Exhibit 99.1
SUMMARY OF
REPORT OF INDEPENDENT INVESTIGATION TO
THE SPECIAL INDEPENDENT COMMITTEE OF
THE BOARD OF DIRECTORS OF ABOVENET, INC.
(FORMERLY METROMEDIA FIBER NETWORK, INC.)
DLA PIPER RUDNICK GRAY CARY US LLP
Counsel
GOLDIN ASSOCIATES, LLC
CORPORATE REVIEW SERVICES, LLC
Forensic Financial and Accounting Consultants
October 7, 2005
The Audit Committee of the Board of Directors of AboveNet, Inc., (“AboveNet” or the “Company”) formerly Metromedia Fiber Network, Inc. (“MFN”), requested that independent counsel conduct an investigation into various accounting and financial reporting issues that preceded an announcement in April 2002 that MFN would not be able to timely file its Annual Report on Form 10-K for the year ended December 31, 2001, and would need to restate its financial statements for the first three quarters of 2001. MFN also announced at that time that KPMG LLP (“KPMG”), which was appointed as MFN’s auditors effective December 2001, was unable to review the quarterly data that is expected to be included in an Annual Report in accordance with professional standards because MFN’s internal control structure and policies and procedures for the preparation of interim financial information did not provide an adequate basis for them to complete such review.
At the culmination of the investigation, independent counsel, DLA Piper Rudnick Gray Cary US LLP (“DLA Piper”), and its forensic financial and accounting consultants, Goldin Associates, LLC and Corporate Review Services, LLC (collectively, “Goldin”), delivered a Report (“Report”) to the Audit Committee and the Special Independent Committee as well as to the full Board of Directors (the “Board”).
Set forth below is a summary of the Report. As part of its continuing cooperation with an ongoing investigation of the Securities and Exchange Commission (the “SEC”), AboveNet has delivered the Report in its entirety to the SEC.
During 2000 and 2001, Wall Street analysts and investors were attracted to telecommunications companies, like MFN, that emphasized revenue growth, earnings before
interest, taxes, depreciation and amortization (“EBITDA”) performance and network expansion. To measure fiber-optic network expansion, analysts and investors considered the number of “operational” and “under construction” cities that were served by the network as well as how many “fiber miles” the network spanned.
MFN did not disappoint. Quarter after quarter in 2001, (like 2000 before it) MFN consistently boasted that its revenue and EBITDA met or exceeded publicly-issued guidance, and it reported steady network growth, measured both by the number of fiber miles its network spanned and the number of “operational cities” served.
In reality, however, MFN’s revenue, its EBITDA and its network did not grow as reported.
MFN began as a start-up company, but grew into a formidable player in the telecommunications industry. Through high-profile acquisitions and organic growth, MFN established four primary business lines by 2001: (1) a fiber optic network (MFN); (2) internet services (through the September 1999 acquisition of AboveNet Communications, Inc.); (3) peering and internet exchanges (through the acquisition of Palo Alto Internet Exchange, which was a subsidiary of AboveNet Communications, Inc.); and (4) acquired or managed services (through the February 2001 acquisition of SiteSmith). At its peak in March 2000, MFN had a market cap of $27 billion and was a heavily traded and well regarded technology stock.
As 2001 began, MFN was trying to secure financing from Citicorp USA Inc. (“Citicorp”) and other prospective lenders. During a due diligence process that stretched into the fall of 2001, the telecommunications industry began to collapse. MFN was not immune to these market forces; several of its customers filed for bankruptcy protection and could no longer pay MFN. Citicorp’s initial commitment of a $350 million credit facility, announced in January 2001, was
2
postponed, then decreased and then postponed again. Ultimately, MFN did not obtain financing until nine months later, on October 1, 2001.
In the interim, an increasingly severe liquidity crisis led MFN to defer, scale back or abandon its network construction efforts as early as the first quarter of 2001. By the summer of 2001, MFN had deferred network routes as it struggled to pay its vendors. By early fall, virtually everyone at MFN knew that its network construction efforts had all but stopped.
The liquidity crisis intensified MFN’s need to obtain additional financing. As negotiations with Citicorp over the financing continued into the second and third quarters of 2001, analysts and investors monitored MFN’s financials closely to determine, among other things, whether it would become insolvent before the financing closed.
Insolvency was a real risk. MFN’s reported $1.1 billion of cash on hand (including its cash equivalents) at January 1, 2001 dwindled to only enough cash to sustain MFN through the end of September. The financing, therefore, was essential to MFN’s survival.
Leading up to the October 1, 2001 closing, MFN was being subjected to a strenuous due diligence process by Citicorp and other potential investors. MFN’s overworked and under-qualified Network Finance and Finance Departments were constantly scrambling to respond to requests for a myriad of detailed financial and operational information regarding virtually every aspect of MFN.
This convergence of events facilitated or led to:
• The setting of overly aggressive revenue and EBITDA targets, with a corresponding overly aggressive emphasis upon meeting those targets;
• Employees’ fear of and a reluctance to challenge management regarding these targets and other issues of importance;
3
• Numerous record-keeping, financial disclosure and accounting errors and irregularities, which were tolerated, dismissed or ignored;
• A dismissal of serious accounting errors as immaterial;
• Journal entries that were largely indecipherable or unsupported;
• Problems in the Network Finance and Finance Departments that were so severe and widely known that other departments within MFN kept their own records of network development expenditures and related liabilities because they did not trust those maintained by the Network Finance and Finance Departments;
• Ultimately, the issuance of quarterly 2001 financial statements that were materially misstated.
The accounting and financial reporting topics discussed in the Report fall into four categories: (1) Manipulated Financial Results; (2) Manipulated Metrics; (3) Inaccurate, Unsupported or Unsupportable Numbers; and (4) Internal Control Problems.
No one admitted that MFN intentionally filed false financial statements. Nonetheless, no one interviewed could provide a credible explanation for certain erroneous accounting decisions and suspicious transactions. Evidence suggests that certain revenue, expense and EBITDA items were manipulated in order to meet or exceed the expectations of Wall Street analysts, lenders or investors, as described below.
