This is important to realize, since most of the policy discussions by the AI believers seem based on the idea that we have never before seen technologies that lead to rapid gains in productivity, which is what AI-caused job loss means. From 1947 to 1973, productivity grew at an annual rate of roughly 3.0%. Instead of causing mass unemployment, this resulted in rapid gains in wages and living standards for the vast majority of the population.

It would be very impressive if AI leads to 3.0% average annual productivity growth over the next decade, but even if it somehow led to even faster growth, with the right institutional structure we would have no problem ensuring the gains were broadly shared. In the post-World War II golden age, many countries in Europe enjoyed even more rapid gains in productivity. These countries also had widely shared prosperity. If AI does produce an unprecedented boom, we do know how to structure the economy so that the benefits are widely shared. We did it before.

Unions: The Key Institution

The most important of the institutions ensuring productivity gains were widely shared was unions. Roughly a third of private sector workers were in unions in the 1950s and 1960s. Unions acted to ensure that workers shared the benefits of higher productivity in higher pay and benefits. Today, just over 6% of the private sector workforce is in unions. As a result, relatively few workers have the bargaining power to secure wage gains in line with productivity growth.

The plunge in unionization rates did not happen because workers suddenly decided they no longer felt they needed representation in dealing with their employers. It happened because the laws were deliberately structured to make it far more difficult for unions to organize and bargain effectively.

Probably the most important change was the Taft-Hartley Act, passed in 1947. This law severely restricted freedom of contract so that companies and unions could no longer sign agreements requiring that everyone represented by a union pay a fee for their representation. This scheme, misleadingly labeled “right-to-work,” eventually allowed workers in most states to freeload on those that chose to pay to support the union. That made it far more difficult to organize and sustain a union.

In addition, the mechanism for enforcing worker rights, the National Labor Relations Board, became a virtual joke in dealing with employer law-breaking. In other areas of law, say a company fails to carry out a contract, there are clear and serious legal penalties. However, even in cases where companies egregiously violate the law, say by firing workers for leading an organizing effort, the penalties are trivial. In a context where there is no serious consequence, companies have become far more aggressive in fighting unionization even when it means violating the law.

It is easy to see the extent to which the decline in unionization is driven by laws and institutions, not technology and workers’ tastes. Canada, which has a very similar culture and economy, has not seen a remotely similar decline in unionization rates over the last half century.

If we want to raise unionization rates in the United States, we can adopt similar policies to Canada, most importantly, requiring first contract arbitration. Many unions never get a first contract because companies just stall rather than negotiate seriously. Having recourse to arbitration will give companies real incentive to negotiate in good faith and ends in a contract if they don’t.

Higher Minimum Wages

From 1937 to 1968 the minimum wage increased roughly in step with productivity growth. This had a huge impact not just on workers getting the minimum wage, but also for workers considerably higher up the wage ladder. This did not lead to mass unemployment. The unemployment rate in 1968 was under 4.0%.

If the minimum wage had continued to rise in step with productivity, it would be close to $28 an hour today. That would put it above the current median wage. This means that more than half of the workforce would be directly getting higher pay. Another 20-30 percent could expect to see wages increased through a spillover effect. This step alone would go far towards addressing the rise in inequality over the last half-century.

Shorter Workweeks and Work Years

Perhaps the most painful aspect of our great minds’ pontifications on the AI-driven job loss is that they seem unable to comprehend that the length of the workweek/work year is not fixed. To put this as simply as possible, 150 million jobs at 40 hours a week, 50 weeks a year, involve equivalent work to 300 million jobs at 20 hours a week, 50 weeks a year. That is probably too simple for people in high-level debates to understand, but most ordinary people can get it.

We made the 40-hour workweek the standard with the 1937 Fair Labor Standards Act. That was almost 90 years ago. It is reasonable that if we really see a massive AI productivity boom, to shorten the workweek further, say to 36 hours, 32 hours, or even less. We can also mandate vacation time, as every other wealthy nation has done. Fewer workdays, like shorter workweeks, spread work among more workers.

AI Productivity Growth Is Not a Problem

It’s entertaining to see the country’s great minds struggling with how to deal with the fallout from an AI productivity boom that may never occur, but the reality is that we have known for many decades how to ensure that the gains from productivity growth are widely shared. We need to teach our elites a little history, not reinvent the wheel.

We also should stop pursuing policies designed to redistribute income upward, like longer and stronger patent and copyright laws, private equity-friendly bankruptcy laws, and special protections for huge social media platforms like Facebook and X. The key point is that the policies needed to reverse inequality and ensure broadly shared prosperity are easily identified. We shouldn’t let the great minds appear to struggle with the problem, as they wring their hands and pretend otherwise.

However, in the Trump era, where the rich can never have enough, it seems Social Security is again on the chopping block. Jeff Bezos’ Washington Post had two hit pieces in the last week, and we can assume many more will follow.

The main reason for supporting Social Security is that it ensures that workers can count on at least a minimal income in retirement, as well as providing protection for survivors after a worker or retiree dies, and insurance against disability. But an underappreciated aspect of the program is that it is incredibly efficient.

I like to focus on this fact, in part because I’m an economist, but also it exposes the lie that conservatives are motivated by the pursuit of efficiency. In fact, people often claim that there is a “neo-liberal” agenda that is about putting market efficiency above everything else. I have long argued that if there is a neo-liberal agenda, it’s about giving more money to rich people, without any concern for efficiency. The effort to cut and/or privatize Social Security provides a great example.

It can be difficult to appreciate the true difference in efficiency between providing retirement income through a public Social Security program and a private system of 401(k)s. This is in part because the latter shows expenses as a share of the stock of money and the former as a share of the flow of benefits. It’s also in part because the difference is so large it strikes many as implausible.

The administrative costs of Social Security are straightforward; they are 0.3% of annual benefits. In 2025, it paid out $1438.5 billion in benefits and had $4.3 billion in expenses.

In the case of 401(k)s, the expenses are typically expressed relative to the stock of money in a fund. As a ballpark number, let’s say expenses are 1%. There are funds with lower fees, but also many with higher ones. It’s also important to remember that the costs are not just what the 401(k) administrator itself charges, but also the fees charged by the funds purchased in the 401(k). These can be very low with a large index fund, but fees on actively managed funds can easily be 0.5-1.0% annually.

The other question is the size of the stock of funds needed to pay a specific benefit. This depends on the rate of return that a 401(k) will typically yield. For simplicity, let’s assume a 4.0% real rate of return. That’s considerably less than the historic return on stocks but far above the real return on government bonds. At present, the interest rate on 10-year Treasury bonds is 4.8%, while the year-over-year inflation rate is 3.5%. This would imply a 1.3% real interest rate, although the expected inflation rate over the decade is likely closer to 2.3%, which would imply a real return on long-term government bonds of 2.5%.

Let’s say the Trump-Bezos gang wanted to replace $500 billion in annual Social Security benefits, roughly one-third of the total. At a 4% real rate (real rates are appropriate, since benefits rise with inflation), it would be necessary to have a stock of assets equal to $12,500 billion. If the annual costs of maintaining this money in 401(k)s are 1.0%, that comes to $125 billion annually. By contrast, the administrative cost of paying this money through the Social Security system would be $1.5 billion annually.

This makes the private 401(k) system more than 80 times more expensive than Social Security. No one who cared about running an efficient retirement system could argue for the 401(k) route.

Of course, if the point is to give money to wealthy donors in the financial industry, 401(k)s are clearly the way to go. There’s no money for these folks in Social Security. And now we can be very clear on why we read so much about how it’s necessary to replace Social Security with private retirement plans.

This boast has absolutely nothing to do with reality. Contrary to what Trump says, there is no boom in auto factories or any other factories coming into the United States. Factory construction has actually plummeted since Trump came into the White House, after booming under Biden.

The looniness about the investment boom should be kept in mind when considering Trump’s deal to take control of 65 billion barrels of Venezuela’s oil. Trump is promising that this deal will both lower gas prices and be used to refill the Strategic Petroleum Reserves that he has largely drained as a result of his war on Iran.

While people are naturally offended by the blatant imperialism in this deal, that is precisely the reason why there is not much cause to be concerned. The basic story is, because the deal is so transparently an effort by the United States to steal oil from a country that it controls militarily, it is virtually guaranteed that as soon as the country is not run by a U.S.-installed dictator, Trump’s deal will be dumped in the trash.

