Class Clowns: How The Smartest Investors Lost Billions in Education

TEXTBOOK PUBLISHING IS possibly the the oldest and largest segment of the for-profit education industry. Unlike other parts of the educational sector, there has never been significant public pressure to limit curriculum development to public or nonprofit entities. This does not mean, however, that the publishers have not been the subject of controversy or attack. Policy makers, parents, and students have often complained about product, content and price, and the historic profitability attracted plenty of competitors. Indeed, the cash-generating ability of textbook publishing despite the perceived inadequacies of the product is precisely what attracted Rupert Murdoch and many others who have sought to disrupt what was long viewed as a smug oligopoly.

Some investors have sought to join rather than beat the incumbent market leaders in textbook publishing. Achieving superior returns from this alternative strategy requires no less of an understanding of the sources of competitive advantage that allowed many of these businesses to perform so well for so long. The failure of some of the newer owners to appreciate the nuance of business and market structure has resulted in a surprising number of bankruptcies in the sector. To understand how this happened, it is worth looking at what has changed and what has stayed the same about textbook publishing. The best way to do that is to look at the full arc of the life, multiple deaths, and rebirth of one the most prestigious names in the industry.

After a relatively calm first hundred years of existence, Houghton’s ownership has changed a half-dozen times since the turn of the twenty-first century. Houghton had actually gone public once before, almost fifty years earlier. At that time, the owners, including descendants of the original partners who had steered the business since its founding in 1880, were looking for new capital to further grow the business. This time, however, all the proceeds would go straight into the pockets of its new hedge-fund owners, who were looking to salvage what they could from their disastrous foray into the arcane world of K–12 publishing.

The great irony of Houghton’s fate at the mercy of a series of financially reckless owners and managers is that the business had originally been the most conservative of its peers. Established as a printer, Henry Houghton became a reluctant publisher when he took possession of the stereotype plates produced for publishing customers who fell behind on their bills during the “Panic of 1857.”² These metal molds developed in the eighteenth century significantly reduced the cost of printing and contained the underlying intellectual property of the publishers. A second irony is that although its acquisition of these plates marked the birth of Houghton as publisher, the failure of its subsequent owners, more than a century later, to appreciate the cash flow implications of plate development for educational works ultimately and repeatedly drove the business into insolvency.

Some of the publishing partners who came together in 1880 to create Houghton had already been operating independently for almost fifty years. Even as an integrated printer/publisher, Houghton generally eschewed large author advances and was constantly looking for ways to repurpose classic texts in anthologies and new editions. At the time, non-American works did not benefit from copyright protection, which created multiple opportunities to pursue such a low-risk publishing strategy. Houghton also used the marketing heft of its stable of magazines, which then included The Atlantic, to attract authors on better terms than would otherwise be available to the publisher.

This cautious approach to publishing, combined with some key assets of Houghton’s, lent itself to educational publishing. Although the educational market at the time was crowded, with well over a hundred competitors,3 a separate education division was established at Houghton in 1882. There were three reasons for this development. First, many of the classic texts published by Houghton could easily be repurposed for use in schools. Second, Houghton had a strong children’s “list”—its established stable of titles and authors—that sold through both traditional trade and educational distribution channels. As early as 1868, the predecessor company, Houghton & Hurd, had convinced Hans Christian Andersen to both write for their children’s magazine and become a staple of their list. Finally, Houghton was able to have an outsized impact on the market because of its unique credibility in the sector. Houghton’s Atlantic magazine had established an education department in 1874 (not long after the federal government established the Office of Education in 1867), which had become a national thought leader in educational policy and pedagogy. It had particular resonance among the local Superintendents of Public Instruction, who were responsible for many school buying decisions.⁴

The launch of the educational publishing business was met with skepticism by George Mifflin, who took over upon Henry Houghton’s death in 1895. By the time Mifflin died in 1921, however, that division had eclipsed the company’s other segments in both revenues and, in particular, profit. This turn of events was driven both by demographics and Houghton’s continuing thought leadership, even after it sold The Atlantic in 1908 at the time of the company’s incorporation.⁵

On the demographic front, the number of schoolchildren had doubled since the establishment of the education division. On the intellectual front, in addition to staying ahead of the nearly constant reassessments of the ideal curriculum as new schools of education emerged around the country, Houghton had also published the modern form of the IQ test. Most notably, the U.S. Army adopted this test after the declaration of war in 1917 to assist in screening and assigning the waves of new enlisted men. By 1921, Houghton was the fourth-largest U.S. educational publisher. That said, it was still a modest business, with less than $4 million in total revenues from its educational, book, magazine, and printing operations (around $50 million in today’s dollars).

The offering itself was a modest affair. The company raised just over $8 million “to help finance its expanding activity.”6 The cash mostly sat on the company’s conservative balance sheet in marketable securities, where it exceeded the totality of its outstanding debt. Going public increased the number of shareholders from 368 to 3,700.

By the mid-1970s, when Houghton began reporting divisional revenues, it was clear not only that the company had continued its growth trajectory but also that the educational division now dwarfed all else. Printing had been closed in 1971, and all trade book publishing, including the children’s list, constituted barely 10 percent of overall revenues. The balance of the business was educational—overwhelmingly K–12, though with small but growing college and testing operations as well.

The overall growth in the still-fragmented United States educational publishing sector attracted frenetic activity not only from the general industrial conglomerates of the era but also from media conglomerates. For instance, in 1962, Time Inc. bought school publisher Silver Burdett and, a few years later, contributed this to a partnership with General Electric called the General Learning Company. The General Learning Company was sold at a loss to educational publisher Scott Foresman in 1974, but Time Inc. then purchased Foresman in 1985, only to turn around and sell it to Rupert Murdoch’s News Corp. three years later.

In 1967, Litton Industries purchased the American Book Company, the publisher that had held the top market share at George Mifflin’s death in 1921. The number two player at that time, Ginn & Company, ended up as part of Paramount after Gulf and Western’s 1975 purchase of Simon & Schuster. Simon & Schuster then spent much of the 1980s rolling up the educational publishing industry.

Houghton resisted active participation in the game of musical chairs being played in the sector. The company’s deep conservatism led it to pursue only modest acquisitions while fiercely guarding its independence.7 In 1978, when Western Pacific Industries announced that it had accumulated a stake in the company, Houghton marshaled the resources of its impressive list of authors and powerful friends to beat back the takeover attempt.⁸ Arthur Schlesinger Jr., John Kenneth Galbraith, and Senator Ted Kennedy, who was chair of the U.S. Judiciary Committee’s Antitrust Subcommittee, all got involved in the dispute, which ended quietly with Houghton buying back Western Pacific’s shares.⁹

Houghton’s fierce leader at the time of the Western Pacific assault was Harold T. Miller. As his retirement approached in 1990, after forty years at the company (with seventeen as chair and chief executive), Miller continued to insist that the company would remain independent and defended Houghton’s cautious approach to acquisitions, leverage, and paying big advances. “I can’t think of anywhere where we wouldn’t have gotten worse results,” Miller insisted.¹⁰ The company’s performance under his leadership, significantly outperforming the industry and the market as a whole, seemed to justify Miller’s confidence.

But the market was not so sure. A number of factors led many to conclude that an imminent takeover was all but unavoidable if Houghton did not do something to change its strategic approach. Houghton was, after all, “the closest thing to a pure play in an industry with 20 percent pretax margins.”¹¹ The slowing in the overall textbook market was expected to put pressure on Houghton’s high trading multiple. The continuing mergers and acquisitions (M&A) activity in the sector, combined with the lack of any debt on Houghton’s highly leverageable balance sheet, would make the business attractive to strategic buyers, conglomerates, and the emerging leveraged buyout kings. Finally, literary culture had shifted so that it now seemed acceptable to be published by an industrial behemoth or even a corporate raider. Houghton author John Kenneth Galbraith, whose uncompromising stand had been so important to the company’s takeover defense a decade earlier, was now quoted as saying, “In case of a takeover, I probably wouldn’t resist the inevitable.”¹² In February 1989, when the Texas billionaire Robert Bass disclosed that he had a 5.6 percent stake in the company, many expected a bidding war to quickly ensue.

