by Gillian Tett
To reinforce that cooperation, the managers at BlueMountain insisted that all the investment ideas be put onto a common database. They also decided that the traders and analysts would receive a significant part of their compensation based not just on the individual investments they had worked on, but the results of the entire team and firm. This collaborative system was very different from the more common “eat what you kill” model at most banks and hedge funds. “We have a distinctive culture, and it doesn’t work for everyone,” Feldstein explained. “But if someone wants to work here, then they need to know that we are taking a team-based approach.”
ONE EXAMPLE OF THIS collaborative bucket-busting approach involved an investment in HanesBrands, a North Carolina–based company that produces “everyday apparel,” as its website says. In plain English, this means underwear. The range includes the Wonderbra, Playtex products, as well as brands such as Champion, Maidenform, and Gear for Sports. “In the United States, we sell more units of intimate apparel, male underwear, socks, shapewear, hosiery and T-shirts than any other company,” the official HanesBrand website declares.26 Four fifths of all households in America have at least one of its products.
In 2011, the company caught the eye of credit portfolio manager Lutova. To her, the company looked like what traders sometimes describe as a “potential bond short,” or the type of company that it was worth betting against, since the price of its bonds looked likely to decline in the future. One reason she thought it might make a bond short trade was that the price of HanesBrands’ bonds were relatively high, even though the company had a high level of debt. Since bond investors typically pay a great deal of attention to the debt profile of companies (because they want to know if they will get repaid), that high leverage level rang a warning bell. Worse still, HaneBrands’ trading margins were falling as a result of a recent increase in the cotton price, which was used in many of the underwear products. The company’s sales of printed T-shirts were suffering too because another company known as Gildan was grabbing market share. Those two factors threatened to undermine another factor that bond investors typically watch very closely: the cash flows of the company.
So Lutova set to work analyzing the company to see whether it could be a bond short, working with colleague Ami Dogra, who was the senior retail analyst at BlueMountain. But as they dug into the numbers together, their perspective started to change. Most analysts who looked at HanesBrand assumed it was a “leveraged cyclical retailer,” or a company that had lots of debt and whose fortunes were likely to go up and down in line with the wider macroeconomic cycle. This designation put it into a particular mental box in the eyes of many investors, since it meant that analysts would compare it to other companies in that definitional box, to work out whether it was a good investment or not. Once a company has been placed into a certain mental box in the investment world, it is often hard for investors to question that; our taxonomies tend to be subject to inertia.
But as Dogra looked more closely at HanesBrands, she started to question the usual definition. She thought it was not really a leveraged cyclical retailer but should be viewed instead as a “stable consumer products” business. After all, Dogra pointed out to her colleagues, HanesBrands dominated the world of American underwear, and was the largest or second largest player in many of the different underwear and clothing niches. Moreover, shoppers tended to buy underwear at a fairly steady rate, irrespective of the economic cycle. That meant that the company’s profits and cash flows were relatively stable—and, crucially, much more consistent than most leveraged cyclical retailers such as fashion outlets. And although these cash flows and margins had been dented by the rise in cotton prices, Dogra thought that HanesBrand would be able to pass the cotton price increases onto shoppers. She also thought that the company could generate new inflows of cash by reducing its working capital. Indeed, when Dogra took note of these factors, she calculated that the company should soon have cash flows worth more than $400 million a year. That made HanesBrands seem like a dramatically more attractive investment bet—particularly since the company was also planning to cut its capital expenditure and move out of the struggling printed T-shirt market. Better still, Dogra argued, although HanesBrands had a lot of leverage the management had promised to cut that debt burden significantly, from 3.6 times the level of earnings before interest, tax, depreciation, and amortization to just twice EBITDA. That alone promised to raise the company’s earnings by 18 percent. So Lutova and Dogra changed their mind: instead of betting against the company with a bond short, they decided they wanted to place a positive bet, or go long.
That conclusion still begged a crucial question: how? An obvious choice would have been to buy the bonds. But Lutova did not think that strategy made sense, since the bonds were structured in a manner that made it hard for investors to benefit from future earnings growth. But what about the shares? Normally, credit portfolio managers such as Lutova would not move into that field. But she asked for advice from David Zorub, an equity portfolio manager. He had already made investments linked to Gildan, the company that was grabbing business from HanesBrand in the printed T-shirt sector, and Zorub’s team knew about the dynamics of the cotton market. That left them initially skeptical about investing in HanesBrands, particularly because its debt burden seemed so high and its trading margins had been falling. But Lutova and Dogra showed Zorub that if you looked at the company as a stable consumer products group, and analyzed its cash flows using the tools that credit—not equity—investors typically used, the picture looked different.
Back and forth, the group tossed their ideas around, testing out Dogra’s thesis. Could the company recover its business margins given the dynamics of the cotton market? Would customers swallow price increases? What would happen to the cash flows and the debt? But eventually a consensus emerged around the idea that HanesBrands should be traded in the markets like a consumer products company, not leveraged cyclical retailer, since its cash flows were stable. That conclusion had a big implication for what a fair value of HanesBrands’ share price should be, since consumer staples were typically deemed more valuable, relative to earnings, than risky cyclicals. Indeed, when the BlueMountain analysts looked at the prospect for HanesBrands’ earnings, and its plans to pay down debt, they predicted that the share price should be twice as high.
