EOE060 Selected articles (temporary package) 15th Sep. 2018 - Traditional vs Newcomers: Who will be the winner?
https://eoesociety.blogspot.com/2018/09/eoe060-selected-articles-temporary.html
The coffee cart’s view
Lehman ten years on: more has changed than meets the eye
In other parts of finance the sense of stasis is yet more striking. Credit-rating agencies provided sweeping and unfounded endorsements of the mortgage-related securities that turned out to be toxically risky, thus allowing fuel for the great conflagration to build up unheeded. This led to calls for heads to roll, more competition and a new regulatory structure. The most recent update from the Securities & Exchange Commission, though, published in December 2017, shows that the industry’s three main firms, Moody’s, Standard & Poor’s and Fitch, still accounted for 96.4% of all ratings.
Automatic for the people
The tide eventually turned. Of the 55 banks operating in Miami-Dade county in 2008, only 28 are still standing. But on the land behind Grove Bank and Trust, the oldest of them, there are two brand-new luxury towers designed by Rem Koolhaas, a famous architect. Flats in the towers sell for more per square metre than anyone in Miami paid before 2008. If many of those who lost their homes have not been able to buy new ones, nationwide the amount of equity in homes has recovered.
The worst excesses of that market—“no-doc” mortgages, fraud, flipping and credit stuffed down the throats of any borrower with a pulse—have largely disappeared. But the “government-sponsored enterprises”, Fannie Mae and Freddie Mac, which operate as conduits between mortgage firms and the capital markets, are still very much in evidence. Both were nationalised the week before the fall of Lehman, their capital far too low to cover the risky assets on their balance-sheets. There had been talk of unwinding them, or privatising them, or dissolving them. Instead, the allegedly temporary “conservatorship” that began in September 2008 continues, with no end in sight.
How banks fund themselves, a process that is good when dull and deadly when interesting, was utterly fascinating during the crisis. Credit markets ground to a halt. A decade later things are reassuringly dull. Banks in Europe and America tracked by the Bank for International Settlements (BIS) have substantially increased their deposits, a particularly stable funding option (see chart 2). They have sufficient amounts on hand, unlike a decade ago, to fully cover lending. Europe’s banks, which binged on dollar funding in the run-up to the crisis and quickly came to regret it, have retrenched. Overseas claims held by the continent’s banks fell by 40% between 2007 and 2016, according to the BIS. European investment banks, in particular, have lost a lot of ground compared with their American competitors.
When capital is a constraint and funding is more expensive, bank bosses behave differently. Highly leveraged and volatile businesses, notably trading, have declined by a lot—and in many institutions they have not been replaced. Areas that have been encouraged, such as wealth management, require little capital and produce adequate returns. That suits the new era. But it cannot approach the pre-crisis peaks of profitability. In 2005-06 euro-area banks were delivering returns on investment of 10.7%; in 2015-16 they were just 4.4%.
As a result, banks have become more cost-conscious. There have been cuts in staff across the board (see chart 3); at Bank of America by a third, at Citigroup by 44%, and at Barclays by half. If you want in-house baristas, go to the Bay Area tech giants; JPMorgan Chase provides its staff with access to coffee by leasing space to Starbucks.
More quietly, the chastened banks have been outsourcing large parts of their operations. Not all that long ago, a common refrain by clever consultants was that banking would endure but banks would not, says Douglas Merrill, head of Zest Finance, a credit-evaluation startup. The industry, it was confidently predicted, would be upended by smart, agile “fintech” operators. The sole company to use technology to gain leadership in a banking business, though, was Quicken Loans, which has become America’s largest home lender. This had as much to do with its superb pre-existing customer-service operation as with the deft technology built around it.
Some startups have changed their business models from competing with banks to supplying them. Zest, for example, initially intended to use its software to build a loan portfolio. Now it licenses its credit-analytics to financial institutions. Three companies with no public name recognition, Jack Henry & Associates, Fiserv and FIS, handle data processing and mobile-banking software for thousands of financial institutions. Banking operations have consolidated a good bit more than the banks themselves have.
In general, the regulatory system that has been assembled in America is complex and opaque. Keefe Bruyette & Woods, a research firm, reckons that over the past decade the 67 large institutions that it monitors paid out $243bn in 219 different fines and settlements—on average 12% of tier-one capital for the 31 American banks and 6% for 36 international ones (see chart 4). This does not include all of the penalties: Keefe does not track fines under $100m. Such numbers now seem so small in these contexts they often fail the relevance tests requiring inclusion in corporate accounts, and are thus invisible to the outside world.