This point was well-made by Venezuelan economist and opposition leader Ricardo Haussman (also cited in a piece by Michael Tomasky in The New Republic). The deal is seen as a joke by pretty much everyone in Venezuela.

If Venezuela doesn’t honor this deal, what will a future administration do, take them to court? Or will they send U.S. soldiers to die for Venezuela’s oil?

In terms of its immediate impact, there is not some spigot that can be turned on that will get massive amounts of Venezuela’s oil flowing. It’s currently producing around 1.25 million barrels a day. While it has massive reserves, its production facilities deteriorated badly as a result of U.S. sanctions. The country was largely unable to get needed parts, and it lacked funding because it was prevented from selling its oil to much of the world.

Oil production can only get back to its 3 million barrel a day turn-of-the-century pace with billions of dollars of foreign investment. It is unlikely that this deal will make companies happier about putting large sums at risk. Any investment will only pay off over many years, long after Trump will be out of office.

It also is not plausible to use Venezuela’s oil to refill the Strategic Reserves. Venezuela’s oil is very heavy and highly acidic. The Strategic Reserve is designed for light, alkaline oil. It could not easily store Venezuela’s oil.

In short, file Trump’s Venezuelan oil deal with the $60 trillion in foreign investment pouring into the country. It exists only in his head.

With the start of the football season, the analogy seems to be a reasonably good description of what we are seeing with AI development. There is no doubt that two or three years ago the Silicon Valley AI team was far ahead, with China nowhere in sight. Then there was the DeepSeek moment last winter, when the startup produced an open-weight model that had performance in the ballpark of the leading Silicon Valley models, and selling at a small fraction of the price.

Since then, Chinese AI companies have produced a string of models that come ever closer, and in some areas surpass, the leading Silicon Valley models. They also maintain their enormous cost advantage, typically selling tokens for less than one-fifth the price, and often less than one-tenth the price.

If this is hard to grasp, imagine you’re Elon Musk trying to sell Teslas for $50k a piece. Suppose there are Chinese models that are every bit as good selling for $5k. How do your business prospects look? That is the question U.S. AI producers, as well as everyone in the supply chain, need to be asking.

And this story should be well known to anyone putting money on the table. That picture doesn’t really seem to be in dispute by people who follow the industry closely. The basic story is the same everywhere. The Chinese AI companies have attained comparable performance to the best U.S. models. In addition to being far cheaper, the open-weight Chinese systems also have the advantage that they can be downloaded and run on a company’s own computers. This means both that the systems can be customized, and they don’t have to worry about OpenAI sharing their customers’ data with the highest bidder.

As a result of the lower prices and greater flexibility, Chinese AI is rapidly gaining the bulk of the world market. According to data from OpenRouter, Chinese AI went from less than 10% of world AI usage at the start of 2025 to more than 60% in early July.

What’s the story that turns this around in the U.S. favor? As much as Donald Trump might like to win the AI race, he doesn’t seem like he is helping the cause. He can’t decide whether he wants China to buy Nvidia chips or not. The answer is likely to depend more on campaign contributions than on the health of the industry.

In the same vein, he tried to kneecap Anthropic, the leading U.S. AI company by most measures, by declaring it a supply chain risk. This would have seriously hampered sales, not just to the military, but to businesses across the economy. Fortunately for Anthropic and fans of U.S. AI, the courts seem to have nixed this effort, although Trump may appeal the ruling.

China has a whole fleet of nimble AI companies that are constantly producing new cutting-edge models. It has a near-limitless supply of electricity (thanks largely to wind and solar), and it vastly outnumbers the U.S. in the scientists working in AI-related areas. In this world, how does the United States prevent China’s near parity from turning into a total rout? Coach Trump seems to be going 180 degrees in the wrong direction right now with his wars on wind and solar energy, and his attacks on immigrants.

Just to remind folks, many of the scientists working on AI are not white and a good number are immigrants. They may not relish working in a country where white nationalism has become the official religion, especially when there are many other countries that are happy to reward their skills.

The Deficit Whiners Disagree with Wall Street on AI

I learned arithmetic in grade school. I will always stand by it since it has always stood by me. As I’ve noted in the past, if we are not seeing a bubble, the huge stock market valuations of AI-related companies imply that the economy will grow far more rapidly in the next decade or so than is projected by most professional forecasters, including official forecasters like the Congressional Budget Office.

If you were one of the people sharing in the national scare over the debt hitting $40 trillion, then you must accept that the economy will continue to grow relatively slowly, and we will not see the massive AI growth dividend implied by the current valuation of AI-related stocks.

If the economy grows at a 3.5% annual rate over the next decade, which assumes that AI raises productivity to around 3.0% annually, the rate we had from 1947 to 1973, then the economy will be 41% larger in 2036 than it is today. The CBO projections assume the economy will be less than 20% larger. The gap between 41% growth and the growth projected by CBO comes to almost $7 trillion annually, measured in 2026 dollars.

Can anyone with a straight face say how we are supposed to be terrified by paying $1 trillion a year in interest on the debt, when AI is giving us $7 trillion to play with? Mr. Arithmetic says that’s crazy.

This is far from the first time where policy debates have been impervious to simple arithmetic. In the 1990s, the central theme in the Social Security debate was generational equity.

One story that the deficit hawks came up with to justify cuts to Social Security was that the Consumer Price Index (CPI) was substantially overstating the true rate of inflation. The usual range was 1.0-1.5 percentage points annually.

The problem with this complaint is that the CPI is our yardstick for comparing real income over time. If our yardstick is broken, and actual inflation is 1.0-1.5 percentage points less than our measures show, then real income is rising 1.0-1.5 percentage points more rapidly than we had thought. It also would rise 1.0-1.5 percentage points more rapidly in the future than the projections show. That would mean that most Social Security beneficiaries had grown up much poorer than we recognized at the time and that future generations of workers would be far better off than we ever could have hoped. How can that be an argument for cutting Social Security?

However, this arithmetic eluded most people in the debate. I was once on a show with Wyoming Senator Alan Simpson, who passed as an intellectual in policy circles. He went on a diatribe where he said some economists tell him that the CPI overstatement could even be 2.0%; and then added that our grandchildren will be living in chicken coops.

Anyhow, however much Washington policy types and Wall Street investors may resist the notion, arithmetic cannot be defeated. If AI is going to deliver a growth boom, there is no reason to be worrying about the deficit. In my own case, I’m not betting on a massive growth boom, but I still don’t think the deficit is much to worry about. But if you do buy the AI boom story, don’t make yourself look foolish by yelling about an unsustainable debt.

The smaller revision makes it less newsworthy than would be the case with a larger revision, but there are still a few points worth noting:

1) The smaller revision could be due in part to less immigration;

2) The drop in manufacturing makes the fall in jobs in the sector even larger;

3) Retail also had a large drop, although construction employment was revised up.

Before going into these, it is worth taking a moment to explain what this revision is. Every month BLS surveys almost 120,000 businesses and government entities, asking how many people they have on their payroll. This survey, Current Employment Statistics (CES), provides the basis for the jobs numbers reported every month. (There is a separate survey of households that provides the basis for the unemployment rate.)

While this is a very large survey, it is still a survey and will not perfectly capture all changes in employment. Notably, it will miss job losses in firms that went out of business and therefore did not answer the survey, and job gains in new businesses that were not included in the survey. And there is always some sampling error because these businesses will not be perfectly representative of the whole economy.

To account for this, BLS compiles data from state unemployment insurance filings. Since nearly the entire workforce is covered by unemployment insurance, these filings provide a near census of employment in the country. The revisions published yesterday are based on these filings, which are compiled into the Quarterly Census on Employment and Wages (QCEW). This revision is preliminary since the QCEW data will still be revised as states compile more complete data, but the changes from the preliminary number to the final number are usually small.

Why Is the 2026 Revision So Much Smaller?

The 2026 revision is actually pretty much in line with most revisions. It comes to just over 0.05% of total employment. Going back to 2022, the revision was a gain of 56,000, and in 2021 it was a decline of just 7,000.

Most years the revisions are relatively small. The exceptions are usually economic turning points. When the collapse of the housing bubble threw the economy into recession in 2008 led to a downward revision of 902,000 in 2009, equal to 0.7% of total employment. In 2006, the peak of the housing boom, the upward revision was 752,000 or 0.6% of employment.

Since the economy was continuing to grow at a healthy clip in 2022 and into 2023 and 2024, there is not an obvious turning point explanation for the large downward revisions. The most plausible story I have heard is that we had massive immigration these years, with many of the immigrants not having legal work status. It’s possible that employers listed these people as workers in the CES, but knowing their status, didn’t file with state unemployment agencies for these workers. That possibly could explain the gap, and then with the plunge in immigration since Trump took office, the decline to a more normal-sized revision.