• First, in the third quarter of 2001, MFN classified its two largest leases as operating, rather than capital leases. MFN apparently did so in order to avoid violating covenants that limited capital expenditures in the third quarter of 2001 imposed by its Citicorp financing agreement. It appears that MFN then made other accounting decisions in the third quarter of 2001 reducing its cost of sales (“COS”), bonus accrual and datatelco acquisition reserves to offset the resulting impact on MFN’s financial statements. In the aggregate, these adjustments
4
effectively produced financial results in line with guidance given to analysts and investors.(1)
• Second, during the second quarter of 2001, MFN inappropriately recognized revenue from a duct sale, and it did not issue or accrue for a significant sales credit because it wanted to avoid a large one-time hit.
• Finally, in the fourth quarter of 2001, MFN proposed a prospective adjustment to defer rent expense for a particular lease because a prior-period adjustment would “have been too big a hit” to its financial results, according to a journal entry.
MFN also exaggerated its network growth at a time when a severe liquidity crisis and a downturn in the telecom industry prompted it to defer, scale back or abandon its construction of additional network routes. Rather than acknowledge the actual state of its network construction, MFN manipulated certain important metrics by which telecommunications analysts judged the value of MFN’s network. More specifically, evidence suggests that:
• MFN continued to insist during 2001 that it was on schedule to deploy 3.6 million fiber miles by the end of 2004 even though it had deferred, reduced or eliminated plans to construct several network routes necessary for MFN to fulfill its plan, and by late 2001, essentially ceased most network construction;
• MFN artificially inflated the number of “operational” cities that supposedly were being serviced by its network in 2001; and
• MFN inaccurately reported in a January 7, 2002 press release that its projected 2002 capital expenditures would decrease to a range of $60 million to $70 million, from $1.3 billion in 2001, as a result of “the Company’s network build-out being ahead of schedule and the completion of substantially all of MFN’s data center build-out.”
(1) In 2002, it was determined that these leases were swap transactions and should have been accounted for as nonmonetary exchanges rather as either operating or capital leases.
5
By and throughout 2001, MFN’s rapid growth had overwhelmed the under-qualified Network Finance and Finance Departments. Virtually all of those interviewed described MFN as having few, if any, effective controls to help it prepare accurate and well-supported financial reports. Internal control problems and an apparent lack of a desire to get the numbers right were rampant. For example, MFN lost control over tracking and accounting for the billions of dollars it was spending on network construction. To compensate, the accrual for capital expenditures was arbitrarily increased each quarter by roughly twenty-five percent of the total.
The internal control problems and the lack of confidence in the Network Finance and Finance Departments were well-known throughout MFN. The problems were so severe, and the lack of confidence so widespread, that other departments within MFN maintained their own records of network development expenditures and related liabilities. Rather than disclosing and effectively addressing the problems, MFN apparently elevated a desire to meet Wall Street’s expectations over its obligations to provide accurate and well-supported financial reports to the public. As a result, MFN’s financial statements during the first three quarters of 2001 contained numbers that were unsupported, unsupportable or inaccurate. Specifically, evidence suggests that MFN:
• Understated its accounts payable (“A/P”) and accrued liabilities;
• Improperly left off its balance sheet substantial liabilities owed to one of its primary vendors which it characterized in its SEC filings as “commitments to purchase materials and contracts” for network construction;
• Wrongly capitalized certain operating expenses; and
• Did not timely undertake an impairment analysis or charge when it was obvious that MFN’s assets had lost value in 2001 and then resisted taking an impairment charge for 2001 after one was recommended.
6
MFN made a number of flawed accounting decisions for which no evidence was found to suggest that they were motivated by a desire to reach a particular result. Instead, these problems appear to be a by-product of MFN’s severe lack of internal controls for which some of these items created an environment vulnerable to fraud, manipulation or management override. The errors include, but are not limited to, MFN’s:
• Failure to timely write-off the loss on its investment in a joint venture;
• Improper recognition of revenue from so-called swap transactions;
• Improper and premature recognition of revenue from customers;
• Inadequate treatment of accounts receivables (“A/R”) and inadequate bad debt reserves;
• Incorrect accounting for sales credits;
• Miscalculated valuation of certain stock options in connection with its acquisition of SiteSmith; and
• Mistaken accounting for deferred rent - goodwill in connection with vacated properties resulting from its acquisition of SiteSmith.
Approximately six months after the Citicorp financing, MFN disclosed on April 17, 2002 that it would not be able to file timely its Annual Report on Form 10-K for the year ended December 31, 2001, that its reported operating results for the first three quarters of the fiscal year ended December 31, 2001 would need to be restated and that KPMG (which replaced E&Y as the independent auditors effective December 2001), was unable to review the quarterly data that is expected to be included in the Annual Report because of inadequate internal controls.
Six days after the initial announcement, on April 23, 2002, MFN provided further information regarding the nature of the proposed restatement. It issued a range of preliminary
7
restated results and disclosed that the restatement involved revenue/sales credit recognition, the timing of expense recognition, non-cash lease accounting and purchase accounting issues. It also announced that MFN would be taking an impairment charge in the range of $3.9 billion to $4.3 billion. The next day, the SEC notified MFN that it had initiated an informal inquiry. On May 20, 2002, MFN announced that it had filed voluntary petitions for bankruptcy under Chapter 11 of the United States Bankruptcy Code and that its stock had been delisted from the NASDAQ Stock Market. On June 3, 2002, the SEC notified MFN that it had commenced a formal investigation into MFN’s accounting and internal control practices leading to the proposed restatement.
To this day, more than three years later, the Company has not filed a Form 10-K for 2001, 2002, 2003 or 2004 nor has it published restated financials for the first three quarters of 2001.
At the suggestion of the SEC, and in an effort to identify the accounting, financial reporting and internal control issues that resulted in the April 2002 announcements, the Audit Committee engaged DLA Piper to conduct an independent investigation (the “Investigation”) shortly after AboveNet’s emergence from bankruptcy.(2) DLA Piper, in turn, retained Goldin to provide forensic financial and accounting advice and support.