But it was not to be. Some of the speculation on the company’s future had stemmed from its failure to name a successor to Miller, who turned sixty-six a few weeks before the Bass stake became public. When Miller announced his replacement later in the year, however, it signaled no radical change in course. Miller selected a twenty-five-year company veteran who, like himself, had started as a junior sales representative. But if Miller hoped that those similarities in résumé would translate into a similarity in approach and outlook, he would be sorely disappointed.

Nader Darehshori, who was named Miller’s replacement, liked to tell the story of walking from his family apple orchard in southern Iran to a one-room schoolhouse with outdated textbooks.13 He would later become a schoolteacher before emigrating to the United States and securing his job at Houghton in 1966. Darehshori was one of a number of new U.S. publishing CEOs born overseas who emerged in the early 1990s.¹⁴ But of this group, none used this background to press the strategic importance of internationalization as enthusiastically as Darehshori did.¹⁵

This internationalization was neither the only nor the most significant break with the conservative traditions of the past. Although soft spoken and deferential, Darehshori had a taste for the limelight. He often appeared in the society pages. As the Boston Globe described Nader and his wife, Cynthia, in an article titled “Partying with Pavarotti”: “They are one of the hardest working social couples in the city and they do not miss many key parties. Indeed, they are often behind many of the best ones.”¹⁶

Darehshori was also prone to grand public pronouncements, particularly around what he viewed as the limitless opportunities that would come from emerging technology and intensified competition. “We are in the very early stages of what promises to be a rebirth of publishing,” he told a crowd at the Bookbuilders of Boston dinner. This brave new world on which he was embarking “will parallel that of the Renaissance, when Gutenberg pulled Europe out of the Dark Ages.”¹⁷ Darehshori had a new age aspect to his futuristic vision, as he took up meditation and sent has senior management to meditative training sessions.¹⁸

Atmospherics aside, however, business results were weak. In the first few years of Darehshori’s stewardship, school publishing revenues actually fell in both 1992 and 1993. Staff reductions and corporate reorganizations were not able to drive significant bottom-line growth. An effort to outsource distribution operations proved disastrous and had to be reversed. Even as core operations struggled, however, Darehshori increasingly focused on diversifying and acquisitions. The buying binge started slowly, with three relatively small acquisitions in 1992. Then, in 1994, Houghton undertook its largest acquisition ever, spending $138 million on McDougal Littell, a Chicago-based school publisher.

In April 1995, in conjunction with the reporting of the company’s anemic financial results for 1994, Darehshori made his grandest pronouncement yet. Houghton, he assured investors, would be a billion-dollar company by 2000. This more than doubling of revenues from the most recent year, he said, would be achieved, in part, by diversifying into entirely new areas, such as network software for colleges and consumer electronic publications.19 A cursory review of the company’s organic growth trajectory made it clear, however, that this goal could not be attained without lots of acquisitions—big acquisitions.

This fact was confirmed a few months later when Houghton announced that it had beat out a half-dozen other bidders to acquire D. C. Heath from Raytheon for $455 million. Houghton’s ability to outbid much larger competitors in a cash deal of this size was not simply a function of its willingness to overpay. Houghton had established a small software division in 1982 that developed spell-checking and reference software. The fast-growing division had $14 million of revenue in 1993, and Houghton had sold a majority stake to the public in 1994 as InfoSoft International. Although InfoSoft would crash and burn within a decade,²⁰ Darehshori was able to realize almost $200 million of cash from its stake long before then, and this money was critical to financing the D. C. Heath acquisition.

The price paid for D. C. Heath was almost double what Raytheon had originally expected to achieve.²¹ Even with the benefit of the InfoSoft offering proceeds, the purchase left Houghton with a level of debt unprecedented for the once-conservative company. The normally sympathetic hometown paper expressed skepticism that the transaction was “historic,” as Darehshori had claimed, given that the “company has had a hard time growing either sales or earnings.”²² The Wall Street Journal was even more critical, pointing to the “whirlwind of activity” that, to date, had only yielded consistently “missed earnings estimates.”²³

This increasing lack of predictability of earnings was partly a function of Houghton’s own operating and strategic missteps, but it also reflected some structural changes in the K–12 publishing market. What was once a relatively steady, franchise-driven business had begun to display significant similarities to the hit-driven movie business. Although so-called open territory states still allowed local districts to select their own materials, K–12 schools were increasingly making critical buying decisions on a statewide basis through a textbook-adoption process that followed a five- to six-year cycle by subject. What had started as an effort by poor Southern states to buy efficiently had turned into a broader high-stakes arms race among publishers to secure big contracts in key jurisdictions.

With reading and math representing upward of two-thirds of so-called basal spending, it was now not unusual for a publisher to spend well over $50 million to try to secure a major adoption in one of the three largest adoption states: Texas, California, or Florida.24 As the industry consolidated around a handful of deep-pocketed diversified players, the ante to win was continually upped. In this context, Houghton was increasingly marginalized, given its capital constraints. In addition, competitors that were part of larger industrial or entertainment conglomerates could mitigate the impact of a missed adoption on overall earnings. Even a more diversified educational publisher with a larger college presence could better manage the inherent gyrations in basal publishing. In higher education, individual professors typically “adopt” a textbook and are averse to switching, which makes for much less volatile results.

The fundamentally capricious nature of the K–12 end market—subject as it is to local political whims, influence peddling, and budget crises—makes for tough sledding as a pure-play public company. A review of those that serve this industry sector and that have tapped the public market in recent decades—whether in publishing, services, or technology businesses—reveals that they are highly prone to missing expectations and that their half-life as independent public companies is brief indeed. That these are usually small-cap stocks (less than $2 billion in equity value) adds to the burden, as small-cap investors are particularly unforgiving. It is hard to think of a single K–12-focused stock that lasted long enough in the public markets to become a large-cap stock (more than $10 billion in equity value).

Houghton exacerbated its structural challenges by spending the latter part of the 1990s buying a dozen other, mostly technology-oriented companies that further diluted earnings. Notwithstanding Darehshori’s unprecedented acquisition spree, as his first decade as CEO came to a close in 2000, Houghton was still the number four K–12 publisher—the same relative position it had held in 1921. Its stock also had significantly underperformed the market over this period. In October 2000, Houghton’s closest public comparable company, Harcourt General, would be sold and split up between two much larger diversified publishing companies—Reed Elsevier, which kept the K–12 business, and Thomson, which added the higher education publishing to its Thomson Learning division.25 A few weeks earlier, Houghton had repeated what one analyst called its “unfortunate history of missing earnings expectations and disappointing investors” by announcing that its “full-year 2000 results would fall dramatically short of expectations.”²⁶

Houghton had been the subject of takeover speculation on and off since Darehshori had taken over, but this speculation now began to take on a life of its own. After collapsing in the face of the bad news about company performance, the stock crept up to the low 40s in early 2001.²⁷ By the time Dahreshori announced his retirement in April 2001, the stock was in the mid-40s. The stock ran another 20 percent in the next month, until the first reports of talks with France’s Vivendi Universal appeared.²⁸ In June 2001, Houghton formally announced that it had reached an agreement to be sold to Vivendi for $60 per share for a total value of $2.2 billion.

For all the increasing speculation about Houghton being an “obvious” takeover candidate as the last independent major educational publisher, no one seemed to have much of an idea who would actually buy it. Once word of the concrete talks with Vivendi surfaced, the New York Times reported that “analysts were hard pressed to think of other possible bidders.”²⁹

It was widely assumed that antitrust considerations would preclude any of the three larger K–12 publishers from taking on Houghton. As recently as the late 1980s, there had been a dozen “major” K–12 publishers. But over time, the four top players had swallowed up almost all the smaller competitors, so that together they now represented almost 90 percent of the basal market. Acquisitions aside, overall market shares were remarkably stable. Pearson and McGraw-Hill each had close to 25 percent of the market, and Harcourt and Houghton were closer to 20 percent.

Even if such a transaction might ultimately be approved, however, the authorities would certainly require the purchaser to divest a significant number of titles in disciplines where the combined business would otherwise dominate. Paying a lofty price for the company only to be required to undertake significant forced sales at low prices was not a particularly attractive proposition. Although there would be real cost synergies in such a pairing—not only overhead savings but also other major benefits from rationalizing duplicative marketing and distribution infrastructure—there would likely be revenue dis-synergies as well. Even in the absence of government intervention, wherever the two publishers had directly overlapping programs, the optimal strategy would be to close or migrate customers of the weaker one.