If the immigration story is true, then we should have seen the largest downward revisions in the sectors that employed the most undocumented immigrants. A quick glance provides some support for that view. The downward revision for hotels and restaurants was 0.9% compared to 0.6% overall; for non-durable manufacturing (food processing and apparel) it was also 0.9%, but it would be necessary to look at more detailed industry data to make a serious assessment of this view. Still, it’s the best explanation I’ve seen.

Manufacturing Employment Is Going in the Wrong Direction

Manufacturing now employs less than 8.0% of the workforce, and the wage premium has largely been eliminated over the last four decades, as the unionization rate in the sector is now only slightly higher than in the private sector as a whole. Still, manufacturing jobs continue to be a major political issue.

In the year from March 2025 to March 2026, the CES showed a loss of 68,000 manufacturing jobs. The revision shows a further loss of 67,000 jobs, bringing the total for the 12-month period to 135,000, roughly 1.1% of employment in the sector.

It’s worth noting that not all sectors were hit equally. Employment in non-durable manufacturing fell by 38,000, a decline of 0.8%. Employment in food manufacturing fell by 3.6%, while employment in apparel fell by 6.6%. This is consistent with an immigration story, as employers may not have been able to fill jobs that had been filled by migrant workers.

The drop in durable manufacturing was 30,000, or 0.4%. There were sectors showing modest gains. Fabricated metals added 9,400 jobs, an increase of 0.7%. Electrical equipment added 7,100 jobs, a gain of 1.7%.

Retail Takes a Big Hit, Construction Shows Better Gains

In the CES data, employment in the retail sector was little changed from March 2025 to March 2026, showing a decline of 22,000, just over 0.1%. The benchmark revision will increase this drop by 154,600, just over 1.0% of total employment.

While we don’t know yet the distribution of these losses, some candidates are building and garden supply stores, which had been reported as losing 39,200 jobs over this period, 2.8% of employment; and furniture stores, which reportedly lost 12,100 jobs, or 3.0% of employment.

On the positive side, the revisions added 62,000 jobs to construction, equal to 0.8% of total employment. Added to the 38,000 gain reported by the CES, this would bring the total growth over the period from March 2025 to March 2026 to 100,000, or 1.3% of employment. While this is decent growth, it’s still less than half than in 2023 and 2024.

One other change worth noting: government employment was revised up by 99,000, or 0.4%. That will be bad news for the government haters.

The Net Story: Not Much Is Different

The large revisions in 2024 and 2025 gave us a notably different take on the economy. Taking the QCEW at face value, employment was growing considerably less rapidly than we had thought. On the plus side, the revisions implied somewhat higher productivity growth. (Fewer jobs means fewer hours worked.)

This report really doesn’t imply much change in thinking about the economy. The job growth is a difference of less than 6,000 a month. The productivity impact is a rounding error. And we have the separate data from the household survey on the unemployment and employment rate. That doesn’t change at all.

I know in the list of Trump outrages this may not rank very high, but what sort of goddamn jerk would pull away people needed in one of our worst fire seasons ever to make Washington look nicer for his birthday? This shifting was especially absurd because the parks were already short-staffed. Elon Musk had reduced personnel by around 25% with his “DOGE” cuts.

It is difficult to muster enough contempt for this destructive act of vanity. Our national parks are a great treasure. They are already endangered due to global warming, which Trump claims to be too stupid to recognize. Just last year, large chunks of the North Rim of Grand Canyon burned, including the historic lodge, because the Park lacked the resources to effectively contain the fire. We are seeing more damage in and around parks this year as high temperatures and lack of rain leave the area vulnerable to wildfires.

By contrast, DC was doing just fine until Trump decided to tear up the whole area behind the White House for his UFC fights, from which he stood to personally profit. He also tore down the East Wing of the White House and left a gigantic construction site for his ridiculous ballroom. He ruined the reflecting pool with his crony flubbing the repair. And he now has the Kennedy Center covered with a tarp so that no one can see his name has been removed from the building.

Anyhow, these points are perhaps secondary to the harm he is doing in so many other areas, but it’s hard not to just say WTF? When it comes to pointless and pathetic actions of a president, it would be difficult to top this one. (Sorry, I wrote this before we got “Lake America.”)

Fast-food spending provides a useful metric since it is unlikely that increases in spending would be driven by rich people eating their stock gains. Again, this is not saying that rich people don’t eat at fast-food restaurants. I’m sure many do. The point is simply that their spending is unlikely to increase in response to stock gains.

Fast-food spending is also useful in that it is very much a discretionary expenditure. If people are feeling stretched, it is easy to cut back on the frequency of eating out at fast food restaurants and instead cook at home.

Here’s the picture.

As can be seen, spending has been largely flat since the fall of 2023. In fact, it is still below the peak hit in September of 2023. There was a sharp dip in April of this year, following the start of the war in Iran. Spending has since bounced back, but it’s still virtually the same as it was three years ago. That doesn’t make it look like the non-rich are feeling very good about their finances.

I realize that spending on fast-food restaurants could be affected by the increased use of Ozempic and related drugs. That is likely the case, but note that this did not stop fast-food spending from increasing rapidly in 2022 and continue rising in 2023. That was a period where spending on sugar and sweets at food stores fell rapidly, dropping almost 7.0% from the second quarter of 2022 to the fourth quarter of 2023. In the last three years, it has edged down by another 2.0%. So, the spread of these weight-control drugs is likely a factor slowing growth in spending at fast-food restaurants, but they cannot be most of the story.

This is how we should view the line being pushed by “moderate” voices that we have to deal with the $40 trillion debt with both spending cuts and tax hikes. The reality is that, apart from the military and Homeland Security, there is little fat in spending to be cut, as even Elon Musk inadvertently acknowledged. Insofar as we have a deficit problem, the issue is on the tax side, as can be easily shown. The rich have been taking an ever-larger share of national income over the last half-century, and they don’t feel like paying taxes on their winnings.

The major media outlets, which are all controlled by rich people, are pretending to be moderate by saying that we need to both raise taxes and cut spending. But there is nothing moderate about saying that we have to cut programs like Social Security, Medicare, and Medicaid because Republicans have given big tax breaks to their campaign contributors.

Republicans pushed these tax cuts, knowing they would increase the deficit, but did not make any corresponding cuts in spending because the cuts would be incredibly unpopular. Now they are using their control over the media to insist that these cuts are now absolutely necessary to offset all the lost tax revenue from tax cuts put in place by Reagan, Bush II, and Trump.

The Jeff Bezos-owned Washington Post gave us a great example of this fake moderate position in its editorial, “To get the national debt under control, start with the retirement state.” The piece makes its case by taking the example of a two-earner couple, with average earnings of $100,000 a year. It shows that the couple, turning 65 in 2025, can expect lifetime Social Security benefits of $739k compared with tax contributions of just $597k. A couple with the same income retiring in 2045 can expect lifetime benefits of $987k compared to tax contributions of $735k.

After laying out this disparity for Social Security (it has a similar story for Medicare, which I’ll come to), it then makes an argument for reducing Social Security for high-income people. This is three-card Monte level deception.

If the idea is that we should reduce the benefits of high-income workers, honest people would look at the relative taxes and benefits for high-income workers. Social Security is explicitly designed to have a progressive payback structure, which means that relatively moderate-income workers, like the ones highlighted in the WaPo editorial, have higher paybacks relative to their taxes.

If the editors were interested in doing an apples- to-apples comparison, here’s what the picture would look like. (This is taken from the exact same source.)

As can be seen, high-income people pay considerably more in taxes than they get back in benefits. For a high-income woman retiring in 2025, the gap is $263k. For a high-income man, the gap is $336k. (The gap is larger for men than women because their life expectancy is shorter.) For a high-income woman retiring in 2045, the gap is $259k. For a high-income man, the gap is $346k.

If the point is to make an argument for reducing the benefits of high-income retirees, then show the taxes and benefits for high-income retirees. No one disputes that Social Security looks like a pretty good deal for more moderate-income retirees, but these people don’t typically have much income in retirement. I guess Jeff Bezos’ paper would have been too embarrassed to argue that we have to reduce the average monthly Social Security benefit of $2,071.

