The World’s Biggest Coincidence

I’ve been reflecting on the spate of outages we’ve experienced over the last week or so. Outlook. Apple. Various banks. As one politician is said to have remarked, if they’re all a coincidence, they’re the biggest coincidence in world history.

I’ve considered a scenario which would not only explain the outages but also the explanations being given for them. My conjecture is that continuous deployment of the sort Amazon, for example, has been doing for years and vibe-coding are not felicitous companions. Consider that Microsoft described its Microsoft 365 outage last week as caused by a “latent software bug”. Microsoft’s explanation is consistent with my conjecture. Public incident reports are unlikely to provide enough information to confirm or refute it.

What is worthy of consideration is that vendors will not voluntarily use some of the gains they achieve from vibe-coding by the extended testing it requires. They would just shrug their shoulders and say “Things happen”. The vendor has an incentive to minimize its own expected costs not necessarily the total social costs of its failures.

I suggest changing that calculation via regulation. I propose establishing reliability requirements and imposing penalties for failures exceeding specified thresholds. Reliability standards should include factors like the duration of the interruption, the number of users affected, and the criticality of the service. Further, I think that vendors should bear some responsibility for the losses their negligence or defective systems cause. This gives them a direct economic incentive to prevent failures while leaving vendors free to determine how to achieve the required reliability. It also avoids micromanagement. I don’t propose telling vendors how to develop software. I propose making them bear more of the costs when their software fails.

One last word. Care should be taken in the regulation to avoid allowing a regulation intended to discipline dominant vendors inadvertently to protect them from competition.

2 comments

Why Do Americans Feel No Better Off?

Nominal hourly earnings: +0.4%
Consumer prices: +0.4%
Real hourly earnings: unchanged

Why do Americans feel no better off?

I think the reasons are multi-factorial but there’s one factor that should be considered. The metrics being used to chart earnings and prices don’t mean what we think they do. Here’s an example.

I would suggest that the Consumer Price Index no longer comports with what most people must spend to maintain their standard of living. In particular it doesn’t keep track of housing, taxes, or healthcare in the way that most people experience them. Those costs either aren’t reflected in the CPI at all or aren’t reflected in the way households actually experience them. Home prices have risen more than five-fold (551%) since 1980. Total taxes paid (federal, state, and local) have fluctuated over the period from 25.6% of GDP in 1980 to 28% of GDP in 2022 before falling to 25.6% of GDP again in 2024. Healthcare spending has risen from less than 9% of GDP in 1980 to 18% of GDP today. Payment for pharmaceuticals has increased 24-fold over the period.

Those aren’t weighted appropriately in the CPI. That’s why people don’t feel better off. They can see how much they spend on the drugs prescribed to them have increased in price every time they go to the drug store. They don’t see how their incomes have risen (about 400% over the period).

And then there’s the dry, mathematical reality that averages conceal as much as they reveal. Without knowing how earnings and expenditures are distributed, an average tells us remarkably little about the circumstances of individual Americans.

The problem isn’t necessarily that the government is measuring inflation incorrectly. It’s that we’re using measures of inflation and earnings to tell Americans how prosperous they are while ignoring much of what determines whether they can afford the lives they want to lead.

1 comment

The Pressures

In the same vein as Sen. Sanders’s op-ed of which I took note, Ruy Teixeira posts on the “pressures” the new Democratic Congress will face in “delivering”:

Yet there are reasons for skepticism, especially in our “politics without winners” era where partisan majorities have been highly unstable. Both of our major parties have established a habit of attaining power and then going far beyond what voters will support. They overreach, then get thrown out on their ear two or four years later.

He goes on to list eleven “pressures”. Here is the first:

The election was a rejection of Trump, not an endorsement of Democrats. In fact, the Democratic party brand is still in deep trouble despite the election results. Smart observers acknowledge this fact, but the temptation will be great to assume voters supported the broad Democratic agenda. In reality, their mandate is simply to be “not Trump,” undo some of the damage he has done, and (somehow) make life more “affordable.” Support for “bold” programs, or really any specific program, is questionable. But that won’t stop many of the loudest Democratic voices from making a racket.

What to do in the face of these “pressures”? That’s why I think the next Congress, to forestall accusations of being a “do-nothing” Congress and actually deliver something, will find the temptation to investigate and impeach Mr. Trump irresistible. They might even remove him.

