingham, Alabama–based HealthSouth became a member of that disreputable group a year later. Moreover, numerous lesser-known companies have also engaged in a host of improper and illegal accounting activities.

The Magnitude of the Problem

On January 23, 2003, the SEC issued its “Report Pursuant to Section 704 of the Sarbanes-Oxley Act of 2002.” Section 704 directed the SEC “to study enforcement actions over the five years preceding its enactment in order to identify areas of issuer financial reporting that are most susceptible to fraud, inappropriate manipulation, or inappropriate earnings management.” The study period began July 31, 1997, and ended July 30, 2002.

Over the study period, the SEC filed 515 enforcement actions for financial reporting and disclosure violations arising out of 227 separate Division of Enforcement investigations. Those investigations fell into three categories:

Revenue recognition, including fraudulent reporting of fictitious sales, inaccurate timing of revenue recognition, and improper valuation of revenue.

Expense recognition, consisting 
of including improper capitalization or deferral of expenses, incorrect use of reserves, and other understatements of expenses.

Business combinations, relating to myriad improper accounting activities used to effect and report combined entities.

All but one of these investigations included revenue-related issues, and many investigations identified violations in two or all three of these categories.


On March 19, 2003, the SEC charged HealthSouth and its CEO, Richard Scrushy, with accounting fraud. The SEC’s complaint alleged that HealthSouth had systematically overstated its earnings by at least $1.4 billion since 1999. Apart from the SEC’s finding, the U.S. Justice Department used information gathered from HealthSouth executives to identify another $1.1 billion of overstated earnings. The PricewaterhouseCoopers forensic study indicates these figures were underestimated, but the specifics of those inaccuracies are not a matter of public record because the company announced that reaudited financials would be disclosed no sooner than 2005.

Fifteen HealthSouth accounting and finance executives pled guilty by the end of 2003, and Scrushy remained under indictment in mid-2004 on 85 criminal charges.

The SEC executed a court order, pursuant to section 1103 of the Sarbanes-Oxley Act, that—

Required the company to place in escrow all extraordinary payments to its directors, officers, partners, controlling persons, agents, and employees;

Prohibited the company from destroying documents relating to financial activities or allegations in the case against HealthSouth and Scrushy; and

Provided for expedited discovery.

The Auditors

HealthSouth paid the Birmingham office of Ernst & Young LLP $3.6 million for its 2001 financial statement audit and related services. Ernst & Young disavowed knowledge of the fraud, citing systemic deception on the part of HealthSouth executives, several of whom have pled guilty to fraud charges. Communication about questionable activities took place between the health-care provider and its auditor, however. For example, in a hearing to decide if Scrushy’s assets should be unfrozen, two Ernst & Young partners stated that the audit firm had received an e-mail from a HealthSouth employee advising them to examine three specific accounts for fraudulent entries related to asset capitalization.

Ernst & Young subsequently contacted HealthSouth’s president and chief operating officer, William T. Owens, and the chairman of its audit committee, George Strong. Owens defended HealthSouth’s capitalization method, but agreed that further investigation was needed. Both Owens and Ernst & Young partner James Lanthron eventually concluded that no costs were improperly capitalized. Ernst & Young did not detect or investigate beyond the scope of normal audit procedures any other substantive questionable activities outside of the capitalization issue.

Meeting Analysts’ Expectations

SEC Director of Enforcement Stephen Cutler stated in a March 19, 2003, press release that “HealthSouth’s standard operating procedure was to manipulate the company’s earnings to create the false impression that the company was meeting Wall Street’s expectations.” This motive is not a new one.

In general, analysts’ expectations and company predictions address two high-profile components of financial performance: revenue and earnings from operations. The pressure to meet revenue expectations is particularly intense and may be the primary catalyst leading managers to engage in earnings management practices that result in questionable, improper, or fraudulent revenue-recognition practices. A Financial Executives International (FEI) study, for example, found that improper revenue recognition practices were responsible for one-third of all voluntary or forced restatements of income filed with the SEC from 1977 to 2000.