While the Investigation initially was guided by the accounting items identified in the announced restatement and by the SEC, the Audit Committee did not limit the scope of the Investigation. Rather, the Audit Committee authorized DLA Piper to explore and pursue any matter related to those periods to be restated that warranted scrutiny in the hope of taking
(2) AboveNet later convened the Special Independent Committee to receive and review the Report.
8
appropriate steps to identify and rectify any problems as well as prevent their recurrence. Accordingly, the Investigation evolved and expanded over time to examine issues raised by: KPMG; DLA Piper; Goldin; claims made in shareholder class actions; documents reviewed during the Investigation; and statements made by current and former MFN officers and employees in interviews.
The Investigation included the review of over one million pages of documents and electronic files gathered from a number of sources, including the Company, Metromedia Company (“MC”), KPMG, current and former MFN employees, various third parties and legal proceedings. DLA Piper and Goldin interviewed over forty current and former officers, employees, and third parties with many of these individuals being interviewed more than once. DLA Piper and Goldin also consulted with KPMG, AboveNet’s current Chief Financial Officer (“CFO”), Vice President of Internal Audit and several consultants. Notably, MFN founder and former Chief Executive Officer (“CEO”) Stephen Garofalo, former President Nicholas Tanzi, former CFO Gerard Benedetto, Citicorp and MFN’s former auditor, E&Y, refused to be interviewed.
The Investigation concluded in or about August 2005 following extensive work performed over thousands of hours by DLA Piper, Goldin, and the Company. As with any private inquiry of this nature, however, the Investigation encountered several limitations, including the lack of subpoena power to compel the production of documents and the questioning of witnesses, several of whom refused to cooperate with the Investigation. Despite the lack of cooperation of some individuals, the Investigation was extensive and the findings and conclusions set forth in the Report are supported by ample evidence, which AboveNet will make available to the SEC. Since the SEC is equipped with subpoena power and other enforcement
9
means, additional facts may eventually be discovered that further inform an understanding of the involvement of certain individuals in the events described in the Report.
The Investigation led to the following observations and conclusions:
• Like many in its industry, MFN grew too quickly, largely through acquisitions, and its accounting and financial reporting functions could not, and did not, keep up.
• MFN operated in an environment where internal controls were nonexistent or ignored. There was a dearth of written policies and procedures outlining accounting, record-keeping, finance or cash-management practices, and accounting rules were applied inconsistently.
• There was little to no communication between departments, and when communication occurred, it was often contentious, inadequate or the information conveyed was inaccurate.
• MFN’s financial systems were inadequate and not fully implemented. This led to extensive manual analyses and calculations using Excel spreadsheets and a predominantly manual journal entry process to record transactions into the general ledger. In addition, the electronic files on the shared server used by the Finance Department were relied upon to analyze and support account balances but were not adequately organized or controlled, resulting in numerous versions of the same schedule and difficulty locating information. Each of these factors increased the risk of accounting and financial reporting errors and irregularities.
• MFN routinely made journal entries for which there was no support, formal approval or evidence of review, and many journal entries were missing. In certain instances, the explanations for the journal entries were inappropriate and reflective of the poor state of internal controls. For example, a July 9, 2001 journal entry indicates that an April-June quarterly invoice should be “reverse[d] for FUFU” and that revenue relating to a customer should be removed because MFN “still [didn’t] know what to do with this!!”
• There was a lack of documentation, including missing invoices, contracts, agreements, billing-authorization forms, service orders, accounting analyses and reconciliations.
• The situation deteriorated to such a degree that different departments within MFN maintained different records of network development expenditures and related liabilities. MFN’s financial statements and other filings with the SEC were
10
prepared based on the general ledger and other accounting data, which frequently did not agree with these records.
• MFN’s Network Finance and Finance Department staff lacked sufficient training and supervision, were overwhelmed and were unable to perform necessary tasks effectively and timely.
• At various moments of crisis, the Company haphazardly established a few internal control policies that were communicated by e-mail before being ignored, unenforced or otherwise rendered ineffective.
• While the deficiencies in MFN’s record-keeping and accounting functions were well-known throughout the Company, they persisted.
• While no one has expressly admitted it, evidence suggests that certain accounting decisions may have been made for the purpose of achieving a particular financial result, to satisfy covenants imposed by the Citicorp financing or to meet guidance.
• Evidence also suggests that MFN manipulated important metrics by which telecommunications analysts judged MFN’s network growth, and it published those inflated metrics in filings made to the SEC and in press releases.
• MFN relied upon data that it knew or should have known was inaccurate, unsupported or unsupportable, and this data was included in financial statements filed with the SEC.
• MFN’s overworked, overwhelmed, under-qualified and under-trained Network Finance and Finance Departments made a number of improper accounting decisions.
• MFN’s lack of controls presented an environment that created the opportunity for management override.
• MFN’s controlling shareholder, MC, exerted significant control over MFN’s financial reporting, efforts to secure additional funding and its communications with Wall Street.
• Neither the Board nor the Audit Committee provided effective oversight.
• Evidence raises substantial questions about the role that E&Y played as MFN’s auditor. E&Y declined to speak with us and only made available access to a very limited portion of its work papers (previously supplied to KMPG.)
In sum, MFN’s quarterly financial statements for the first three quarters of 2001 were materially misstated. Further, there is evidence to suggest that, during certain quarters of 2001, MFN
11
manipulated its reported financial results and metrics in order to meet guidance and paint a falsely optimistic picture of its revenue, operating results, network growth, and solvency. In addition, MFN made many flawed accounting decisions due to a lack of awareness or understanding of the relevant accounting principles and as a result of MFN’s severe lack of internal controls.
The Investigation included a comprehensive analysis of numerous accounting and financial reporting topics which are grouped into four categories: (1) Manipulated Financial Results; (2) Manipulated Metrics; (3) Inaccurate, Unsupported or Unsupportable Numbers; and (4) Internal Control Problems.
While no one admitted that MFN intentionally filed false financial statements, there is evidence that suggests that MFN manipulated revenue, expense and EBITDA in order to meet or exceed the expectations of Wall Street analysts, lenders or investors in three areas, which are described below.