But though the legal environment has been rich in claims it has been poor when it comes to facts. Cases have invariably been settled rather than going to trial. Why fines have been the size they have been, and what exactly each is a penalty for, has often been left unclear. Banks defended settling by saying they had no choice—merely going to court would lose them their licences to operate. Individuals justified the opposite approach with a similar logic—anything short of total exoneration would end their careers. With rare exceptions, when defendants fought back, prosecutors either backed down or lost.
“That”, says Peter Wallison of the American Enterprise Institute, a think-tank, “is a time bomb”—the resurrection of a system of poorly supported mortgages that could prompt another crunch, though not as severe as the one in 2008.
Even for those who believe that the regulatory expansion has worked well, there is the risk that market participants now believe that if conditions sour, another successful bail-out will be forthcoming, thus producing tolerance, if only quietly, for excessive risks. Provisions included in the new rules that prohibit the bail-out of individual institutions were designed to blunt just this sort of expectation. But this limitation, among others, is under attack by many of the individuals who were involved in Lehman’s demise a decade ago on the ground that, in an emergency, all tools should be available.
Such concerns may be excessive. What the crisis showed is that, in a pinch, authority can expand. More difficult is taking the heat out of problems before they boil over. If it is time to believe the crisis that began in 2008 has really ended, it is past time to wonder how the new conditions which have come about in its wake could contribute to the one that comes next.
Until then, Masood will continue to caffeinate the remaining bankers of 745. They are, he says, better than the construction workers who put up the building in the first place, and tended to bring their own food to work. But nothing could replace the glory years before 2008. Masood got great tips—including, once, a cheque for $10,000, to take his family on holiday. He returned it, saying he could not accept so much that he had not earned; his would-be benefactor said he understood the sentiment. If such understanding ever spreads, it would be no bad thing.
Peerless
Funding Circle, a British tech unicorn, plans an IPO
It is not hard to see why policymakers are keen on a firm that provides alternative sources of capital to small businesses. In the past, especially in Europe, banks have been such companies’ only option—and they have not always been a good one, since most banks see small-business lending as a minor and not particularly lucrative part of their business.
Funding Circle boasts of having channelled £5bn in loans to over 50,000 companies, in Britain, America, Germany and the Netherlands, since its launch in 2010. Lenders range from retail investors to private institutions like Citibank and public bodies like the European Investment Bank and Germany’s state development bank. For Anton Ruddenklau of KPMG, a consultancy, Funding Circle’s IPO is a sign that its business model “is here to stay.”
The offering, which could value the firm at more than £1.65bn, is a bellwether of the health of London’s financial technology scene. Fintech, as it is known, continues to attract venture capital, including a record-setting $1.8bn in 2017 and over $850m so far this year, compared with just over $700m in the whole of 2016, according to PitchBook, a data provider. Of Europe’s eight fintech unicorns (privately held firms valued at more than $1bn), six are British, according to CB Insights, another data firm. One of them, Acorn OakNorth, announced on September 6th the completion of a new round of fundraising that values it at $2.3bn.
Brexit will bring obstacles. Fintech bosses are particularly worried that it could get harder to hire talented foreigners. Firms like Currencycloud and Azimo have been looking into setting up offices in continental Europe. Mr Desai, for his part, says he trusts the British government will continue to allow skilled workers into the country. For his firm the lead-up to Brexit, which has crimped lending by Britain’s banks, has been an opportunity, he says.
Still, peer-to-peer lending is a tricky business. The industry is so young that it has yet to weather a recession. Gaining new customers is expensive: in 2015 Funding Circle spent over 60% of its revenues on marketing (it has since reduced the share to 40%). And regulators are paying more attention. In America, Lending Club, a rival to Funding Circle, has been sued by a regulator for charging hidden fees. In Britain, the Financial Conduct Authority is looking into barring ordinary retail investors from peer-to-peer lending, and limiting it to sophisticated or wealthy investors, thereby rather straining the definition of “peer to peer”. Nice words from the chancellor do not guarantee smooth sailing.
To hear an interview with Samir Desai, Funding Circle’s chief executive, subscribe to Economist Radio on your podcast app and search for Money Talks
The deciders
How social-media platforms dispense justice
The forum has become a frequent stop on the company’s publicity circuit for journalists. Its working groups recommend new guidelines on what to do about, say, a photo showing Hindu women being beaten in Bangladesh that may be inciting violence offline (take it down), a video of police brutality when race riots are taking place (leave it up), or a photo alleging that Donald Trump wore a Ku Klux Klan uniform in the 1990s (leave it up but reduce distribution of it, and inform users it’s a fake). Decisions made at these meetings eventually filter down into instructions for thousands of content reviewers around the world.
