Michael Burry is shorting some tech companies. With the market as frothy as it is, that’s not exactly prescience. Unless you’re as good a market gambler as Burry, I wouldn’t recommend it. (And if you were as good as Michael Burry – you would already have a lot more zeros in your net worth). It is still true the market can stay irrational longer than you can be solvent. But what Burry isn’t just pointing out the emperor has no clothes. He is pointing to financial engineering. Why is that important? Why does presenting the information in a slightly better fashion matter?
The pressure is on to show something. All the public companies in the AI orbit, with elevated stock prices because they’re part of the “AI-play,” need to show earnings. The non-public AI startups do not need to show earnings. Oracle, Broadcom, Micron, etc. need to show revenue. Immediately they do not need to show revenue, as they sign contracts. That’s future revenue, and the stock price goes up as a multiple of earnings. With expected future earnings rising, the value of the company increases, even though current earnings may not have moved. A company that trades at 15 times their earnings begins trading at 30 times their earnings, based on the expectation of making more money in the future. But at some point, the imaginary future money needs to become real money in the present to justify that multiple.
Could companies like Palantir and Oracle be over-stating their income by altering the way they treat depreciation? Maybe as much as 20%? That’s what Burry sees. When companies structure their earnings to provide a better light than what would otherwise be the case, we refer to that as lower quality earnings. It may be legal and within the GAAP (generally accepted accounting principles), but it suggests the actual earnings are inflated. This is completely legal, as long as it is disclosed. Eventually, the lower quality earnings should result in a lower multiple. But in the short term, investors may ignore it or simply accept the statements of the companies that the new accounting practices make more sense. Longer term, investors tend to give companies that do a lot of financial engineering side-eye. Eventually reality will set it and the fundamental reasons they aren’t doing well will overtake the financial engineering.
But where there’s that much pressure to push earnings, it means there is building pressure to fake earnings. This can be done by either aggressively booking sales when the sale isn’t really complete and moving liabilities off the balance sheet. I would suspect the former is already unfolding. When everyone is desperate for more data center space, more power, more networking, and more processors, booking a sale early may not seem like a big deal. You feel the actual sale will almost certainly close in the very near term. Or you can call the next firm in line, waiting to snap up the same scarce resource. So why not report it in this quarter to juice your numbers a little? But it doesn’t take long before some firms start booking speculative sales, to keep the line on the sales chart going up and to the right. One of two things will happen, either the auditors will stumble over this and realize there’s fraud going on, or (more likely) a short seller will sniff it out. The former is bad enough when firms are forced to restate their prior earnings, as some executives go ‘spend more time with family,’ and shareholders bring suits. But the latter is devastating, usually resulting in obliteration, with the fraud investigations coming later.
The other approach is to engage in balance sheet engineering. A loan or an obligation to make payments in the future are recorded as liabilities. There are ways to move the liabilities off the balance sheet of the parent company, for example, by using special purpose vehicles to actually carry the obligation. Company X doesn’t owe the money. The money is actually is actually owed by Able Baker, a joint venture between X and Y. Company X doesn’t record the liability, even though the counter-party (the lender) can collect from Company X, should Able Baker default. The auditors may miss this if Company X misrepresents the true nature of the obligation (commits fraud). No one will notice a thing as long as the market that props up Able Baker is healthy. Once that changes, and Able Baker defaults, Company X may find itself illiquid.
On overstated earnings or engineered balance sheets you can quickly build other frauds. For example, understating the risk of loans to those companies. Lenders may be aware that something fishy is going on, but continue to lend to the companies, collecting fees on deals that should never have been closed. Even in the best possible light, it means suspicious insiders put aside suspicions to chase the deal. After all, the entire market can’t be wrong. And everything looks good for now. If the demand for the underlying market dries up, the loans held either by the direct lenders or the positions investors have in that lender, are worth pennies on the dollar. (Or nickles, now that we’ve stopped minting pennies). Suddenly, the lender (now likely to be a private equity firm rather than a bank) is exposed to losses large enough to wipe out its equity. Investors that invest or lend to the private equity firms suddenly find their positions wiped out as well, creating significant counter party risk. Which can ripple through other sectors of finance through reinsurance products. And liquidity dries up as everyone becomes unsure of any of their counter-parties actual financial health. What threatened to bring down the entire house of cards in 2007/2008 was the overnight lending market between banks was shutting down.