Historically, MFN classified several leases as operating that should have been capitalized. In 2002, it was determined that these were part of swap transactions that should have been accounted for as nonmonetary exchanges rather than operating or capital leases. The Company’s historical treatment as operating leases, not capital leases, had the effect of: (a) increasing MFN’s COS; (b) decreasing depreciation and interest expense; (c) increasing EBITDA loss; and (d) decreasing the book value of MFN’s network assets. Critically, however, this accounting treatment also caused MFN to reduce the debt reflected on its balance sheet.
12
During its audit for the year ended December 31, 2000, E&Y noted that MFN had improperly treated one lease as operating rather than capital, and it requested that MFN review all of its leases to determine whether similar errors had been made. Short of resources, MFN did not complete this review until late October 2001, when it concluded that five leases required capital lease treatment.
Shortly thereafter, MFN prepared a draft Form 10-Q for the third quarter of 2001 dated November 2, 2001, which reflected the capital treatment of these five leases as well as a series of other adjustments related to COS. As a result of these adjustments, MFN’s EBITDA loss appeared to be $34.4 million, just within the guidance range for the third quarter of a $31 to $35 million loss. This November 2, 2001 draft Form 10-Q for the third quarter of 2001 was circulated in advance of a November 6, 2001 Audit Committee meeting.
Shortly before, at or after the November 6, 2001 Audit Committee meeting, a decision was made to change back to operating the two largest leases that just had been capitalized. Evidence suggests that the change was made after questions were raised about whether right-of-way (“ROW”) contingencies in these leases potentially limited their terms and thus supported the operating lease classification. These ROW provisions did not provide an adequate basis for treating these leases as operating because, among other reasons, the possibility of a shorter-than-expected lease term was remote given other lease terms and MFN had capitalized other leases with substantially similar ROW provisions. Moreover, had these not been nonmonetary exchanges, such a position would have been inconsistent with the relevant accounting guidance under Generally Accepted Accounting Principles (“GAAP”).
MFN’s decision to recharacterize these leases may have been motivated by a desire to avoid violating certain financing covenants that became effective as of the third quarter of 2001.
13
Specifically, the Citicorp financing agreement placed strict ceilings on “acquired indebtedness” and “expenditures.” Evidence suggests that MFN was closely monitoring covenant compliance and was aware that lease reclassification could cause MFN to violate these financing covenants. It appears, however, that MFN consulted with E&Y, which approved this accounting treatment, at least for the Form 10-Q for the third quarter of 2001.
This change back to operating lease treatment had an impact upon MFN’s COS, among other things. A comparison of schedules generated in October 2001 identifying “items to be considered for booking in September” to versions of the schedule dated November 6, 2001 (the “Schedules”) suggests that MFN made certain decisions and accounting adjustments to offset the effect of this change on its financial statements.
Specifically, in the third quarter of 2001, MFN recorded a series of adjustments relating to three categories – expenses charged to cost of sales that should have been capitalized in the first four months of 2001 (“capitalization catch-up”), a reduction to the bonus accrual (“bonus accrual reduction”), and a reduction to the datatelco acquisition reserve (“datatelco acquisition reserve reduction”). With these adjustments, MFN’s reported EBITDA loss for the third quarter of 2001 ($35 million) exactly met the low end of its published guidance range for the third quarter of 2001. These adjustments were recorded in two phases. First, MFN recorded certain adjustments to these and other categories in late October 2001, prior to circulation of the November 2 draft Form 10-Q for the third quarter of 2001. Second, it then recorded additional adjustments to these categories after the November 6, 2001 Audit Committee meeting. MFN apparently tracked these adjustments on multiple iterations of the Schedules.
The Schedules list categories of accounting issues that the Finance Department identified in connection with its preparation of the third quarter 2001 financial statements. Specifically,
14
they depict the effect that various contemplated accounting adjustments would have on line items on MFN’s income statement and balance sheet. The capitalization catch-up, bonus accrual reduction and datatelco acquisition reserve reduction adjustments apparently were designed to offset the impact of other adjustments, including those made for the reclassified leases. Significantly, the combined effect of these adjustments allowed MFN to just meet EBITDA guidance for the third quarter of 2001 in both the November 2 draft Form 10-Q and the final Form 10-Q even though the amounts to be adjusted for certain items changed from one version of the Schedules to the next.
For four reasons, the adjustments recorded in the Schedules appear to constitute a manipulation of MFN’s financial results. First, MFN contemplated certain adjustments in October (reflected on “the October Schedules”) which were recorded prior to the circulation of the November 2 draft Form 10-Q (with the leases newly classified as capital). These adjustments allowed MFN to meet guidance. When the two leases were reclassified as operating, that increased MFN’s COS. After the reclassification, MFN considered and recorded additional adjustments (reflected on “the November Schedules”) which, while some were small individually, offset the impact to MFN’s COS and allowed MFN to meet guidance.
Second, for each of the capitalization catch-up, the bonus accrual reduction and the datatelco acquisition reserve reduction, MFN recorded a different adjustment in the November 2 draft Form 10-Q than in its final reported financial statements. No one provided any credible explanation regarding what happened between the November 2 draft Form 10-Q and the reporting of MFN’s final financial results for the third quarter of 2001 on November 7, 2001 that warranted the changes nor was there adequate support for these adjustments.
15
Third, adjustments were made as part of the capitalization catch-up and to the bonus accrual in June 2001, at the end of the second quarter of 2001, that left additional adjustments which were then taken in the third quarter of 2001. In June 2001, MFN apparently concluded that it failed to capitalize a portion of the departmental expenditures incurred by the Operations Department and therefore had overstated COS in the first five months of the year. Instead of recording an adjustment for all five months, however, MFN only adjusted COS for the month of May prior to its finalization of the Form 10-Q for the second quarter of 2001. Then, in October 2001, MFN took a portion of the adjustment for the first four months leaving the balance as a “remaining adjustment to take,” according to the Schedules. Similarly, in June 2001, MFN reversed its bonus accrual by nearly $2.2 million and its bonus accrual for AN by almost $1.0 million, but did not reduce its monthly bonus accrual going forward. Instead, MFN continued to record approximately the same amount in the bonus accrual. Then, in the October Schedules, the bonus accrual was identified as an area for a “[p]ossible take back,” the amount of which was “TBD.” No credible explanation was provided for these decisions.