Such considerations only corporate conglomerates or an opportunistic financial investor as possible suitors for Houghton. But this was 2001. The era of conglomerate buildups in the 1960s and 1970s had already been largely repudiated and had unwound in the 1980s and 1990s.30 Although the media industry did not get the memo about conglomerates for a few years, and Time Warner had recently announced what would become the most value-destructive conglomerate merger of all time, even this industrially retarded sector had concluded that educational media added nothing of value to a consumer media conglomerate. Time Warner, News Corp., and, most recently, Viacom³¹ had all exited educational publishing by this point, having realized how awkwardly the locally focused institutional sales channel fit within its mass consumer franchises.

Luckily for Darehshori, one media mogul was content to ignore the historic lessons of his peers. Former investment banker Jean-Marie Messier was brashly committed to turning Vivendi, the once-sleepy French water utility he ran, into “the world’s preferred creator and provider of personalized information, entertainment, and services to consumers anywhere, at any time, across all distribution platforms and devices.”³² If there was any question about the scope of Messier’s ambition and self-regard, one only had to look at the title of his memoir cum manifesto: J6m.Com. The appellation stood for “Jean-Marie Messier Moi-Meme Maitre du Monde,” which translates to “Jean-Marie Messier, Myself, Master of the World.”³³

Messier would become a broader symbol for the excesses of an era in general and of the media industry specifically.34 In the context of the more than a hundred transactions totaling more than $100 billion perpetrated by Messier during his mercifully brief tenure as Vivendi’s CEO, the Houghton deal “does not really stand out in either absolute size or strangeness.”³⁵ That said, many of the allegedly strategic justifications for the purchase provided by Messier dovetail nicely with the strategic imperatives championed by Darehshori, albeit on a much smaller scale.

Messier marketed the transaction to the public largely based on rights to Curious George that either Houghton did not own or had already been licensed to Messier’s Universal Studios before the transaction.³⁶ With respect to the educational core of the business, however, Messier echoed Darehshori’s earlier pronouncements on the importance of globalization and technology. Messier emphasized that, although nothing about the deal would move Houghton from its stubbornly held fourth-place position in the U.S. market, when combined with its own largely francophone publishing business, Vivendi would now be the number two educational publisher globally. The problem was that, given the intensely local nature of educational publishing, a strong position in one national market does not translate into a meaningful benefit in another.

Messier had earlier spent $1 billion on a U.S. educational software company, and the suggestion was that this digital element would create synergies by spreading development costs across the global platform. But the essentially local nature of the business limited the possibility of combining these efforts within print or electronic media. The lack of any meaningful savings was underscored by the fact that Vivendi provided assurances that there would be absolutely no job losses at Houghton.³⁷

Messier stressed the boundless unexploited global opportunity in the realm of what he termed “edutainment,” even touting the potential educational software applications of the film The Mummy Returns. Peter Jovanovich, the CEO of Pearson Education, the market leader both globally and in the United States, described this proclamation as “utter nonsense,” stating, “We have had twenty years of these movie execs exclaiming that ‘edutainment’ is the future.”³⁸ Research analysts were no more credulous. On the record, the most that could be mustered was that the deal was “not totally illogical.”³⁹ Off the record, they complained, “There are no synergies with Universal Pictures, Universal Music or Canal Plus, which Messier has said would be his emphasis.”⁴⁰

Former Houghton Mifflin chief Harold Miller could only look on in horror. The sale to a foreign conglomerate was only the icing on the cake. Unable to do anything about the dismantling of his legacy, Miller had long before begun a series of more than a hundred interviews with former colleagues and peers to create a testament to the disappearing publishing culture he had tried to preserve as CEO. In 1970, Houghton itself published a book chronicling the company’s formative years through 1921. Miller’s angry effort, which focused on the more recent decades through his retirement, would need to find a different outlet. The final chapter of Miller’s book is titled “Independence: How We Protected It and Why.” He ends the chapter by listing dozens of once-independent U.S. publishing houses that had been subsumed by foreign conglomerates. After detailing the various synergistic defenses of these transactions, he rejected these as “boiler plate explanations [that] soon submerge discussion into the deep swamp of confusion and self-justification.”41 Although many of Miller’s criticisms of conglomerate publishing acquisitions are well founded, his complaints at times take on an almost nativist tone. “We were Boston and independent publishing,” he wrote. “That’s what the entire project and this book seek to make known.”⁴²

Barely a year later, Vivendi would sell Houghton at a loss to a Boston-based consortium of private equity buyers. This next chapter in Houghton’s history does not provide much support for Miller’s belief in independent, hometown ownership as a panacea for what ails educational publishing.

Vivendi’s acquisition of Houghton closed on August 1, 2001. Within six weeks, the 9/11 tragedy would bring the global economy to its knees and heightened investor scrutiny to the expansive strategy pursued by Messier. Although it is hard to imagine a scenario in which Messier’s Vivendi would not ultimately be dismembered, the terrifying events of late 2001 undoubtedly accelerated that process.

Within months of the closing, Messier’s CFO sent him a private handwritten note describing “the unpleasant feeling of being in a car whose driver is accelerating in the turns and that I am in the death seat.”43 The company would report a staggering $12 billion loss for 2001, and Messier would be fired in July 2002.⁴⁴ Messier’s successor rushed to hire lawyers to radically pare back the $18 billion of debt left in his wake. Houghton was high on the list of noncore assets to monetize.

Vivendi’s bankers scoured the landscape for potential buyers of Houghton Mifflin, but no “strategic” buyers emerged. As the final bid date approached, a who’s who list of well-known names in private equity were rumored to be interested. When the smoke cleared, the winner was a group led by two of the largest Boston-based firms. Together, Bain Capital and Thomas H. Lee Partners represented 80 percent of the equity required to fund the deal, which, at just under $1.7 billion, reflected a discount of 25 percent to what Vivendi had paid the previous year.⁴⁵ But just because the private equity buyers were paying less than Messier had didn’t mean they were getting a bargain. In a broad auction like the one run by Vivendi’s investment bankers, competition ensures that the clearing price is relatively full.

When a number of private equity firms compete against each other for the same asset—and when none of them contributes any strategic synergies through an existing portfolio company—what determines which one will pay the highest price? Contrary to what is often assumed, the answer is not differences in which firm has better access to leveraged financing. Major private equity firms usually are offered very similar debt terms, and any differences are rarely significant enough to change the outcome. Instead, two related factors overwhelmingly determine which fund wins an auction conducted exclusively among private equity participants—who takes the rosiest view of the company’s prospects and who is willing to accept the lowest financial returns for their investors. Each of these factors, in turn, has its own drivers.

A relatively optimistic view of an acquisition target’s future can be based on a perspective on broader trends affecting the sector, as well as specific opportunities to enhance revenue and improve margins in an undermanaged asset. This conviction can be justified based on industry knowledge gleaned from previous companies owned or by working with an experienced executive. The level of confidence a particular fund has regarding its base case projections also affects what level of returns is required for the investment. The riskier the investment, the higher the returns required to justify it. Deep sector knowledge can sometimes lower perceived risk and provide a rationale for accepting lower returns. More parochial considerations, however, can have an even greater impact on the required level of returns.

Other factors of course also come in to play. Who knows if Bain and Thomas H. Lee shared, at least to some extent, Harold Miller’s romanticism around hometown ownership of this Boston publishing icon. What is clear, however, is that they could not have been relying too much on existing management. Just before putting the business up for sale, Vivendi had hired an executive without educational publishing experience.46 This CEO was initially kept on and was positioned as the organization’s leader going forward on the highly successful road show to raise $1 billion in junk bonds to finance the deal. He left the company, however, in June 2003. After a brief period with an internal interim CEO, Houghton hired Tony Lucki, the internal candidate at Harcourt who Reed Elsevier had passed over for their global education CEO role.⁴⁷

Private equity firms typically hold their investments for between three and seven years. So, from the very start, they are focused on one burning question—how will I exit the investment? There are three primary means to monetize private equity investments: going public in an IPO, selling to another private equity investor, or selling to a strategic buyer. Many of the most successful private equity investments have been achieved from the last of these. Strategic buyers are willing to pay over to the seller at least a portion of the synergies they believe are achievable in a combination. This explains the sometimes significant “control premium” present in such deals. In an IPO, however, there is no control premium, because only a minority of the shares is being sold to the public. In a sale to another private equity firm, control is being sold, but there are no operational synergies to share with the seller.

