Tag: financial disaster stories

Posted on
by
The main entrance of the Federal Reserve Bank of Chicago

I don’t usually worry much about investment bubbles. There have been a lot of them over the past few hundred years, and most of them (railroads, telegraph, dotcom…) were expensive disasters largely only for the people who invested in them. Some though, such as the Great Financial Crisis of 2007–2009, were expensive disasters for lots of other people as well. So it’s worth thinking a bit about whether the current AI bubble is of the former sort or the latter—and how to protect your finances in either case.

Bad just for investors

One big difference between bubbles that are going to be wretched for everybody when they pop and those that’ll end up mostly okay except for the foolish investor’s portfolio, is whether the excess investment got spent on something of enduring value.

For example, railroad lines got enormously overbuilt in the 1840s in the UK and in the 1880s in the US, leading in both cases to a stock market bubble, followed by a stock market crash and a banking panic. But (and this is my point), the enormously overbuilt railroads were of some value. As the firms went bankrupt, the people who had over-invested lost a lot of money, but the railroad tracks, rights-of-way, and rolling stock all still existed. The new firms that got those assets, free of the excess debt, were often viable firms that went on to be successes—hiring workers, providing transportation, and eventually providing a return to the new investors. The people who got screwed were the old investors. (And not even all of them, as the original investors often saw the overbuilding happening early and sold out just as the clueless people who knew nothing about running a railroad, but just saw stocks soaring and wanted to get in on it, started piling in.)

Much the same was true of part of the dotcom bubble. A lot of money got spent on a lot of things. To the extent that it was spent on buying right-of-way and burying fiber, there was something of enduring value that ended up owned by somebody, making it one of the less-bad bubbles.

The key to avoiding catastrophe in bubbles of this sort is largely just a matter of not investing in the bubble yourself.

Bad for the economy

But some bubbles have produced horrible, wretched, prolonged difficulties for the whole economy. The other part of the dotcom bubble, besides the dark fiber build-out, was the bubble in companies with no profits and no prospect of ever having profits, whose stock prices went up 10x based on nothing but a story that sounded compelling until you thought about it for 10 seconds. As usual, that ended up being very bad for the people who invested in those companies, but it also was bad for the whole economy, because when those firms went bankrupt, they left behind nothing of enduring value.

The result was that the imagined wealth of those companies just vanished. The stock market went down, which was bad for (almost) everybody, and it produced a general economic malaise, because post-dotcom crash it became hard even for legit companies with real assets, a real profit, and a real business plan for growth, to raise money, which made actually producing that growth much harder.

Really bad for the economy

There is, however a step beyond just pouring a bunch of money into a bubble that doesn’t actually produce anything of enduring value, like a fiber optic network or a railroad. That’s when the money is raised with leverage (i.e. debt).

The 1929 stock market crash was a rather drastic example. People invested in stocks not because there was an underlying business that was worth what the investors were paying for it, but purely because the stocks were going up. That might have been okay in other times, but stock brokers had recently started allowing ordinary people (as opposed to just rich people) to buy on margin—where you just put up a fraction of the price of the stock you want to buy, and the broker lends you the rest.

In the 1920s you could buy on 90% margin, where you only put down 10% of the price of the shares. That meant that, if the stock price went down by just 10% your whole investment was wiped out, and the broker would sell you out to raise money to pay off (most of) the loan. And of course, all those sales into a falling market produced more losses, leading to the crash.

Since the 1930s you could only buy stocks on 50% margin, making it much less likely that your broker will sell you out into the teeth of a general stock market crash—although it can still happen.

Bubbles with leverage

A great example of a bubble with leverage is the Great Financial Crises of 2007. (Most people date it from 2008, because that’s when Lehman Brothers collapsed. I date it from 2007 because that’s when my former employer closed the site where I worked and I ended up retiring rather earlier than I’d planned.)

That was a particularly bad bubble. A whole lot of money was raised, with leverage, to buy housing. But very little of the money ended up being spent to build more housing (which would have been something of enduring value that would have lasted through the subsequent collapse). Instead, the money was spent bidding up the prices of existing housing, which then fell in value after the bubble popped.

So we had two of the classic producers of bad bubbles: Nothing of enduring value created, and leverage. The whole things was made even worse by the structure of the leverage in question.

This is getting rather far from my main point, so I won’t go into much details, but to raise the large amount of money that was going into houses, the rules on housing market leverage were being eased over a period of time. It used to be that you had to put 20% down on a house. Then you still had to put 20% down, but only half of it had to be cash, with the other half being funded with a second mortgage on the property (at a higher interest rate). Then they started letting people put just 3% down. Then they started letting people with good credit put nothing down. Then they started letting people with no credit put nothing down. At the same time, “structured finance” obscured just how risky all those mortgages were, meaning that when the bubble went pop lots of “mortgage-backed securities” ended up being worth zero.

Which kind is the AI bubble?