Investigations, hearings, subpoenas, and impeachment proceedings are highly visible Congressional activities. They generate headlines, satisfy partisan constituencies, and demonstrate opposition to the president. But none necessarily improves the material circumstances of voters.

I wish the newly-elected Congress well in addressing the country’s manifest problems, from which I think the president is largely a distraction. That is not to say I support him; I don’t. That is not to say I do not think he is corrupt; I do. They have bigger fish to fry.

2 comments

Delivering

There is one thing in Vermont Sen. Bernie Sanders’s Fox News op-ed with which I agree wholeheartedly:

Democrats now have the votes to pass legislation that will greatly improve the lives of millions of working families. And that’s exactly what they must do. No more empty rhetoric and poll-driven messaging.

No more telling people, over and over again, how terrible Donald Trump and the Republicans are. No more excuses. The new Democratic majority must deliver for a long-neglected working class.

The balance of the op-ed consists of his agenda for Democrats. I agree with many of the objectives he outlines. Healthcare is very expensive. The gap between the very rich, the merely rich, and the ordinary person is too great. ICE’s enforcement practices have been quite harsh over the last several years and we need immigration reform.

There is one agenda item with which I completely disagree. We don’t need a moratorium on building data centers. There are legitimate questions about their electricity consumption, water requirements, siting, and effects on local infrastructure but a blanket moratorium is a remarkably blunt instrument for addressing those concerns and one that threatens a major engine of economic growth and jeopardizes an important American competitive advantage.

But there is a connecting thread linking his proposals for remedying the ills he identifies. They are mismatched. To cite some examples the ultra-rich didn’t become that way simply because taxes were too low. In at least one case they did so by exploiting the incentives that Congress gave them. Ordinary people’s incomes didn’t stagnate because the minimum wage was too low. Healthcare did not become expensive because of insurance companies. Not to mention that Congress’s track record on controlling the growth of healthcare spending is weak. Sen. Sanders’s “Medicare for All” depends on that control.

The message here isn’t “be careful what you wish for”. It’s “be careful how you try to get it”. Good intentions are no substitute for understanding incentives, anticipating consequences, and learning from experience.

3 comments

My Mother’s Birthday, 2026


Today’s is my mother’s birthday. I don’t have a lot of pictures of my parents together for a rather simple reason. From about 1940 to 1968 my father took most of the pictures of my mom. I don’t know who took this picture or even where it was taken. Clearly, my parents attended together and my father handed his camera to somebody else.

So, here they are together. He was around 50 and she was in her 40s when this photo was taken.

0 comments

Ground Stoppage and Nvidia’s Headaches

The two leading topics on the “talking heads” programs this morning were the national “ground stoppage” and Nvidia’s headaches relating to its CEO’s affair with a woman allegedly a Chinese spy. Many of those interviewed, Democrats and Republicans alike, laid both problems at the feet of the Trump Administration, pointing both at lack of attention and corruption. I think that both incidents deserve substantial investigation.

The irony of the ground stop was at least some analysts attribute the problem to the very process that others are complaining about—the upgrading of the communications system from copper wiring to fiber optics. That has been in process for decades. By 2024 75% of the FAA’s air traffic control systems were already classified as unsustainable or potentially unsustainable.

I found Donna Brazile’s remark during the “Powerhouse Round Table” thought-provoking:

Right, right. But look, they — they really want to set the agenda. There are progressives inside the party. They want to set the agenda.

They want to basically control what happens now that Democrats have a great team of people both in the house as well as the Senate. They just want this agenda to be about getting things done. To your point, they want to return confidence, rebuild trust in institutions. But I’m not afraid of the fight on the House or the Senate side. What I’m concerned about is that we don’t spend the next 100 days talking about, you know, internal stuff, we have to deliver for the American people because they did not elect Democrats to come to Washington, D.C. and figure out who’s going to sit in the high chair.

touching, as it did, on a story I thought was under-reported, the “What’s Next?” meeting in Denver. In that meeting the issue of “affordability” was front and center.

I wish them the best of luck not the least because I suspect the reflex will be to subsidize consumption. Without an increase in supply when you subsidize consumption, all else equal prices rise. That’s true whether you’re talking about housing or healthcare.

Consider for example, “high speed rail” in California. That project has been in progress for decades, has cost billions, and hasn’t produced a single mile of track or reduced the cost of travel in the Golden State. It only took five years to build Hoover Dam.