What makes this especially troubling in the current environment is a confluence of factors. First is the inability to actually jail corporate executives of very large companies. Even if there is fraud, we fine the corporation rather than hold its officers criminally liable. Let me do the math for you. Let’s say you put together a 300,000,000 dollar deal where your bonus is 5%. If you commit the fraud necessary to close the deal, you will be paid 15,000,000 dollars. Should you even get caught, you will have to give most of it back but won’t go to jail. And you keep all the other bonuses you also received. It’s likely the government will settle with your former employer. And if you’re not caught, and the company goes under, they still owe you the 15,000,000. You can sue for it in bankruptcy court, or from the company that acquires your old employer. If the company gets bailed out with public money, contractually, they will still need to pay you the 15,000,000. But what if it isn’t fraud and it’s just making a bet you wouldn’t otherwise make? There is are incentives to take outsized risks. After all, they’re losing other peoples’ money. So you will likely make 15,000,000, or maybe as little as 3,000,000 on the off-chance you’re caught. You won’t go to jail. And you won’t lose any of your houses just because you lost your job.
But coupled to that is a president who is willing to pardon anyone who is a supporter. He has commuted or pardoned people for political advantage. Like CZ to make good with the crypto crowd. Or the violent protester from January 6. It could be that even though there is criminal fraud, having donated to the campaign, the ballroom, the inauguration, and made statements pleasing to the ear of the administration, is sufficient to insulate your from Federal prosecution. And if you’re in a state such as Texas, it might also insulate your from state prosecution. We may find that explicit fraud was committed, people knew and traded on the fraud, but the fraudsters supportive of the administration are pardoned. In other words, they get to walk away with a lot of zeros in their bank account and no accountability.
If it were “free money,” I wouldn’t care. But what happens when a PE firm, lending money for data center construction, suddenly finds itself the proud owner of a bunch of half built data centers? Or even finished data centers filled with useless, expensive, and rapidly depreciating assets because AI demand isn’t what many expected? And what happens to the pension fund that put 250 million into that PE firm? And multiple other PE firms who also couldn’t resist deals around AI? Or the bank that provides liquidity for the PE firm? It’s not “free money,” it’s coming from somewhere. And that somewhere could be a wide-spread, systemic problem. How does the Federal government backstop PE firms, who are not insured or regulated like banks? Does the government step in and buy stock in the fraudulent company, injecting good money after bad? How do we put possibly trillions of dollars of bailout into firms but let the fraudsters walk away with all their money? Does the government step in and buy the data centers? Do the fraudsters stay in charge if they made enough dulcet noises of support for the administration?
Let’s put together a package that “doesn’t cost the taxpayers a dime.” It involves backstopping the loans, purchasing shares in troubled companies, and buying some data centers. All with money is effectively printed by the federal reserve or raised from “investors” with guaranteed loans. Essentially, this injects a pile of money into the economy, which will fuel inflation. It would also expand the debt, causing even more worry about the US debt burden. A burden the US has every incentive to ease by devaluing the dollar and inflating its way out of the crisis. And like the COVID relief bills, a lot of that money will go into creating even more income disparity. Not only will the wealthy (including fraudsters) walk away with the money they made on the way to the crash, eventually they will reap the reward of the stimulus injected to moderate the economic damage. And while the previous administrations tried to put some limits on how either the post 2008 or COVID stimulus could be used, I doubt this administration will suffer that burden.
I don’t know if or when the AI trade unwinds. I suspect it’s ‘when,’ and the longer it goes on, the more I suspect it will unwind badly. There is little I’m hearing that makes me sanguine about an orderly end to this. On the spectrum there will be true believers to outright fraudsters. Like flies to shit, fraudsters are drawn to environments where people making money would rather not look too closely as long as money is being made. If anyone did look closely, the party’s over no one is making money. After all, a little ‘wiggle room’ makes the market possible. To repurpose Mao, this is the water in which the fraudster swims. Whether its repacking bad home loans, creating accounting practices at suspiciously rock-star energy companies, or the sales figures at ‘world leading’ telecom companies, no one wants the gravy train to come to an end. But rest assured, the longer the massive (and frankly stupidly large) sums of money are changing hands (or not actually changing hands) over various AI deals, the more openings fraud will find.