New bond king Gundlach warns: If the Federal Reserve holds rates steady, long-term interest rates will rise sharply.
On the eve of the Federal Reserve's interest rate meeting, "New Bond King" Jeffrey Gundlach issued three warnings: if the Fed does not raise interest rates, long-term interest rates will soar; inflation trends are strikingly similar to those of the 1970s; AI bond spreads have nearly doubled, while US stock valuations are so high that they are "like a hotel minibar—there are no bargains."
On the eve of the Federal Reserve's interest rate decision next Wednesday (FOMC meeting day), Jeffrey Gundlach, CEO and Chief Investment Officer of DoubleLine Capital, systematically reviewed the current macroeconomic landscape in the latest episode of his "Gundlach Unlocked" online live broadcast, and issued a series of warnings on interest rates, inflation, credit markets, US stock market concentration, and the trend of the US dollar.
Gundlach stated frankly on the program that he is skeptical about the Federal Reserve raising interest rates next week, even though the market has priced in a probability of about 60%. He clearly stated: "I wouldn't be surprised if the Fed doesn't raise rates next week. If that happens, I expect long-term rates to rise quite significantly thereafter." Conversely, if the Fed does raise rates, the bond market may remain unchanged.
Behind this assessment lies his deep concern about the current inflation situation and his continued vigilance regarding the trajectory of US fiscal policy. He uses a wealth of detailed data to point out that against the backdrop of persistently high inflation and a severely out-of-control US fiscal deficit, the US Treasury market, US stock valuations, and the booming AI credit bond market are all at an extremely fragile historical juncture.
If the Federal Reserve holds rates steady, long-term interest rates could face significant volatility.
Regarding next week's FOMC meeting, Gundlach believes that the Fed's pace has once again become disconnected from bond market pricing. He points out that based on current pricing at the short end of the yield curve, the market perceives a roughly 60% probability of a Fed rate hike, but he remains skeptical.
“I wouldn’t be surprised if the Fed doesn’t raise rates next week. But if they do, I expect a fairly significant rise in long-term rates after the Fed meeting. If they do raise rates, the bond market will likely remain at its current level,” Gundlach said.
He presented his benchmark model for the 10-year US Treasury yield (based on the 7-year average of the German 10-year bond yield and US nominal GDP). Data shows that the model currently indicates the 10-year US Treasury yield should be around 4.71%, while the actual yield is 4.78%, suggesting the yield is within a reasonable range. However, he emphasized that after a massive 500 basis point rate hike, the market has not seen a substantial rebound, and " the path of least resistance is likely to continue rising. "
Inflation is far from over; the CPI trajectory bears a striking resemblance to that of the 1970s.
Gundlach ruthlessly criticized the market's optimism about cooling inflation. He pointed out that both core PCE and overall PCE had higher annualized growth rates over six months than over twelve months, indicating that "inflation has not really improved and is still far from the 2% target."
Even more worrying is the possibility of history repeating itself. Gundlach overlaid the inflation trajectory since 2014 with the period of soaring inflation from the 1960s to the early 1980s in his charts. He warned:
" The current trajectory is strikingly similar to the inflation disaster that occurred before and during Volcker's tenure. It will be very interesting to see if we continue to repeat the trajectory of that inflation disaster."
Regarding data mining, Gundlach emphasized his most valued unadjusted, non-seasonal indicator—the import and export price index. Currently, US export prices are up 8.25% year-on-year, while import prices are up 5.95% year-on-year.
“If you average them out, based on this purest measure of inflation, the actual inflation rate is around 7%. ” Coupled with Brent crude oil near $100 a barrel and global crude oil inventories at historically low levels, he believes that the bottom support of oil prices will make inflation more stubborn than the Federal Reserve would like.
In addition, he cited multiple signs of inflationary pressures:
- The Bloomberg Commodity Index has risen 34% since the start of the war and has recently rebounded from its 200-day moving average, approaching its highest level in more than 10 years.
- Residential electricity prices have risen from 12.5 cents per kilowatt-hour about eight years ago to 18 cents, an increase of more than 50% , and the upward trend shows no signs of slowing down.