Seeing how each company moderates content is encouraging. The two firms no longer regard making such decisions as a peripheral activity but as core to their business. Each employs executives who are thoughtful about the task of making their platforms less toxic while protecting freedom of speech. But that they do this at all is also cause for concern; they are well on their way to becoming “ministries of truth” for a global audience. Never before has such a small number of firms been able to control what billions can say and see.
Politicians are paying ever more attention to the content these platforms carry, and to the policies they use to evaluate it. On September 5th Sheryl Sandberg, Facebook’s number two, and Jack Dorsey, the boss of Twitter, testified before the Senate Select Intelligence Committee on what may be the companies’ most notorious foul-up, allowing their platforms to be manipulated by Russian operatives seeking to influence the 2016 presidential election. Mr Dorsey later answered pointed questions from a House committee about content moderation. (In the first set of hearings Alphabet, the parent of Google, which also owns YouTube, was represented by an empty chair after refusing to make Larry Page, its co-founder, available.)
Scrutiny of Facebook, Twitter, YouTube et al has intensified recently. All three faced calls to ban Alex Jones of Infowars, a conspiracy theorist; Facebook and YouTube eventually did so. At the same time the tech platforms have faced accusations of anti-conservative bias for suppressing certain news. Their loudest critic is President Donald Trump, who has threatened (via Twitter) to regulate them. Straight after the hearings, Jeff Sessions, his attorney-general, said that he would discuss with states’ attorneys-general the “growing concern” that the platforms are hurting competition and stifling the free exchange of ideas.
Protected species
When the online industry was limited to young, vulnerable startups this approach was reasonable. A decade ago content moderation was a straightforward job. Only 100m people used Facebook and its community standards fitted on two pages. But today there are 2.2bn monthly users of Facebook and 1.9bn monthly logged-on users of YouTube. They have become central venues for social interaction and for all manner of expression, from lucid debate and cat videos to conspiracy theories and hate speech.
At first social-media platforms failed to adjust to the magnitude and complexity of the problems their growth and power were creating, saying that they did not want to be the “arbiters of truth”. Yet repeatedly in recent years the two companies, as well as Twitter, have been caught flat-footed by reports of abuse and manipulation of their platforms by trolls, hate groups, conspiracy theorists, misinformation peddlers, election meddlers and propagandists. In Myanmar journalists and human-rights experts found that misinformation on Facebook was inciting violence against Muslim Rohyinga. In the aftermath of a mass shooting at a school in Parkland, Florida, searches about the shooting on YouTube surfaced conspiracy videos alleging it was a hoax involving “crisis actors”.
In reaction, Facebook and YouTube have sharply increased the resources, both human and technological, dedicated to policing their platforms. By the end of this year Facebook will have doubled the number of employees and contractors dedicated to the “safety and security” of the site, to 20,000, including 10,000 content reviewers. YouTube will have 10,000 people working on content moderation in some form. They take down millions of posts every month from each platform, guided by thick instruction manuals—the guidelines for “search quality” evaluators at Google, for example, run to 164 pages.
Although most of the moderators work for third-party firms, the growth in their numbers has already had an impact on the firms’ finances. When Facebook posted disappointing quarterly results in July, causing its market capitalisation to drop by over $100bn, higher costs for moderation were partly implicated. Mark Zuckerberg, the firm’s chief executive, has said that in the long run the problem of content moderation will have to be solved with artificial intelligence (AI). In the first three months of 2018 Facebook took some form of action on 7.8m pieces of content that included graphic violence, hate speech or terrorist propaganda, twice as many as in the previous three months (see chart), mostly owing to improvements in automated detection. But moderating content requires wisdom, and an algorithm is only as judicious as the principles with which it is programmed.
At Facebook’s headquarters in Menlo Park, executives instinctively resist making new rules restricting content on free-speech grounds. Many kinds of hateful, racist comments are allowed, because they are phrased in such a way as to not specifically target a race, religion or other protected groups. Or perhaps they are jokes.
Fake news poses different questions. “We don’t remove content just for being false,” says Monika Bickert, the firm’s head of product policy and counterterrorism. What Facebook can do, instead of removing material, she says, is “down-rank” fake news flagged by external fact-checkers, meaning it would be viewed by fewer people, and show real information next to it. In hot spots like Myanmar and Sri Lanka, where misinformation has inflamed violence, posts may be taken down.
