We're forever blowing bubbles

Much talk at the recent Kilkenomics festival (the world’s first festival of economics and comedy - an excellent combination!) focussed on the current boom in AI investment; whether it was a bubble and if it was, what might be the implications for the wider economy of it bursting.

Such exuberance wouldn’t be a first, human history is littered with examples as people try to get rich quick and become frightened of missing out - from the south sea bubble and tulip mania, through the railway boom of the 19th century and onto the 1929 crash, the dot-com boom and the sub prime bubble which resulted in the financial melt down in 2008.

Every bubble has its own peculiar features, along with a number of similarities, such as the availability of relatively easy credit and the feeling that ‘it will be different this time’, so the lessons of history can be ignored. On each occasion many investors lose a lot of money, often those who came late to the party and can least afford it.

The wider economic implications of any crash also vary, depending in part on how much the results of the investment are still useful, even though the original investors don’t benefit - think railway lines and fibre optic cable.

There are varying schools of thought as to value of the current AI investment, much of which seems to be going into data centres. While the physical infrastructure and any associated renewable energy supplies could be of some lasting value, there are some concerns that the chips at the heart of the centres my have a shorter life that expected, due to physical burn out and obsolescence. Indeed the balance sheet values being booked for these chips may well be a source of financial vulnerability in themselves as they are reassessed.

There is also a big question about how the application of AI will actually impact on the economy in the longer run, which is still very unclear despite all the investment. In a recent blog Brad de Long highlighted a number of different possibilities, from relatively simple natural language interfaces to databases to an artificial general intelligence with human like cognitive abilities. The current investment boom reflects a variety of bets as to which model or combination of models will prevail.

The degree to which mainstream financial institutions are caught up in the investment frenzy is also crucial for the wider impact of any investment bubble - either as direct investors or credit providers. The credit crunch caused by banks and other financiers unwinding their positions was a key contributor to the 2008 crisis. The current boom features a lot of relatively unregulated private credit and there are concerns about the degree to which banks are exposed to these institutions.

In a recent history of the 1929 crisis, told from the perspective of a number of the key players, Andrew Ross Sorkin concludes with a lesson which applies to all bubbles:

“..the story of 1929 is not about rates or regulation, nor about the cleverness of short sellers or failures of bankers. It is about something far more enduring - human nature. No matter how many  warning are issued or how many laws are written, people will find new ways to believe that the good times can last forever….we need to remember how easily we forget.”

Human nature is an often overlooked factor in economic analysis. In her recent book (‘In this Economy - How money and markets really work’) and in her regular Substack articles Kyla Scanlon is at pains to remind us of a very simple fact, the economy is not some abstract, number driven, entity - people are the economy (and therefore we should make greater efforts to make the economy about people).

“The economy is a circuit board of people, interacting with each other and other things, money is the electric current, lighting up to create the system as we know it. Our interactions are the economy.”

A few years ago she coined the term vibecession to describe an economy which isn’t looking too bad as far as the traditional numbers were concerned but where people feel that things are much worse and this drives their behaviour, which in turn can become self fulfilling. This disconnect is exacerbated by macro indicators which in large part are made up of averages and the wide variety of experience which goes to make up the average - made worse by the growth in income, wealth and intergenerational inequality.

Another feature of the AI boom is the associated increase in demand for energy to power and cool data centres, which is having a wider impact by increasing prices for other companies and consumers more generally. This is adding to what’s been termed an affordability crisis, which may well have wider political implications.

This crisis is in many respects a manifestation of the vibes that Scanlon is talking about. In a recent Substack article Paul Krugman begins to take an in depth look at what’s behind the crisis. On the face of it affordability should simply be about the relationship between prices and income, Krugman argues it goes deeper to include economic inclusion, security and fairness. As he defines them all involve feelings:

• economic inclusion - the ability to purchase the goods and services that allow you to feel a full standing member of society
• security - a feeling based not just on current real income, but also an assurance that severe hardship isn’t just a stretch of bad luck away
• fairness - people feeling upset about high prices when they feel that they are being taken advantage of

The affordability crisis could also be influenced by a cognitive bias (sometimes called the money illusion) in economic transactions where people pay more attention to nominal prices rather than real prices and incomes and tend to be ‘anchored’ to old nominal prices.

Digging deeper to have a clearer understanding of what’s important for people and what their real thoughts, feelings, needs and interests are is a key part of any mediation process. Mediators know only too well that peoples’ understanding of situations and the decisions they take are driven by a complex mix of reason, emotion and cognitive bias. Understanding the impact of this mix is key to developing options and deciding on ways forward to resolve issues. A similar depth of understanding is critical in the development of economic policy - appreciating the impact of bubble mentality is just one element of this.

As the founder of Kilkenomics David McWilliams puts it at the start of each of his podcasts: “To understand the economy you have to understand human nature.”