Fourth, MFN’s decision to reduce excess amounts in the datatelco acquisition reserve by reducing COS violated GAAP. The accounting guidance provides that MFN should have reduced goodwill, not COS, because it had recorded goodwill as part of the cost of the acquisition and an adequate amount of goodwill remained on MFN’s balance sheet at the time of the adjustment. MFN apparently consulted with E&Y and mistakenly concluded that MFN could not adjust goodwill because more than one year had passed since its acquisition of AboveNet Communications.
In addition, MFN’s disclosures regarding the Company’s COS in its press release and Form 10-Q for the third quarter of 2001 were misleading because they did not disclose that the
16
reported COS amount included reductions to eliminate excess amounts in the datatelco acquisition reserve and to correct amounts charged to COS that should have been charged to balance sheet accounts in the first two quarters of 2001. Instead, the November 7, 2001 press release indicated that MFN’s gross margins had improved due to “the reduction in headcount and decreases in the Company’s expenses for data circuits together with increases in revenues.” The Form 10-Q similarly stated that the decrease in COS as a percentage of revenue as compared to the year before “reflect[ed] headcount reductions, as well as [MFN’s] decreased cost of data circuits coupled with revenue growth.”
A review of MFN’s revenue recognition practices identified certain irregularities, most of which appear to have been the result of poor internal controls or flawed accounting decisions, as further discussed below. MFN’s accounting for two transactions, a duct sale, and a sales credit, however, appears to have been motivated by a desire to reach a particular result.
In the case of the duct sale, on June 8, 2001, MFN issued an invoice to the customer. It then recognized revenue on the sale as of June 30, 2001. At that time, however, the customer had not paid the invoice within five days of acceptance of the ducts, as required by the contract, and was thus in material breach of the agreement. The customer had still not paid the invoice when the Form 10-Q for the second quarter of 2001 was filed on August 15, 2001, and the evidence suggests that collection of the revenue was not reasonably assured at that time. If MFN had not recognized the revenue from the customer in the second quarter of 2001, it would not have met its revenue guidance.
With regard to the sales credit, MFN acknowledged that it owed a customer credits for two months, due to over billing. Because MFN did not want to issue the full amount of the
17
credit in the second quarter of 2001, it instead agreed to provide the customer with two free months of service, one in late 2001 and another in late 2002. Evidence suggests that MFN intended to accrue for these credits on a monthly basis, which was improper under GAAP. In any event, it did not begin to accrue for these sales credits until five months later. These actions resulted in an overstatement of revenue in the second quarter of 2001.
In the summer of 2001, MFN discovered errors in the amortization schedules for certain of its leases with periods of free rent or escalation clauses and created new amortization schedules. One of the newly-created amortization schedules showed that a multi-million dollar adjustment was required. Evidence suggests that MFN decided to adjust the numbers prospectively over the remaining life of the lease, rather than restating the prior periods or making a cumulative entry in the fourth quarter of 2001 because, according to the journal entry, an “Adjustment Going Backward [would] have been too big a hit.”
Evidence suggests that MFN intentionally exaggerated its network growth by manipulating certain important metrics by which telecommunications analysts judged the value of MFN’s network. MFN reported these misrepresentations in SEC filings and press releases.
As early as October 25, 2000, MFN announced that it intended to construct a nationwide network spanning 3.6 million fiber miles by the end of 2004. MFN then updated and reported, at the end of each quarter, the number of fiber miles deployed. The Investigation discovered three problems with MFN’s reporting of its fiber miles.
18
First, MFN employed two fiber-mile counting practices that could be viewed as improperly double-counting and otherwise inflating the total number of deployed fiber miles. MFN often counted fiber installed in the ground even if it was not fully tested or spliced as well as those that were delivered to a construction site, but remained on the reel. In both cases, the fiber was not ready for service. MFN also counted those fibers that it provided carriers in swap transactions as well as those fibers that it received in return.
Second, MFN’s 3.6 million fiber miles forecast included 750,000 miles for “Future Overbuilds and Network Expansions” and 250,000 fiber miles for “Customer Builds.” Evidence suggests that there was no factual support for these figures; rather, they were plug numbers used to support MFN’s 3.6 million fiber miles target.
Third, as early as December 2000, MFN found itself in a severe liquidity crisis, which prompted it to defer, scale back or abandon the construction of certain network routes throughout 2001. Evidence suggests that MFN may have purposely delayed the completion of certain construction projects as early as February 2001 to avoid incurring payments that would become due once certain routes became operational.
One analyst seemed to pick up on this development in April 2001 when he reported that MFN appeared to “have scaled back” its build out. While the analyst viewed this approach as “prudent,” he also cautioned that it now was difficult to value MFN since “the revised plan [had not] been quantified or communicated to investors.”
On April 30, 2001, MFN issued a press release announcing its first quarter results stating that it had deployed “approximately 1.7 million fiber miles” by the end of the first quarter and advising that, with the anticipated Citicorp financing, it was “fully funded to complete its
19
planned build out of 3.6 million fiber miles.” The analyst whose “field checks” indicated that MFN had “scaled back its network build” questioned MFN’s representations.
As the summer of 2001 progressed, MFN continued to defer the construction of certain routes as its struggled to pay its network construction vendors. Notwithstanding the deferred construction plans, MFN disclosed in an August 15, 2001 press release that it had deployed “approximately 2.0 million fiber miles” by the end of the second quarter — “an increase of 18%” over the previous quarter.
By the early fall of 2001, virtually everyone within MFN knew that network construction efforts had all but ceased. Yet, MFN continued to cling to its fiber miles target and reiterated at the Citicorp closing on October 1, 2001 that it could reach that target. MFN never revised its network growth targets.
By the end of 2001, the Network Development Department was being disbanded and MFN was constrained by tight capital expenditure limits imposed as part of the Citicorp financing. The only construction that MFN planned to undertake heading into 2002 consisted of providing connections from existing network rings to the customer location; it did not contemplate building out entirely new cities or adding substantial fiber miles. As discussed in greater detail below, however, MFN issued a press release on January 7, 2002 stating that its network development efforts were “ahead of schedule.”