YouTube’s moderation system is similar to Facebook’s, with published guidelines for what is acceptable and detailed instructions for human reviewers. Human monitors decide quickly what to do with content that has been flagged, and most such flagging is done via automated detection. Twitter also uses AI to sniff out fake accounts and some inappropriate content, but it relies more heavily on user reports of harassment and bullying.
As social-media platforms police themselves, they will change. They used to be, and still see themselves as, lean and mean, keeping employees to a minimum. But Facebook, which has about 25,000 people on its payroll, is likely soon to keep more moderators busy than it has engineers. It and Google may be rich enough to absorb the extra costs and still prosper. Twitter, which is financially weaker, will suffer more.
More profound change is also possible. If misinformation, hate speech and offensive content are so pervasive, critics say, it is because of the firms’ business model: advertising. To sell more and more ads, Facebook’s algorithms, for instance, have favoured “engaging” content, which can often be the bad kind. YouTube keeps users on its site by offering them ever more interesting videos, which can also be ever more extreme ones. In other words, to really solve the challenge of content moderation, the big social-media platforms may have to say goodbye to the business model which made them so successful.
https://eoesociety.blogspot.com/2018/09/eoe060-selected-articles-temporary.html
EOE060 Selected articles (temporary package) 15th Sep. 2018
The coffee cart’s view
Lehman ten years on: more has changed than meets the eye
But is the restructured financial sector safe?
Print edition | Briefing / Sep 6th 20ㅊ18| NEW YORK
MASOOD, who uses no other name, is the proprietor of the Express Café coffee cart that sits outside 745 Seventh Avenue in midtown Manhattan. He thus had a front-row seat for the global financial crisis. In 1998 he was selling his wares to the workers at the building site where a new 32-floor building for Morgan Stanley was going up. In 2001 he greeted the employees of Lehman Brothers, whose downtown headquarters had been damaged and polluted by the attacks of September 11th; having been working shifts in a nearby hotel, they were delighted to move into the building Morgan Stanley had ceded to them. He was there on September 15th 2008 when Lehman’s employees filed out with soon-to-be-iconic cardboard boxes. And he was there in the days that followed, when many of them came back, some with the same boxes, to work on cleaning up the mess under the auspices of Barclays.
The fall of Lehman was not the beginning of the global financial crisis, nor the moment when risks were highest: that came in the days and weeks that followed, as more and more institutions started to teeter. But it was the point at which the previously unthinkable became real. A big bank with a portfolio of dodgy securities that dwarfed its capital could no longer fund itself—and it was allowed to fail. The world’s credit markets froze with fear. A complete collapse of the financial system seemed plausible, and with it a global depression much deeper than the parlous recession which actually followed.
But in many ways the past ten years have seen little change at 745—though Masood points out that the strangely glitzy zipper of display screens that runs across the building’s façade has gone from Lehman green to Barclays blue. The sky did not fall in. Nor has the sustained rebound in prices which is now setting record stockmarket highs produced the sort of investment-banking buzz that it would have in decades past. The excitement has shifted elsewhere: in finance, from banks like Lehman to private-equity firms, algorithmic traders and cheap, automated exchange-traded funds; in business, from finance to technology. Masood misses the animal spirits. “Lehman”, Masood recalls wistfully, “was good days. They didn’t care about spending—they bought what they want.”
Banking may have lost its buzz, but much looks the same as it ever did (see chart 1). Back in 2007, the five leading institutions in terms of global investment-banking revenue received 32.6% of all revenues in the sector, according to Dealogic, which tracks such matters. That is precisely the same share they get today, though one of the old big five, UBS of Switzerland, has been replaced by Bank of America Merrill Lynch.
In commercial banking, the five largest participants in America are the same today as they were 11 years ago. There were more casualties in the industry’s lower tiers: in America, Bear Stearns, Washington Mutual, Wachovia and Countrywide. In Europe, Britain’s HBOS and Germany’s Dresdner were absorbed, respectively, by Lloyds and Commerzbank.
Many more large institutions were kept going through bail-outs—most notably the vast recapitalisation authorised by the Troubled Asset Relief Programme in America—or, in the case of RBS and AIG, outright nationalisations. Lots of small banks were consolidated. In Spain 55 of them have been clumped together into just a dozen or so. As a result the biggest banks have got bigger; in both Europe and America the percentage of assets held by the biggest five has increased.
