Weekly Global Economic Update

Week of August 11, 2026

• A weakening US job market
• Mexico depends on AI
• Poorer countries could benefit from AI

A weakening US job market

• The US government’s jobs report for July surprised many investors. The government reported a decline in payroll employment, the first drop since February. Moreover, the government also downwardly revised job growth for the previous two months. As such, it now appears that, for the past several months, there has been a steady deterioration in the labor market. Still, despite a decline in employment, the unemployment rate fell, as a large number of workers dropped out of the labor force. Finally, the report found that inflation-adjusted wages declined in July as wage growth failed to keep pace with inflation. Let’s look at the details.

The US government releases a monthly report on the labor market that encompasses two surveys: a survey of households and a survey of establishments. The establishment survey found that, in July, employment declined by 23,000 from the previous month. This followed downwardly revised growth of only 20,000 in June. In July, the decline was due, in part, to a 53,000 decline in government employment. That, in turn, was mostly due to a decline in employment at local schools. 

Excluding government, private sector employment grew by 30,000. This included gains of 22,000 in construction and 18,000 in durable goods manufacturing. Thus, aside from these two categories, private sector employment declined.

Several sectors saw a significant decline in employment. These included retailing (down 19,400), financial services (down 14,000), non-durable goods manufacturing (down 13,000), and leisure and hospitality (down 40,000). Within the last category, restaurants and bars were down 26,100. Meanwhile, employment was up only modestly in most other categories. The only category that saw significant growth was professional and business services (up 18,000).

Also, the establishment survey found that average hourly earnings of workers were up only 0.1% from June to July, the lowest increase since April 2025. Even on a year-ago basis, earnings were up only 3.2% in July. This was the lowest annual increase since May 2021, when the pandemic was receding. Moreover, the consumer price index was up 3.5% in June and will likely rise even faster when the July numbers are published. As such, wage growth is not keeping pace with inflation. This means that workers are losing purchasing power. Consumer spending has held up well, but that was largely due to a decline in the personal savings rate. Given that the savings rate is now historically low, it seems unlikely for spending to continue growing indefinitely. There are reasons to expect a slowdown in its growth.

A separate survey of households, which includes self-employment, found that the labor force declined by 264,000 from June to July. The labor force participation rate fell to 61.4%, the lowest since February 2021 during the pandemic. The survey also found that the number of people employed declined by 87,000. Consequently, the unemployment rate fell from 4.2% in June to 4.1% in July. These numbers suggest a significantly weak labor market, especially given the very low rate of participation. The participation rate measures the share of the above-16 population who is either working or seeking work. This includes retirees, too. Some of the decline may reflect slower net immigration, which shrinks the pool of the working-age population available to the labor force. It also means a lower unemployment rate, as supply of labor is falling faster than the demand for labor. Consequently, a low unemployment rate becomes somewhat misleading.

What does the jobs report mean for monetary policy? The Federal Reserve has a dual mandate to seek low inflation and high employment. Inflation has been seen as the primary challenge given that it has persistently been above the Fed’s 2% target. Yet, the weakening labor market could become cause of concern for the Fed. Plus, a weakening labor market will likely lead to less inflationary pressure. As such, today’s report likely reduces the probability of a rate hike anytime soon. Indeed, following the release of today’s report, the futures market’s implied probability of a rate hike in September fell from 55% yesterday to 41.9% today.

The United States faces AI-driven layoffs

• Job dismissals in the United States are down this summer, according to a survey by Challenger, Grey, and Christmas, an outplacement firm: “The pace of layoffs fell dramatically this summer.”

In July, there were 33,429 job cuts, down 46% from a year earlier. This was the lowest number of job dismissals in two years. For the first seven months of 2026, job reductions were down 41% from a year earlier. Recall that, in early 2025, there was a large number of job dismissals by the US government. 

Meanwhile, the technology sector accounted for 29.5% of all job reductions in July and 31% for the first seven months of 2026. Technology sector dismissals in the first seven months were 149,023, up 67% from a year earlier.

