JUST outside Stanford University’s campus sits the headquarters of Symphony, one of the myriad tech companies that sprout like weeds in Silicon Valley. After a lunch break exercising in a nearby park, a dozen fit-looking employees, still in workout clothes, help themselves from buckets of fruit, energy bars and the food of the day (Indian), before plopping themselves in front of monitors in an airy room bathed in natural light. For the sought-after engineers making up most of the company’s 200-strong workforce, this sort of environment is the norm.

Work is supposed to be healthy and relaxed—a far cry from the terrors of a New York bank with its incessant pressure to sell and complex internal politics, not to mention often unappetising, pricey food.

Across the continent, in a newly opened tower within the World Trade Centre, Kensho, a three-year-old company, has a similar feel. Like Symphony but a bit smaller, it is stuffed with talented engineers. In a New York approximation of the West Coast, it boasts “vertical gardens”—rectangular patches of vegetation like framed paintings—and a pool table.

Symphony is a messaging platform, owned by a consortium of investment firms. It offers a critical function at present almost monopolised by Bloomberg: the seamless incorporation of data and communication that makes the terminal the most important conduit in finance since Wall Street went from thoroughfare to metaphor. Kensho screens vast amounts of information—speeches, earnings, earthquakes and on and on—to help investors find correlations among all these data that might move prices. 

If the two companies succeed—a big if—their products could become pervasive. They are tiny entities with vast potential. And they are examples of technology firms backed and used by Goldman Sachs, a big investment bank, in its efforts to transform itself, and indeed its industry, at a time when its core business is being pummelled by technology and regulation.

In 2014 Goldman spun out a messaging technology developed internally as a new company, Symphony. Kensho was formed with backing from Goldman in 2013. Early on, the investment bank had a contractual right to be the sole user of its products among brokers. Goldman continues to be the only outside investor with voting rights on the company’s board, but many other banks have taken stakes in it and are customers.

It is possible that these two companies will provide little benefit to Goldman. Cynics are entitled to wonder whether these and similar efforts are merely a way of putting a modern veneer on an old structure. Tech companies are fashionable and widely perceived as helpful; banks are unfashionable and seen as parasitic. The non-cynical take is that Goldman understands that answers to the challenges it faces will have to come, at least in part, from outside its mirrored-glass headquarters in downtown Manhattan. It may have many flaws; a failure to grasp corporate vulnerability is not among them.

Goldman, with its enormous influence, lavish compensation and alumni network in pivotal political roles looks anything but embattled. But the firm—derisively dubbed a “great vampire squid” by Rolling Stone magazine—is in the process of seeing its tentacles severed.

Lost prop
 
Since 2009 revenues have dropped by a quarter; they remain below where they stood a decade ago (see chart). Even in a good quarter, such as the one just completed, its return on equity barely exceeds single digits. “Principal transactions”, ie, proprietary trading and investments, produced $25bn in revenues in 2009 and $18bn in 2010 but only $5bn in 2015. The decline is a result of new rules that limit these activities—and regulators threaten more.
.

Fixed income, commodities and currency (FICC), the once immensely lucrative niche that nurtured the careers of Goldman’s chief executive, Lloyd Blankfein, and its president, Gary Cohn, has also been hit hard. Revenues reached $22bn in 2009. In the first three quarters of this year they totalled $5.6bn.

Richard Bove, an analyst at Rafferty Capital Markets, concludes Goldman has just one superb business left among its distinct parts: its traditional niche of providing advice on important transactions, notably mergers. Goldman remains the global leader, despite having missed out on a role in the year’s biggest proposed deal—AT&T’s attempt to take over Time Warner.

Even its M&A business is in some difficulty—as exemplified by its big cutbacks in Asia, where governments in China and elsewhere still favour local institutions. Goldman also has a good business in wealth-management, which thrives on sophisticated schmoozing and does not require much capital.

Its other businesses, which collectively still account for about 60% of revenues, face unrelenting pressure from regulatory and technological change. None of this is unique to Goldman. A recent survey of 35 global investment banks by the Boston Consulting Group implied a long-term, industry-wide contraction: over the past five years, revenues have declined by 20%, return on equity has slipped from an inadequate 9% to 6%, and almost every business area has shrunk, with the exception of the advisory work that represents only about one-tenth of the overall pie.

Regulators want investment banks to reduce risk, and to do so by cutting out businesses that directly support their own returns as opposed to those of clients. That means they cannot hold large inventories of securities, must reduce proprietary trading and must take on ever more capital (diluting returns). They should, in short, be intermediaries.