Relatively, the banks have not done so well; the banks’ share of S&P 500 market capitalisation has declined from 10% to 6%. But a notable aspect of the decennial anniversary is that the clouds may, finally, be lifting. Genuinely strong results from resurgent American banks, especially Morgan Stanley, suggest decent returns have become feasible. Jamie Dimon, chief executive of JPMorgan Chase, has gone so far as to say that a “golden age” of banking beckons. Profits are beginning to improve, economies are expanding, credit quality is good, regulation is ebbing.
Even Europe’s banks, so slow to put the crisis behind them, are finally looking to the future. In August Jes Staley, the boss of Barclays, reported one of the bank’s “first clean quarters” in years, free of write-downs and fines. The salaries for high-ups remain phenomenal. In 2017, AIG’s new boss, Brian Duperreault, was paid $43m, Mr Dimon $29.5m, Goldman Sachs’s Lloyd Blankfein $24m and Bank of America’s Brian Moynihan $23m.
In other parts of finance the sense of stasis is yet more striking. Credit-rating agencies provided sweeping and unfounded endorsements of the mortgage-related securities that turned out to be toxically risky, thus allowing fuel for the great conflagration to build up unheeded. This led to calls for heads to roll, more competition and a new regulatory structure. The most recent update from the Securities & Exchange Commission, though, published in December 2017, shows that the industry’s three main firms, Moody’s, Standard & Poor’s and Fitch, still accounted for 96.4% of all ratings.
Perhaps the most remarkable non-change is in the market where the crisis was centred: American mortgages. In the years leading up to the crisis, fed by unsupportable credit, home ownership rose sharply, from 64% to 69% of households. Prices shot up, too. This suited the people selling the mortgages, the people buying them, the people hiding the risks in securities and the politicians who wanted people who could not really afford houses to be able to buy them anyway.
Automatic for the people
Once the crisis hit, according to an estimate by the St Louis Fed, 9m homeowners in America were forced from their homes, about 10% to 15% of the total. The impact on housing markets was particularly vivid in southern Florida, with Las Vegas and southern California having some claim to a similar experience. Between 2008 and 2011 the price of an apartment in Miami fell almost by half, from $3,720 a square metre ($346 a square foot) to $1,830.
The tide eventually turned. Of the 55 banks operating in Miami-Dade county in 2008, only 28 are still standing. But on the land behind Grove Bank and Trust, the oldest of them, there are two brand-new luxury towers designed by Rem Koolhaas, a famous architect. Flats in the towers sell for more per square metre than anyone in Miami paid before 2008. If many of those who lost their homes have not been able to buy new ones, nationwide the amount of equity in homes has recovered.
If the landscape looks similar, and the prospects brighter, the changes to the financial system have been wrenching, nonetheless. Two areas stand out: the balance-sheets, funding and business models of the banks themselves; and the increased clout of the regulators.
Core tier-one capital, a gauge of the equity that banks use to fund themselves, has gone up a lot. Measured as a percentage of risk-weighted assets, it has risen by roughly two-thirds in the euro zone (from 8.8% to 14.7%) and by roughly a third in America, from 9.8% to 12.9%. As a result, returns on equity have fallen, and thus profitability.
As a result, banks have become more cost-conscious. There have been cuts in staff across the board (see chart 3); at Bank of America by a third, at Citigroup by 44%, and at Barclays by half. If you want in-house baristas, go to the Bay Area tech giants; JPMorgan Chase provides its staff with access to coffee by leasing space to Starbucks.
Some startups have changed their business models from competing with banks to supplying them. Zest, for example, initially intended to use its software to build a loan portfolio. Now it licenses its credit-analytics to financial institutions. Three companies with no public name recognition, Jack Henry & Associates, Fiserv and FIS, handle data processing and mobile-banking software for thousands of financial institutions. Banking operations have consolidated a good bit more than the banks themselves have.
Cultures have changed, too. Lehman was driven, in the end fatally so, by employees who were encouraged to take initiative and create opportunities—an approach that worked well in equities but failed disastrously in the fixed-income business, which collapsed under the weight of bad property-related debt. The new managerial order for every bank is less free-spirited, with tight controls and, too often, dispiriting bureaucracy.
As the banks have pulled back, more credit has been coming from outside the banking system. That carries its own risk, but may be more resilient to runs. Issuance of long-term corporate bonds has been consistently strong since the crisis. Private-equity firms have entered the lending business, with assets under management up from nothing in 2006 to $200bn, according to Preqin, a research firm. That is less than 1% of banking assets, but in the markets most targeted, notably smaller and heavily leveraged businesses, it probably represents a lot more.