The Post’s editorial makes the push that while cutting Social Security, we should expect people to be more reliant on private 401(k)s. In addition to increasing risk, this is also enormously inefficient. Private 401(k)s cost more than 40 times as much to administer per dollar of benefits as Social Security. It is understandable that Mr. Bezos would be happy to see more money going to his rich friends in the financial industry, but most of us would rather see the money going to ordinary workers.

Medicare Benefits: Big Bucks to Hospitals and Drug Companies Are Not Benefits to Workers

The Post’s graphs do show a huge imbalance between the taxes paid out for Medicare and the cost of the benefits received. This is also deceptive.

In the United States, we pay almost twice as much per person for healthcare as the average for other wealthy countries. This is not because we get more or better healthcare. Our life expectancy ranks near the bottom for wealthy countries.

The big bucks for healthcare go to the income of drug companies, insurers, hospitals, medical equipment makers, and doctors. In each case, we pay two times as much, or more, than people in other wealthy countries. A paper that was not answerable to one of the richest people in the world would suggest bringing our payments in line with the rest of the world. But instead, the WaPo wants to beat up on the country’s retirees.

No one should be confused: cutting Social Security and Medicare benefits to reduce the deficit is not a moderate position. It’s one that attacks hundreds of millions of ordinary workers to avoid taxing the rich or reducing waste in our health care system. That is extreme, but the rich media owners pushing this position will do everything they can to convince us they are being fair and balanced.

Unfortunately, confusion on the debt can spread even among the sane. Paul Krugman and Jared Bernstein, two top-notch economists (also friends) had a discussion of the debt in a podcast posted on Saturday. At one point, they noted that interest rates have risen, creating a situation where the interest rate is higher than the rate of growth, which could lead to explosive debt growing ever larger as a share of GDP.

That’s all straightforward arithmetic, but the confusion stems from the reason they see interest rates as going higher. Part of the reason is Trump’s craziness, which we all recognize undermines confidence in the economy and the debt. However, part of the reason is the investment boom from building AI data centers.

While I agree that this is a big factor in pushing rates higher, the assumption motivating this investment is that there will be some massive payoff in the form of higher profits stemming from huge gains in productivity. If these gains in productivity materialize, then we will be far richer than current projections show, and growth will have far surpassed interest rates. In other words, no explosive debt problem. Also, if we grow an extra 10-20 percentage points over the next decade, it seems a bit silly to be whining about debt, a point I will come back to.

There is also the possibility (likelihood in my view) that the productivity gains do not materialize. In that scenario, the AI-related stocks will crash at some point, the boom will bust, and we will likely be looking at a serious recession. In that world, our problem will be boosting the economy back to full employment, not the deficit. And for those who care about such things, interest rates should again be very low.

It seems this is a point that Paul and Jared should have brought into their deficit/debt discussion. I’ll give them a whack on the wrist for missing this point, but come to the more general issue that is usually lost in the tirades about gigantic debt.

Government Debt Is Not a Measure of Generational Equity

It is just atrociously bad economics to imagine the debt is some sort of measure of generational equity. At the most basic level, who owns the debt? The deficit hucksters talk like we send interest on the debt to people on another planet. In fact, the vast majority of the debt is held by U.S. citizens and U.S. corporations. We pay the debt to ourselves.

It’s true that ownership of the debt hugely skews upward, so the interest is mostly paid to higher-income people. But we can tax it back. In fact, let’s take an extreme case and say we tax back 100% of the interest.^[1] In that case, is the debt a burden on future generations? We will of course not tax back 100% of the interest, but we can and should have progressive income taxes. Which, if they are actually collected, will mean that much of the money paid in interest will go right back to the government. And if we apply progressive taxes more generally, we can certainly offset any negative impact that interest payments on the debt have on income distribution.

And for trivia buffs, the rich people who will collect the interest in 30-40 years are part of future generations, the rich part. The story is again a class issue, not a generational one.

Growth and the Debt

Faster economic growth reduces the burden of the debt in the sense that it makes interest payments smaller relative to the size of the economy. But the more important point is that, in principle, faster growth makes us richer.

If the economy grows 10% more (in real terms) over a decade than had been expected (imagine AI actually pays off), then we would be roughly $3.9 trillion richer in 2036 than is now projected. Suppose in this story we had to pay another $500-$600 billion a year (again, in real terms) in interest compared to what we do today. Would we have done some injustice to our kids with a faster growth and higher debt story? The size of the economy is what will determine the well-being of our kids, not the amount of debt owed by the government.

But GDP growth is not everything, especially when it leads to destruction of our environment, as we are seeing now as a result of global warming. It would take some pretty batshit crazy accounting to be concerned about the debt we are passing on to our kids, but not about the environmental destruction we have caused by our use of fossil fuels.

And this is not just destruction of the lands where people like to be able to hike, fish, or hunt. It is also very much dollar-and-cents damage. The recent wildfires in Spokane are estimated to have caused over $1 billion in property damage. The damage done by the fires in Los Angeles last year may have been as high as $250 billion (0.8% of GDP). With global warming leading to more and bigger fires, as well as more frequent and severe hurricane and flooding events, we will be seriously damaging the life prospects of future generations even if we were to pay off the national debt tomorrow.

Trumpian Corruption

I know I have said this before, but the point deserves to be constantly thrown in the face of the deficit hawks: Trump’s corruption poses an infinitely greater threat to economic stability than the $40 trillion debt. Trump has been given a green light by the Roberts Supreme Court to corrupt every agency in the country (except the Fed) to enhance his power and his pocketbook. As we have seen, this means having the Justice Department and FBI focus on prosecuting his political opponents. It means selling antitrust rulings to campaign contributors. It means allowing political allies to sell unsafe food and likely drugs.

Trump has also had the Office of the Comptroller of the Currency greenlight his new crypto bank and had the Securities and Exchange Commission look the other way on his insider trading. He even ordered the Census Bureau to cook up a nonsense report on non-citizen voting to claim that, somehow, he actually won the 2020 presidential election that he lost in a landslide. (That’s the Trumpian usage of the word, where he says he won the 2024 election in a “landslide.”)

Anyhow, these are the sorts of things that happen in “shithole countries,” not advanced economies expecting the trust of investors. With Trump continually pressing his corruption further and his supporters in Congress saying it’s all good, investors should fear putting their money in the United States even if we had zero debt. That might be too difficult for the deficit hawks to understand, but most of us have parents who taught us not to do business with crooks.

^[1] I know, no one would hold bonds if we taxed back all the interest, but I’m making a point. Trade economists have often assumed in their modeling that we replace tariffs with lump-sum taxes to show the benefits of trade. Unlike tariffs, lump-sum taxes don’t create economic distortions. Lump-sum taxes also don’t exist in the world, but trade economists felt it useful to assume them to make a point.

To show my appreciation for AI (and also out of laziness), I decided to ask it what it thinks about the bubble. I gave three leading AI systems the following prompt.

“If a pension fund manager knows there is an AI bubble, but they also know they will suffer no career consequences if the bubble bursts and costs the fund a large share of its assets, how will they manage the fund’s portfolio?”

Here’s what they had to say. (The answers are all literally verbatim. I just cut and pasted.)

Google Gemini:

A pension fund manager in this situation will continue to buy AI stocks and ride the market wave to capture short-term gains, rather than trying to avoid the crash.

This behavior happens because of standard financial incentives and psychological safety. Here is how and why they will manage the portfolio this way:

Chasing the Boom

• Buy popular assets: They will overweight their portfolio with leading AI companies, chips, and tech giants.

• Ignore high prices: They will ignore traditional valuation warnings (like high price-to-earnings ratios) because prices can keep climbing for a long time.

• Match the benchmark: They will track major market indexes closely so they do not fall behind other funds while the bubble is growing.

Avoiding Short-Term Pain

• Fear of missing out: If a manager avoids AI and the bubble keeps growing for two more years, their fund will look terrible compared to competitors.

• Peer pressure: In finance, it is safer to fail with the crowd than to succeed or fail alone. If everyone loses money together when the bubble bursts, it is blamed on market conditions rather than poor individual skill.

The Logic of “No Consequences”

• Asymmetric payoff: If the bubble grows, the fund wins prestige and high returns. If the prompt states there are zero career consequences when it bursts, the manager has zero incentive to protect the long-term capital.

• Surviving the present: A manager’s immediate danger is underperforming today. Since the crash carries no penalty, they have every reason to maximize the upside right now.

