Ruins, Imperial Valley, California. Photo: Jeffrey St. Clair.
As we all wait for reality, and/or China, to catch up with the Silicon Valley AI boys, it might be a good time to go back in time a bit and see what it looked like in the past when our bubbles burst, specifically the tech bubble in 2001 and the housing bubble in 2008. Fortunately, it is easy to find the key data.
The Tech Bubble Deflates; the 90s End with the Millennium
The general view among economists in 2000 was that the strong growth of the 1990s would continue and the country would be highly prosperous for the foreseeable future. This was despite the fact that the stock market had already started to head down in March of 2000.
In January of 2001, the Congressional Budget Office (CBO) projected that GDP would grow 2.4 percent in 2001, 3.4 percent in 2002, and 3.3 percent in 2003. The CBO projections are useful here both because they provide the framework for federal policy and also because they are generally near the center of professional forecasts. This is by design. CBO tries not to be an outlier.
The growth for those three years turned out to be considerably worse. For 2001, it was just 1.0 percent, for 2002, it was 1.7 percent, and for 2003, it was 2.8 percent.
But growth is an economic abstraction; it is not something that people directly see. They do see jobs, and here the story is more dramatic. CBO projected the economy would create 1.0 million jobs in 2001, 1.7 million jobs in 2002, and 1.8 million jobs in 2003.
The actual numbers for jobs turned out to be a loss of 1.7 million jobs in 2001, an additional loss of 500,000 jobs in 2002, followed by a gain of 100,000 jobs in 2003. That’s a cumulative miss between forecast job growth and actual growth of 6.6 million jobs. That’s big even by Washington standards.
The gap between projected and actual job growth had a huge impact on the lives of millions of people. Not only were many more people unemployed. Tens of millions of people were stuck in jobs that they didn’t like or prevented from moving up the career ladder they were envisioning during the 1990s boom. While the unemployed may be just 4.0-5.0 percent of the labor force, in a normal month, 5-6 million lose or leave their jobs.
In a weak labor market, many fewer workers will voluntarily quit their jobs, and those who do leave a job, either voluntarily or involuntarily, find that their prospects for new jobs are far worse. In this context, it is not a surprise that the strong real wage growth we had at the end of the 1990s quickly came to an end after the tech bubble collapsed.
The Housing Bubble Bursts: Economists Learn that Nationwide House Prices Can Fall
When I was trying to warn of the risks of the housing bubble as its size was getting ever larger, several economists confidently dismissed my concerns by saying we had never seen a nationwide drop in house prices. We actually had seen a nationwide drop in prior decades, and we certainly did from 2007 to 2010. (We might again see some house price declines when the AI bubble bursts.)
This collapse in prices derailed the housing sector, which had been the driving force in growth following the collapse of the tech bubble. It also led to the financial crisis, as tens of millions of mortgages were suddenly underwater, meaning that the value of the home backing them was less, and often much less, than the debt on the mortgage. The resulting wave of defaults led to the collapse of hundreds of banks, including Bear Stearns and Lehman Brothers, the giant investment banks.
The resulting recession again meant a huge divergence between projected growth and actual growth. In January of 2008, CBO projected GDP growth of 1.7 percent, 2.8 percent, and 3.5 percent for 2008, 2009, and 2010, respectively. The actual growth was 0.1 percent, -2.6 percent, and 2.7 percent, respectively for the three years. Annual growth rates continued to lag the 2008 projection until 2015.
Again, the story was even more dramatic in terms of jobs. CBO projected the economy would add 1.0 million jobs in 2008, 1.3 million jobs in 2009, and 2.2 million jobs in 2010. The economy lost 3.5 million jobs in 2008 and 5.0 million jobs in 2009. It gained back just 1.0 million jobs in 2010. The gap between the projected job growth for these three years and the actual job growth was more than 12 million jobs. There is a reason this is known as “The Great Recession.”
How Bad Will a Bursting AI Bubble Be?
It’s difficult to say in advance how bad the hit from a burst bubble will be, just as it is difficult to predict the timing of the collapse. It is almost inevitable that there are some sectors that are seriously vulnerable in ways that are hidden by the bubble. For example, there were few people who recognized that the nation’s largest insurer, AIG, would face bankruptcy when the housing bubble burst because it had issued hundreds of billions of dollars of credit default swaps against mortgage-backed securities.
The severity of the downturn will depend not only on reduced demand from the AI industry and its suppliers and the negative wealth effect of a plunging stock market, but also on other sectors that may be caught up in the fallout. The part of the story that we can be certain about is that, as was the case with the last two bubbles, the economic forecasters will miss it.
This first appeared on Dean Baker’s Beat the Press blog.