What the banks have lost in swagger since 2008, the regulators, particularly in America, have gained. This is the second big change: a new and intrusive regulatory structure. The Federal Reserve is the one institution that emerged from the crisis with more authority and little damage; its power as a single and powerful regulator is one of the reasons that American banks did better quicker than European ones—especially in recapitalising.
The Fed has increased the number of people in supervision and regulation from 3,000 to 4,800, the tip of a massive expansion in oversight. It is often claimed that the financial system itself is made safer by the Fed’s annual “stress tests”, which are meant to assess banks’ resilience to shocks and which provide gainful employment to battalions of mathematical modellers at the central bank. It is also said that a new philosophy of “forward-looking” supervision can take into account what might unfold, rather than merely providing a static measure of a bank’s position based on data that, by the time it is reported, are already out of date.
Not everyone is reassured. John Allison, a former chief executive of BB&T, a large regional bank, and a board member at the Cato Institute, a libertarian think-tank, says the sorts of models the Fed uses have four problems: their methods miss outlying risks; they rely on historical data and thus often miss changes (an example being pre-crisis changes to bank-capital requirements by regulators favouring the kind of home loans that soured); they force different institutions to take similar positions, aggravating overall risk; and they do not work well with small, idiosyncratic kinds of loans—meaning those for small businesses, a clientele that matters.
But though the legal environment has been rich in claims it has been poor when it comes to facts. Cases have invariably been settled rather than going to trial. Why fines have been the size they have been, and what exactly each is a penalty for, has often been left unclear. Banks defended settling by saying they had no choice—merely going to court would lose them their licences to operate. Individuals justified the opposite approach with a similar logic—anything short of total exoneration would end their careers. With rare exceptions, when defendants fought back, prosecutors either backed down or lost.
The new dispensation also puts ever-greater authority in the hands of the political appointees in charge of regulatory agencies. This means the institutions will inevitably be more politicised—an outcome many who initially supported these innovations have come to regret since the change in administration.
Nowhere is this more true than in housing finance. Fannie and Freddie have been granted the right to issue loans on the basis of down-payments as low as 3% and debt-to-income ratios of up to 50%. The sharpest increases in recent housing prices have been in low-cost homes, 84% of which are guaranteed by those two institutions and the Federal Housing Administration, which provides insurance.
“That”, says Peter Wallison of the American Enterprise Institute, a think-tank, “is a time bomb”—the resurrection of a system of poorly supported mortgages that could prompt another crunch, though not as severe as the one in 2008.
Even for those who believe that the regulatory expansion has worked well, there is the risk that market participants now believe that if conditions sour, another successful bail-out will be forthcoming, thus producing tolerance, if only quietly, for excessive risks. Provisions included in the new rules that prohibit the bail-out of individual institutions were designed to blunt just this sort of expectation. But this limitation, among others, is under attack by many of the individuals who were involved in Lehman’s demise a decade ago on the ground that, in an emergency, all tools should be available.
Such concerns may be excessive. What the crisis showed is that, in a pinch, authority can expand. More difficult is taking the heat out of problems before they boil over. If it is time to believe the crisis that began in 2008 has really ended, it is past time to wonder how the new conditions which have come about in its wake could contribute to the one that comes next.
Until then, Masood will continue to caffeinate the remaining bankers of 745. They are, he says, better than the construction workers who put up the building in the first place, and tended to bring their own food to work. But nothing could replace the glory years before 2008. Masood got great tips—including, once, a cheque for $10,000, to take his family on holiday. He returned it, saying he could not accept so much that he had not earned; his would-be benefactor said he understood the sentiment. If such understanding ever spreads, it would be no bad thing.
Peerless
Funding Circle, a British tech unicorn, plans an IPO
The peer-to-peer lender is part of a strong London financial-technology scene
Print edition | Britain / Sep 6th 2018
IT IS rare for the government to shower praise on just one startup. Not so with Funding Circle, a London-based peer-to-peer lender that allows institutional and retail investors to lend to small businesses, which announced on September 3rd that it would seek an initial public offering. The British government has lent £100m ($129m) directly through its platform. Philip Hammond, the chancellor, has hailed the firm as a “real success story”. Its co-founder and chief executive, Samir Desai, was awarded a CBE in 2015 while still in his early 30s.
It is not hard to see why policymakers are keen on a firm that provides alternative sources of capital to small businesses. In the past, especially in Europe, banks have been such companies’ only option—and they have not always been a good one, since most banks see small-business lending as a minor and not particularly lucrative part of their business.