GROK

They will continue holding (or even overweighting) AI-related assets, riding the bubble rather than exiting or hedging it away.The decisive factor is the manager’s personal incentive structure, not the fund’s long-term interests or the manager’s private knowledge that AI valuations are unsustainable. With no career penalty for large absolute losses when the bubble bursts, the dominant risk for the manager becomes relative underperformance while the bubble is still inflating.

• Staying invested (or matching the peer/benchmark exposure) means the manager keeps pace if prices keep rising. When the eventual collapse occurs, the losses are shared with other managers who behaved the same way; the absence of career consequences removes the usual personal downside.

• Exiting or underweighting early creates immediate relative underperformance if the bubble continues. In institutional settings this typically triggers client outflows, board scrutiny, or job risk—risks that remain even when absolute losses later carry no personal cost.

The result is classic agency-driven herding: the portfolio is managed to minimize the manager’s career variance rather than the fund’s long-horizon risk-adjusted return. Exposure to the bubble assets is maintained (or increased to stay competitive), diversification or hedges that would reduce participation are avoided, and any “prudence” language is subordinated to benchmark-relative positioning. The fund therefore remains vulnerable to a large drawdown once the bubble deflates, precisely because the manager’s payoff does not penalize that outcome.

CHAT GPT

If the manager knows an AI bubble exists and also knows that being wrong will not hurt their career, the portfolio can look quite different from the portfolio of a manager who is personally accountable for the downside.

The key is that the manager has an asymmetric incentive: the upside from participating in the bubble may accrue to the manager through performance, reputation, or compensation, while the downside is borne primarily by the pension beneficiaries.

The basic logic

Suppose the manager believes:

• AI-related assets are substantially overvalued.

• There is a meaningful probability of a large crash.

• But not participating in the boom creates career risk if the bubble continues to inflate.

• Participating in the bubble creates little or no career risk if it subsequently collapses.

Then the manager has an incentive to remain heavily invested in AI even while believing it is a bubble.

Why? Because their relevant objective isn’t simply:

“What portfolio maximizes the pension fund’s long-run risk-adjusted return?”

It is closer to:

“What portfolio maximizes my expected career outcome, given that I won’t be punished if the risky bet goes badly?”

A simple example

Imagine two possible strategies:

Strategy

AI bubble continues

AI bubble bursts

Underweight AI

Manager looks foolish / may lose career opportunities

Manager looks prescient

Hold lots of AI

Manager looks successful

Manager suffers no career penalty

If the second column has a career benefit and the third column has no career penalty, the manager rationally chooses the second strategy—even if they privately believe the bubble will eventually burst.

This is a classic principal-agent problem. The pension beneficiaries are the principal; the portfolio manager is the agent. The manager’s incentives aren’t aligned with the beneficiaries’ interests.

What would the portfolio actually look like?

The manager might:

1. Maintain a large allocation to AI-related equities rather than aggressively shorting them.

2. Avoid being dramatically underweight the benchmark, because that creates tracking-error and relative-performance risk while the bubble continues.

3. Own the broad market, which gives them substantial indirect exposure to AI through the largest technology companies.

4. Describe the position as a strategic allocation rather than making an explicit “AI bubble” bet.

5. Potentially buy some downside protection, but not enough to sacrifice much of the upside if the bubble continues.

6. If they have discretion, shift risk toward assets where the consequences of being wrong are less visible or less attributable to the manager.

There’s an important distinction here between knowing something is overvalued and having an incentive to bet against it.

A manager can rationally think:

“I give AI assets only a 20% chance of producing the returns the market expects, but I am still going to own them.”

That isn’t necessarily irrational from the manager’s perspective.

The really interesting implication

If many institutional managers face this incentive structure, you can get a situation where almost everyone privately thinks an asset is overvalued but nobody wants to be the person who gets out first.

That can help explain why bubbles can persist.

Each manager thinks:

“Yes, this is probably a bubble. But if I sell and the bubble continues for another three years, I’ll underperform everyone else. If I stay invested and it crashes, I’ll be able to say that nobody could have predicted the timing—and I won’t lose my job.”

In other words, the absence of career consequences for the downside can actually make the manager more willing to hold the bubble asset, not less.

This is closely related to the phenomenon sometimes described as “career risk” or “benchmark risk” in institutional asset management: a manager may prefer a portfolio that is privately believed to be wrong but is conventionally wrong, rather than one that is unusually positioned and therefore exposes the manager to being blamed for underperformance.

And if the pension fund’s compensation system rewards relative performance against other pension funds or a benchmark, the incentive can become particularly strong: being wrong together is safer than being right alone.

AI Seems to Get the Bubble Right

Okay, I did load the deck by specifying there will be no career consequences for fund managers that sink their funds when the AI bubble bursts, but does anyone think this is wrong? To be as specific as possible, if AI-related losses cause pensions, endowments, or other funds to lose 30-40% of their value, does anyone believe that there will be mass firings of the managers who determined the holdings?

We sure didn’t see anything like that after the collapse of the tech bubble in the 90s or the housing bubble and the financial crisis of the 00s. If there is a mass firing (good thing in my view) it would be radically different than what happened the last two times.

My bet is that all the highly paid fund managers will get a collective “who could have known?” amnesty and leave others to suffer the consequences of their failures. In any case, my guess is that they are operating with that expectation and therefore the bubble persists, as AI says.

After insisting, in complete defiance of all the evidence, that tariffs are paid by the exporting country, Trump is apparently now acknowledging that the tariffs get passed on to U.S. consumers. When it comes to trying to keep Republicans in control of the House and Senate, Trump is willing to wade into reality.

However, he is not doing a full-fledged plunge into the real world; he’s just getting his toe wet. In addition to weakening his quotas on beef, he also pronounced in his Truth Social tweet:

“We have a commitment that this beef will be sold at 25 percent below current market prices.”

This raises the question of a commitment from who? And how will this price reduction be enforced? At the risk of being rude, I will bring some context into the picture.

According to the United States Department of Agriculture, we consume a bit over 28 billion pounds of beef a year. That would come to about 7 billion pounds over a 90-day period. That means Trump is proposing to increase the beef supply by a bit less than 10%.

Most estimates find that the demand for beef is relatively inelastic. This means that a 10% increase in supply (which assumes no corresponding reduction in domestic supply) would lead to a drop in beef prices that is considerably smaller, say a 5% drop. This means that Trump’s decision to relax beef quotas won’t come close to reducing the market price by 25%.

It may be the case that Trump has a plan to ration low-priced beef, sort of like what was done under Soviet communism, where selected people (MAGA party members?) get coupons for their pound of low-cost beef. Or alternatively, Trump may have no idea of what he’s talking about and the 25% price reduction is imaginary, just like the $20 trillion in foreign investment coming into the country and his Great Healthcare Plan.

I guess we’ll just have to wait. But in the meantime, a whole new generation will be asking, “Where’s the beef?”

I’m used to people complaining that I don’t take debt and deficits seriously enough, so I guess I would be disappointed if my post didn’t prompt some criticism. Anyhow, this gives me the opportunity to explain again why I’m not especially troubled by the $40 trillion debt. I will also explain why people really should be hair on fire over Trumpian corruption.

Debt Needs to be Understood as Part of a Larger Economic Picture

We all know that $40 trillion is a really big number, but after we give it a big salute, the question is what does it mean for the economy and for our pocketbooks? Most of the debt discussion didn’t really do this, or at least do this in a way that makes much sense.

First, we should express the debt relative to the size of the economy. The $40 trillion debt is equal to about 125% of GDP. Arguably the more appropriate measure is the publicly held debt, which excludes bonds held by the Social Security trust fund and various other public funds. This is roughly $32 trillion, or just over 100% of GDP. That’s large, but the debt-to-GDP ratio was considerably larger just after World War II, and that didn’t prevent us from having the most prosperous quarter century in our history. It also didn’t have investors here and elsewhere fleeing the dollar and government debt.

If we needn’t be scared by the government debt, what about the trillion dollars (3.3% of GDP) that we are paying out each year in interest payments? That’s a good chunk of change, but it is worth putting on our thinking caps instead of panicking.

Donald Trump just proposed, and may well get, a defense budget for fiscal 2027 of $1.5 trillion, roughly 5% of GDP. The last Biden budget for 2025 called for spending $864 billion for the military. This means Trump wants us to spend $636 billion more for the military each year, roughly 2.0% of GDP, than we did under Biden.

Virtually everyone thought the level of military spending under Biden was fine to meet our defense needs. If the new real or imaginary enemies that Trump has us combatting requires spending another 2% of GDP on the military, how is that better than spending another 2% of GDP on interest on the debt?

