HP and Autonomy: how to lose $8.8bn

Making $8.8bn disappear is not easy. Hewlett-Packard managed it, and quickly, when it bought the information management software company Autonomy in 2011 for $11.6bn and wrote off 80 per cent of the purchase price a year later.

HP’s history is rife with self-inflicted injury and the Autonomy affair is, in part, a depressingly familiar story: a company in crisis overpaying for an acquisition it can tout as transformative. Heads have already rolled. Léo Apotheker, HP’s boss at the time of the deal, departed almost immediately afterwards and Raymond Lane, the board’s chairman at the time, and two other directors followed him last month.

HP, however, attributed more than $5bn of the writedown to “accounting improprieties, disclosure failures and outright misrepresentations”. It alleges that low-margin hardware sales were disguised as high-margin software sales and that products were sold into the distribution channel when there was no buyer.

Mike Lynch, Autonomy co-founder and former chief executive, argues it is hard to believe that malfeasance on a scale to merit a $5bn writedown could have slipped through the due-diligence process. And so it is. Yet it is equally hard to see how, as Mr Lynch contends, the damage was done in a single year of poor management by HP. Autonomy was run as an independent unit within HP, led by Mr Lynch for the first seven months, and the trouble began soon after the deal was completed. In its last quarter as an independent company, Autonomy reported $250m in revenues. Yet in HP’s second quarter with Autonomy as a subsidiary, HP’s software division, which had been growing nicely on its own, added just $175m in new sales, and Autonomy software sales were down from the year before. (Mr Lynch argues that the drop in revenues is due to a combination of accounting changes and mismanagement by HP.)

Neither is it easy to see why HP’s current CEO, Meg Whitman, on the board at the time of the deal, would make a very serious accusation in which she was not confident.

See you in court

HP shareholders filed a lawsuit, expected to be strenuously contested, against the company’s board and its advisers – Barclays Capital and Perella Weinberg – in a US District Court last week. It alleges that HP ignored warnings of accounting improprieties and weak growth at Autonomy, then ran an abbreviated due-diligence process, and is claiming to have discovered the improprieties only after the fact to cover for its grotesque overpayment.

Read for free
If you enjoy in-depth analysis like this, register today on FT.com to see up to eight free stories a month

The issues of overpayment and of the financial impact of the alleged misrepresentations are inextricably linked. A company bought at a fabulous price (HP paid 26 times trailing cash flow for Autonomy) must suffer a big drop in value if expectations for its growth are reduced. If HP expected a 20 per cent medium-term growth rate in Autonomy’s cash flows – which it would have had to expect to justify the price – cutting that expectation to 10 per cent could easily justify a $5bn writedown. To overpay is to wager that high hopes will be met.

So the core question is whether Autonomy’s growth was, in quantity or quality, not what it seemed to be, and whether HP had any reasons to suspect as much. During Autonomy’s latter years as an independent company, a small group of analysts – notably Paul Morland of Peel Hunt and Daud Khan of JPMorgan Cazenove – argued that growth looked overstated. Autonomy responded in detail. That debate takes on new resonance now.

Buy, buy, buy

The first area of contention was acquisitions. From 2004 through to its purchase in 2011, Autonomy reported after-tax cash flows from operations of $1.2bn. It spent $1.9bn on acquisitions over that period. This is striking for a company that described itself as “one of the least acquisitive players in the software universe”.

Autonomy reported strong organic growth even in periods following acquisitions but these reports are hard to assess in the standard way because of its unique acquisition model. It would take the acquired company’s products and ram Autonomy’s own technology – the Intelligent Data Operating Layer (Idol) – into them. As a result of this integration, Autonomy argues that it is impossible to show results from acquired businesses separately from the results of the combined group. Furthermore, the parts of acquired businesses that did not fit into Autonomy’s high-margin model (a pure software business with an indirect sales channel and no services component) were promptly discontinued. Thus, organic growth was calculated as total growth minus the discontinued operations of the acquired companies.

Take, for example, Autonomy’s acquisition of US software company Interwoven for $775m in March 2009. Interwoven had $260m in revenue in 2008, a 16 per cent rise in growth compared with 2007. Assuming that Interwoven’s 2009 revenue did not grow, one might expect it to contribute roughly $200m to Autonomy’s sales in the nine months till the end of that year. That would account for about 85 per cent of Autonomy’s 2009 growth, leaving an organic growth rate of under 6 per cent. But Autonomy reported that $66m of Interwoven’s revenue had been discontinued. Organic growth in 2009 was 16 per cent.