- Brent crude oil is approaching $100 a barrel;
- Strategic petroleum reserves fell from a peak of 750 million barrels to 287 million barrels, a drop of more than 50% , marking the lowest level since the reserve was established; global oil inventories are also at their lowest point since 2018—"This will continue to provide a floor for oil prices, making inflation stickier than the Fed would like."
Beware of secondary risks: AI corporate bond spreads widen dramatically.
In the credit bond market, Gundlach astutely observed a significant divergence that had been overlooked by the market: AI-related corporate bonds were facing substantial selling pressure.
Data shows that while the spreads in the broader investment-grade bond market have not changed substantially, the spreads in the AI sector have surged from 50 basis points to about 125 basis points, widening by 75 basis points. The situation is even more dramatic in the high-yield (junk bond) sector, where the spreads in AI-related sectors have jumped from about 180 basis points to around 325 basis points.
“This is a huge divergence we should really be paying attention to,” Gundlach stated bluntly. “Given the staggering demand for AI and AI-related businesses, this will undoubtedly put further spread pressure on the AI sector… The market is clearly struggling to absorb such a massive supply, and the supply of AI sector bonds will continue to be an avalanche. ”
US stock valuations are nearing extremes; "an extremely concentrated market means extreme danger."
Regarding the stock market, Gundlach issued his most severe bearish warning. He pointed out that the current Shiller PE ratio of the S&P 500 is as high as 42, a figure even higher than that of the bubble period before the Great Depression in 1929.
"At this price-to-earnings ratio level, the real return over the next 10 years has never been positive. In fact, it is often a significant negative real return, ranging from -5% to -9% annually. Therefore, broadly buying the market capitalization-weighted S&P 500 at this Shiller price-to-earnings ratio level means facing huge real losses. "
He specifically pointed out the "abnormal" boom in technology stocks. Currently, the information technology sector accounts for a record 38% of the S&P 500, far exceeding the concentration during the dot-com bubble of 1999 and before the 2008 financial crisis.
" This is an extremely concentrated market, which means it's an extremely dangerous market. Therefore, I would not recommend any market capitalization-weighted stocks. "
US Dollar and Emerging Markets: Bearish on the US dollar, bullish on emerging market equities and local currency bonds.
Gundlach holds a clear bearish view on the US dollar. The US Dollar Index (DXY) has fallen below 100 from its high of 110 at the end of 2024. "The dollar has hardly made any meaningful moves in the past year or so, and it looks almost like it has been manipulated," but he expects the dollar to continue to weaken.
He cited historical data to demonstrate a strong correlation between a weaker dollar and emerging market (EM) assets outperforming US assets. In the comparison chart he presented, the dollar trade-weighted index and the S&P 500 relative to the EM index showed a highly consistent trend—"If the dollar falls, the S&P 500 is likely to underperform emerging markets."
Since the end of 2024, the S&P 500 has underperformed emerging markets by approximately 20% .
He is also optimistic about the performance of emerging market local currency bonds relative to US corporate bonds, a logic also based on the expectation of a weaker dollar.
US Debt and Fiscal Policy: With $40 trillion in debt looming, long-term TIPS offer no protection.
Gundlach points out that the total U.S. public debt has reached $40 trillion (including the portion held by the Federal Reserve and the Social Security system), and based on the current trajectory, it may exceed $50 trillion by 2032 .
Meanwhile, CBO forecasts indicate that the fiscal deficit as a percentage of GDP will continue to expand, and current forecasts are based on relatively optimistic assumptions such as "interest rates below current levels, deficit ratios below current levels, and continued positive real GDP growth"—"once you put pressure on these assumptions, it is clear that we are on the path to a deficit of 7% to 8% of GDP within less than a decade."
He also specifically clarified a market misconception: some believe that those who dislike nominal government bonds should turn to long-term TIPS as a hedge. Gundlach explicitly disagrees:
"The yields on long-term TIPS (30-year) and 30-year nominal Treasury bonds have moved in exactly the same direction since the end of 2021, with the difference between them remaining virtually unchanged over the past five years. If you don't like nominal long-term Treasury bonds, there's no reason to believe that 30-year TIPS will protect you. Don't buy long-term TIPS thinking it will hedge your nominal Treasury bond risk."