In 2000, MFN announced its plan to expand its network to 67 operational cities by the end of 2004. As with fiber miles, MFN refused to budge from that target even as it abandoned plans to construct some of those cities. Instead, MFN continued to project network growth by changing its method of counting “operational cities”; it did not, however, disclose that change,
20
resulting in an artificially inflated jump in the number of operational cities from its 2000 Form 10-K to its Form 10-Q for the first quarter of 2001.
Specifically, in December of 2000, MFN decided to change the way it counted “operational cities.” At the time, MFN counted each “market” as one city, but it appeared that competitors were defining operational cities more narrowly, thereby creating the impression that MFN’s network was not as vast as theirs. For example, competitors may have been counting the San Francisco market as three cities, while MFN counted the same communities as one.
In its Form 10-K for the year ended December 31, 2000, MFN reported that it had 14 operational cities. In contrast, with the new definition of “cities,” the Form 10-Q for the first quarter of 2001 reported that MFN had 27 cities operational. Absent the change in counting the number of “cities,” MFN would have reported an additional six operational cities, for a total of 20, as of the first quarter of 2001. The seven additional “cities” that were reported as operational previously had been considered part of other cities already counted as fully operational as of the end of 2000. MFN did not disclose in its Form 10-Q for the first quarter of 2001 or elsewhere that it changed its definition of “operational cities.”
After internal logistical problems made the new counting method impractical, MFN reverted back to its old definition. By switching back to the old method of counting, however, the total number of operational cities actually decreased to 25 for the second quarter of 2001 from the 27 reported for the first quarter of 2001. Yet, rather than explaining the change, MFN counted as operational four additional cities that were considered “under construction” by the Network Development Department, and reported that number (29) in its Form 10-Q for the second quarter of 2001.
21
In a January 7, 2002 press release, MFN announced that its 2002 capital expenditures would be dramatically reduced from the $1.3 billion incurred in 2001 to a range of $60 million to $70 million because its “network build-out [was] being completed ahead of schedule.” The release further provided that MFN “still expect[ed] to have approximately 3.6 million fiber miles by the end of 2004.” These statements were not accurate.
At the time of the issuance of the press release, the Operations Department was the only department that was charged with, or had a budget to do, network construction. The construction for which the Operations Department had budgeted, however, was limited to customer-driven build-outs to existing space; it did not encompass new network construction that would lead to additional cities. In addition, the financial covenants present in the Citicorp financing agreement severely limited MFN’s ability to spend additional sums on network construction.
The January 7, 2002 press release also reaffirmed MFN’s previously issued guidance of more than $100 million to $103 million in revenue for the fourth quarter of 2001 and provided revenue guidance for 2002. Before the press release was issued, questions were raised about the appropriateness of reaffirming or providing guidance given various accounting issues that were being identified and the state of the network construction.
MFN relied upon data that it knew or should have known were inaccurate, unsupported or unsupportable, and included the data included in filings made with the SEC. Those amounts related to:
a. A/P and accrued liabilities;
b. Amounts owed to a primary vendor;
22
c. Capitalization of departmental, interest and other expenses; and
d. The timing of impairment charges.
MFN had no accurate, up-to-date record of how much it was spending on network construction, and it did not follow with uniformity any particular A/P or accrued liabilities process or policy. Network construction work often was authorized without the approval or issuance of a purchase order and the A/P Department often never received invoices. When the A/P Department did receive invoices, they often did not reference any job cost number and could not be matched with any purchase order.
MFN’s liabilities would have been significantly understated if the total was based only on the invoices and purchase orders for which the Network Finance Department had records. The Network Finance Department attempted to capture the expenditures of which it was not aware by arbitrarily increasing the accrued capital expenditures account by roughly 25 percent of the total throughout 2000 and 2001. For example, the accrual in December of 2000 was increased by $85.5 million (roughly 25 percent) through a single journal entry for which there was no support or explanation. Similar increases (and some decreases) were made without any support or explanation in March, June and September of 2001. Although these figures were based on estimates, there was no support and the Company nonetheless reported them in its SEC filings and press releases.
As MFN’s Network Finance and Finance Departments attempted to operate in this environment, a liquidity crisis prevented the payment of many invoices. In response, the A/P Department began cutting and then holding checks made payable to its vendors and suppliers as it did not know which bills to pay. The held checks pile grew larger and larger. Between
23
January and July 2001, MFN lost control of the process; internal records suggest that the combined total of the held checks fluctuated between $70 and $91 million. As payments were withheld, the interest charges and penalties for late payments accumulated and MFN’s relationships with its suppliers and vendors deteriorated.
In the summer of 2001, MFN’s A/P problems were highlighted as it began negotiating outstanding balances with its vendors and suppliers. During this process, it became clear that MFN’s A/P records for its largest vendors were severely understated or altogether missing. When the Network Finance Department could not support its own figures, an entirely new vendor payables database was constructed to record MFN’s vendor payable accounts. By late July or early August 2001, it was clear that MFN’s A/P information was unreliable and severely understated. Nonetheless, it appears that MFN later included in its Form 10-Q for the second quarter of 2001 the inaccurate A/P data. Despite the many settlements that were reached between MFN and its vendors as part of the Citicorp financing closing on October 1, 2001, evidence indicates that MFN’s records remained incorrect afterwards.
During the first two quarters of 2001, MFN apparently failed to properly accrue for tens of millions of dollars in liabilities it owed under an agreement with one of its primary vendors. Under this agreement, as amended, the vendor invoiced MFN only for a “down payment” of 25 percent of the value of those products delivered by the vendor, with the balance to be paid upon “acceptance” but no later than January 31, 2003. MFN did not record the 75 percent balance (the “pay as you grow” amounts) as a current or long-term liability, but rather characterized certain of the amounts in its Form 10-Qs for the first and second quarters of 2001 as a “commitment to purchase materials and contracts for the construction of fiber optic network systems.” These
24
“pay as you grow” amounts should have been included in MFN’s balance sheet as liabilities. Furthermore, MFN should have disclosed its agreement with this vendor and its material terms including the amount of the commitment that was due for its historical purchases and the amount committed for future purchases. It did not.