The industry that had the second largest number of dismissals was transportation with 41,748 dismissals in the first seven months, up 303% from a year earlier. According to Challenger, the industry has had to absorb rising costs and shifting trade patterns.

The US economy remains dependent on China

• Even after nearly a decade of US tariffs on China, potential tariffs, and other restrictions on trade, evidence suggests that the US dependence on Chinese goods has not significantly changed. This comes from a study conducted by the Peterson Institute. It found that, although US direct imports of Chinese goods has fallen significantly as a share of total imports, the Chinese share of the value added in US imports has barely changed. That “includes Chinese products, parts, and other content shipped from other countries.” The latter reflects the decision by suppliers to “reroute supply chains through third countries to avoid shipping directly from China to the US.” The countries that most benefited from this shift were Taiwan, Vietnam, and Mexico. Notably, the US government is seeking to limit the ability to transship through Mexico. 

This study demonstrates that, despite the intention of both the Biden and Trump administrations to de-couple the United States from China, it remains a challenge. In fact, the Peterson study predicts that the newest round of tariffs will fail at this de-coupling.

The challenge for the United States is that China has developed a massive capacity to produce and distribute some of the most important products in the world. While this might change in the future, especially if India continues to grow and move up the value chain, it is not likely to change anytime soon. Moreover, China’s role has been maintained by shifting supply chains to avoid direct economic interaction between the United States and China.

Mexico depends on AI

• Much of the discussion about the economic impact of investment in AI has centered on the United States and several countries in East Asia. However, Mexico has also played a big role in the rollout of AI. Specifically, Mexico has become the second largest supplier of servers to the United States after Taiwan. In May, Mexican deliveries of servers to the United States were actually greater than deliveries from Taiwan. The two countries combined account for more than 80% of servers imported into the United States. In part, Mexico’s role grew due to the investments in the country made by several Taiwanese companies. One notable result of this is that Mexican exports of servers now exceed its exports of automobiles. Indeed, servers and related equipment now account for about one-fifth of Mexican exports. All AI-related exports are roughly 33% of Mexico’s exports.

US imports of server-related equipment have reached about US$25 billion per month, up from about US$6 billion per month in early 2025. Plus, Mexican exports of automative data-processing equipment have nearly tripled in the past two years. Not only is AI-related investment fueling economic growth in the United States; it is also fueling growth for Mexico, Taiwan, South Korea, and Japan. For Mexico, which continues to grow at a very modest pace, it is likely that the economy would not be growing at all without the surge in AI-related exports. 

On the other hand, because the production of server technology is far less labor intensive than producing automobiles, the current surge is not having as large an impact on employment as would be the case if it involved other products. Also, server production mostly involves importing and then assembling parts. As such, it has been estimated that Mexican content in locally produced servers is only about 3% to 7% compared to 39% for locally produced automobiles. 

The US government recently chose not to extend the trade agreement with Mexico and Canada for another 16 years, especially while it talks to China. Rather, the trade agreement continues but remains in limbo. The United States is keen to avoid Chinese goods entering its territory through Mexico. Yet when it comes to servers for AI, the parts are coming from Taiwan. Moreover, if the United States wants to continue developing the AI industry, it may have no choice but to import those parts, at least in the short run. Thus, it is not likely that the United States will be averse to imports of Taiwanese servers made in Mexico. Meanwhile, the surge in Mexico’s AI-related exports comes at a time when exports of automobiles have been faltering, in part because of trade tensions with the United States.

Poorer countries could benefit from AI

• AI might be less disruptive to labor markets in poorer countries than in affluent countries, according to a study by the World Bank. It estimated that 16.2% of jobs in developing countries could “see productivity meaningfully boosted by AI.” In addition, an estimated 18.7% of jobs in high-income countries will see such gains. 

The chief economist of the World Bank says: “AI has thrown developing economies a lifeline, and they should seize it. By adapting small, low-cost AI tools to local conditions, they can bring better medical care, education, judicial services and agricultural extension within reach of millions.” 