Funding Circle boasts of having channelled £5bn in loans to over 50,000 companies, in Britain, America, Germany and the Netherlands, since its launch in 2010. Lenders range from retail investors to private institutions like Citibank and public bodies like the European Investment Bank and Germany’s state development bank. For Anton Ruddenklau of KPMG, a consultancy, Funding Circle’s IPO is a sign that its business model “is here to stay.”
The offering, which could value the firm at more than £1.65bn, is a bellwether of the health of London’s financial technology scene. Fintech, as it is known, continues to attract venture capital, including a record-setting $1.8bn in 2017 and over $850m so far this year, compared with just over $700m in the whole of 2016, according to PitchBook, a data provider. Of Europe’s eight fintech unicorns (privately held firms valued at more than $1bn), six are British, according to CB Insights, another data firm. One of them, Acorn OakNorth, announced on September 6th the completion of a new round of fundraising that values it at $2.3bn.
Brexit will bring obstacles. Fintech bosses are particularly worried that it could get harder to hire talented foreigners. Firms like Currencycloud and Azimo have been looking into setting up offices in continental Europe. Mr Desai, for his part, says he trusts the British government will continue to allow skilled workers into the country. For his firm the lead-up to Brexit, which has crimped lending by Britain’s banks, has been an opportunity, he says.
Still, peer-to-peer lending is a tricky business. The industry is so young that it has yet to weather a recession. Gaining new customers is expensive: in 2015 Funding Circle spent over 60% of its revenues on marketing (it has since reduced the share to 40%). And regulators are paying more attention. In America, Lending Club, a rival to Funding Circle, has been sued by a regulator for charging hidden fees. In Britain, the Financial Conduct Authority is looking into barring ordinary retail investors from peer-to-peer lending, and limiting it to sophisticated or wealthy investors, thereby rather straining the definition of “peer to peer”. Nice words from the chancellor do not guarantee smooth sailing.
To hear an interview with Samir Desai, Funding Circle’s chief executive, subscribe to Economist Radio on your podcast app and search for Money Talks
The deciders
How social-media platforms dispense justice
Their armies of content moderators are expanding
Print edition | Business / Sep 6th 2018| MENLO PARK
EVERY other Tuesday at Facebook, and every Friday at YouTube, executives convene to debate the latest problems with hate speech, misinformation and other disturbing content on their platforms, and decide what should be removed or left alone. In San Bruno, Susan Wojcicki, YouTube’s boss, personally oversees the exercise. In Menlo Park, lower-level execs run Facebook’s “Content Standards Forum”.
The forum has become a frequent stop on the company’s publicity circuit for journalists. Its working groups recommend new guidelines on what to do about, say, a photo showing Hindu women being beaten in Bangladesh that may be inciting violence offline (take it down), a video of police brutality when race riots are taking place (leave it up), or a photo alleging that Donald Trump wore a Ku Klux Klan uniform in the 1990s (leave it up but reduce distribution of it, and inform users it’s a fake). Decisions made at these meetings eventually filter down into instructions for thousands of content reviewers around the world.
Seeing how each company moderates content is encouraging. The two firms no longer regard making such decisions as a peripheral activity but as core to their business. Each employs executives who are thoughtful about the task of making their platforms less toxic while protecting freedom of speech. But that they do this at all is also cause for concern; they are well on their way to becoming “ministries of truth” for a global audience. Never before has such a small number of firms been able to control what billions can say and see.
Politicians are paying ever more attention to the content these platforms carry, and to the policies they use to evaluate it. On September 5th Sheryl Sandberg, Facebook’s number two, and Jack Dorsey, the boss of Twitter, testified before the Senate Select Intelligence Committee on what may be the companies’ most notorious foul-up, allowing their platforms to be manipulated by Russian operatives seeking to influence the 2016 presidential election. Mr Dorsey later answered pointed questions from a House committee about content moderation. (In the first set of hearings Alphabet, the parent of Google, which also owns YouTube, was represented by an empty chair after refusing to make Larry Page, its co-founder, available.)
Scrutiny of Facebook, Twitter, YouTube et al has intensified recently. All three faced calls to ban Alex Jones of Infowars, a conspiracy theorist; Facebook and YouTube eventually did so. At the same time the tech platforms have faced accusations of anti-conservative bias for suppressing certain news. Their loudest critic is President Donald Trump, who has threatened (via Twitter) to regulate them. Straight after the hearings, Jeff Sessions, his attorney-general, said that he would discuss with states’ attorneys-general the “growing concern” that the platforms are hurting competition and stifling the free exchange of ideas.