To put this simply, there are a lot of people getting hysterical about spending 3.3% of GDP on interest, who were fine, or at least not equivalently alarmed about Trump’s plan to increase military spending by 2% of GDP. If we make enemies in the world that require us to spend another 2 percentage points of GDP on the military, that is at least as bad as spending another 2 percentage points of GDP on interest on the debt. But only the latter will get the deficit hawks excited

In the same vein, our decision to ignore global warming is imposing high costs in dealing with wildfires and flooding. Most likely this is not close to 2% of GDP on an annual basis, but if the current path continues, we might be there before too long. In any case, if we are worried about the ongoing burden on the economy, the additional spending required to deal with the problems created by climate change is every bit as much a burden as interest on the debt.

Also, if we’re being serious, we have to recognize that spending money is not the only way the government pays for things. It also issues patent and copyright monopolies to support innovation and creative work. The annual sums transferred as a result of these government-granted monopolies are easily over $1 trillion a year. In the case of prescription drugs and other pharmaceuticals alone, it is over $500 billion.

The idea that we should worry about the debt and the interest we pay on it, but ignore the higher prices we pay for drugs, medical equipment, software, and other products because of the government-granted monopolies is something only a policy pundit could take seriously. If we’ve paid off the government debt, but a whole range of items cost twice as much as they would in a free market because we gave out patent monopolies lasting a century, would our kids have reason to thank us? In that case, they don’t have to worry about paying interest on the debt, just about paying massive patent rents on everything they buy.

Trump Corruption Threatens the Dollars, not Government Debt

For more than a century, investors, both foreign and domestic, felt comfortable having their money in the United States because they knew it was governed by the rule of law. This meant, for example, if there was a contract dispute, they could reasonably expect it would be decided in a court based on the merits, not based on political connections. The same is true for regulatory rulings, such as the Securities and Exchange Commission’s (SEC) ruling on corporate filings or the Federal Drug Administration’s (FDA) approval of drugs.

Under Trump, this is no longer true. If a company lied about its profits on its financial statements, investors could present evidence to the SEC and expect that it would impose sanctions if there was a solid case, but not under Trump. If the company is a big Trump contributor, it’s likely he will instruct the SEC to ignore the complaint or rule in favor of the perp.

The same applies with all the bureaucracies, as the Supreme Court has said Trump can tell them to do whatever he likes. The FDA can refuse to approve the drug of a company whose CEO has criticized him. The Federal Aviation Administration can refuse to recognize the safety of an airplane manufactured by a company that Trump dislikes.

The story goes on. Trump has a green light to use all parts of the government to punish companies and individuals that anger him and to favor those who hand him money.

This is an infinitely greater threat to the credit of the U.S. government and the strength of the dollar than the debt. That fact should be obvious to anyone who is not an elite pundit.

Every Democratic president of the last half century has left with a deficit that was lower, measured as a share of GDP, than the one they came in with, except Obama, who left it unchanged. By contrast, every Republican president has left with a considerably higher deficit than what they inherited.

Starting with Carter, the deficit for fiscal year 1976 was 4.1% of GDP. When he left office in 1980, it was down to 2.5% of GDP, despite a recession that year. That was the starting point for Reagan. (These figures refer to fiscal years, which end on October 1 of the year.) Reagan’s tax cuts, along with a big military buildup, were most of the story of higher deficits. When his successor, George H.W. Bush, left the White House in 1992, the deficit was 4.5% of GDP.

For better or worse, Clinton took deficit reduction seriously. He was helped by an explosion of tax revenue associated with the tech bubble, but he both made budget cuts and increased taxes. When he left office in 2000, the government was running a surplus equal to 2.3% of GDP.

George W. Bush quickly reversed the picture. A big part of the story was the collapse of the tech bubble in 2001-2002, which both led to a recession and a plunge in tax revenue from capital gains. He also had big tax cuts and a military buildup associated with his invasions of Afghanistan and Iraq. When Bush left office after 2008, he handed Obama a deficit equal to 3.1% of GDP, as well as a financial crisis and severe recession, resulting from the collapse of the housing bubble.

The deficit initially exploded in 2009 under Obama, as the country faced the worst recession since the Great Depression. As the economy gradually recovered, the deficit came down, falling back to 3.1% of GDP in 2016, just as Obama was leaving the White House.

Trump’s tax cuts caused the deficit to rise again. It hit 4.6% in 2019, but it really took off the following year, as a result of the pandemic. It reached 14.7% of GDP in 2020, the largest since World War II. The recovery and some modest increases in tax collections brought the deficit down to 6.3% of GDP in 2024.

Taking the cumulative changes from Democratic and Republican presidents, Democratic presidents have reduced deficits by 16.7 percentage points of GDP during their terms in office, while Republican presidents have raised them by 18.9 percentage points. Somehow, many people talk about Republicans as the party of fiscal responsibility.

As I said earlier, I am not hugely troubled by the debt. It would be better to be paying less money in interest, but 3.0% of GDP going to interest is not a disaster. The more important issue is to have a healthy economy with solid growth.

Here is where the big failure is. Trump’s war is leading to shortages, most importantly of oil, but also fertilizer and other products. His tariffs have led to higher prices for a wide range of products, as has his mass deportations. Perhaps most importantly, Trump’s open corruption and self-dealing undermine confidence in the U.S. financial markets and business system more generally.

In the past, investors could view the United States markets as relatively clean and stable. Unlike in some other countries, getting your investment back didn’t depend on staying in the good graces of the political leadership. Under Trump, this is no longer true. He has openly threatened companies and their management for saying and doing things he does not like. That is not a good recipe for a stable economy with solid growth.

If there is a run on the dollar, and interest rates soar higher, it is far more likely to be the result of Trump’s corruption and incompetence than the high debt. This is what people should be losing sleep over, not the debt crossing the $40 trillion mark.

The jump in rates had the predictable effect on the housing market. Starts in July fell to 1,239,000, the lowest level since the early days of the pandemic. The monthly data are erratic, but a three-month moving average tells the same story.

The monthly average for the period from May to July was 1,279,000, the lowest three-month average since May-July 2020, when the economy was largely shut down due to the pandemic.

The plunge in construction is bad news for those hoping for lower house prices. As I noted recently, real house prices have been drifting downward, following the pandemic surge. Affordability fans might have hoped that a more rapid pace of construction would accelerate this decline.

However, an uptick in construction does not seem a likely prospect for the near future. There is no obvious end to the war any time soon, and as long as the war continues, rates are more likely to go higher than lower.

The other big factor is immigration. A large share of the workers in construction are immigrants. Research shows that a reduction in the number of immigrants results in a decline in construction. This means that Donald Trump’s mass deportation drive is likely to slow construction even if the war ends and interest rates fall back to their pre-war level.

Therefore, we are not about to see robust construction numbers any time soon. Real house prices may still drift somewhat lower, as the pandemic surge fades into the distance, but we are not likely to see large price drops in the next couple of years, barring an economic collapse.

Trump has zero respect for the commitments of prior administrations or even treaty obligations. He doesn’t even care about his own commitments, as Canada and Mexico have discovered, as he freely ignores the trade agreement he negotiated in his first term. He does whatever he feels like.

For this reason, hopefully South Korea was not surprised by Trump’s decision to downsize the joint military exercises with South Korea, as he announced his renewed friendship with Kim Jong Un. For folks with poor memories, Trump has repeatedly touted the “love letter” that Kim sent him. He even stole the letter from the White House when he left office in 2021. (The letter is the property of the U.S. government, not Donald Trump.)

The story gets worse. Trump complained that South Korea refused to join the war that he and Israel started with Iran, even though he said he really didn’t need its help. Trump also again claimed the war was to keep Iran from having a nuclear weapon, even though no one in the world, except Trump, seems to think the country is anywhere close to having a nuclear weapon.

The irony of this gets even worse when we consider that Trump’s friend, Kim Jong Un, actually does have nuclear weapons, and likely has the means to deliver them to the United States. The Clinton administration had a multi-nation agreement designed to keep North Korea from getting nuclear weapons, which George W. Bush then abandoned.

There were efforts at a new agreement under Obama, but Trump quickly abandoned these after getting the love letter from Kim. This undoubtedly continues to be a great source of laughter in Pyongyang.

In addition to downgrading the military exercises, Trump also wants more money from South Korea for its defense assistance. On top of seriously overestimating the number of U.S. troops in the country (39,000 in Trump’s head, compared to 28,500 in reality), Trump apparently still has no understanding of the idea of collective defense.