Making somebody else pay for something doesn’t make it more affordable. Neither does borrowing the money and leaving the bill to our children. Affordability isn’t achieved by changing who pays or when. It’s achieved by reducing what things actually cost.

1 comment

The City’s Finances

Last week as the City Council prepares to debate Chicago’s budget, Ashlee Gabrysch, the city’s Chief Financial Officer (CFO) resigned. She is the third Chicago CFO to resign in less than a year. Today she explained her reasons in an op-ed in the Chicago Tribune. Here’s her explanation:

I knew the numbers were hard when I took the job. I left because I wouldn’t present them as anything less.

expanding:

Here is the problem: in the budget address and in the press conferecnce afterwards, the mayor said 76% of this year’s budget fix is “structural”, meaning it solves the problem for good, rather than for one year. That’s only true if you leave out the money the city is borrowing to pay legal settlements, about a quarter of the total fix. When I testified, I counted that borrowing. I still do.

and I subscribe fully to the principles she outlines:

  • Show in every gap-closing plan how much of the gap is truly fixed for good, using the standard of the Government Finance Officers Association.
  • Budget for settlements based upon actuarial analysis.
  • Don’t borrow to pay everyday bills, even if we’ve done so in the past.
  • Save money for the rainy day that will come.
  • Treat every new commitment as a weight on the scale, not a line in a speech.
  • Lift the thumb. Let the scale tell the truth.

I think I would have said “Don’t borrow to pay everyday bills, especially since we’ve done so in the past”. Chicago’s history of borrowing for operating expenses isn’t a precedent that justifies continuing the practice. It’s evidence that the practice has failed. And IMO the “rainy day” is already here. Chicago can’t pay its pension obligations or legal settlements without borrowing or raising taxes and repeated tax increases risk driving away the very individuals and businesses on whom the city depends for revenue.

I wouldn’t be a bit surprised if many members of the City Council recognize those principles. I suspect they simply don’t know how to get re-elected if they subscribe to them, which explains a good bit of the fix we’re in.

2 comments

Post Title of the Day

I recently ran into what I consider the post title of the day, possibly of the week:

Lame Duck a L’Orange

It is from the Doomberg Substack newsletter, to which I give full credit. Well done.

The topic of the post is the pace at which China is able to build data centers which contrasts with our own. Indeed, it has been suggested that at least some of the growing American public opposition to data center construction is due to a Chinese psyop.

0 comments

Real Healthcare Spending Since 2012


The graph above illustrates the real per capita national healthcare spending in the United States from 2012 to 2025. The following graph illustrates the second derivative (acceleration) of national per capita healthcare spending over the same period:

A lot of things might be said about that and people do. It has something for everyone. Defenders of the Affordable Care Act point to the moderation in percentage growth rates between 2014 and 2019 compared to historical pre-2010 averages, arguing the law bent the cost curve. Others say there has been an uninterrupted rise in total and real per capita dollars ($11,135 in 2012 to $12,743 in 2019 in constant 2024 dollars) and that the law did not reduce costs. Supporters of the first Trump administration highlight the flat second derivative and modest real growth from 2017 to 2019 (averaging roughly 1.6% annual real growth before COVID-19). Critics point to the massive 2020 spending spike (+$1,140 real per capita in a single year) during the pandemic response. Supporters of the Biden administration emphasize the temporary real per capita decline in 2021 and 2022 (dropping from $13,883 to $13,393 in constant 2024 dollars) as economy-wide inflation outpaced medical price growth, alongside expanded coverage. Critics point to the 2023–2024 re-acceleration (reaching $15,474 per capita) and the steep rise in nominal spending.

What cannot reasonably be said are that healthcare spending is declining or that there has been a durable reduction in spending. Even when the second derivative turned negative, that did not necessarily mean spending was declining. Indeed, real per capita spending increased in most years, and the declines in 2021 and 2022 proved temporary.

4 comments

The Outlook Outage

It may well be that the most consequential aspect of the ongoing outage in Microsoft Outlook won’t be the outage itself. It will be the fallout. That may have a number of diverging aspects.

First, managers should begin wondering if they are too dependent on Microsoft. That dependence may go beyond email to identity and authentication, productivity tools, and infrastructure itself via Azure. That has proven to be a risk.

Echoing my earlier post, there may be heightened regulatory scrutiny on Microsoft. That’s been delayed for far too long.

Finally, maybe cloud architecture isn’t such a good idea after all.

2 comments