He also expressed skepticism about the Treasury's recently announced bond buyback program, believing it was unlikely to have a meaningful impact on long-term Treasury yields.
The full text of the latest episode of DoubleLine Capital CEO and CIO Jeffrey Gundlach's "Gundlach Unlocked" webcast is as follows (translated with AI assistance).
Thank you for participating. This is the third installment of our "Gundlach Unlocked" webinar, where I'll be sharing some macro topics and occasionally touching on micro-level issues. Interestingly, over the past three months, the S&P 500 has actually outperformed the NASDAQ . If you were to build a 60/40 portfolio using the NASDAQ instead of the S&P 500, the return would only be around 8%.
Let's begin. The screen displays the worst-to-worst yield of the Bloomberg Aggregate Bond Index , with data dating back to the late 1990s. We can see that, starting about four years ago, the index's yield has been trading in a sideways range : around 4% at the low end and around 5% at the high end, except for a brief breakout in 2024.
The chart has several horizontal dashed lines representing average returns : - Blue line : 4.03% average over the past 30 years - 3.25% average over the past 20 years - Interestingly, the average over the past 10 years is slightly higher than the average over the previous 10 years.
Therefore, current yields can no longer be described as "suppressed." The Bloomberg Composite Index reflects real interest rates, which is certainly a good thing. Many of our funds are even trading at a premium to this level, with many currently yielding 6% . If you choose higher-risk fixed-income products, such as emerging market bonds denominated in local currencies or bank loan indices , the yield is around 7% .
This is quite competitive for the stock market—as we will see later, the Shiller cyclically adjusted price-to-earnings ratio (Shiller CAPE) is basically at an all-time high .
We are currently in an environment of rising interest rates, a trend that has lasted for six years and is about to enter its seventh. We can see that, with the exception of Switzerland, interest rates in almost all countries are rising in tandem. Most notably, Japan 's interest rates, which had been suppressed to near zero for many years, have now risen to 3.97% , less than 150 basis points behind the 5.24% yield on the 30-year US Treasury bond .
Interest rates in all these developed countries are rising in tandem, with the UK experiencing the most significant increase .
As we can see, the 30-year US Treasury yield bottomed out in 2020 , which was also the bottom of this upward channel. The red line in the chart represents the position two standard deviations away from the center line. From the incredible historical low of 27 basis points in 2020, it has risen all the way to today's 5.24% , meaning the price of 30-year Treasury bonds has fallen by more than 50% to date , and we are still near the high point, around 5.25% .
When the market trades sideways for an extended period without rebounding, it often signifies one thing: we've experienced a sharp surge in bond yields, but prices have barely corrected. Typically, if the market fails to rebound to correct a nearly 500-basis-point increase in interest rates, it likely means the next move will be a continuation of the upward trend.
I developed this model a long time ago to provide a benchmark for the 10-year US Treasury yield . In the chart , the brown-yellow line represents the real yield of the 10-year Treasury bond, the dark line is the model fit value, and the yellow line is the model's forward forecast value.
The model is constructed by taking the yield on 10-year German government bonds and combining it with the 7-year average of US nominal GDP . This combination surprisingly provides a reliable reference point for the yield on 10-year US Treasury bonds. Note the box at the bottom of the chart— the R² (goodness of fit) for these two lines is an astonishing 0.93 . If the calculation starts from 1990 instead of going back to 1986, the R² would be even higher.
The current model shows that the expected reasonable yield for the 10-year US Treasury bond is 4.71%, while the actual yield is exactly 4.78% , which is very close to the model's prediction range. However, the path of least resistance still seems to be upward , and we will explain why later.
I often talk about the relationship between the two-year Treasury yield and the Federal Reserve , and they are now showing some degree of "asynchrony" again. In 2022 , we experienced a very significant asynchrony—the two-year Treasury yield was well above the Fed's near-zero interest rate, exceeding the federal funds rate by 200 basis points . That was the biggest gap I've seen in my 42-year career .