This brings us to the current AI bubble. A whole lot of money is pouring into building two things:

  • Data centers (buildings filled with computer chips of the sort used to train and run AI models)
  • Large language models (non-physical things that are basically just a bunch of numeric weights of a bunch of tokens which can be used to produce streams of plausible-sounding text)

Each of those may have some enduring value.

Data centers will have some. They will probably have a lot less than a network of fiber optic cables, which can be buried and will have value for decades with minimal cost or maintenance. Since newer, faster chips are coming out all the time, a data center is well behind the cutting edge as soon as it’s finished. Plus, training or running an AI model runs those chips hard, meaning that they probably only last a couple of years (due to thermal damage on top of regular aging).

Large language models probably have even less enduring value, because so many people are training new ones all the time. People are always trying to make them bigger (trained on more data) while also making them smaller (so they can run without a giant data center). All that means that your two-year-old LLM probably isn’t worth what you paid to build it, and a four-year-old LLM probably isn’t worth anything.

That’s how things looked a year or so ago—a perfect example of a bubble that would burn the people who sank money into it, but leave the broader economy untouched.

Sadly, that’s been changing.

First, the structure of the leverage has been changing. It used to be rich people and rich companies were building data centers and hiring software engineers to build LLMs. But lately that’s been getting screwy. Those large companies are raising off balance-sheet money with Special Purpose Vehicles (small companies that big companies create and provide some capital to, that then borrow a bunch of money to make something, with the loans collateralized by the things they’re making—but importantly, not an obligation of the big company that created them). Any particular SPV can blow up, if it turns out that the things it built don’t earn enough to pay the interest on the money the borrowed to build them. And large numbers of SPVs can blow up if financial conditions change to make it harder for all the SPVs to roll over their debts as they constantly have to keep their data centers running.

Second, they’re also engaging in weird circular investing and spending arrangements, where company A buys stock in company B which then turns around and pays all that money back to company A to buy chips, letting company A treat it as both income and an investment, while company B can pretend it got its chips for free.

Finally, there’s all the non-financial obstacles that may well throw a wrench into the whole thing. The fact that LLMs are all built on copyright violations. The fact that running data centers requires huge amounts of power and water (that has to be produced and paid for). The fact that producing that water and power brings with it horrible environmental impacts.

What to do

So, if AI is a bubble, and its one of the bad sort that will produce a panic and a recession when it pops, what should you do?

There are a lot of little things you can do that will help. I wrote an article with suggestions at Wise Bread called Are your finances fragile? It talks about what financial moves you can take to put yourself in a better position if there’s a general financial crisis. (If you’re interested in my writing about this stuff more broadly, I wrote a overview of my perspectives on personal finance and frugality called What I’ve been trying to say, that includes a bunch of links to other of my posts at Wise Bread.)

Besides that general advice, there are also a few things to strictly avoid. In particular, strictly avoid thinking that you can find some very clever investment strategy that lets you make money off the popping of the bubble. Yes, after the fact there will be some investments that make a lot of money, but no amount of keen insight will let you find and make those investments, as opposed to the thousands of very reasonable-seeming investments that will blow up just like all the rest.

Along about the end of the Great Financial Crisis I wrote an article called Investing for Collapse, which explains why any such effort is pointless. It holds up pretty well, I think.

Short version? Avoid debt. Keep your fixed expenses as low as possible. Build a diversified investment portfolio that limits your exposure to the most obviously stupid investments, but doesn’t do anything too weird or wacky in an effort to get them to zero—it’s pointless, and will probably do more harm than good.

Good luck when the AI bubble pops!

Posted on
by

Assets are called “safe” when they’re free of default risk. But that doesn’t mean their prices can’t drop, or that the financial system is safe if systemically important institutions buy them on margin.

What appears to be a liquidity issue will ultimately become a financial stability issue as investors discover their “safe assets” are not safe.

Source: Solvency Constraints – Fed Guy

Posted on
by

With no card number, CVV security code, expiration date or signature on the card, Apple Card is more secure than any other physical credit card.

Source: Apple Card launches today for all US customers – Apple

While @jackieLbrewer was working at the bakery there was a cash register glitch. For several days they took credit card payments on paper, writing the number down by hand, and then entering them manually at the end of the day.

Those customers would have been totally secure from being able to buy bread.

Marketing image courtesy of Apple

Posted on
by

This sort of thing was a fantasy of mine, back in the 1980s. I could see things getting worse in the US, and the idea that a foreign passport and foreign residency could provide an escape if things got too bad was pretty appealing.

Nowadays not so much. It’s not that things have gotten better in the US; it’s that things have gotten worse other places at least as quickly. More to the point, things getting worse in the US seems to make things worse other places, so the conditions that make the idea appealing are the same conditions that make it pointless.

One book that substantially influenced my thinking in this area is Emergency: This book will save your life by Neil Strauss. I recommend it highly. In entertaining and informative prose, he documents his transformation from just the sort of kook I was in the 1980s into somebody with a much more practical perspective.