Once Bain/Lee had put into place a new team, they quickly turned to assuring employees, customers, and the Boston community that they were committed to “reinvigorating” Houghton for the long term. Lucki, a local boy who had started his career at Houghton almost thirty years previously before moving to Harcourt, exulted that their “goal is to be leaders in everything we publish.”48 As proof of the private equity firms’ commitment, they pointed to the fact that the new owners had “poured about $155 million into expanding the company’s product line” in the first eighteen months of ownership.⁴⁹

On closer examination, however, remarkably little had really changed under new management. The money that was “poured” into the business was not done so at any greater rate than had been done previously, even under Vivendi’s ownership. And that money was essential to the operations of the business.

“Plate capital expenditures” (plate) are the lifeblood of any educational publisher. This is the modern version of the physical “plates” that turned Henry Houghton into a publisher in the 1850s. The mammoth investments required to compete in state adoption processes are the contemporary form of plate development. Ongoing spending on the publisher’s core intellectual property, though required every year, is treated as a capital expense because its useful life spans several years. Darehshori had conceded that the continuing escalation of the scale of investment required to be competitive in key state adoptions was a major reason he had sold the business when he did: “Frankly, I was finding it difficult to compete with the big boys.”⁵⁰

As a result, a clear picture of the business’s cash flow generation can be achieved only by looking at operating profit minus plate. At Houghton, plate capital expense was about $100 million. Although the precise annual amount varied based on timing of state adoption cycles and a general upward secular creep, something on that order of magnitude needed to be spent every year. As Darehshori lamented, “If you don’t invest, you don’t have products in three or four years.”⁵¹ And that fact would be perfectly obvious to any potential buyer. Accordingly, simply stopping the plate spend to milk the business was not an option.

This characteristic of educational publishing makes it very different from almost all other publishing and information businesses. The industry’s financial attributes are more reminiscent of the movie business—if the film library is not continuously refreshed, the business will slowly atrophy. The other analogy to the movie business noted earlier is its hit-driven nature. These qualities make movie studios, like K–12 education businesses, extremely difficult stand-alone public companies. This combination of high investment requirements and volatility is why all the major studios are embedded within much larger companies with cash flows dominated by the much more predictable cable channel business.

When Bain/Lee looked to the junk markets to finance their acquisition, they accordingly pointed to EBITDA as the relevant metric, and their “bankers pitched the Houghton Mifflin bonds on a similar business profile to two large Yellow Page bond deals that were successfully sold last year.”52 But those two high-margin Yellow Page businesses had no plate expense and almost no capital expenditures to speak of. With 2002 EBITDA of just over $250 million, the extraordinary success of the junk deal was due, in part, to successfully positioning it as being leveraged at a relatively conservative four times EBITDA. But if the correct number to look at was EBITDA minus plate, Houghton was actually leveraged at almost seven times—nosebleed levels that would leave very little margin for error.⁵³

So enthusiastic were Bain/Lee about the apparent willingness of the junk markets to ignore the plate spending inherent in their business that they quickly returned to put even more debt on the company. Within a few months—even before their new CEO was in place—Houghton had raised an additional $150 million. This money was not to be “poured” into the company; it was simply to serve as a dividend straight back to the private equity firms in order to reduce their exposure to the business. Although the credit ratings agencies quickly downgraded or revised to negative their outlook on the acquisition debt that had been issued earlier that year, the offering remained a complete success.⁵⁴

Houghton continued to aggressively express the view that its significant investments would soon yield dramatic results. But, of course, all of their competitors were also making investments. The best indicator of strong barriers to entry in an industry is the stability of market shares. Ironically, if Houghton’s claims of the ability to quickly shift market shares through investment were true, it would have shown that the industry was less attractive overall.

The good news and the bad news was that Houghton’s claims were not true. By the end of 2004, Moody’s had downgraded the junk bonds to reflect “heightened concerns regarding Houghton Mifflin’s financial results, which continue to fall short of Moody’s expectations and the risks associated with the company’s strategy to expand its product offering.”55 Faced with challenging results, a lack of strategic buyers in the K–12 publishing sector, and inadequate cash flow to pay down debt, the private equity owners looked to whether there was a way to bulk up their small higher-education publishing business. Two other companies, John Wiley and Macmillan (the latter owned by the private German company Holtzbrinck), had similar-sized higher-education businesses. The private equity owners engaged in strategic conversations with both about some kind of combination but to no avail. Although there were clear benefits from combining the college textbook businesses, neither was as interested in the more volatile K–12 business, and they certainly were not willing to give Houghton credit for its continued optimistic view of future results.⁵⁶

With no strategic buyers for the business, stagnant results, and increasingly high leverage, a sale to another private equity buyer seemed unlikely. This left very few liquidity options for the investors, beyond trying to continue to pull out dividends as long as the junk market allowed and potentially taking the company public. The Bain/Lee group pursued both of these avenues aggressively in 2006.

Goldman Sachs was directed to prepare a prospectus to take the business public. Not wanting to leave anything to chance, however, the private equity owners also sought to pull even more cash out of the business in advance—twice as much as had already been taken out in 2003. This move was all the more remarkable because the company had not managed to pay down any of its debt since the 2002 buyout. Indeed, as Moody’s pointed out, leverage had already increased to almost eight times 2005 EBITDA minus plate. The leverage, after the Bain/Lee group took out another $300 million, would approach an eye-popping ten times.

The dividend was successfully financed in May 2006, and Goldman was preparing to file IPO documents in June. At that point, what was truly remarkable about the transformation in Houghton Mifflin during the almost four years of much-touted hometown ownership was that it hadn’t transformed at all. EBITDA minus plate in 2002 had been $158 million, and by mid-2006 EBITDA minus plate was $159 million. What had changed was the amount of debt on the company. At the time of the original LBO, the private equity buyers put $1.097 billion on the company. Now Houghton boasted $1.6 billion in debt. This difference was due, in part, to the $450 million in dividends that the private equity owners had taken. In addition, not only hadn’t the company been able to pay down any of the principal, but also the net cash flows from the business had not even covered the interest and other expenses.

Barry O’Callaghan was a self-described “professional middle-class boy, likely to do well in his Leaving Cert” from the Jesuit-run boarding school he had attended in his native Ireland.57 After graduating with a law degree from Trinity in Dublin in 1991, he spent the 1990s as a M&A banker, bouncing from Morgan Stanley to Salomon Smith Barney (now part of Citigroup) to Credit Suisse in 1997. Then in 1999, an Irish schoolteacher turned serial entrepreneur, Pat McDonagh, decided to make O’Callaghan—then thirty years old and with no operating experience—CEO of one of his companies.⁵⁸ The company was called Riverdeep. McDonagh had already made his first fortune taking CBT Systems public in the early 1990s.⁵⁹ Although billed as an educational software company, Riverdeep had barely twenty employees, significant losses, and scant revenues.⁶⁰ Thus, at the height of the first Internet boom, McDonagh may have concluded that what he needed was not an operator but a banker to exploit the window of opportunity.

Just as NASDAQ topped 5,000 for the first time in 2000 Riverdeep went public.⁶¹ By then, Riverdeep had achieved losses of over $20 million on barely $2 million of revenues. So, O’Callaghan went about using the public sale proceeds of almost $125 million and the highly valued stock currency—at one point, Riverdeep’s stock was worth more than $2 billion—to acquire seven companies for more than $300 in quick succession over two and a half years.⁶²

Notwithstanding O’Callaghan’s deal-making frenzy, when the music stopped, Riverdeep’s stock collapsed and was delisted from NASDAQ in late 2002. At the time, the board considered replacing O’Callaghan as CEO because of his “fractured relationship” with investors.⁶³ O’Callaghan then hired the same bank that had orchestrated the sale to the public only a few years earlier to find a way for him to buy back the company at a fraction of that price.⁶⁴ It would also come to light that those same bankers had purchased shares at a steep discount in the months leading up to the IPO and had sold many of them before the share collapse, even as their research analyst touted the stock’s upside.⁶⁵

Before the 2003 go-private transaction, O’Callaghan owned 6.3 percent of Riverdeep. McDonagh was so appreciative of O’Callaghan’s efforts, however, that he agreed to pool his shares and split their combined stake of 42.4 percent equally with O’Callaghan. Soon thereafter, McDonagh and O’Callaghan bought out two private equity investors who had backed the delisting in the company rather than take the original buyout.⁶⁶ Then, in 2006, O’Callaghan purchased McDonagh’s remaining stake for $120 million—still less than the original IPO back in 2000 and a fraction of the value O’Callaghan would shortly tell new investors it was worth.