Then we saw that in 2025 the Federal Reserve clearly veered to the other extreme (deviated from its course). And now, judging from the current position of the two-year Treasury yield, the federal funds rate looks set to be about 50 basis points higher than it is now . The Federal Reserve's policy meeting will be held next Wednesday; we'll wait and see.
The "warp function," which measures the probability of the Federal Reserve adjusting interest rates—and is determined by the shape of the yield curve—shows that, based on the pricing of short-term Treasury bonds, the probability of the Fed raising interest rates is about 60% .
However, there are some things about Kevin Warsh that I don't completely trust, so I tend to disagree with the 60% probability , although I'm not entirely certain about it.
I wouldn't be surprised if the Fed doesn't raise rates next week.
If that's the case, I expect long-term interest rates to rise quite significantly after the Fed meeting ; while if the Fed does raise rates, the bond market will likely remain around its current level.
The next chart, which I also used in my last online live stream , is from JP Morgan Asset Management :The vertical axis (Y-axis) represents the ISM Manufacturing Prices Paid index . When "prices paid" rise, people naturally expect the Federal Reserve to be more inclined to raise interest rates than to lower them.The horizontal axis (X-axis) represents the ISM Manufacturing Employment Index. When the index is above 50 , people expect the Federal Reserve to be more likely to raise interest rates ; when it is below 50 , it is more likely to cut interest rates .
The chart is dotted with many small dots: blue dots represent the Federal Reserve cutting interest rates (easing) , and orange-red dots represent the Federal Reserve raising interest rates (tightening) .
Based on the original diagram of JPMorgan Asset Management, I drew a few rectangles and added some information :
In the rectangle at the bottom left , almost all the dots are blue, with only three or four exceptions. I pointed to those three or four dots with an arrow— those were Paul Volcker's actions in early 1982. At that time, he completely deviated from the bond market's lead, instead taking proactive , sometimes even impulsive, actions , suddenly announcing interest rate adjustments before meetings were held. The most famous example was one Saturday night when he raised interest rates by hundreds of basis points in one fell swoop , the infamous "Saturday Night Massacre."
The rectangle in the upper right corner presents a different picture—in this range, one should expect to see more tightening , given the high price-paid index (indicating inflation) and the high employment index. Both aspects of the Fed's dual mandate point to tightening monetary policy , and the chart is indeed almost entirely dominated by orange-red dots indicating tightening, with only about four blue exceptions . Those exceptions occurred during Arthur Burns' term, when he was pressured by the then-president to keep interest rates artificially low . This, of course, largely contributed to the US entering an era of high inflation . We will see related charts later.
The chart shows a large orange dot located above the 70 line on the vertical axis and to the right of the 50 line on the horizontal axis. This suggests, to some extent, that if the Federal Reserve were to take action, it should be tightening interest rates, not easing them .
However, if you observe all the smaller dots around that large orange dot, you'll find some are red and some are blue ; there's no definitive conclusion. For this specific stage, while there's no definitive answer, I believe there are slightly more compact dots than looser blue dots .
We are now beginning to see some changes in bond spreads. On the left side of the chart, the light blue line represents the spreads in the US corporate investment-grade bond market excluding the AI sector , while the dark line represents the spreads only for the AI sector. One thing is quite clear: there hasn't been any significant widening of spreads in the broader investment-grade bond market, but the AI market has seen a substantial widening of spreads relative to the investment-grade sector . We see the AI sector spreads widen from 50 basis points to approximately 125 basis points, an increase of 75 basis points over the period , while investment-grade spreads have remained unchanged.
On the right side of the chart, we performed the same analysis on high-yield bonds, and the situation became even more dramatic— the spread for AI-related bonds widened significantly, from approximately 180 basis points to approximately 325 basis points . Meanwhile, for high-yield bonds outside the AI sector, the light blue line is actually close to this year's historical low. Therefore, we are seeing a huge divergence, which is what we really need to pay attention to.
As everyone knows, the US Treasury is borrowing heavily with a fiscal deficit of 6% to 7% of GDP . Now, AI and AI-related businesses are generating substantial financing needs, which will undoubtedly put further pressure on the yield spreads of the AI sector. I'm really not sure who is buying these AI-related bonds. Perhaps it's insurance companies held by private lending firms, which in turn are held by private equity firms, which are then controlling the investment behavior of their subsidiaries. But the market clearly can't absorb such a large supply , and the supply in the AI sector will continue to surge like an avalanche.