So they put their idea to work. In early 2012, Lutova and Zorub started to buy HanesBrands’ shares on a large scale, betting that their analysis was correct. By the summer of 2013, the stock had indeed doubled—just as they had hoped. “It was a very profitable strategy,” Lutova later commented. Or as Zorub observed: “We wouldn’t have done this if we hadn’t had the type of joint collaboration that BlueMountain makes possible. That was what made the investment work.”
ON FEBRUARY 5, 2014, BlueMountain held a conference for about 100 of its investors at the Council on Foreign Relations building in New York. It was an elegant, opulent setting: the council headquarters are tucked into a gracious tree-lined street next to Manhattan’s exclusive Park Avenue. The conference room reeks of history and gravitas. However, the hedge fund’s organizers were not content with using the normal background of dark wood panels or curtains. Instead they cut a collection of gigantic silver buckets in half and fixed these to the council’s conference wall. These emitted a ghostly flow under the stage spotlights. The installation would have blended well into the displays in the Museum of Modern Art.
“What we do at BlueMountain is bucket-bust,” Andrew Feldstein explained from the podium to the assembled crowd, gesturing at the big silver buckets. “The financial system is divided into buckets and these are often very artificial. We like to break those boundaries down.” Stephen Siderow was in the room with Feldstein. So was Jes Staley, the man who had formerly run the JPMorgan investment bank. As the audience listened, Siderow, Staley, and Feldstein explained their philosophy and investing approach. They talked about their bet on HanesBrands. Then BlueMountain analysts described how they had
combined equity and debt analysis to invest in companies such as NRG Energy, Valero (another energy group), Eastman Kodak, Lexmark, and Scripps (publishing and television). Not all of these investments had paid off as handsomely as HanesBrands or the whale trades. Indeed some had barely produced gains. But the direction was clear. “The financial system is divided into buckets and these are often very artificial,” Feldstein declared. “We like to break those boundaries down.”
The audience listened respectfully. Some observers seemed to be impressed. By 2014, the ideas being advanced by BlueMountain were starting to find favor elsewhere. On the other side of the world, groups such as the New Zealand Superannuation Fund (a sovereign wealth group) and the Government of Singapore Investment Corporation (another sovereign wealth fund) were investigating similar ideas about mixing up bond and equity analysis, both in terms of how they organized themselves and handed money to outside investors. Canada’s Pension Plan Investment Board was moving in that direction too. And as these mighty sovereign wealth funds jumped into this sphere, it was starting to provoke more interest among smaller groups.
But not everyone was impressed. On the contrary, some of the investors who were sitting in the Council on Foreign Relations room, staring at the ghostly display of silver buckets, seemed distinctly wary, if not baffled, by what they saw. Many of these came from traditional asset management groups, such as pension funds, small endowments, or local government offices. Like big banks, these investors lived in a bureaucratic world, where there were clear rules about how anyone could invest money. They generally expected and wanted neat, familiar labels to be attached to trades and institutions. When they decided which hedge funds to invest in, they usually did so by measuring these funds against a box of similar funds. They did not know how to judge success in a world without clear boundaries, or the categories that seemed familiar. The type of bucket-busting that BlueMountain was pitching made them feel lost.
“This all sounds very clever, but it’s harder to see how it all works in practice,” one pension fund manager in the audience observed. Or as Feldstein admitted: “People don’t know what to make of us, because we just don’t conform to what they expect. They ask us if we are a fixed income fund, or an equity fund, or something else. When we try to explain our strategy, some get confused.”
In one sense, this presented a problem for BlueMountain. Precisely because the idea of bucket-busting was so alien, the hedge fund sometimes struggled to attract as many new clients as it would like. In a world of boxes, it was easier to pitch your wares to investors if you fit into the usual categories. Potential clients might applaud the results in theory. But some were not willing to take the plunge. But in another sense, the fact that BlueMountain was an outlier was also the secret of its success. The more that silos were ingrained in the other parts of the financial system, the more opportunity that created for institutions that were willing to challenge the artificial boundaries. Time and again, price distortions kept appearing in the markets because different teams of financiers had peculiar patterns of incentives or simply did not talk to each other or swap information. Organizational boundaries were rigid, but money was not. And that created a never-ending set of opportunities to make profits, not just for BlueMountain but for any investor who was smart enough to look at the system as a whole. Or, more accurately, for any financier who was able to think about finance not merely in terms of statistics and spreadsheets, but through the lens of social patterns—and silos—too.
* * *
I. Feldstein’s strategy, above all, was to make trades linked to how the prices of the different tranches (or slices) of the CDOs were moving in the markets. This was linked to the ratings that they were being given by the credit-rating agencies.
Conclusion
CONNECTING THE DOTS
“The real voyage of discovery consists not in seeking new landscapes but in having new eyes.”