But that intermediary role is also under attack. Big fixed-income investors say they can underwrite many debt offerings directly. Fewer companies want to issue public shares. New competition has emerged in the shape of more than 300 “fintech” companies, a broad term for entities using technology to do what existing banks do with more people and at higher cost.

Monetary stimulus
 
So far-reaching are these changes that it is surprising bank revenues have not fallen further.

The most common explanation is that repressed interest rates have stimulated many borrowers to refinance their debt more cheaply. If so the positive news will be transitory; the pressures will endure.

In deference to these trends, Goldman describes its strengths in terms of characteristics—superior contacts and execution—rather than specific franchises (which may be imperilled). That provides a framework for four intertwined strategies.

The first involves collaborative efforts or strategic investments that gave rise to Symphony, Kensho and a number of similar ventures. “Orbit”, for example, is a suite of smartphone apps Goldman developed that enable e-mailing, browsing and file-saving within an environment controlled by an employer (and thus accessible by a regulator). It was spun off last October to another publicly traded company, Synchronoss, in exchange for a minority stake. Such ventures are more valuable if used more broadly than just by Goldman. If it had retained control, potential customers might be unwilling to allow a competitor access to sensitive information.

The second prong is automation. Not all that long ago, 600 people worked on a vast floor trading shares. Traders yelled and phones were slammed (though perhaps with more decorum at blue-blooded Goldman than elsewhere). Obscured by the din were 66 distinct actions, many of them amenable to mechanisation. Now, Goldman has two people who trade equities and another 200 software engineers who work on systems that, in effect, do the job on their own.

Traditional investment-banking is ripe for change as well. Goldman has mapped each of the 146 steps of an initial public offering in 51 charts that appear in proper sequence on a five-foot long roll-out. Costly, redundant steps are being cut or, once again, automated.

The next big change is in the bank’s sources of funds and its lending. Goldman pays just under 5% interest on its long-term debt, the most stable component of its funding. Its competitors, JPMorgan Chase and Bank of America, pay a fraction of one per cent on trillions of dollars of government-insured deposits. It is not feasible for Goldman to open branches. Nor, these days, is it necessary. In April, it acquired GE’s internet bank with $16bn in savings accounts, on which it pays an average of 1% in annual interest.
.

On October 13th, as expected, it launched an online lending arm to match, named Marcus after the firm’s founding Goldman. Clients will pay from 6% to 23% a year for loans of up to $30,000, to be used to repay more expensive credit-card debt. The clients are those huddled masses previously not affluent enough to afford a human Goldman account-manager, but now, apparently, an attractive market for a Goldman machine.

And that leads to the fourth change—in how Goldman interacts with clients. Not long ago, it was almost entirely through phone-calls, e-mails, electronic orders and presentations delivered in person.

Now, a client portal named “Marquee” gives access to tools such as Goldman’s risk analytics for trading shares or arranging hedges (named “Studio”) or for corporate clients to create strategies for executing large share buy-backs (“Athena”).

Behind the paywall
 
Among the largest challenges for this effort at reinvention is how to charge. The old methods—large fees on deals, commissions on trades, extraction of spreads (often in opaque ways) between the price paid by buyers and that received by sellers, the use of information gained in transactions for proprietary trades—are somewhat compromised. Clients know too much and can do too much on their own. New methods are being considered, such as a fee based on the number of employees at a firm, or number of users, or some form of subscription-based remuneration.

The change in environment is accompanied by a change in the Goldman kind of person. One-quarter of its employees now have a background in some facet of technology, be it a degree in mathematics, engineering, computer science or the like. That is up from 5% not long ago (a number it believes is still common for other banks). And the number of internal engineers underestimates the change since it does not include outside investments, such as Symphony and Kensho.

Perhaps the oddest aspect of this transformation is how little evidence exists of a payoff.

Athena, the firm says, has been used in many share buy-backs. Another tool named “Simon” is widely used by customers who want to create customised “structured notes”, or debt instruments. Kensho is profitable. Symphony has many adopters. But it is early days and in many ways these are just experiments.

Further transformation is still to come and if, as seems probable, it enhances efficiency, then the Goldman of the future may do as much as it does now with far fewer people and smaller costs. In the past, Goldman’s rise to the pinnacle of the investment-banking stack was a consequence of besting its rivals. The challenge now is less from them than from a difficult external economic, technological and regulatory environment. As for every other bank, change is not a matter of choice.