So what we are facing is this: the Ministry of Finance is borrowing too much, and companies seem to have an endless need to issue bonds. The market—as can be clearly seen from the dark line in the chart— is starting to demand higher returns .
I often hear comparisons between Treasury Inflation-Protected Securities (TIPS) and nominal bonds, and we like TIPS, holding them in some low-risk funds. We prefer short-term TIPS because we believe the inflation expectations implied by comparing nominal bonds to TIPS are too low. They essentially imply that the Fed will immediately reach its 2% target and maintain it there, which I think is highly unlikely, so I believe short-term TIPS are undervalued.
What I'm showing on my screen now is long-term TIPS, specifically a comparison between 30-year TIPS and 30-year nominal Treasury bonds . Many people say they like TIPS; I've even seen some frequent guests on financial media mention that they now prefer long-term TIPS because they are not optimistic about nominal long-term interest rates due to the large scale of Treasury borrowing. But it's obvious that the two lines are very similar. Just look at the bottom of the chart, the difference between the two, and you'll find that this difference has been completely stable over the past five years .
Therefore, TIPS cannot hedge your risk . If you are bearish on nominal Treasury bonds, there is no reason to believe that 30-year TIPS will protect you, as its interest rate has risen in exactly the same direction as nominal bonds since the end of 2021. Do not buy long-term TIPS thinking it will somehow hedge your risk—if you are bearish on 30-year nominal Treasury bonds.
Now let's look at inflation. Kevin Walsh made it clear at his last press conference that 2% is their target and they will achieve it . He pledged to use the PCE deflator to measure inflation. He cited the 12-month PCE deflator, correctly stating that the figure is 3.7% . He further pointed out that the 6-month annualized rate of change of the PCE deflator is actually higher , meaning that the increase in the past six months has actually been faster than in the previous six months. So the PCE deflator hasn't really improved much. The 6-month annualized rate of core PCE is higher than the 12-month annualized rate, and both are far from 2%.
Let's look at the year-over-year data, including core and overall indicators. Core inflation is at 3.3%, and overall inflation is at 3.7% . Both appear to be on an upward trend since mid-2024, although the rise has stalled in recent reports. What the next inflation data will bring will be very important to watch, as I believe it will significantly influence the direction of the Federal Reserve's future policy.
The next chart is more for entertainment purposes. We've overlaid the inflationary experiences (measured by overall CPI) from the 1960s to the early 1980s. In the 1970s and 80s, the CPI rose to 12.5% , then broke through further in the early 1980s, approaching 15% . Then we have the recent round of interest rate hikes from January 2014 to 2026. Surprisingly, the blue line (the recent experience) and the red line (the experience before and after the Volcker era) are strikingly similar in shape . At least the blue line has now turned downwards, but whether we will continue to repeat the trajectory of that inflationary disaster will be very much worth observing.
As everyone knows, my favorite inflation gauge is the import and export price index because it's unadjusted, not seasonally adjusted, and simply reflects pure price data. Currently, export prices are up 8.25% year-on-year, and import prices are up 5.95% year-on-year, both at fairly high levels. Averaging them, we get about 7% . So, based on this purest measure of inflation, inflation is actually running at around 7%. No wonder consumer confidence is so low.
This is the Bloomberg Commodity Index, which experienced a mid-year correction before rebounding, right from its 200-day moving average (red line) . It currently appears to be breaking through highs of the past 10 years or even longer. Regarding inflation, let's look at electricity retail prices for residential users. I'm not particularly focused on year-over-year data; I'm just looking at this dark line, which represents the price per kilowatt-hour in cents. About eight years ago it was 12.5 cents ; now it's 18 cents, a 50% increase . And this trend doesn't seem to be slowing down, which is another reason for the dampened consumer confidence and why the current officials' poll numbers are disappointing.