If Trump wants to be able to project force against China, which may not be a good idea, but something he seems to want to do, then having an ally like South Korea is incredibly valuable. Just as the base in Bahrain, which Iran destroyed, was crucial for supporting Trump’s attack on Iran, the bases potentially provided by Korea would be of huge value in a military conflict. But this is a point that is likely too complicated for Donald Trump and his Fox News host Defense Secretary to understand.

Anyhow, the key takeaway for South Korea and the rest of the world is that the United States under Donald Trump cannot be counted on as an ally. And it would be very foolish for any country to view it as one.

Also, when it comes to trade, Trump’s commitments are absolutely meaningless. South Korea pledged $350 billion in investment in the United States in exchange for a modest reduction in the tariff rate that Trump imposed on imports from the country. Given the fact that Trump could raise his tariffs any time he feels like it, South Korea would be wise to keep its investment as a pledge. Only a fool would give something real in exchange for a Donald Trump promise.

Whatever tariff Trump ends up charging on our imports from South Korea is likely to have more to do with how he feels on a given day than any agreement between the two countries. As he continually tells the world, Trump does not care about his commitments. That’s the way he did business and the way he runs the country.

This matters, because at the end of the day, if the AI companies are not making money, the whole thing breaks down. To use a common analogy, suppose that steel companies are making huge bucks producing steel for rails, and construction companies are making money laying the rail, but the companies that run the railroads are all going broke. That doesn’t look like a story of long-term prosperity. In the great minds think alike category, Ed Zitron jumped on the same point in his excellent newsletter.

Anyhow, I take a somewhat different tack than Ed and focus on the Chinese competition. I realize that even if there was no competition from China, it is unlikely that AI would ever have the massive payoffs the hyperscalers are banking on, but the existence of that competition makes the story considerably less likely. And developments in the last couple of weeks seem to make the case for American AI even weaker.

Chinese AI Is Cheap and Getting Cheaper, U.S. AI Less So

As I have frequently noted here in the past, Chinese AI costs far less per input or output token than U.S. AI. For the cutting-edge models, the Chinese AI sells for one-fifth or even one-tenth of the price of U.S. AI. One response I have seen is that, even though the Chinese AI costs less per token, it can still end up being more costly because the systems are less efficient and require more tokens per task.

I am not sure that the measure of cost per task is a sufficiently standardized metric to allow it to be compared in a meaningful way, but insofar as it can be, it looks like the U.S. advantage has gone away. According to the Korean electronics industry publication, The Elec, the leading Chinese AI model is now cheaper on cost per task than the leading U.S. model, and performance gaps continue to narrow.

In the same vein, both Google and DeepSeek released new flash models last week. The DeepSeek model scored better on several benchmarks. And it sells for less than one-tenth the price.

If that makes the picture look bleak for U.S. AI producers, don’t worry, it will likely get worse. Alibaba reports having developed a modular design that will allow it to build data centers in 100 days. Compare to 12-18 months in the United States. This should mean lower costs and greater capacity for Chinese AI producers. That means the flood of low-cost, high-quality Chinese AI is likely to get even larger in the months ahead.

Chinese AI Is Finding New Customers

Given its huge cost advantage, it’s not surprising that Chinese AI models are gaining ground rapidly at the expense of U.S. models. I’ve noted before that Chinese AI seems to be winning out by large margins in most regions of the developing world. However, it also seems to be gaining ground in Europe. There are political considerations that could make European companies reluctant to rely on Chinese AI; however, given the erratic behavior of Donald Trump, it’s not clear that going with U.S. provides greater security.

And it looks like Chinese AI is continuing to gain ground in the U.S. market. It seems that Apple is looking to Chinese AI as a cheaper alternative to the Silicon Valley producers. Apple by itself is potentially a huge market, but perhaps more importantly, it is a company that has been at the cutting-edge of innovative technology for more than a quarter century. Its decision to go with Chinese AI is sending a serious message.

And remember, the question for those expecting really big bucks for the AI makers is not just whether Anthropic, OpenAI, and the rest can hang onto a large share of the market. It’s whether they can do so while selling at prices that give them the huge profits the stock market is banking on.

Can Creative Financing Overcome the Problems?

As mortgage issuers sold ever more dubious mortgages to further inflate the housing bubble, the wizards of Wall Street assured us that their financial magic would make it all work. This attitude was best conveyed by former Treasury Secretary Larry Summers at an academic conference in 2005, where he dubbed a critic of the growing house of cards as a “financial luddite.” Somehow, Summers thought innovative finance would make the millions of underwater mortgages issued to people with weak employment prospects and no reserve assets all work out fine.

We might be getting the same story with the AI bubble. Getting back to Torsten Slok’s point about the chipmakers making big bucks, while AI producers are making big losses, it seems Nvidia is looking to address the problem. It has just arranged $500 billion in financing from major banks for the hyperscalers that buy its chips. Fans of markets everywhere are asking the obvious question: if there is so much money to be made in building the data centers, why does Nvidia have to arrange the financing?

The details are not clear at this point, like whether Nvidia will in any way be on the hook for the financing, but there is a suggestion that it could involve securitization with tranches carrying different levels of risk, sort of like mortgage-backed securities or collateralized debt obligations. It could be lots of fun!

Most immediately, it will be harder for people to buy a home, as mortgage interest rates have risen by roughly the same amount. For someone looking to borrow $300,000 to buy a house, Trump’s war has cost them an extra $2,100 a year in mortgage payments.

It will also make it more expensive for state and local governments to pay for infrastructure or borrow for other needs. And corporate borrowers will have to pay more, notably the hyperscalers who are spending massive sums for building out their network of data centers. If they are carrying another $1 trillion in debt after a couple of years, a jump of 70 basis points will add $7 billion a year to their interest payments, in addition to the impact of any increase in the risk premiums they may face.

It’s worth noting that none of this increase in interest rates is due to an expectation of higher inflation. If anything, the rate of inflation expected now is slightly lower than it was when Trump started the war.

The rise in interest rates means that investors are expecting higher real interest rates now than they did six months ago. There are two stories as to why we might have expectations of higher real interest rates now.

The first is simply that there will be greater demand for borrowing. Part of this could be expectations on data centers are now higher than they were six months ago. That could be true to some extent, but investors were already anticipating massive spending on AI six months ago. It’s hard to imagine that expectations have increased that much.

The other reason expectations might have changed is that investors now anticipate considerably larger budget deficits than they did six months ago. Trump is requesting a $1.5 trillion military budget for fiscal 2027, which starts in October. That is up from $860 billion in the last Biden budget. This is an increase of more than $600 billion, which is nearly 2% of GDP, enough to have paid for extending the Biden ACA subsidies for twenty years. In other words, it’s real money.

I don’t find the demand for borrowing explanation entirely plausible, since the interest rate is not set directly by the supply and demand for savings. It’s set by the supply of money. And here there is a story that fits. The Federal Reserve Board is now projecting that the Federal Funds rate will be 0.5 percentage points higher in 2027 than they had before the war and 0.3 percentage points higher in 2028 than they did before the war. Assuming that the Fed’s open market committee members reflect attitudes of investors more generally, this could easily explain the jump in interest rates since the start of the war.

One item that is worth noting in this respect is that long-term interest rates on other countries’ debt have risen by a similar amount. While most have long-term rates that are lower than in the U.S., for example, Germany’s 10-year rate is 3.2%, and France’s is 4.1%, the increase has kept the gap roughly constant. This means that Trump’s war has raised interest rates for the whole world. Yet another reason for them to be thrilled with the United States under Donald Trump.

When Biden came into office, in January of 2021, gas was very cheap, selling for just under $2.40 a gallon. That was primarily because the economy was still feeling the effects of the pandemic. Unemployment was 6.4%, and we were down 6 million jobs from the pre-pandemic period.

The price rose quickly as the U.S. and world economy recovered, reaching $3.40 by February of 2022. Part of the reason for the rapid rise was that oil that had been taken out of production could not be instantly restarted once demand increased. Removing oil from production during the pandemic was a political decision taken in part by Donald Trump, in order to prevent oil producers from taking large losses.

Prices rose much further when Russia invaded Ukraine in February, and Biden responded by putting sanctions on Russian oil. The price peaked at over $5 a gallon in June but then fell back quickly as more oil came online. By the end of 2022, it had fallen to around $3.20 a gallon. It hovered around this level until Biden left office.