The drive to meet analysts’ expectations was brought to public attention in 1998 by then SEC Chairman Arthur Levitt in his landmark speech “The Numbers Game.” Levitt expressed his concern that too many corporate managers, auditors, and analysts let the desire to meet earnings expectations override good business practices. He called for a fundamental cultural change on the part of corporate management and the entire financial community. Levitt identified several examples of what he called accounting “hocus pocus”: “big bath restructuring charges, creative acquisition accounting, cookie-jar reserves, immaterial misapplications of accounting principles, and the premature recognition of revenue.”

Early Warning Signs

The systematic and substantive HealthSouth fraud raises certain questions:

Should the auditors have suspected fraud?

More important, could the auditors have detected the financial statement manipulations and exposed them?

Given that the auditors did not detect fraud, could the investment community have done so through careful examination of the financial statements?

If so, what tools would help financial statement users detect the fraud?

The CPA Journal article “Abusive Earnings Management and Early Warning Signs,” by Lorraine Magrath and Leonard G. Weld (August 2002), distinguished between earnings management activities that are simply good business practices and abusive earnings management intended to deceive the financial community. Good business practices include the following activities:

Careful timing of capital gains and losses;

Use of conferencing technology to reduce travel costs; and

Postponement of repair and maintenance activities when faced with unexpected cash flow declines.

On the other hand, abusive earnings management results from actions such as those cited in the SEC section 704 Report:

Improper revenue recognition;

Improper expense recognition; and

Using reserves to inflate earnings in years with falling revenues.

Magrath and Weld identified six relationships that investors and auditors should consider as early warning signs of abusive earnings management:

Cash flows that are not correlated with earnings;

Receivables that are not correlated with revenues;

Allowances for uncollectible accounts that are not correlated with receivables;

Reserves that are not correlated with balance sheet items;

Acquisitions with no apparent business purpose; and

Earnings that consistently and precisely meet analysts’ expectations.

Analyzing HealthSouth’s Disclosures

If analysts, investors, or auditors had examined these relationships, would there have been reason to suspect abusive earnings management at HealthSouth? In retrospect, the answer is a qualified yes. Investors or auditors might have detected abusive earnings management if they had understood the context of the financial statements as well as their content, and if they had thoroughly analyzed specific early warning signs of earnings management.

As HealthSouth’s credibility unraveled in full public view, it became apparent that for many years its financial disclosures had neither represented economic reality nor conformed to GAAP. The company acknowledged as much in its Form 8-K filed with the SEC on March 26, 2003:

The company also announced today that, in light of the recent Securities and Exchange Commission and Department of Justice investigations into its financial reporting and related activity calling into question the company’s previously filed financial statements, such financial statements should no longer be relied upon.

Whether the auditors should have detected a scheme that originated at upper management levels remains a matter of conjecture; regardless, investors and other financial statement users lacked reliable data.

Correlation analysis would not have revealed “accounting hocus pocus” on the part of HealthSouth. Exhibit 1 presents correlation coefficients for 1993 through 2001 (the last year HealthSouth filed a Form 10-K). Exhibit 1 indicates a positive correlation between each pair of accounts. Although their relationships vary to a certain extent, HealthSouth’s accrual-based numbers were positively related to each other, and cash flows mirrored income.

Lack of information precludes correlating reserves with balance sheet items, the fourth early warning sign of abusive early management. HealthSouth met the final indicator of earnings abuse: The company matched analysts’ expectations for 48 consecutive quarters through mid-1998. That unerring track record could have been attributed to predictable operations and acceptable income-smoothing techniques, but hindsight proved otherwise.

Upon examination of the techniques used by HealthSouth, the authors offer two techniques that investors and auditors can use to raise red flags about abusive earnings management practices: a more detailed analysis of receivables, and a link between cash flows and an array of performance measures.

Accounts Receivable Analysis

A detailed analysis of credit sales can help unearth abusive earnings management, even if the primary accounts create an illusion of successful performance. This technique requires an investigation of all of the accounts comprising net receivables, as well as their interrelationships, in order to identify potential problems. Public disclosures, even in the most reliable financial statements, do not provide such data in a user-friendly format. Often, companies provide information about bad-debt expense, the amount of accounts written off as uncollectible, and estimates of doubtful accounts in an Item 14 valuation schedule toward the end of the 10-K, far removed from the financial statements and the notes in Item 8. Alternatively, some companies disclose estimated uncollectible accounts as a parenthetical disclosure to net receivables, and break out bad-debt expense on the income statement. This method, however, neglects to report the dollar amount written off as uncollectible.