MFN was required to capitalize certain departmental and other expenses so long as it could establish that these amounts were directly related to network development. MFN, however, had no policies or processes in place to ensure that adequate documentation was maintained to support this accounting. As a result, MFN did not adequately support the departmental (such as labor and travel costs) and other expenses it capitalized during the first three quarters of 2001 and the system in place was vulnerable to fraud, manipulation and management override.
With respect to departmental expenses, MFN capitalized the expenses incurred by those departments devoted to network construction during the first three quarters of 2001. Although MFN could appropriately capitalize those costs incurred to construct its network, it failed to maintain any empirical or tangible data that demonstrates a clear connection between the expenses it capitalized and the construction of the network. As a result, it is virtually impossible to confirm that the departmental expenses capitalized were directly related to network development due to the lack of systems to track and control these expenses.
In addition, MFN capitalized certain departmental costs incurred by the Operations Department, which devoted only some resources to network construction, based upon a “percent of completion method.” This policy, however, was inconsistent with GAAP because MFN failed
25
to maintain any tracking mechanisms that linked the capitalized expenses directly to network routes being developed.
Finally, MFN offered no training, instruction, formal policy or guidelines that helped its employees determine whether to capitalize or expense an invoice. As such, expenditures were subject to individual assessments made by different personnel with different backgrounds. The Finance Department had no consistent oversight or quality-control role in this regard despite receiving occasional requests for guidance. As a result, MFN capitalized expenses improperly in at least two recurring circumstances: (1) maintenance expenses; and (2) expenses attributed to fully constructed network routes.
These errors could have resulted from poor communication between departments and MFN’s failure to develop reasonable policies and procedures to facilitate the consistent treatment of common expenses. There is evidence, however, that MFN may have capitalized certain repair and maintenance expenses to hide budget overages. Moreover, MFN maintained several schedules identifying the amount of quarterly expenditures it needed to capitalize in order to reach a desired result. These schedules calculated MFN’s results of operations both before and after capitalization of departmental expenses and quite easily could have served as a roadmap to manipulate the treatment of expenditures. Nonetheless, no one admitted to knowing about such a scheme.
On April 23, 2002, MFN disclosed that it would take a charge of between $3.9 and $4.3 billion for impairment to its network and other assets and goodwill. By March 2003, the estimate increased. While the exact amount of the impairment charge is not yet known, assets
26
that MFN previously valued at $6.4 billion as of December 31, 2001 appear to be worth substantially less than that.
Based on a review of “trigger events” and a comparison of MFN’s cash flow projections to the book value of long-term assets, net of accumulated depreciation, it appears that MFN had sufficient information to trigger an impairment analysis by the third quarter of 2001 and that such an analysis likely would have indicated substantial impairment of its assets. These trigger events include: a “meltdown” in the telecom industry during late 2000 and 2001; the bankruptcy of a number of MFN’s customers; the falling stock prices of telecom companies, including MFN; and the decline of MFN’s bond prices and credit ratings.
Although MFN’s disclosures in its 2000 Form 10-K and 10-Qs for the second and third quarters of 2001 state that an impairment analysis had been or would be conducted, little evidence was found to indicate that MFN undertook one prior to November 2001. Even then, there was significant resistance to the effort to commence an impairment analysis.
MFN made a number of flawed accounting decisions. These include the:
a. Delayed write-off of the loss on its investment in a joint venture;
b. Improper recognition of revenue from so-called swap transactions;
c. Improper and premature recognition of revenue from customers;
d. Inadequate treatment of A/R and bad debt reserve;
e. Incorrect accounting for sales credits;
f. Miscalculated valuation of certain stock options in connection with its acquisition of SiteSmith; and
g. Mistaken accounting for deferred rent - goodwill in connection with vacated properties resulting from its acquisition of SiteSmith.
27
In the fourth quarter of 2001, MFN wrote off several million dollars as the loss on its investment in a joint venture formed in November 1997. MFN considered closing down the joint venture as early as March 2001. Evidence suggests that, by June 2001, MFN apparently believed the net realizable value of the joint venture was negative and not expected to recover. While MFN should have written down its investment at that time, it instead waited to do so until the fourth quarter of 2001.
So-called swap transactions were typical in the telecommunications industry in the late 1990s and early 2000s. Because these transactions involve the exchange of similar productive assets, they should be accounted for as nonmonetary exchanges and no revenue should have been recorded. In several high-profile instances, however, telecommunications companies improperly recorded upfront all of the “revenue” on these agreements.
MFN entered into seventeen transactions in which it exchanged long-term leases of fiber capacity with other telecommunication companies. Of the seventeen, MFN improperly recorded revenue from eleven of these transactions from 1999 through 2001. Unlike certain other telecommunication companies, however, MFN did not record revenue upfront on the total purported value of the agreements. Instead, it appears that, as a general matter, MFN improperly recognized revenue on the transactions evenly on a monthly basis over the life of the lease commencing once the fiber was in use.
Evidence indicates that all of MFN’s swap transactions had a legitimate business purpose. Virtually everyone interviewed considered them a necessary and important component of the Company’s growth aspirations since they allowed MFN to trade its abundant inventory of
28
metro fiber for much-needed long-haul fiber. This strategy was disclosed in MFN’s public filings and allowed it to increase capacity and efficiently build a nationwide network without incurring the expense of constructing a long-haul network. However, MFN never disclosed how the swap transactions were accounted for or the impact of these transactions on its balance sheet, results of operations or cash flows.
Evidence does suggest, however, that MFN endeavored to support its (improper) revenue recognition practices by changing the structure of its swap transactions from a single agreement in which no money was exchanged to two agreements with a cash component. These structural changes were intended to strengthen MFN’s (mistaken) belief that the transactions could be considered monetary exchanges (and thus revenue could be recognized) if the lease agreements were not coterminous, cash was involved in the transfer and there were two separate agreements. This belief may have been based on advice given to MFN by E&Y.