Another inflationary issue is, of course, oil. Brent crude, the true global benchmark, is currently approaching $100 a barrel . We can see that strategic petroleum reserves have declined significantly since the outbreak of the war, currently at their lowest level since the reserves were established in the 1980s. Reserves have now fallen to 287 million barrels, down from a peak of 750 million barrels, a decrease of more than 50%.
When the Strategic Petroleum Reserve begins to replenish—which will inevitably happen at some point—this will provide a floor for oil prices and make inflation more sticky than the Fed would like to see. But the problem isn't just about U.S. oil reserves. Looking at global oil inventories, going back roughly 10 years to 2018, we can see that the horizontal line drawn with dots represents the latest level, which is essentially at an all-time low, comparable to levels seen in 2025. This further amplifies the pressure on the floor for oil prices.
This is a very interesting chart. It shows the performance of various assets since the outbreak of war at the end of February this year . The results we are seeing are quite remarkable. Commodities, especially the energy sector, have achieved outstanding returns. The Bloomberg Commodity Index has risen by 34% since the outbreak of war . The stock market has also performed quite well, especially emerging market stocks, and the Japanese market has also performed well. Almost all assets have performed well, all achieving double-digit gains. The worst performers seem to be the MSCI Europe and UK, but basically all assets have achieved double-digit or even more than 20% gains. All commodities are rising, and the Bloomberg Commodity Index has risen by 34% as mentioned above.
However, the bond market fared poorly . The best-performing bond category was leveraged loans, which rose by only 3.1%. Emerging market sovereign bonds also saw slight gains, while investment-grade categories—government bonds, mortgage-backed securities, and corporate bonds—all recorded negative returns, with mortgage-backed securities experiencing the smallest decline. This is quite peculiar. We are seeing a huge "donut hole"—the outer ring assets offer substantial returns, while the fixed-income sector yields negligible returns.
Debt growth is clearly a problem. The US nominal GDP is represented by the blue line, and the total public debt of the US Treasury is represented by the red line. We can see that the red line is growing much faster than the blue line, especially since the global financial crisis , and this trend shows no signs of abating; the trajectory is only getting steeper. Currently, the total debt, including the portion held by the Federal Reserve and the Social Security system, has reached $40 trillion, and at the current rate, it could reach $50 trillion by 2032 —this is almost certain.
More notably, even the Social Security Administration itself has stated that under the current funding and benefits system, Social Security will run out of funds by 2032. Of course, their assumptions have historically been overly optimistic, meaning we could actually face this problem as early as 2029 or 2030 —at which point Social Security must be reformed, or face a reduction of approximately 22% in payments. This would likely be unacceptable to the surviving baby boomers who have contributed for many years.
But we'll see. We face a very serious problem. And the situation is clearly not getting any better. This is the federal annual deficit, calculated by fiscal year, going back to 2021. We've set a new record here—for a period this year, the deficit for fiscal year 2025 was slightly lower on a year-to-date basis. But ultimately, it still hit a new high. The current "frontrunner" is fiscal year 2026, and that year is almost over. We'll soon be entering fiscal year 2027, and it looks like this year will set another record.
This chart shows the federal budget deficit as a percentage of GDP, based on projections from the Congressional Budget Office extending to 2035. The yellow line represents interest payments , and the gray vertical line to the right shows future projections, which are not optimistic. These projections are also based on rather optimistic assumptions—that interest rates will be lower than current levels, the deficit-to-GDP ratio will be lower than current levels, and that real GDP will continue to grow positively throughout the projection period. Once you question these assumptions even slightly and apply a little pressure, it becomes very clear that , following the current trajectory, the deficit-to-GDP ratio will likely reach 7% or even 8% within 10 years, but will still be below 10%. This is by no means a good thing.
Gold's price action mirrors that of other commodities. Gold experienced a strong rally in the first quarter of 2026, followed by a significant pullback, falling below $4,000. Now it's starting to rise again. I believe gold should be part of every portfolio. And it's quite evident that as the dollar weakens, central banks and institutional investors are generally starting to favor holding gold over fiat currencies.
Let's look at this chart again— the Shiller PE ratio is currently 42. It was higher in 1999, but not by a significant margin. As you can see, we've traced this data back to the 1870s, and the current level is far higher than during the 1929 bubble. Therefore, the stock is definitely not cheap. This is a very interesting study.