Protected species
This turn of events signals the ebbing of a longstanding special legal protection for the companies. Internet firms in America are shielded from legal responsibility for content posted on their services. Section 230 of the Communications Decency Act of 1996 treats them as intermediaries, not publishers—to protect them from legal jeopardy.
When the online industry was limited to young, vulnerable startups this approach was reasonable. A decade ago content moderation was a straightforward job. Only 100m people used Facebook and its community standards fitted on two pages. But today there are 2.2bn monthly users of Facebook and 1.9bn monthly logged-on users of YouTube. They have become central venues for social interaction and for all manner of expression, from lucid debate and cat videos to conspiracy theories and hate speech.
At first social-media platforms failed to adjust to the magnitude and complexity of the problems their growth and power were creating, saying that they did not want to be the “arbiters of truth”. Yet repeatedly in recent years the two companies, as well as Twitter, have been caught flat-footed by reports of abuse and manipulation of their platforms by trolls, hate groups, conspiracy theorists, misinformation peddlers, election meddlers and propagandists. In Myanmar journalists and human-rights experts found that misinformation on Facebook was inciting violence against Muslim Rohyinga. In the aftermath of a mass shooting at a school in Parkland, Florida, searches about the shooting on YouTube surfaced conspiracy videos alleging it was a hoax involving “crisis actors”.
In reaction, Facebook and YouTube have sharply increased the resources, both human and technological, dedicated to policing their platforms. By the end of this year Facebook will have doubled the number of employees and contractors dedicated to the “safety and security” of the site, to 20,000, including 10,000 content reviewers. YouTube will have 10,000 people working on content moderation in some form. They take down millions of posts every month from each platform, guided by thick instruction manuals—the guidelines for “search quality” evaluators at Google, for example, run to 164 pages.
Although most of the moderators work for third-party firms, the growth in their numbers has already had an impact on the firms’ finances. When Facebook posted disappointing quarterly results in July, causing its market capitalisation to drop by over $100bn, higher costs for moderation were partly implicated. Mark Zuckerberg, the firm’s chief executive, has said that in the long run the problem of content moderation will have to be solved with artificial intelligence (AI). In the first three months of 2018 Facebook took some form of action on 7.8m pieces of content that included graphic violence, hate speech or terrorist propaganda, twice as many as in the previous three months (see chart), mostly owing to improvements in automated detection. But moderating content requires wisdom, and an algorithm is only as judicious as the principles with which it is programmed.
At Facebook’s headquarters in Menlo Park, executives instinctively resist making new rules restricting content on free-speech grounds. Many kinds of hateful, racist comments are allowed, because they are phrased in such a way as to not specifically target a race, religion or other protected groups. Or perhaps they are jokes.
Fake news poses different questions. “We don’t remove content just for being false,” says Monika Bickert, the firm’s head of product policy and counterterrorism. What Facebook can do, instead of removing material, she says, is “down-rank” fake news flagged by external fact-checkers, meaning it would be viewed by fewer people, and show real information next to it. In hot spots like Myanmar and Sri Lanka, where misinformation has inflamed violence, posts may be taken down.
YouTube’s moderation system is similar to Facebook’s, with published guidelines for what is acceptable and detailed instructions for human reviewers. Human monitors decide quickly what to do with content that has been flagged, and most such flagging is done via automated detection. Twitter also uses AI to sniff out fake accounts and some inappropriate content, but it relies more heavily on user reports of harassment and bullying.
As social-media platforms police themselves, they will change. They used to be, and still see themselves as, lean and mean, keeping employees to a minimum. But Facebook, which has about 25,000 people on its payroll, is likely soon to keep more moderators busy than it has engineers. It and Google may be rich enough to absorb the extra costs and still prosper. Twitter, which is financially weaker, will suffer more.
More profound change is also possible. If misinformation, hate speech and offensive content are so pervasive, critics say, it is because of the firms’ business model: advertising. To sell more and more ads, Facebook’s algorithms, for instance, have favoured “engaging” content, which can often be the bad kind. YouTube keeps users on its site by offering them ever more interesting videos, which can also be ever more extreme ones. In other words, to really solve the challenge of content moderation, the big social-media platforms may have to say goodbye to the business model which made them so successful.
https://eoesociety.blogspot.com/2018/09/eoe060-selected-articles-temporary.html
Comments
Post a Comment