O’Callaghan had also already bought out many smaller shareholders who remained in the company rather than take the original buyout. Where he got the money to make these various purchases, as well as subsequent ones, is the subject of much speculation and debate.67 After subsequent deals led to the company’s insolvency, a political uproar ensued over the prospect that the Irish taxpayers would foot the bill for bad personal loans made to O’Callaghan by the then-nationalized Allied Irish Bank. The controversy led O’Callaghan to make public assurances that he would honor his commitments without specifying what precisely those commitments were.⁶⁸

What is beyond question is that by 2006, O’Callaghan had secured majority control of Riverdeep. That summer, O’Callaghan met with Houghton CEO Tony Lucki for dinner in Boston to discuss a strategic alliance. Lucki had briefly served on the public Riverdeep board of directors after Reed Elsevier had invested $25 million as part of a strategic alliance with the Harcourt division Lucki had then still headed. O’Callaghan learned over dinner about the planned Houghton IPO and began in earnest to come up with a way to offer the private equity owners a more attractive alternative that would leave him in control of both companies.⁶⁹

But how to engineer such a transaction was not obvious. Riverdeep was a fraction of the size of Houghton. In 2005, the company experienced the third consecutive year of revenue declines and had operating profit of just $10 million. Houghton’s 2005 revenues, at $1.29 billion, were almost ten times greater than Riverdeep’s, though its operating profit margin was similarly anemic. In the end, O’Callaghan agreed to pay $3.33 billion in cash to the private equity owners of Houghton Mifflin. This represented an unprecedented multiple—more than twenty times trailing EBITDA minus plate—for a business that essentially had been flat for years. It is not surprising that the private equity buyers took this offer, which ended up delivering them almost four times their original investment. What is surprising, even shocking, is that O’Callaghan was able to secure funds to consummate the deal.

Where did the money come from? The answer is a mix of gullible and sophisticated equity and debt investors around the world. In 2005, Riverdeep and Houghton together produced operating profit of less than $100 million. With the help of some of the world’s most respected financial institutions, O’Callaghan was able to raise $4 billion of new equity and debt based on a valuation of $5 billion for the combined enterprise. This, in turn, was justified by looking not at the modest operating profit but rather at the almost $500 million in something described in deal marketing materials as “pro forma adjusted EBITDA with full synergies.”

The $500 million included $100 million in cost ($70 million) and revenue ($30 million) synergies from the combination in 2007. But fully a quarter of Riverdeep’s modest revenues had nothing to do with the school business; rather, they came from selling productivity and other software through retail, direct-to-consumer, and computer equipment manufacturer channels. Another 16 percent of revenues were sold internationally, mostly through joint ventures and resellers, with little opportunity for cost savings. In addition, the $70 million in promised total cost synergies represented more than Riverdeep’s entire expense base. The rest of the inflated number came from neglecting to subtract “plate” or other capital expense and making adjustments for various supposedly one-time expenses, including more than $20 million in “special” bonuses received by Houghton management.

Of course, if someone believed that $500 million was the relevant financial metric, then at around ten times, $5 billion might seem like a reasonable number. Based on press reports, the presentations did not stop at rosy positionings of various pro forma adjustments and accounting definitions. Davy apparently also was not shy about suggesting that synergies could really be $200 million and that investors could double their money in two years.70

The presentations to debt investors were similarly aggressive. The capitalization summary showed the debt-to-EBITDA ratio of a full but apparently manageable level of six times, giving credit for everything Davy had marketed except the $30 million of revenue synergies. Buried on page 37 of the lender presentation, however, was a summary of the business’s actual cash flow. EBITDA (with all the previously discussed add-backs included but no credit for “synergies”) for the past twelve months was $391 million, but this did not account for $149 million of capital expenditures (“capex”)—mostly in the form of ongoing plate needs. In addition, a footnote stated that this amount excluded $49 million in costs associated with online development and a major enterprise software implementation. If the full $198 million actually spent were subtracted from the $391 million in EBITDA, then the cash flow generated in the past year was actually $193 million, at best.

The required infusion of $660 million of new equity was nonetheless accomplished, attracting not only rich Irish interests but also rich Middle Eastern interests, including the Saudi Obeikan Investment Group.71 One tool O’Callaghan used to overcome potential skepticism was the assurance that he would be personally contributing millions in the deal. As noted, where precisely this money came from is a subject of continuing speculation. More relevant, however, is that although he and some co-investors contributed about 10 percent of the new money, O’Callaghan and his team would control more than 50 percent of the equity. Part of the difference in equity came from rolling his stake from Riverdeep into the new entity at the outsized valuation described. But the biggest single source of his equity—roughly half—was simply given to him as a “promoter and management incentive,” supposedly “designed to incentivize and lock-in Mr. O’Callaghan.”⁷² Although some preferred equity is often reserved for management in a buyout, the 25 percent provided here is more than twice as much as would be typical.

It is less difficult to comprehend why the major investment banks agreed to sponsor this questionable capital structure in the marketplace—there were hundreds of millions of dollars in fees and significant bragging rights on the line. In particular, Credit Suisse as an institution benefited, as did the individual bankers involved,73 due to their long-standing association with O’Callaghan. Banks make money from transactions, and a deal-junkie client like Barry O’Callaghan was the gift that keeps on giving. Thus, there seemed to be every reason for the bank to continue to milk the relationship for all it was worth.

That said, the deal was not without risk for the investment banks. In debt offerings, the lead banks usually agree to hold a certain amount of the offering, and if it trades badly, the bank can lose millions. Furthermore, in this deal, $1 billion of the debt was in the form of so-called bridge financing. Because sellers want to close transactions quickly, they are often unwilling to wait for bonds to be marketed to investors. In such instances, the lead banks may offer to “bridge” the deal themselves, with the expectation that their temporary loan will be replaced with permanent financing shortly thereafter.

When the private equity buyers originally bought Houghton, they had used a bridge. That deal closed on December 31, 2002, and was followed by a very successful refinancing in the junk bond market a few weeks later. When it works out, as in that deal, a bridge is a very useful tool—it provides speed and certainty to the seller, while giving the buyer time to determine and secure the best long-term capital structure.

But bridges don’t always work out, and they can quickly become the proverbial bridge to nowhere. These instruments are structured to have the interest rate increase every month that they are outstanding in order to incentivize a quick refinancing. But if there is no market to refinance the instrument, this increasing interest rate is almost certainly not enough to compensate the investment banks for holding it on their books. The result is that the investment bank is losing millions, while the company’s cash is being increasingly devoured by the ballooning interest payments. This phenomenon is called a “hung bridge.” Eventually the hung bridge automatically turns into a bond at the highest possible interest rate, often placing a crushing permanent burden on the company, and the investment bank is left to sell off the bonds at a huge loss.

Goldman Sachs had been working closely with Houghton’s private equity owners to either find a buyer or take the company public. In the early reporting on the deal that fall, Goldman was described as being on board as a leader of the financing.74 By late November, however, when the lender presentations were made, Goldman had quietly disappeared from the underwriting group, leaving only Credit Suisse and Citigroup. Both would soon wish they had followed Goldman’s lead and left Barry O’Callaghan to pursue his grand vision without their help.

When Reed Elsevier, the Anglo-Dutch professional information conglomerate, purchased Harcourt General for $5.65 billion in 2001, it knew almost nothing about the U.S. K–12 educational market. The Harcourt auction had attracted its attention because the company also owned a well-regarded scholarly publisher called Academic Press, which was highly strategic to their Elsevier Science division. When Harcourt rebuffed efforts to break up the business, Reed pursued partners for an entire company bid. After securing an agreement with Thomson to take Harcourt’s higher-education assets for $2.1 billion, Reed had to decide whether it was willing to keep the K–12 assets as the price of getting its hands on Academic Press. Reed concluded that managing a leading educational business in an oligopolistic market like the United States couldn’t be that hard. They were wrong.