This report is welcome news given that, on average, developing countries are now experiencing relatively poor economic growth. AI evidently provides an opportunity to boost growth later this decade.  However, it is not guaranteed. As the World Bank noted, “the most advanced AI systems are being built by a small number of countries and companies, while many developing economies still lack the power, internet access, data, skills, and institutions needed to use AI effectively.” As such, the World Bank offered some policy suggestions for developing countries. 

Finally, the World Bank said: “There is a huge upside for doing things that would otherwise have taken decades, maybe even a century.” For example, it talked about how better weather forecasting could dramatically improve food output, thereby boosting agricultural productivity and freeing up workers to perform other tasks. In other words, the potential is vast.

Week of August 4, 2026

• Rising cost of insuring tech-company debt
• The US Federal Reserve waits
• US economic growth decelerates while consumer and business spending are strong

Rising cost of insuring tech-company debt

• Remember credit default swaps? They had played a role in the near collapse of the financial system in 2008, when there was a sharp rise in defaults on mortgage-backed securities. Now CDSs are back in the news, and this time, it has to do with technology companies.

First, a quick primer: When investors purchase a security—such as a corporate bond, a government bond, or a collateralized debt obligation—they often want to purchase insurance against the possibility of default. They purchase such kinds of insurance in the form of CDSs, which are derivative products usually sold by financial institutions. The price of a CDS reflects the perceived risk of default.

Currently, the notional value of the market for CDSs issued by a single debtor is about US$9 trillion. CDSs are quoted in the form of a spread priced in basis points. Currently, an index of CDSs for investment-grade corporate bonds is trading at 53 basis points.

Lately, the prices of CDSs issued by many tech companies have risen sharply. Tech companies, many of which are flush with historically high levels of cash, are investing so much in artificial intelligence that they’ve chosen to go to the bond market for financing.  As this has taken place, and as perceived risks have increased, the cost of these CDSs has also risen.

There are several potential explanations for the rise in CDS prices. First, there is increasing concern about tech companies’ ability to generate sufficient cash to cover debt-servicing costs. A major tech company reported its first quarter of negative free cash flow since it went public two decades ago. Moreover, bond yields have risen, and further tightening of monetary policy could push them higher, increasing the cost of servicing debts.

Meanwhile, the sharp decline in the equity prices of semiconductor companies could be a signal that investors are concerned about a sharp slowdown in the buildout of AI capacity, which, in turn, could reflect concerns about excess capacity.

Second, there is increasing concern in the United States about competition from AI companies in China. The selloff of tech shares this week was, in part, attributed to concerns about the rise of Chinese AI companies. Many can offer good-quality AI services at relatively low prices. The challenge for US-based AI companies is that the massive investments they are making as first-movers could be undermined by cheaper second-movers—in this case, Chinese companies.

Third, Nikkei Asia reported that the volume of off–balance sheet debt incurred by big tech companies is now significantly large: Off–balance sheet debt had quadrupled in the past four years at five major US-based tech companies, hitting US$1.65 trillion. This is greater than the volume of debt appearing on their balance sheets.

Finally, there is increasing concern about so-called “circular financing.” An example would be where a semiconductor company provides funding to an AI company to build tech capacity. In return, the AI company purchases the semiconductor company’s chips.  The main concern with this arrangement is that, if the AI company is unable to generate strong cash flow, it becomes not only a concern for the AI company but also for the semiconductor company. This is reminiscent of the dot-com bubble 26 years ago when telecom companies invested in internet companies that were buying telecom equipment.  When the internet companies had issues, so did the telecom companies.