HealthSouth disclosed its receivable-related activities on the face of its financial statements. The relatively large percentage of receivables estimated as uncollectible is not surprising, given an industry dependent on third-party payments from Medicare and insurance companies. This component’s volatility, ranging from 38.9% of gross accounts receivable to 12.2%, is troubling, however. Exhibit 2 extends the analysis of HealthSouth’s receivables, presenting the company’s allowance balance and expense disclosures and using those numbers to determine the dollar amount written off as uncollectible.

In addition to this volatility, HealthSouth’s 1999 provision for doubtful accounts is an outlier when compared to other years’ bad-debt expense. It may be only coincidental that 1999 was the first full year in which investors knew the company would not meet previously announced earnings projections. Exhibit 3 presents bad debt expense as a percentage of annual sales.

Annual write-offs for uncollectible receivables lacked any consistency whatsoever. Moreover, the amount of the accounts written off in any given year did not correlate with the allowance established for them. To the extent that these disclosures were reliable, these data indicate that HealthSouth used bad-debt reserves to manipulate earnings. This lack of correlation could be an indirect indicator of the fourth warning sign of abusive earnings management: reserves that are not correlated with balance sheet items.

Consider two related items related to uncollectible accounts, to further understand possible earnings management. First, there were disproportionately large allowance for doubtful accounts balances at the end of 1994 and 1995: nearly 40% of gross receivables. These existing balances could have been drawn down without the need to record an accurate provision for doubtful accounts. Consequently, the year-end charge to bad expense, required to replenish the depleted contra-asset account, could have been less than normally expected if the unadjusted allowance balance jibed with economic reality. By understating expenses in this manner, HealthSouth could have manufactured earnings beginning in the mid-1990s. These data provide some evidence of the classic “cookie-jar reserve” ploy.

The related issue is the amount of bad-debt expense matched against revenues in 1999. As noted, that unusually large charge to earnings was made at the time when it became publicly known that HealthSouth could no longer hit its earnings target. Company officials may have decided to replenish the balance in the allowance account (add cookies to the cookie jar) or recognize previously understated levels of bad-debt expense. In either case, the large charge (8.4% of revenues) occurred when Wall Street diminished its earnings expectations for HealthSouth. Taking this “big bath” for bad debts merely exacerbated 1999’s already poor financial performance, information that was already discounted in the marketplace. The question arises: Did HealthSouth bury this apparently inflated expense amount within a sea of red ink in an attempt to manage earnings?

An analysis of receivables composition sheds light on the potential to overestimate cash collected from sales. Exhibit 4 shows the trend of cash collected from customers, determined by two different methods. The first metric is the conventionally simplified version used to convert indirect cash inflows to direct ones. It merely adjusts revenues for changes in net accounts receivable to determine cash collected from customers in a given reporting period. The deficiency in this method is that net receivables result from the interaction of account write-offs and bad-debt expense recognition. Taken together, these entries affect the allowance’s ending balance, which, in turn, determines net receivables. Bad-debt expense, however, is a noncash charge, and has no bearing on operating cash inflows; nor has bad-debt write-offs. They must be subtracted from revenues, however, along with the adjustment for changes in gross receivables, to accurately determine cash collected from customers. The difference between cash flows determined by the net and gross methods can be material if write-offs are significant, as they were in 1999. Cash collections should be constant over time, regardless of the method used, inasmuch as accounts written off and the bad-debt expense recognized are usually comparable in any given reporting period.

HealthSouth’s cumulative cash inflows as a percentage of sales equaled 93.5% on an adjusted (gross-receivable) basis, as opposed to 97% on an unadjusted (net- receivable) basis. This translates into a reduction of $725 million in cash collections from 1994 through 2001, which matches the write-offs from Exhibit 2. Annual percentages, however, differ from those geometric means, as evidenced by the data points in Exhibit 4. Moreover, a visual inspection of the graph reveals variability between the two methods, which is attributable to HealthSouth’s history of inconsistently writing off uncollectable accounts and recognizing bad debts. It is also interesting to note that percentages of revenues realized in cash improved after HealthSouth announced its inability to meet earnings projections in 1998. In addition, the two measures of cash flows were more correlated from that point forward. Could eliminating the need to meet earnings goals have affected managerial behavior?