By 2007, Reed and its shareholders had long tired of its educational foray. Multiple earnings misses tied to the lumpy and unpredictable nature of the business came as something of a shock to the owners of remarkably steady subscription businesses characterized by high renewal rates and regular price increases.⁷⁵ In February 2007, Reed announced that it planned to sell Harcourt. Once it sold the student assessment and international businesses to Pearson in May for almost $1 billion, the company turned to off-loading the core Harcourt K–12 business.

Before making adjustments or subtracting plate, Houghton and Harcourt had almost identical, basically flat EBITDA of about $350 million in both 2005 and 2006. Harcourt generated this EBITDA on about $250 million less of revenue, resulting in higher profit margins. But Harcourt had also been investing significantly more in plate than Houghton, suggesting that, contrary to the claims of O’Callaghan and the private equity investors before him, Houghton had actually been relatively starved of curricular development investment for some time. From Reed’s perspective, this meant that potential buyers would be looking at EBITDA minus plate of around $150 million. If buyers paid Reed the same multiple of EBITDA minus plate that the private equity investors had given Vivendi in that auction, Harcourt would end up attracting about a $1.6 billion purchase price. Reed knew it would be very lucky to get $2 billion.

Furthermore, almost as soon as the Houghton acquisition closed, O’Callaghan’s auditors quit, “[a]s a result of incorrect representations made to us by the company’s parent.”76 The departing Ernst & Young had been replaced by KPMG in 2003 but had returned as the company’s auditors in 2004. Houghton would now try out a third set of auditors, PwC—hardly a good omen for investors, lenders, or business dealings of any kind. In any case, the conventional wisdom was that the antitrust authorities would require so many divestitures before it would permit such consolidation, that any deal would be uneconomic. Reed and its shareholders wanted Harcourt gone quickly and cleanly. Any proposed deal that had regulatory risk—or even significant regulatory delay—would be dead on arrival as far as Reed was concerned.

And yet, none of these impediments struck O’Callaghan as overwhelming. In the past, he had found that enough money could overcome any objections. But how could he get enough money to interest Reed in a deal with him over any simpler alternative? The answer was that he had to convince the investors and lenders in the Houghton deal that if they didn’t double down on the Harcourt deal, they would risk losing it all.

Reed was highly skeptical. Reed Elsevier is a highly conservative Financial Times Stock Exchange (FTSE) 100 company created through the merger in 1993 of two century-old publishing houses—Reed in the UK and Elsevier in the Netherlands. But O’Callaghan was correct in that enough money could overcome that skepticism. However, the price required was not a generous $2 billion but a staggering $4 billion.77 Only O’Callaghan’s own purchase of Houghton the previous year had even approached the multiple reflected in this transaction. In addition, in order to secure Reed’s assent, O’Callaghan still had to agree to pay Reed a $550 million reverse breakup fee if it did not receive regulatory clearance within a year.

The Harcourt financing made the Houghton transaction that preceded it seem almost conservative. To take out the hung bridge, the banks agreed to replace it and the term loan with new permanent facilities more than twice as large. Of course, they would benefit from the more than $250 million in financing fees and transaction expenses generated. (In combination with the Houghton deal before it, this added up to well over half a billion in less than a year.) The existing equity investors had to come up with an additional $235 million. Reed also ultimately agreed to take $300 million of the $4 billion purchase price in equity—but internally, they assumed this would be worth nothing, knowing that even at $3.7 billion, they were being paid an astounding twenty-five times EBITDA minus plate.

Closing the transaction required more than $7 billion in new financing raised through a new Cayman Islands holding company called Education Media & Publishing Group. The marketing materials the banks used pointed to the potential for $377 million in cost savings. These were claimed to cost an unbelievably low $100 million to achieve.⁷⁸ Typically, one-time costs of achieving permanent savings are one to two times the projected level of annual benefits. Furthermore, O’Callaghan asserted that the company had already exceeded the $75 million in cost savings promised in the Riverdeep/Houghton combination. According to the documents, $90 million in cost savings had already been realized or “actioned” with the realization to come later. Yet, suspiciously, for the first nine months of 2007, actual EBITDA remained stubbornly flat, and plate expense actually increased, which suggested that any synergies supposedly generated had been offset either by dis-synergies or by business declines.

Although the banks launched the deal in late October 2007, within a few weeks, the debt markets had turned against them, as the beginnings of what would become the 2008 subprime financial crisis began to manifest themselves. By the time the deal closed in mid-December, Houghton had announced the sale of its higher-education business for $750 million to support the troubled financing. Despite these efforts, the offering was downsized to a fraction of its original size, and on much less favorable terms. The banks were left holding the balance of the debt, and a five-year downward spiral of the newly rechristened Houghton Mifflin Harcourt (HMH) had begun.

Despite the difficulty in finding buyers for its debt or its products, O’Callaghan and his bankers continued to hype the story. In the summer of 2008, O’Callaghan sent shareholders a letter assuring them of their ability to repay the debt, meet covenants, and “comfortably” remain within their credit lines. Synergies in year three could total $340 million, and an incremental $100 million of product development savings, he assured investors, would yield “total cash synergy of $420 [million] to $430 [million].”79 Analysts from Davy, his Irish brokers, were even more expansive, suggesting that the latest investors to put new equity behind O’Callaghan could “double their investment over the next two years.”⁸⁰ This outcome was assured because there was “likely to be a deal with a large international publishing company such as Newscorp or Viacom”⁸¹—despite the fact that both alleged suitors had divested their educational publishing businesses long ago and had made clear their intention never to return to the sector.

Within months of these words of encouragement, the company underwent a series of ratings downgrades to the most speculative levels possible based on liquidity concerns and the high likelihood of default.82 S&P named HMH one of the European companies at the highest risk of default, while the company publically challenged the latest Moody’s downgrade.⁸³ Even as the company touted its “ample liquidity to meet its needs” and predicted market share gains and double-digit growth, Houghton’s trade imprint, the storied publisher of Curious George, stopped buying new books altogether and put itself up for sale.⁸⁴ The business was taken off the market when it became clear that the multiple of cash flow offered by potential buyers was lower than the debt multiple that burdened the overall business.⁸⁵

In March 2009, just months after the letter from O’Callaghan expressing confidence in the ability “to meet all covenant requirements,” the company reached an agreement with lenders to pay a fee and higher interest in return for the loosening of covenant restrictions.⁸⁶ Because Credit Suisse served as agent and itself still owned a significant portion of the debt, securing the majority approval of the holders was not challenging. But this did not come close to bridging the fundamental disconnect between the company’s cash needs and what it could generate. That accord was followed by further negotiations to convince some debt holders to exchange their debt for equity in the company and find some new investors. This tactic was also successful, with the Dubai royal family entering the fray, but the equity was at half the price of earlier rounds.⁸⁷ In the end, $800 million of the debt became equity, and another $1.7 billion deferred interest payments for a higher payout later. The result was that the company’s $500 million annual interest bill was reduced to $400 million.

It became almost immediately apparent that there was no way the company could service even this reduced debt load and that the dramatically lower equity price was still far too high. The shareholder base and the list of debt holders had quite significantly transformed. The balance sheet restructuring, though inadequate to solve the fundamental issues the company faced, had diluted the existing shareholders by 45 percent. O’Callaghan himself had gone from an effectively controlling position of nearly 40 percent to just 22 percent. In addition to the new shareholders from the Middle East, the debt holders who converted into equity now had a significant say in the company’s affairs. These were overwhelmingly not the original lenders; instead, they were highly opportunistic, distressed-debt hedge funds that had seen a chance to exploit the banks’ desperation to get out of the credit at almost any price.

One surprising aspect of the restructuring was that these highly sophisticated and aggressive investors did not insist, as a condition of accepting the proposed deal, on securing management with experience in operating and integrating these kinds of businesses. Tony Lucki was a generally respected publisher, but he had no experience with anything of this scale or with operating with these kinds of liquidity constraints. O’Callaghan, a former banker, did hire a president at the holding company level in 2007, between the Houghton and Harcourt deals, to complement the team.88 But this executive, Jeremy Dickens, was a former corporate lawyer from Weil Gotshal & Manges. Whatever his faults, Lucki remained the only credible senior publishing executive associated with the now massive enterprise. But then he announced his retirement in April 2009, and his replacement was the now thirty-nine-year old Barry O’Callaghan.⁸⁹ Then, even Dickens would announce his departure shortly after the restructuring, leaving no meaningful bench of talent at either the operating or the holding company level.