• Meanwhile, the concern about a potential AI bubble stems from the possibility of overinvestment. That is, if companies are building more capacity than can be profitably used over a reasonable period of time, then challenges can crop up, including a reversal of investment spending, a drop in equity prices, and even defaults on bonds. In a recent report, the Bank for International Settlements, which advises the world’s leading central banks, described the concern as follows:

“Technological breakthroughs are typically accompanied by investment booms and buoyant macroeconomic activity. Exuberance about the promise of new technologies intensifies competition among firms eager to capture a share of the revenues. The race to get ahead can result in excessive investment that makes the boom unsustainable and prone to a disruptive ending. This fragility is further aggravated by the leverage that accompanies the rapid ramp-up of investment. This boom-bust pattern recurs across history, from the US canal mania in the 1830s and the British railway mania in the 1840s, to the roaring ’20s, and the dot-com boom in the late ’90s. These episodes all ended in sharp corrections, with wider economic fallout.”

The Bank for International Settlements goes on to note the massive scale of investment taking place. It said that “the potential demand for AI services is clearly vast and could justify a substantial expansion in computational power. Yet, relative to its pre-boom trough, the current build-out is on track to outgrow every previous episode only three years in.” It also said that the huge increase in leverage to finance the buildout, along with a lot of circular financing, has increased the risk of potential troubles on the road ahead.

To better understand what could happen, the Bank for International Settlements developed a simple theoretical model, in which, numerous AI firms compete in a “winner take most” environment. With most of the reward from the investment accruing to a small number of players, the end result is excess capacity. Ultimately, “an AI boom creates fragility that undermines itself. The more capacity the sector builds, the higher the productivity bar it must clear to sustain the boom, so a larger boom is both more likely to disappoint and more damaging when it does.” The Bank for International Settlements concluded that “the larger the boom, the deeper the eventual bust. The race to commit early through debt and circular financing also makes a bust more likely.”

There are three important things to note about this analysis. First, it does not imply that a debilitating correction is likely. It simply implies that such an outcome is a realistic possibility. Second, even if a correction is likely, it is impossible to estimate the timing.  That is, in past episodes, naysayers accurately predicted doom only to find that the doom came much later than anticipated. A boom can go on for a prolonged period before trouble emerges. Third, even if a correction comes, it does not imply that the investments made were foolhardy. It simply implies that the path toward a revolutionary change is not a straight line. There was a sharp correction during the dot-com bubble.  It did not mean that investment in the then burgeoning internet was wrong. After all, the internet eventually changed everything—which will probably be true for AI, as well.  But along the path to that idyllic future, there could be tears.

The US Federal Reserve waits

• In recent years, it has usually been the case that announcements by the Federal Reserve on interest-rate policy were not a surprise. Usually, the direction of monetary policy was widely anticipated, often having been signaled by the Fed itself. This time is different, however. Prior to the announcement by the US Fed last week, the futures market was signaling an implied probability of 33% that the Fed would raise its benchmark rate that day, and a probability of 67% that the Fed would leave the rate unchanged. That indicates there was significant uncertainty in the minds of investors, which likely reflects the fact that Fed Chair Warsh did not previously offer a hint vis-a-vis his thinking.

On July 29, 2026, the Federal Reserve’s Open Market Committee announced that the benchmark interest rate will remain unchanged. The 12-member committee voted 9-to-3 to keep the rate unchanged, with three members voting to boost the rate by 25 basis points. The committee commented that “economic activity is expanding at a solid pace despite elevated uncertainty that owes, in part, to the conflict in the Middle East. Productivity growth and capital investment are strong. Job gains have kept pace with the workforce, and the unemployment rate has changed little. Inflation remains elevated relative to the committee’s 2% goal, in part reflecting supply shocks that have driven price increases in certain sectors, including energy. The committee will deliver price stability.”

In his press conference, Fed Chair Kevin Warsh reiterated that the Fed intends to maintain the 2% target for inflation, which he deemed the definition of price stability.  Moreover, he said he wants to anchor investor expectations, that is, he wants to convince investors that the Fed intends to maintain the 2% target. Warsh also said that, if inflation remains elevated, interest rates could be “part of the equation” for the Fed in targeting lower inflation. However, Warsh was repeatedly asked by reporters why the Fed was not simply acting now to raise interest rates. His answer was that such impatience was not warranted.