Linking Disclosures with Performance

HealthSouth acquired numerous rehabilitation clinics and outpatient surgery centers during the 1990s. Rolling up the industry, however, did not translate into greater returns on investment. Exhibit 5 presents the company’s return on total assets and equity during the period analyzed. The purpose and viability of business acquisitions is the fifth early warning sign of abusive earnings management.

Although one cannot infer causality to HealthSouth’s acquisitions, an auditor or investment analyst could question the soundness of an aggressive acquisition strategy given the diminishing rates of return. In addition, the data indicate a pronounced decrease in returns from 1998 to 2001, as compared to the preceding four-year period. Again, these declines coincide with the year that HealthSouth admitted its inability to meet earnings forecasts. Although not exhibited in the graph, all of the components of investment return decreased in the later four-year period. Net profit margins, asset turnovers, and the leveraging of the asset base were all lower from 1998 through 2001 than they were from 1994 through 1997.

Cash flow analysis also calls into question HealthSouth’s acquisition strategy. Operating cash flows inadequately met the company’s requirements for sustainable operations. The only year examined in which HealthSouth’s annual cash flow adequacy ratio (defined as cash flows from operations divided by the sum of payments for fixed assets) exceeded 1 was 1996. One might argue that the annual adequacy measures were generally below the level needed to cover fixed commitments because HealthSouth was expanding during this period. This was undoubtedly the case, but the goal of financing activities that deplete cash in an expansionary period is to build sound investments that produce acceptable investment returns. Such was not HealthSouth’s experience, as discussed above. What end did these acquisitions serve?

Operating cash flows and operating income were positively correlated, but there was a lack of comparability between these two amounts. The operations index measures their degree of correspondence by dividing cash flows from operations by operating income. An ideal ratio is 1, indicating that income from HealthSouth’s core business activity was being realized in cash. HealthSouth’s annual operating index averaged about 0.5 from 1994 through 2001, even after adjusting operating income for the noncash charges of depreciation and bad-debt expense. These cash flow measures do not prove that HealthSouth’s acquisitions did not serve a legitimate purpose or were solely attributable for faltering cash flows. They do, however, point to pressure on management to achieve earnings targets, something which was not realized in operating cash. Making the numbers would have benefited the company in its attempts to secure external financing. As evidenced by the cash flow adequacy ratio and the operations index, outside funds were needed because internal sources of cash were insufficient to sustain operations.

Warning Signs Have Limitations

Manipulating earnings requires disclosure of plausible managerial assertions. Markets readily identify and punish ill-conceived fraudulent financial reporting schemes that result in improbable financial scenarios. In other words, annual disclosures must pass muster to deceive auditors, regulators, and investors. HealthSouth’s financial disclosures succeeded on this point: They metaphorically looked, walked, and quacked like a duck. With the benefit of hindsight, however, the company’s financials were more foul than fowl.

Analysis of inaccurate financials is a dicey proposition. Auditors and investors are unaware that they are erroneous when first encountering them. Well-conceived and -executed financial frauds plausibly articulate duplicitous financial statement items to other related accounts. Such was the case as HealthSouth exhibited highly correlated earnings, revenues, and receivables. Nothing specific stood out to trigger alarm in the minds of investors and auditors—the usual warning signs were not apparent. Perpetrators of fraud, however, sometimes fail to logically articulate subcomponent account disclosures. HealthSouth’s bad-debt expense, allowance for uncollectible accounts, and changes in overall receivables failed to match the symmetry portrayed in net receivables. Moreover, investment returns and cash flows could have called into question management’s operations. The six early warning signs of abusive earnings can point to potential evidence of abusive earnings management, but users must be very diligent in examining all components of the financial statements and their relationships to each other.

Leonard G. Weld, PhD, is a professor of accounting and head of the department of accounting and finance, at the Harley Langdale, Jr., College of Business Administration at Valdosta State University, Valdosta, Ga, and Peter M. Bergevin, PhD, is a professor of accounting, at Redlands University, Redlands, Calif. and Lorraine Magrath, PhD, is an associate professor of accounting at the Sorrell College of Business at Troy State University, Troy, Ala.

Editor’s Note: The referenced August 2002 CPA Journal article written by Professors Weld and Magrath won the 2002 Max Block Distinguished Article Award, which each year recognizes the most outstanding work published by The CPA Journal.