Still, an EU passport might have its upsides. Anybody got a spare quarter million euros and an interest in learning Greek? (If money is no object, a half million euros will let you buy in to Ireland, and I expect learning Gaelic is optional.)

Graphic from Citizenship for Sale on the International Monetary Fund’s IMFDirect blog.

 

Posted on
by
Prairie plants

Prairie plants
Prairie plants

One utterly predictable consequence of climate change is that the price of northern farmland will rise as growing regions shift north.

Tobias Buckell yesterday shared a report that just this sort of price shift is now occurring—interesting to me because this result is not merely predictable: I predicted it my own self, way back when I was in high school.

Global warming was still pretty speculative then (in the 1970s), but people were already talking about the greenhouse effect and trying to figure what the result would be. At the time, I was mainly thinking about the geopolitical implications of  shifting the growing regions north—how things would change if Canada and the (then) Soviet Union were suddenly way more productive of food, while places like the United States, China, and France suddenly less so.

What I discovered, though, was that those northern regions aren’t nearly as fertile as places like Illinois, where 8,000 years of tall grass prairie left an incredibly thick layer of rich soil.

No matter how perfect the climate is, Saskatchewan is not going to produce the bushels per acre of Illinois or Kansas. Their soil is not only less fertile, it’s also much more fragile than the soil of the tall grass prairies. The fertile layer isn’t as deep, so the land must be plowed with greater care, and it will in any case be more quickly depleted.

I’m sure there’s a lot more and better data available now than there was back then, but I doubt if it changes the fundamentals. Shifting growing regions means winners and losers, but it also means less total food production.

Posted on
by

The tea-party right was willing to risk the hard stop in spending that would have resulted from running up against the debt limit—a game of chicken that neither the Democrats nor the sane fraction of the Republicans could take the risk of losing.

The fiscal cliff looks a little similar, but it’s much less dangerous. It’s a game that lends itself to playing through to the end, because the risk of losing isn’t nearly so bad.

Suppose we did go over the fiscal cliff. What would happen?

First, tax rates would go up for everybody. That’s bad, but it’s not terrible. Actually, taxes at those rates would produce revenue roughly equal to the amount of government people seem to want.

Second, spending would be cut, with the cuts falling on almost everything except Social Security. A lot of good stuff would be cut, but that might not be such a stiff price to pay, considering that a lot of the things that ought to be cut (such as defense spending far beyond our needs) would otherwise be very hard to cut.

The result would be a rough year or two, hard on everybody from working-class folks to defense contractors, but all those problems would be fixable. In fact, Congress would love to fix those problems! Congress could cut taxes! (Just not as much as Bush did.) Congress could boost spending! (Just not to current levels.) Really, there’s nothing congressmen like better than cutting taxes and spending money on stuff.

The other details are similar. The AMT would strike middle-class folks hard, but that could be fixed, too. In fact, having to fix it would be an opportunity to improve it—turn it back into what it was supposed to be, a minimum tax rate that applies to everyone, no matter how many tax shelters they have or how many special preferences they qualify for. The end of the “doc fix” would hurt health care providers, but that could be pretty easily fixed too. (We’ll no doubt have to make a lot of small changes to healthcare stuff, once health care reform goes into effect and we run into the inevitable glitches.)

It’s always hard to raise taxes and cut spending, so it’s hard to do what needs to be done. But that’s why the fiscal cliff is so perfect. Once we go over the edge, we won’t need to raise taxes and cut spending—we’ll need to cut taxes and raise spending, and that’s dead easy.

Dive over the fiscal cliff, then fix things. It’s not perfect, but it wouldn’t be nearly as bad as what we’ve got now.

Posted on
by

One of the things I try to do in my fiction is show any collapse scenario as a process.

Your standard “if this goes on” story is all about looking ahead to see the result of current trends. But trying to see the endpoint of climate change, peak oil, habitat destruction, environmental degradation, or any similar process is going to be misleading. There is no endpoint—things keep going on.

I think the destruction wrought by Sandy is a good example. I’ve read a lot of stories set in a world where climate change has inundated the coasts. What aren’t nearly as common are stories set in the world we’re approaching: A world where the coasts are inundated only 1% of the time (or, a bit later, 2% of the time). A world where we see a 100-year storm every 8 or 10 years. A world where all our infrastructure spending is going for repairs, and yet we keep falling behind.

I say this is a world we’re approaching, but we may have already reached it. How many 100-year storms can you have in a decade before you have to admit that it’s not just a statistical anomaly, but rather is the new reality?

Posted on
by

I’ve got a guest post up at Asta Lander’s blog Simply Living, about the trade-offs that people make when they choose to work for wages or a salary, and how you can get most of the benefits while avoiding most of the downside.

It’s called Choosing Freedom.

There was a time when most people were self-sufficient. They acquired what they needed through some mix of hunting, gathering, fishing, farming, raising animals, and making things themselves. Not many people do that any more.