—Marcel Proust1
IN LATE 2014, AS I was finishing this book, I met again with Mike Flowers, the lawyer-turned-computer-geek who had pioneered the big data experiments at Michael Bloomberg’s City Hall. By then his life had moved on in several ways. At the start of that year, Bloomberg’s term as mayor had come to an end, and he had been replaced by Bill de Blasio. That had sparked an overhaul of the senior staff at City Hall and Flowers had left for new pastures. So, as Flowers and I sat in a cheap Italian neighborhood café in downtown Manhattan, he told me he was working at New York University, teaching data science and government to a new generation of kids. He liked to think of it as another type of silo-busting, trying to bridge the gap between the public sector and academia. He was also working with an open data start-up, Enigma, and offering advice to other governments who wanted to create their own silo-busting skunkworks. Soon after our lunch he headed to Paris to work with French officials there.
By then he had lots of juicy stories about silo-busting to tell his students, the French, or anyone else. There was the tale of how City Hall had reduced the scourge of yellow grease dumps. There was the saga of the struggle to spot deadly fire traps. But one of Flowers’s favorite stories about silo-busting revolved around ambulances. Soon after Flowers joined City Hall, the Health Department had noticed that there was a wide variety in the length of time that it took ambulance crews to respond to 911 emergency calls. So Flowers asked his skunkworks team to look at the issue, and they stumbled on something odd: in New York, like most American regions, there were no fewer than six different bureaucratic systems involved in handling 911 emergency calls. Nobody had ever tried to connect the data behind these separate structures to get an effective overview of how the process worked, which made it impossible to monitor the system. So one of the skunkworks kids, Lauren Talbot, tried to link the statistics. After a long and painful struggle she created a central monitoring process that prompted city officials to change the scripts that telephone operators use when they handle 911 calls. That shaved several seconds off the response times.2
“This kind of thing makes my job worthwhile,” Flowers explained over lunch. “It didn’t require big changes. We just had to bring the data together. And think.”
As I listened, I realized that this is the essence of what this book is about. Most of us have an uneasy sense that our world is marred by silos. We might not use that specific word to describe the problem. However, we encounter it all the time: in bureaucracies where one department does not talk to another; at companies where teams are fighting each other or hoarding information; in societies where rich and poor or different ethnic and political groups live in separate social and intellectual ghettos, side by side. Technology should help break these barriers down. In theory, the Internet could connect us all. However, social media will not do this automatically, or even easily. Silos exist in cyberspace too. We live in a world that is hyper-connected, yet often we barely know what is happening around us.
That begs the question: what can we do? We cannot entirely abolish silos, any more than we could abolish electricity and maintain our modern lifestyles. We need to have specialists in the twenty-first-century world to create order in the face of extreme complexity and an ever-swelling deluge of data. Facebook could not operate as a company if everybody was trying to write the same piece of code all the time. Some autonomy and accountability is essential. Similarly, Cleveland Clinic would not be an effective hospital if everybody tried to treat the same patients. Central banks would not be able to conduct their monetary policy operations unless somebody inside the institution knew how economic models worked. Silos, if you define this concept as narrow, specialist groups, are inevitable.
But as this book shows, when our classification systems become excessively rigid, and silos dangerously entrenched, this can leave us blind to risks and exciting opportunities. The story of Sony, in Chapter Two, shows those perils. So does the tale of UBS, or the story of the economics profession before 2007. These stories are not necessarily the worst examples out there: silos have caused problems a
t numerous other institutions, such as Microsoft, General Motors, the White House, Britain’s National Health Service, the BBC, BP. To name but a few.
So is there anything we can do to mitigate this problem? I believe there is. In the second half of this book, I presented some stories where ordinary people have tried to master their silos, instead of being mastered by them. These stories should not be viewed as finished, neat success stories. Mastering silos is not a task that is ever truly completed. It is always a work in progress. But the stories in the second half of this book do, I hope, offer some varied ideas about what we can do to ameliorate the silo syndrome. One lesson is that it pays to keep the boundaries of teams in big organizations flexible and fluid, as Facebook has done. Rotating staff between different departments, as in the Hackamonth program, makes sense. Creating places and programs where people from different teams can collide and bond is also a good idea, be that through hackathons, off-sites, or other types of social collisions. It can also be beneficial to design physical spaces that funnel people into the same area, forcing constant, unplanned interactions. The corridors at Cleveland Clinic do this well. So do the squares at Facebook. Either way, people need to be mixed together to stop them becoming inward-looking and defensive.
A second lesson is that organizations need to think about pay and incentives. When employees are rewarded purely on the basis of how their group performs, and when groups are competing with each other internally, they are unlikely to collaborate—no matter how many expensive off-sites an institution holds, or open plan offices it creates. A key reason why UBS was so fragmented, as I described in Chapter Three, was that it had an “eat what you kill” incentive structure. The same problem besets most large financial groups. It also affects medicine, where the “eat what you treat” approach has raised health care costs in America. Collaborative pay systems, of the sort seen at Cleveland Clinic or BlueMountain Capital, are needed—at least in part—if people are going to think as a group.