The largest single outside buyer of the distressed debt, at what seemed to him to be a bargain-basement price, was one of the largest beneficiaries of the financial collapse: John Paulson. Paulson’s hedge funds had earned more than $15 billion in 2007, when he bet against subprime mortgages, 20 percent of which he and employees of Paulson & Co. pocketed themselves.⁹⁰ As 2009 drew to a close, Paulson and the other distressed-debt owners—notably, Guggenheim Partners—negotiated a transaction in which more than $4 billion of the remaining debt was converted to equity. The transaction left about $2.5 billion of debt, even as $650 million of new equity—secured from the largest debt holders themselves—were scheduled to come in for working capital. Annual interest payments would now fall to $200 million.

This time, the almost $1 billion of equity previously invested along the way would be largely wiped out. The restructuring caused political repercussions in Ireland, as many of the local investors wooed by Davy had been financed by the Anglo-Irish bank, which was now owned by the Irish taxpayers after being privatized during the financial crisis. O’Callaghan, as usual, was unapologetic, complaining that “I can’t be blamed for things I can’t control” like state budgets and insisting that “operationally and strategically it has all worked out.”⁹¹ Despite the obvious evidence to the contrary, O’Callaghan asserted: “I don’t think you can say that the investment thesis was flawed.”⁹² As evidence for his blamelessness, he repeatedly pointed out that “no one has lost more than me.”

O’Callaghan’s claim that no one had lost more than he was not strictly accurate. On the one hand, he had paid far less than others for his stake by virtue of the enormous financial “promote” that he had secured in the transaction financings. On the other hand, unlike other equity holders who were being offered a small share in any eventual value accretion above $10 billion, O’Callaghan would negotiate a huge incremental financial incentive to stay on with the company.

Paulson’s thinking remained something of a mystery to the other bondholders.93 He was usually represented by an inscrutable young partner, Sheru Chowdhry, in discussions over the company’s fate. It soon became clear, however, that O’Callaghan had managed to convince one other important bondholder that he was a visionary with the key to unlocking long-term value at HMH. In public statements, the face of Guggenheim Partners was managing partner Todd Boehly. During one critical bondholder debate to decide O’Callaghan’s fate, however, the principal whose interests were being represented by Guggenheim emerged from behind the curtain to make the case for keeping the current CEO: Michael Milken.⁹⁴ Milken was apparently the largest investor in the Guggenheim fund that had made the investment in Houghton.

Although Milken’s track record in education, which is the subject of chapter 4, is undistinguished, at the time investors knew of his interest in the sector but nothing about his returns. And high-yield investors viewed him with awe due to his foundational role in the sector. Milken’s relationship with Guggenheim in general and Boehly specifically would later become the subject of a Securities and Exchange Commission investigation. During the bondholder negotiations, the other stakeholders simply assumed that Boehly spoke for Milken.⁹⁵

Others at the meeting pressed that even if O’Callaghan’s vision were credible, his reputation in the marketplace would undermine the long-term value of their stake. After all, given how unlikely a strategic sale was, the most likely exit was clearly through the public markets. In light of how many public debt and equity investors had been disappointed by O’Callaghan in the decade since the original Riverdeep IPO, he was hardly an ideal public company CEO.

Paulson did not know Milken but had agreed to meet with him privately before the fateful bondholder session. By virtue of the size of their collective holdings, Guggenheim and Paulson together would control the outcome of any bondholder vote. Milken argued forcefully to Paulson that, as the mastermind of the Houghton-Harcourt combination, it would be a mistake to fire O’Callaghan when so much of the bad news was due to market forces outside of his control.

Paulson indicated no second thoughts. He publically expressed “great admiration” for O’Callaghan. Speaking to employees, Paulson assured them that, in his view, with the team in place and the restructured balance sheet, “Houghton Mifflin Harcourt is well positioned.”96

Paulson’s professed optimism was short lived. Throughout 2010, the bad news just kept coming. Although O’Callaghan kept up a busy schedule announcing feel-good initiatives, such as free books for public libraries and a $100 million global education innovation fund, revenues continued to head downward. O’Callaghan had attracted new capital from the distressed-debt investors in the restructuring on the promise that the business had bottomed out and was poised to benefit from the promised massive synergies. But even with a reduction of 60 percent in the outstanding indebtedness, it became clear that the company could not service even these much-diminished obligations.

In early 2011, this became even more apparent, as the auditors finalized their accounts. Just a few days after O’Callaghan mused on the future of education at an expo that HMH underwrote, along with Goldman Sachs, News Corp., and AT&T for the glamorous Aspen Institute,⁹⁷ PwC signed papers highlighting “material uncertainty” regarding the company’s standing as a going concern.⁹⁸ In March, after feverish negotiations, O’Callaghan quietly stepped down—but not before negotiating the effective forgiveness of an $11 million loan and a one-year consulting deal that would pay $2 million up front, with potential for $3 million more.⁹⁹ The CFO responsible for developing the unreliable projections, a long-time Milken associate, was promoted to CEO on an interim basis.¹⁰⁰ The company insisted that the sudden departure was “not a commentary on our financial situation” and pointed to unspecified increases in market share and profit, despite the continuing overall revenue declines.

A company spokesperson assured investors that the year 2011 had “started strongly,” despite the unexplained removal of the chief executive. But, in fact, revenue for 2011 would soon fall another 14 percent. In addition, during the treacherous first half of the year, when working capital needs peak in the seasonal business, the company had to scramble to pay its bills.

To provide a more permanent solution to its cash needs and to refinance its shorter-term borrowings, the company launched a bond deal for $1.35 billion, even in the absence of a permanent CEO. The offering attracted scant investor demand and closed in May 2011 after being downsized by more than a billion dollars to a mere $300 million.101 Even this modest result was achieved only after the existing lenders agreed to buy $115 million of the $300 million deal.¹⁰² Adding insult to injury, as the company struggled to make ends meet while it searched for a new CEO, Paulson had to read that O’Callaghan was poised for a $150 million payday as he prepared to take a Chinese English-language-learning subsidiary of EMPGI public, leveraging the exclusive license for HMH’s e-learning software in China.¹⁰³

Six months after O’Callaghan’s departure, the company announced his replacement. In the interim, the company had reported disastrous results, with EBITDA in the second quarter down 67 percent.¹⁰⁴ Linda Zecher was a respected Microsoft executive who had no education experience. She had been in charge of Microsoft’s $8 billion global public-sector business, but she did not cite this experience with selling to governmental entities as a key driver for being attracted to the job. “I love reading,” she later said. “So I couldn’t pass up the chance to work there.”¹⁰⁵

Within months, Zecher announced that she would be cutting 10 percent of the staff and would replace both the head of the education group and the CFO.¹⁰⁶ She was soon thrust into negotiations with the owners and continuing lenders over the terms of a prepackaged bankruptcy. The key questions were not only how to divvy up the ownership among the various interests but also, critically, how much debt the company could support. The company once supported close to $8 billion of obligations. This amount was reduced by around $1 billion in the first restructuring and another $3 billion in the second. After much debate, an optimal level of debt was determined—none whatsoever. The company would have ample credit lines to manage the significant working capital swings inherent in the business, but no permanent debt.

When the company made its official bankruptcy filing on May 21, 2012, it looked remarkably like the stand-alone Houghton Mifflin before it had been combined, first with Riverdeep and later with Harcourt.107 Revenue was almost identical, and profit had shrunk somewhat. Revenue in 2011 and 2005 was $1.295 and $1.289 billion, respectively. The corresponding EBITDA figures before plate were $238 and $297 million, respectively. It was almost as if the deals had never happened at all. After plate spend, EBITDA remained not much more than $100 million.