In January of 2025, when Donald Trump took office, the price was around $3.10 a gallon. It stayed around this level through most of the year, falling somewhat below $3.00 in December and bottoming out at $2.80 in January of 2026.

Then the decision to attack Iran sent gas prices soaring. They hit $4.00 a gallon by the end of March and peaked at over $4.50 a gallon in early May. Since then, they have bounced around in response to reports of peace agreements and new threats but have mostly been near $4.00 a gallon. (They are $4.08 today.)

Taking the averages to date, prices are still somewhat lower under Trump than under Biden: an average of $3.35 a gallon under Trump compared to $3.46 in the four years of the Biden administration.

However, this gap is being whittled away with gas over $4.00 a gallon due to the largely closed Strait of Hormuz. If gas prices are not brought down, Trump will soon be able to boast of having higher gas prices on average than Joe Biden.

As always, there are a few seeming anomalies. Prescription drug prices fell 0.8% in July and are down 3.1% over the last year. Nonetheless, people are spending about 2.0% more on drugs this year than last year. Computer prices jumped 3.5% in July. This is the data center story.

Rent and owners’ equivalent rent both rose 0.3% in July, somewhat faster than in prior months, but this is mostly due to rounding. Over the year, the indexes are up 2.9% and 3.2%, respectively. Food prices fell 0.1% in the month but are still up 2.7% year-over-year. Lettuce prices plunged 16.4%. Any ideas how that could have happened?

One real anomaly was a 0.3% drop in the car insurance index, leading to a year-over-year decline of 4.5%. This sort of drop is unprecedented outside of the pandemic. There were some modest declines in 1998 and 1999, but other than that, the index has always risen and typically far outpaced the rest of the CPI.

I have noted the falling car insurance index before and waited for it to turn around, but it has continued to be on a downward path since the start of the year. I’m betting for the insurance index to turn around and start rising again, but I have been making that same bet for many months. It accounts for 2.6% of the index, so it matters. It was a major contributor to inflation in 2022 and 2023 when there were double-digit increases.

But stepping back from the specifics, this is a bad story for the economy. Inflation is not about to soar out of control, assuming Trump doesn’t do anything too crazy, but it is outpacing wages. Over the last year, the average hourly wage increased 3.2%. The annualized rate over the last three months compared with the prior three was just 2.5%. This means that workers, who had already been feeling pressed, are falling further behind.

This is sort of good news from the standpoint of the Fed. It doesn’t have to worry about a wage-price spiral, but it does mean that we have an economy that will not be powered by workers’ consumption. With job growth having slowed to a crawl and real wages trending downward, workers will not have the means to increase consumption. This means that growth will be driven by AI investment and wealthy people spending based on stock gains and capital income. That does not look like a very solid basis for growth.

There is some truth to this story. On its current path, the Social Security Trust Fund will only be able to pay a bit less than 79% of scheduled benefits by 2033. The standard story puts the problem as too many baby boomers being supported by too few workers. As I pointed out last month, this seriously misrepresents the issue.

The problem of Social Security is one of income distribution, not demographics. The upward redistribution of income over the last half-century has hugely worsened the program’s finances.

There are three ways upward redistribution hurts the program:

1) A smaller share of wages is subject to the payroll tax;

2) Benefits are higher relative to wages;

3) It is difficult to ask workers to pay more to support the program when their wages are not rising.

More Wage Income Is Going Over the Tax Cap

In 1982, the last time the program was restructured in a major way, only 10% of wage income went over the cap on taxable wages, currently $184,500. Now, because so much income goes to high-end wage earners, such as CEOs, Wall Street types, and highly paid doctors and lawyers, 18% of wage income escapes taxation. If there had not been this massive upward redistribution, the current income of the program would be almost 10% higher, and the trust fund would be able to support full benefits much further into the future.

Wage Income Has Not Kept Pace with Productivity

In the 1980s and 1990s, inequality was primarily the result of income going from ordinary workers like retail clerks and assembly line workers to highly paid workers. Since the turn of the century, inequality has been primarily a story of more income going from labor to capital.

This affects Social Security because benefits are only indexed to inflation. If wages rise faster than inflation, then taxes, which rise in step with wages, rise relative to benefits.

To get an idea of the size of this impact, suppose the average beneficiary has collected benefits for 15 years. If wage growth was 1.0 percentage point faster over this period, then Social Security’s revenue would be 15% higher, while benefits would be the same. In this story, the higher revenue would be almost enough to close the projected shortfall, in addition to the trust fund being considerably larger today.

Workers Are Willing to Pay More for Their Retirement If Wages Are Rising

From 1960 to 1990, the Social Security tax rate rose from 6.0% to 12.4%, a 6.4 percentage point increase. The tax rate has not increased at all in the last 36 years. There was no massive revolt against the Social Security tax increases in the prior thirty-year period, both because people valued Social Security and, at least through the first part of this period, wages were rising rapidly.

If wages had kept pace with productivity growth, they would be more than 60 percent higher than they are today. In that scenario, it seems plausible that workers would be willing to pay 1-2 percentage points more in taxes for Social Security.

The Elites Broke Their Bargain on Wages: Don’t Let Them Steal Our Social Security

Social Security was set up as a program that would be supported by a tax on wages. Workers were collectively paying for their own retirement. This was totally feasible in a context where ordinary workers got their share of productivity growth.

It becomes less feasible in a context where wages are nearly stagnant, and most of the gains from productivity growth go to highly paid workers and capital, as has been the case for the last half-century. In that context, cutting Social Security benefits is just rubbing our noses in it.

We easily can afford to pay full benefits. The projected shortfall in 2033 is roughly 1.5% of GDP. That’s less than the increase in military spending that President Trump is requesting for 2027 compared with the last Biden budget in 2025.

In fact, as an economic matter, we are already paying most of this shortfall. When the government repays the bonds owned by the trust fund, it has exactly the same impact on the economy as if it was just paying the benefits directly, without the bonds held by the trust fund. The bonds are a promise, but they are not real economic resources. They don’t make it easier to pay benefits than would otherwise be the case.

Upward Redistribution Was Done to Us: It Didn’t Just Happen

The elites like to pretend that all the money going to Elon Musk, Mark Zuckerberg and the rest of the billionaire clique was just a happy coincidence. It’s a story of “look what the markets have done.”

That is 100% crap. The system was rigged to give this crew all of our money. The most obvious way the rich redistributed money to themselves was with government-granted patent and copyright monopolies. In the case of pharmaceuticals alone, this raises the cost by close to $600 billion a year or $4,500 per household. The money goes to the drug companies and their shareholders.

Larry Ellison and Bill Gates are among the richest people in the world because of the patents and copyrights the government gave them on the software produced by Oracle and Microsoft.

Patents and copyright monopolies come from the government, not the market. There are other mechanisms for financing innovation and creative work. In any case, we could have made these monopolies shorter and weaker, instead of longer and stronger as we did, which would have meant less money going to the rich and very rich.

The financial system is incredibly bloated by conscious choice. We could downsize it hugely by applying the same sort of sales tax on financial transactions that states impose on sales of shoes and clothes. That would mean many fewer fortunes there. We also don’t have to bail out banks when they put themselves into bankruptcy, as they did in the financial crisis and more recently when the crypto boys put the Silicon Valley Bank over the edge.

We also have written and enforced labor laws to the detriment of unions and workers. Most obviously, most states ban contracts that require all workers who are represented by a union to pay for that representation. While these contracts are not enforceable in most states, contracts that prevent workers from working for a competitor are enforceable. Also, unions are prohibited from supporting strikes as a union. This means that the steelworkers cannot refuse to produce steel for an auto company that is facing a strike by its workers.

These and other policies were designed to redistribute income upward. They have had their intended effect of taking money from the rest of us and giving it to the rich. And now that their upward redistribution has had the effect of undermining the financing of the country’s most important social program, they want to cut Social Security. It’s essential that people stand up to the lies. The problem is the rich taking too much of our money, not overly generous Social Security benefits.

Polling consistently shows that Social Security enjoys solid support across the political spectrum. (See “Support and Strengthen Social Security” in CEPR’s Majority Agenda.) The only way the elite will get their cuts to Social Security is if they get us to buy their lies. Towards this end, they will produce “bipartisan” plans where they find Democrats who will go along and legitimate the lies.

The basic point is incredibly simple. As a society, we have plenty of money to pay for Social Security. The rich have just taken so much from workers that it is difficult for them to pay the full tab out of their wages. The answer is to take back some of the money we have given the rich over the last half-century.