FRANCE STUMBLES TOWARD CRISIS
Student marchers across France are ostensibly protesting against decaying and understaffed schools. However, schools are just one of the nation's great problems that have been left to metastasize for more than twenty years due to incompetent governance. As with all western countries, those who get elected to French public office get there by promising more government benefits to voters. No-one wins an election promising austerity or a balanced budget. And so the nation's fiscal crisis continues to expand.
France is currently deemed to be the "Sick Man" of the EU. Its government debt to GDP is near 120%. The cost of "servicing" that debt (paying interest on its government bonds) is expected to increase nearly 60% by 2030, just four years from now. Its 10-year borrowing costs are rapidly rising - now over 5%, the highest in over twenty years. It has nearly $1 trillion of debt coming due in the next four years and must sell $380 billion of new debt next year.
If interest rates remain elevated, or rise further, the interest cost, already one of the nation's biggest budget expenses, threatens to collapse France's fiscal house of cards. The country's annual fiscal deficit is over 5% of GDP despite EU rules strictly limiting member deficits to 3%. France has exceeded that limit twenty out of the last twenty-five years with evident impunity. The French government's tax burden now consumes more than half of the nation's entire GDP.
Realizing that elected representatives in the National Assembly will never agree to meaningful budget cuts to reduce the deficit, French Finance Minister Roland Lescure threatens to utilize "emergency powers" set out in clause 49.3 of the French Constitution, and/or executive decrees, to rein in public spending. That threat allows the political left (Jean-Luc Mélenchon) and right (Marine Le Pen) to stir up public unrest in the hope of seizing power in the spring elections. Sadly, neither of them have come forth with any rational answers to the nation's problems. Mélenchon's solution?
"All we have to do is take the 18% [of the government debt] held by the Bank of France and chuck it in the fire."
That means defaulting on nearly one-fifth of the nations' debt. But defaulting comes with consequences. Who thereafter would buy the French bonds that have to be sold to fund the government's ongoing operations? Buyers would reasonably fear their bonds might soon be "chucked into the fire". Adapting Martin Niemöller's famous post-WWII confession, "First they came for the government held bonds and I did not speak out because I was not the central bank...." Simplistic/populist solutions always come with unexpected results.
In April, 2023 President Macron bit the bullet and raised the retirement age from 62 to 64 because the cost of the government pension program was unsustainable. People are now living far longer than when the program was enacted. It was the rational thing to do but was wildly unpopular. His popularity rating is now an abysmal 15%. Just as in the US, England and elsewhere, the government cannot afford to support retirees for twenty or thirty years - unless wage withholdings to cover that cost are dramatically raised. For her part, National Rally's Le Pen has proposed lowering the retirement age to 60 without any explanation as to how that extra cost could possibly be funded. Her popularity rating is now 36% - more than double Macron's.
France has not had a balanced budget since 1974 - fifty-two years ago. There are limited solutions: raise taxes, cut spending, or a combination of the two. But there is no popular or political will to do so. So long as France's fiscal problems remain unaddressed, its borrowing costs will likely rise, rapidly escalating the crisis.
Ambrose Evans-Pritchard at the Telegraph recently outlined the history behind France's current plight.
When France joined the euro – or more accurately snookered Germany into giving up the Deutsche Mark, thinking it could subjugate the almighty Bundesbank by that means – it gave up its sovereign economic defences. It no longer has a central bank at its command, acting as lender-of-last-resort in the national interest. Its debt is denominated in a currency that it does not control. It cannot print its own money, bail itself out or inflate away its debt burden unilaterally. Nor can France devalue to boost exports and cushion the contractionary shock of the coming fiscal reckoning.
He then added the quiet part about what France clearly hopes to achieve: get bailed out, once again, by the European Central Bank, at the expense of all of the EU taxpayers.
The raging debate among insiders is over how long the European Central Bank (ECB) will wait before stepping in to rescue Paris and suppressing another festering eurozone crisis. “There can be absolutely no suspense on how this ends,” says Robin Brooks, the former chief economist at the Institute of International Finance and now at the Brookings Institution. “The ECB will intervene to cap yields. It’ll do so because it thinks sovereign debt defaults are an existential threat to the euro and therefore its own survival. It’s done so ever since Mario Draghi said ‘whatever it takes’ in 2012. France and Germany both had public debt ratios near 60pc of GDP in 2007. It is what made monetary union viable. Today, Germany is still near 60pc while France has doubled to 120pc and is on its way to 140pc. That divergence of the EU’s two core economies is the widening structural crack in the euro edifice.