It is hard to believe that only five years earlier, O’Callaghan had painted the picture of a combined entity with $2.5 billion in revenues and well over $1 billion in “pro forma adjusted EBITDA with full synergies.” To be sure, the industry had undergone significant retrenchment as states reduced their educational spending during the financial crisis. But HMH’s most direct competitors, Pearson and McGraw-Hill, had not experienced anything like the deterioration reflected in these shocking results.¹⁰⁸

When the company emerged from bankruptcy a month later, Paulson was by far the largest owner, with 26 percent, and the group of a half-dozen of the largest hedge fund holders together controlled two-thirds of the equity. And yet, one of the previous major hedge fund owners—the third largest, in fact—was conspicuously absent from this group as Houghton emerged from bankruptcy. Guggenheim and Milken had been the chief proponents of the misguided strategy of keeping and paying Barry O’Callaghan after the previous restructuring. Representing almost 10 percent of the equity at the time, they had almost immediately started liquidating their position in the company as soon as the radically downsized bond offering in May 2011 had closed.

Now the question was how Paulson would make a return on the investment. It is impossible to calculate how much Paulson actually paid for his stake without knowing precisely when his fund purchased the debt and at how great a discount. Based on the size of Paulson’s position and the overall debt trading levels, along with his contributions to new equity in the second restructuring and his purchase of new senior notes shortly before the final bankruptcy, it seems very unlikely that the ultimate investment was less than $1 billion.

Hedge funds like Paulson’s look for annual returns in excess of 30 percent to justify the risk they are taking in distressed securities. Paulson had built up his position over many years at this point. Given that the latest restructuring had eliminated all the remaining $3.1 billion of debt, it was clear that the company was worth less than that. The restructuring advisor presented a value range of $1.6–$1.9 billion.109 Thus, under any plausible scenario, Paulson had not only failed to meet his return threshold, but he had also lost hundreds of millions on an absolute basis.

The options for liquidity had only gotten worse since the previous private equity owners had faced the same issue a decade earlier. The absence of any potential truly strategic buyer had led them to prepare for an IPO, while hoping for someone to make them a better offer than what would be available from the public. Lightning struck for the Bain/Lee group in the form of Barry O’Callaghan. Given the tragic and very public history of all that ensued, however, the chances of that happening again were slim. At a minimum, any private deal for Houghton would need to involve much lower leverage than had made any of the previous mergers and acquisitions transactions possible.

Unfortunately for any potential public offering, Houghton revenue fell again in 2012. In the first half of 2013, however, Houghton showed a modest top-line gain for the first time in years and quickly filed to go public. Public offerings are usually to raise capital to fund a young company’s growth plan. This “primary” capital goes directly into a company. Even when private equity investors own a company, the owners rarely sell their own stake in the IPO—so-called “secondary” shares—because of the negative signal it sends to new investors. Such a signal almost always results in investors paying a lower price than they otherwise would.

In the case of HMH, however, the distressed investors were so exhausted from their experience to date that the entire offering consisted of “secondary” shares, with the existing investors trying to sell as much of their position as they possibly could. As sophisticated market participants, they were well aware of the potential adverse impact on the offering, but they were willing to risk it. The result was that although underwriters proposed a pricing range of $14–$16 per share, the market for the IPO shares cleared at just $12 per share. Instead of the nearly $300 million anticipated, the offering in November 2013 raised just over $200 million. The entire value of Houghton Mifflin in the public market initially was almost exactly the price of $1.7 billion paid by the Bain/Lee consortium more than a decade earlier, before the assets of Riverdeep and Harcourt had been added to the mix.

Paulson, as a result of the anemic offering, was only able to pull out about $65 million. He continued to own 22 percent of the company for almost two years. After the IPO, the HMH stock steadily climbed, and by the end of 2014, it had settled at around $20 per share. Paulson was able to exit the balance of the investment in May 2015 after a brief uptick in the stock to $23 before it returned to below $20 within a few months. Even with the fortunate timing of this sale, total proceeds of Paulson’s liquidation of his position were less than $350 million. Although it is impossible to calculate precisely, it is clear that he left the investment many hundreds of millions in the red.110

Meanwhile O’Callaghan has managed to continue to thrive personally, even as the list of unhappy partners lengthens. The planned IPO of EMPGI’s Chinese operations never happened. The Dubai state investment company that had invested $50 million to acquire one-third of HMH’s international business in 2008 completely walked away from its investment in 2013.¹¹¹ This was all part of a mysterious liquidation of the Cayman Islands holding company that resulted in a new Cayman-based entity with O’Callaghan squarely in control. O’Callaghan has attracted a board that includes not only former HMH stalwarts like Tony Lucki but also celebrities, such as Chariots of Fire producer Lord David Puttnam. Also included for added credibility is former U.S. Education Secretary Bill Bennett, who ironically made a name for himself by lecturing the public on managing excess in their personal behavior.

The restructured EMPGI now goes to market with its network of owned and franchised English language learning centers across Asia and the Middle East under the RISE brand. The RISE website describes Chairman O’Callaghan as “a high profile entrepreneur with a strong educational pedigree.” The site touts the fact that “foundational content is provided by Houghton Mifflin Harcourt,” and it prominently displays the logo of the company O’Callaghan had led into bankruptcy. Given this outcome of their previous efforts, it is remarkable that RISE further brags that “its leadership [has] overseen an investment in excess of US$1 billion in [e]ducational content and IP.”¹¹²

Although the HMH stock performed well in the first year after the reintroduction to the public market, the underlying realities of the industry dynamics had not changed. After an initial burst of enthusiastic utterances about growth in adjacent markets113 and the digital future,¹¹⁴ new CEO Zecher seemed to settle into the hard slog of selling educational materials to states and localities, which continues to be HMH’s core business. The good news was that 2014 was a banner adoption year, with Houghton reporting market share gains. Although reported revenue was essentially flat again, both billings and cash flow were significantly up. The bad news was that this set the stage for an inevitably much more challenging 2015.

Zecher has always described generating organic revenue growth as her number one priority. Early on in her tenure, she said that, to complement this, she would “perhaps make some small acquisitions.”¹¹⁵ But Houghton’s first high-profile smaller acquisition as a public company should give pause. On May 13, 2014, Houghton purchased the carcass of what was once Chris Whittle’s Channel One. Channel One was strategic, Zecher explained, because it “shares our commitment to high quality digital content.”¹¹⁶

As 2015 approached, however, she was mentioning the possibility of “larger strategic” transactions.¹¹⁷ Given the history of Houghton, despite an activist investor’s pressure to releverage the balance sheet, Zecher was sensitive to the issues of which large deals are truly “strategic” and how much leverage the company could actually support. That said, the large deal came quickly, with the purchase of Scholastic’s educational technology unit for $575 million in April 2015.¹¹⁸

Zecher had approached Scholastic’s long-time CEO and controlling shareholder, Richard Robinson, about buying all of Scholastic. Robinson had no interest in selling the business, which had traded for years at a single-digit EBITDA multiple, notwithstanding its storied children’s publishing and education brand. Robinson was frustrated. On the one hand, he couldn’t afford to buy the “ed tech” assets that came on the market and went for what he thought of as crazy prices. On the other hand, his anemic stock price failed to give credit for the significant number of digital educational assets Scholastic had developed over time. At the price Zecher offered just for Scholastic’s mostly homegrown technology businesses, which reflected her belief that she could find $20 million of cost savings, Robinson netted his desired “crazy” price even after paying taxes.

HMH stock ran to over $25 per share the summer after the deal— this is the period during which time Paulson took the opportunity to sell his remaining holdings. Hugely disappointing earnings announcements, however, quickly followed, and full-year expectations were revised downward.119 Having already releveraged the balance sheet with the Scholastic acquisition and doubled the buyback program from $100 to $200 million, the company finally acceded to activist demands and announced a massive $1 billion buyback program. This “aggressive usage of debt to fund share buy-backs” led Moody’s to downgrade the rating on the company’s credit facility in early 2016.¹²⁰ The full year 2015 ended with “one final miss to end a difficult year,” and the share price remained stubbornly in the teens.¹²¹ Billings for 2015 were down, even with the addition of the Scholastic technology assets, and overall market share had fallen from the impressive levels achieved in 2014.

Zecher now faces a choice. She can focus on the effective management of her core educational assets in a sector that is relatively low growth with significant volatility but in which market shares shift slowly and scale is a distinct advantage. Or she can double down on her embrace of the digital future through additional expensive higher-growth technology acquisitions and aggressive transformative investments to fundamentally change HMH’s business mix and model. The financial devastation that Houghton left in its wake was due to the wrong capital structure, exacerbated by a lack of operational focus—it was not due to slow growth. Thus, before following the latter path, she would do well to examine both this history and Rupert Murdoch’s experience.