Meanwhile, investors evidently now expect the Fed to raise the benchmark rate at its next meeting in September. The futures market’s implied probability of a rate hike in September is now 68%. Plus, there is an implied probability of 89% that the Fed will raise the benchmark rate at least once before the end of the year, as well as a roughly 48% likelihood of two or more hikes before the end of the year.

The expectation that the Fed will tighten monetary policy this year reflects concern that oil prices are likely to remain elevated or even rise further, mainly due to the risk of continued conflict in the Middle East. Following an Iranian attack on US facilities this week, the United States and Saudi Arabia attacked Iran-allied facilities in Iraq. These actions led to a sharp rebound in the price of oil.

Following Warsh’s press conference, yields began to soar, with the yield on the 30-year bond hitting the highest level since 2007. Plus, the gap between the yields on 30-year and two-year bonds shot up by almost 20 basis points.

Why did this happen? First, many investors were hoping that the Fed would increase the benchmark rate and were disappointed, especially given current inflationary pressures.  Moreover, in his press conference, Warsh indicated that inflation remains too high. Yet, when asked several times why the Fed did not raise rates immediately, his answers may not have been satisfactory to many investors. He largely indicated that there is no need for impatience.

Second, Warsh did not indicate an intention to raise rates. When discussing how the Fed would respond if inflation remains elevated, he said that interest rates were one “part of the equation,” suggesting that there might be other inflation-fighting tools—although he did not say what those tools might be. Moreover, he said that there could be indicators other than the personal consumption expenditures deflator for measuring inflation, yet he did not indicate what those indicators might be. Finally, he suggested that, by boosting yields, the market is already tightening monetary policy, thereby suggesting that the Fed might not need to do anything.

Third, he reiterated his opposition to forward guidance. And, unlike in the past, there were no dot plots. Thus, investors have no guidance as to the Fed’s future intentions, thereby creating a perception of greater risk.

Normally, a Fed meeting that leaves the benchmark interest rate unchanged means no news and no significant movement in asset prices. But this time, the situation was different. Although the benchmark rate remained unchanged, asset prices moved a lot, as comments from the chair evidently created uncertainty.

US economic growth decelerates while consumer and business spending are strong

• The US economy decelerated in the second quarter, although underlying growth in consumer spending and business investment remained strong. These were offset by a sharp increase in imports and a sharp decline in inventories. The strength of consumer spending reflected a continuing decline in the rate of personal savings—something that cannot be sustained indefinitely. The strength in business investment was largely due to the continuation of massive AI-related investments. Let’s look at the details:.

In the second quarter, real (inflation-adjusted) gross domestic product was up at an annualized rate of 1.5% from the previous quarter, down from 2.1% in the first quarter.  Real consumer spending grew at a very strong rate of 3.2%, accounting for more than 100% of GDP growth. This was partly offset by a strong 14.7% rise in imports.

Households appear to be keen to maintain a high level of spending despite the sharp rise in the price of gasoline. Consequently, the personal savings rate (share of disposable income that is saved) fell from 3.5% in March to 2.7% in June—the lowest rate in four years and one of the lowest rates on record. This probably cannot continue indefinitely, in which case, the pace of spending growth should weaken in the months to come absent offsetting factors. Meanwhile, spending growth was largely fueled by demand for durable goods.

Meanwhile, real nonresidential fixed investment grew at a rate of 8.4% in the second quarter. Notably, growth was strong despite investment in structures falling at a rate of 5%. Investment in equipment grew at a rate of 15.2%, while investment in intellectual property was up 8.8%. These numbers were likely fueled by AI. Investment in AI largely involved investment in information technology equipment as well as software. These two modest categories accounted for 46% of real GDP growth in the second quarter. These categories had accounted for 129% of GDP growth in the first quarter. In other words, absent investment in AI, the economy would barely have grown in the first half of the year.