In what appears to be a bizarre act of surrender, the heads of major US companies providing AI services to the government, industry and the public are calling for their businesses to be heavily regulated by the government. The stated reason is that AI has the ability to do things it should not be doing. They alarmingly claim that AI has the potential "to destroy humanity". The media has headlined stories of how versions of AI in testing mode have "escaped" their controlled domaines, broken into outside data centers and collaborated with one another to cheat on tests and found means to avoid human detection of their errant ways. What has not been well-publicized is that a recent "escape" occurred because developers left back doors open to the AI program, either intentionally or negligently.
The industry's united outcry is that someone must step forward to protect us from these risks - and who better to do that than a new government bureaucracy made up of politically appointed "experts." The government has no such experts, so these people will have to be drawn from the AI industry (Meta, Microsoft, Anthropic, OpenAI, et.al.) - people who have actual knowledge of the workings of AI. Thus, the "controllers" will be people coming from the "controlled" industry. Much like the Interstate Commerce Commission during the US railroad buildout in the 1800's. It was staffed by political appointees coming from or supported by the railroad owners. This is referred to as "regulatory capture" or "crony capitalism".
Should you be inclined to think that regulatory capture is a fiction concocted by libertarians, we encourage you to read Patrick Newman's Cronyism - Rise of the Corporatist State 1849-1929. Crony capitalism is as old as the US itself. You will be dismayed to learn of the pervasive, corrupt nexus among businesses (banking, meat packing, railroads, cotton producers, iron and steel producers, oil drillers, refiners and distributors) and government officials who purported tor regulate them. Consumers were left to pay the cartels' fixed prices, industry-solicited tariffs that raised prices, eliminated consumers' choices, leading to shortages, and the "government regulations" used to favor rich, politically connected businesses. This is not the history any Americans were taught in school. They may have been briefly exposed to a few of the "Robber Barons" of yesteryear but were assured they were long-ago reined in by anti-trust laws. It is hoped that Newman will come out with a Volume II, 1930-Present, refuting that widely-held but mistaken belief.
There is no doubt that advanced AI has the ability to cause great mischief. There is also no doubt that AI from Palantir Technologies and other companies is currently being used by governments to spy on their citizens, political opponents, foreign states, develop the means to attack adversaries' electrical systems, communication networks, data bases, defense systems, aviation control centers and to steal government and commercial secrets and confidential personal information.
Thus, the risks of AI are real. The question is whether the proposed solution is the answer to the problem or is, instead, a Trojan horse. When any industry pleads for heavy government regulation "to save the world", we immediately suspect ulterior industry motives. The major US AI players have spent hundreds of billions of dollars building out their AI products. It is in their interest to protect those investments from competition.
How better to do that than to subject the industry to massive regulation - regulations that the large existing firms can afford to comply with, while upstart competitors will be overwhelmed by them - to the obvious benefit of the existing industry cabal. Our suspicions are heightened by the fact that at the very moment industry insiders raised the alarm, elected officials just happened to have on hand already drafted laws to regulate the industry. Bernie Sanders' legislation proposes to ban any further development of AI until a federal regulatory body can be set up to oversee the industry. Either Sanders was incredibly prescient (a premise unsupported by history) or there was joint planning by industry insiders and government officials working together to benefit the industry insiders.
Extensive regulation will freeze out new players in the industry thereby reducing competition and innovation. Regulators will create detailed requirements that must be complied with before the release of any new AI iteration. This leads to our biggest concern: legislation setting up a compliance program will either directly or indirectly protect AI providers from legal liability from law suits seeking damages that their products inflict on others. When sued, AI providers can be expected to argue that they fully complied with the federal regulations and therefore they have met their duty of care to the public - as a matter of law. This outcome would grant indirect immunity to the AI majors.
Historically, providers of goods and services are subject to lawsuits seeking damages for bad outcomes from use of their products. But there are a few exceptions to that rule. One example is the PREP Act (Emergency Immunity) that created broad immunity for manufacturers of the Covid vaccines from all bad outcomes - except in cases of "willful misconduct" - a burden of proof nearly impossible to meet. This is direct immunity for damages. Coupled with that was the Countermeasures Injury Compensation Program that provided for limited monetary awards to those injured by the vaccines. The funds for that program come from annual congressional appropriations - i.e., taxpayers' money, rather than money out of the pocket of the pharmaceutical company that caused the injuries.
Tort liability properly puts the onus on the company to ensure that its product is safe. This is where the risk belongs. AI companies can protect against that risk by self-insuring the safety of their products or they can buy risk coverage from insurers and other entities. If AI companies are allowed to avoid the financial consequences of their products damaging people, that is a monumental win for them and a huge loss for society at large.
AI companies have another problem. Their capital expenditures are soaring but their revenues are falling. This chart shows actual and projected AI costs in the vertical bars and revenues in the falling blue line. This is not a recipe for success. As more users switch to less expensive Chinese open AI programs, the losses may grow even faster.