It may seem crazy to buy a competitor at a high multiple of earnings only to shut down big parts of the target’s businesses. It makes more sense if, as Autonomy argues, the discontinued operations had low or no margin. And the majority of analysts following the company found the Autonomy acquisition model more than plausible. The market agreed: Autonomy’s share price rose briskly in 2009.

Cash conversion

Questions have also been raised about cash. One of the great features of the software business, as normally structured, is that companies’ cash flows are often equal to, or in excess of, reported profits. Customers generally sign maintenance contracts at the same time as they buy software licences, paying cash up front. The company books only part of that cash as revenue, recognising the rest over the life of the contract. This means that operating profits as reflected on the income statement – in earnings before interest, taxes, depreciation and amortisation – can be surpassed by money the company takes in, as reflected on the cash flow statement.

Yet in 2006 and in 2008, years of strong revenue growth, Autonomy saw no growth in revenue that it deferred on to its books. Instead, deferred revenue grew in big leaps in 2005, 2007 and 2009, all years where the company made acquisitions.

Autonomy’s cash conversion ratio (the measure of operating cash flow generated relative to ebitda) from 2004 to mid-2011 was about 80 per cent, relatively low by the standards of growing software companies. Over that period Autonomy grew from $65m in annual revenues to nearly $900m. Over similar periods of rapid growth Sage Group, for example, converted 85 per cent of ebitda to cash, Adobe 90 per cent and Informatica almost 100 per cent. Low levels of cash conversion raise the concern that a company may be recognising profits aggressively.

. . .

Nothing to see here

Autonomy offers a straightforward explanation for the patterns in its deferred revenue account. Over the years leading up to the Interwoven acquisition, it had been moving away from licence sales and towards a software-as-a-service model where customers paid in arrears, generating no deferred revenue. Yes, peers such as Salesforce.com have billing models that generate significant deferred revenues – but these cater to smaller clients. Autonomy’s larger contracts are not suited to pre-payment.

On cash conversion, Autonomy’s explanation is more complex. The company accounts for much of the cash gap by pointing to $172m in working capital deficits that came with the companies it acquired. Autonomy often funded its acquisitions in shares issued at a discount. In order to limit the dilution associated with the share issue, Autonomy would strike an agreement with the target company making it accumulate as much cash on its balance sheet as possible by collecting aggressively from clients and delaying payment to creditors. This accumulated cash in effect helped fund the acquisitions without the dilution associated with selling shares. Ultimately, of course, the negative working capital has to be replaced – but that could be done over several quarters as the newly enlarged Autonomy generated cash from its operations. It is an internally coherent explanation but also a description of a highly unusual form of merger financing.

Water under the bridge?

These controversies – as esoteric as they may seem to non-accountants – could prove as important as HP’s claims about recognition and classification of revenue. Those specific claims have grabbed most of the attention but HP described them as examples and made broader reference to Autonomy’s “core growth rate” and business mixture. Both sets of claims paint a picture of a company unable to sustain high levels of organic growth while keeping up appearances with aggressive accounting.

Autonomy’s accounts and its acquisitive history raise legitimate questions about the quality of its growth. Suggestive patterns are not, however, to be confused with evidence of wrongdoing. And two crucial points cut against HP. First, the longstanding controversy about Autonomy’s growth was played out publicly in analyst reports, company earnings calls, the media and the investment community. HP simply cannot claim to have been taken by surprise by issues involving Autonomy’s organic growth.

Second, HP’s accusations came – or should have come – after a deep due-diligence process where it would have had access to much more than just the publicly disclosed financials. It is hard to believe that sales mischaracterisations and channel stuffing could have survived such a process. The shareholder lawsuit claims that the process was compressed. If that is true, it is a terrible indictment of both companies and the bankers and auditors working for them.

Where does this leave HP? If the UK Serious Fraud Office, or the US Securities and Exchange Commission and Department of Justice, find substance in the accusations, HP will hardly look good. But if the accusations are refuted, it should spell the end of Ms Whitman and raise the prospect of expensive settlements with aggrieved shareholders. It is, ironically, the possibility that no crimes were committed against HP that its investors should fear the most.

Copyright The Financial Times Limited 2017. All rights reserved. You may share using our article tools. Please don't cut articles from FT.com and redistribute by email or post to the web.