I was at after work drinks a month or so ago, and one of my work colleagues invited his daughter and wife along, who had been in the area. I guessed that his wife might be more interesting to talk to than my work colleagues, so asked a few conversational questions. She had worked at NatWest, but left a couple of decades ago to become a full time mother/housewife. I thought this was interesting, because it was just about the same time that banks began to lose their way. Although people remember the financial crisis in 2008, banks had started underperforming the wider stock market long before then, and around 2002-3 there was nearly a similar crisis, which no one remembers, because it was just about avoided. I was interested in her thoughts, because she was well informed having worked in a bank for years, but distant enough to have a different perspective. I was interested to know where she thought it had all gone wrong for banks.
She thought for a while. And then suggested it was technology. She said: “A couple of decades ago it just wasn’t possible to do all the clever things banks now do. Credit scoring. High speed processing trading in such large volumes and so quickly. Or price complex derivatives.”
And I think she was right. Bankers’ hubris clearly played a part in the financial crisis. But bankers have always been driven by money, and arrogantly convinced of their way of thinking. In other words bankers’ attitudes had not changed. But technology had changed a lot, and had a real influence on banking in the last few decades.
A straightforward example is that technology made credit scoring possible. To a certain extent that drove lending growth, because banks could feel confident in analysis, building Markov chain models, and Partial Differential Equations to give them confidence to lend to people they wouldn’t normally have lent to. That certainly played a part. These models always seemed to assume that future behaviour would look similar to the past, yet the models encouraged banks to lend to people who previously had not had access to so much credit. Broadening access to credit is normally seen as a good thing. But not if technology allowed more and more households to get into financial difficulties.
Really all that happened, was that the sophisticated mathematics and technology made bank management overconfident that they understood the risks that they were taking.
But that is not the full explanation. Credit scoring models don’t explain why bank assets went from 0.5x GDP to 5x GDP (see chart from the Bank of England). That is a huge change.
Instead that was driven by lending between banks. Since before the Bank of England was founded in 1694, banks business model have kept to the sample simple business model. That is to keep something in their vaults (originally gold, now customer deposits) and then lend it out. Banks know that it is unlikely that everyone will ask for their deposits back at the same time. Previously they were constrained on how big their balance sheets could become. But technology removed this constraint. Of course, for as long as banks have been lending out deposits, and creating money that hadn’t existed, people have recognised the risks. But technology enabled banks to push this model of creating more and more intrabank exposure (both assets and liabilities) to another level, and eventually push it too far with the resulting 2008 crisis. So technology and mathematical models were also responsible for rampant expansion in bank balance sheets.
Technology that had helped the banking industry grow, ended up making banking far more risky. Investing in a banks technology didn’t improve things, it destroyed the reputations of management, bankrupted some customers and ruined the economic returns for shareholders.
This isn’t the normal narrative of technology and competition, where a new entrant (Microsoft, Google) destroys the business of an existing company (IBM, Yahoo). Instead, what she thought had happened to banks is, that the shareholder returns of an industry had been destroyed by technology that they themselves have introduced. We are familiar with this in fiction. From the computer HAL in “2001”, to the post nuclear apocalypse world, caused by the SkyNet starting a nuclear war. But rarely do shareholders or management think of a technology that lowers costs, or increases customers, as damaging to the long term prospects of the business. Perhaps they should.
Back to our conversation in the pub. I suggested an analogy with newspapers. And she nodded her head in agreement.
In the internet bubble, many people believed internet distribution would benefit newspaper economics. It would reduce the cost of distribution, and widen the readership after all you wouldn’t need to print so many newspapers and send them out early in the morning all over the country. Many optimistic analysts claimed that “content is king” – all you needed was valuable content that the public wanted to read, and you had a valuable business model.
Except the internet technology not only lowered the distribution costs – it also made it harder for newspapers to charge their readership. For the last few decades newspapers have experimented with “paywalls”, mostly to abandoned them when so few people sign up. The Sun is the latest to do this. According to ABC data quoted by the FT, in September The Sun had just 1.1m readers versus Mail Online 13.4m. The Mail Online has 13x more readers, but is still loss making.
Formerly highly profitable newspapers have invested in technology that destroyed their own economics. They did it to themselves.
And I am beginning to think that the same process is going on at the FT’s former owner, Pearson. Most of the revenues come from selling educational courses, clearly the internet lowers the costs of distribution and makes the business much more scale-able. But it also means that there is a lot more competition from MOOCs (Massive Open Online Courses). Pearson warned last month that earnings would disappoint, but said this was due to the strength of the US economy, and less people signing up for training and courses, because they could get jobs directly. That is a cyclical explanation, and suggests earnings will recover.
Some of my worst mistakes have been to confuse cyclical and structural stories. Either to buy into a company that I thought was only going through temporary problems (Cleardebt) but which never recovered. But the more expensive mistake is to sell too early companies that had recovered from distressed valuations, but would double again as they benefitted from long term structural trends (Legal and General, Cineworld).
Companies rarely admit they have structural problems, and one of the hardest things for investors to work out is whether a company’s problems are temporary (ie cyclical) or permanent (ie structural.) Pearson has invested heavily in digital technology, the idea being that online learning is more effective than old fashioned text books. Over the last few years the companies net debt has quadrupled to now £1.6bn. Meanwhile sales have not increased and returns on invested capital have fallen from 10.3% to 5.6%. Perhaps returns will rebound strongly. But I wonder if the the old fashioned “text book” might have been a better business model, as opposed to online learning. The threat is not the MOOCs, but instead the technology that Pearson itself has invested heavily in has made Pearson a more risky investment. The company’s own motto is.
“An investment in knowledge always pays the best interest.”
But technology and knowledge are not quite the same thing. The newspaper that has coped best with the threat of the internet is Private Eye. Ironically. They have very little online content, you have to buy the hard copy. The circulation is the highest in 3 decades. Sometimes the best response to new technology, is not to invest in it, but to stick to what you know works. Certainly banks, who did not invest in knowledge, have very poor returns.