In a number of customer transactions, MFN recognized revenue in violation of GAAP. The revenue recognition irregularities fall primarily into four categories: (1) revenue recognized when collectibility of the amount was not reasonably assured; (2) revenue recognized in the wrong period; (3) revenue recognized on swap transactions; and (4) revenue recognized without proper consideration of relevant contractual terms. In several instances, no adequate support or explanation was provided for MFN’s accounting decisions. For example, in April 2001, MFN identified an unbilled revenue opportunity in connection with its customer contracts. Accordingly, MFN issued invoices to customers for costs incurred for services dating back several months or longer. As collectibility was not reasonably assured, it should have waited to recognize this revenue until such collection concerns were resolved or when the payment was
29
received. Instead, however, MFN recognized the invoiced revenue as of July 2001. As a further example, in January 2000, MFN acquired M.I.B.H., Inc. (“MIBH”), a network outsourcing provider. When MIBH’s financial records were consolidated with those of another MFN business unit, a balance of deferred revenue was attributed to MIBH. As of the first quarter of 2001, MFN recognized as revenue the entire amount of the MIBH deferred revenue without any support or explanation.
At two of MFN’s three business units, there was a significant deficiency in establishing, monitoring, increasing, and reconciling the reserves for uncollectible A/R. In 2000, these business units had non-existent or severely inadequate bad debt reserves. The bad debt reserves were increased in 2001, but those additions were insufficient and calculated without supporting documentation or any discernible formula.
During this time, there were few, if any, written policies or procedures governing the setting or monitoring of reserves. As a result, the monitoring and review of A/R and the reserves were haphazard. MFN began to increase its reserves more steeply, document those increases better, and attempt to formalize its policies after the Citicorp financing closed on October 1, 2001. In practice, however, MFN still failed to follow its policies consistently, and the reserves were still too low in light of, among other things, the deterioration of the internet and telecom industries and the increasing numbers of MFN customers which filed for bankruptcy protection. The presence of insufficient reserves, arbitrary increases and sloppy documentation raises doubts as to whether MFN’s receivables were fairly stated for the first three quarters of 2001.
30
MFN regularly relied upon sales credits to resolve customer disputes and internal discrepancies caused largely by poor internal controls. For example, the lack of communications between several departments frequently resulted in billing errors which often were corrected by the issuance of sales credits. While there is evidence that MFN prepared policies relating to the issuance and accounting for sales credits, the processes set forth in these policies do not appear to have been well-known or consistently followed.
MFN issued sales credits when: improper amounts were billed; invoices were issued to customers before installation of services; invoices were issued to customers after services were terminated; or they were necessary to give customers an incentive to renew its contract or to resolve a customer dispute. Since its record-keeping practices were sloppy, MFN was at a disadvantage in negotiations with its customers and it was common for MFN to make up for service problems by issuing its customers sales credits to resolve disputes. Often, the Finance Department was not notified when sales credits were issued. When the Finance Department was notified (often by the customer), the credits were not consistently verified as legitimate or accounted for properly. These problems were compounded by the facts that all three of MFN’s business units (MFN, AN and SiteSmith) had different systems for monitoring, recording and accounting for sales and the post-acquisition integration of these systems was not smooth. The lack of effective controls over the issuance and accounting for sales credits created vulnerability to fraud, manipulation and management override.
On October 9, 2000, MFN entered into an agreement to purchase SiteSmith. At that time, MFN’s stock price was $21.8125. The transaction did not close, however, until February 8,
31
2001, when MFN acquired all of the outstanding shares of SiteSmith common stock for approximately $1.1 billion. In the interim, MFN’s stock price dropped by $9.75, to $12.0625.
MFN calculated the intrinsic value of the unvested SiteSmith stock options as the amount by which the market price of the underlying stock exceeded the exercise price of the unvested options. It then recorded this amount as deferred compensation costs. Contrary to GAAP, however, MFN improperly used the market price ($21.8125) on October 9, 2000, the day the “definitive agreement” was reached, rather than the market price ($12.0625) on February 8, 2001, the day the transaction was consummated. As a result, MFN overstated deferred compensation and understated the goodwill associated with the deal. This accounting error resulted in additional amortization of deferred compensation, which increased MFN’s net loss. In addition, non-cash deferred compensation was excluded from MFN’s calculation of normalized EBITDA as reported in its press releases. Evidence suggests that the decision to use the stock price on the day a definitive agreement was reached was nothing more than a mistake.
In connection with its acquisition of SiteSmith, MFN decided to vacate five SiteSmith properties that it believed would provide no future economic benefit to the combined company. In accordance with GAAP, MFN established reserves and increased goodwill for the remaining lease obligations for these properties between August and December 2001.
Contrary to its expectations, however, MFN continued to use four of the five properties for up to twelve months after the SiteSmith acquisition closed. Nonetheless, it charged the rent payments during these periods to the reserve. GAAP provides that MFN should have accounted for the rent payments as an expense because it continued to derive an economic benefit from
32
those properties. It did not. As a result, MFN overstated its goodwill and understated its expenses. No adequate explanation was provided for the delay in vacating the properties or MFN’s accounting decisions.
On September 8, 2003, the Company emerged from Chapter 11 bankruptcy as AboveNet, Inc. Most of senior management has changed or been replaced. AboveNet’s current CEO was formerly its Senior Vice President of Operations; its General Counsel remains in that role. Its CFO, Controller and Assistant Controller as well as much of its Finance Department staff are new. AboveNet also added the newly-created position of Vice President of Internal Audit and instituted a new Code of Conduct. All but one of the Company’s directors and every member of its Audit Committee is new. Further, the Company no longer has a management agreement with MC, formerly its controlling shareholder. Moreover, the auditor during the period under investigation, E&Y, has long been replaced.
AboveNet has spent significant resources investigating the events that occurred in the past at MFN, reconstructing its financial statements and working to implement improved internal controls. In addition, AboveNet has cooperated with the SEC, by, among other things, producing documents and witnesses, conducting the Investigation, submitting the Report and making available the supporting evidence. The Company’s efforts to reconstruct its financial statements and improve its internal controls are continuing, but it is still not in a position to fully comply with all of its financial reporting obligations.
33