This is a scatter plot covering data from 1965 to 2015, showing the actual returns over the next 10 years based on the CAPE ratio (cyclically adjusted price-to-earnings ratio). The plot includes a downward-sloping regression line. It's clear that when the CAPE ratio is at its current level (currently 42), there has never been a positive actual return over the next 10 years. In fact, the actual return is significantly negative, approximately -5% to -9% per year. This means that buying stocks at this CAPE ratio level using a market capitalization-weighted S&P 500 index would result in substantial actual losses. Interestingly, historically, there have also been many instances of significantly negative actual returns at lower P/E ratios, which seems somewhat anomaly. However, over the past 15 to 20 years, we have become accustomed to higher P/E ratios than in the past. But in any case, this is absolutely not an endorsement of a portfolio heavily weighted by market capitalization-weighted stocks; quite the opposite.
In fact, I don't recommend any of the above. Interestingly, the stock market has a very high concentration, a fact well-known as the technology and AI sectors have grown. Here we can see a light blue line representing the weighting of the information technology sector in the S&P 500 index, which then abruptly disappears. The dark blue line represents the sector with the largest weighting. This means that since 2008, the technology sector has consistently been the largest weighted sector in the S&P 500, currently reaching a concentration of 38% . This level is not only higher than the highest sector concentration in 1999 but also far higher than before the global financial crisis. Therefore, there aren't many truly bargains in the S&P 500 market capitalization-weighted index, much like finding few good deals in a hotel minibar.
Here we see the historical trend of stock market concentration, tracing back to the railroad era. I can't guarantee the accuracy of data from around 1840, but if we look at the "Nifty Fifty" of the 1920s and early 1970s, we can see the stock market bubble of 1987, the situation before the bursting of the dot-com bubble in 1999, and the current market conditions brought about by the top ten AI companies . This is an extremely concentrated market, which also means it is an extremely dangerous market. Therefore, I will not recommend any market capitalization-weighted stocks.
The internal workings of the stock market have also changed. Here we see the correlation between the AI sector and the S&P 500 excluding AI in its rolling 120-day returns. From 2021 to 2025, and even the first half of 2026, the correlation was quite high. But in the past few months, this has changed significantly. If we calculate based on 20 trading days per month, this roughly corresponds to a six-month average. Now the two are negatively correlated.
This is interesting—when the AI market performs well, the rest of the market does the exact opposite. Currently, the two show a slight negative correlation, declining from around 0.5 earlier this year to the current -0.14, and this trend is strong. Therefore, I don't think this situation will reverse anytime soon.
It's worth noting that the S&P 500 equal-weighted index has begun to outperform the market capitalization-weighted index . This chart starts from 2017, but the equal-weighted index began outperforming approximately one to one year and three months ago. When a trend begins to show signs of reversal, we can look back at 2020 for reference—we can see that the relative performance of market capitalization-weighted and equal-weighted indices began to consolidate before a significant correction occurred, with equal-weighted indices significantly outperforming. Now that equal-weighted indices have begun to outperform, while it's not enough to be entirely certain that this is the start of a major trend, at least it's no longer lagging behind.
Furthermore, US stocks are no longer outperforming other global markets . When this line is upward, it indicates that US stocks are outperforming non-US stocks; when the line is downward, it indicates that non-US stocks are outperforming. Over the past year and a half, this line has essentially remained flat, but US stocks have clearly stopped outperforming. Looking at a shorter timeframe, the same chart shows that this actually began two years ago—the relative strength of US stocks peaked nearly two years ago, and a significant relative underperformance occurred between mid-2025 and the first quarter of 2026. I believe that, from a trend perspective, this line will continue to decline in the future. Therefore, I think it makes sense to think from a long-term perspective , not just a short-term one. Regarding foreign stocks , I have been investing in them previously. But now, I want to turn my attention back to the near term because I don't like the current risk profile in the market.
The US dollar has been declining since the end of 2024 , when it was at 110 on the Dixie Index. It then fell below 100 and is currently hovering below that level. Its movement has been unusually smooth, almost appearing manipulated. I mean, for over a year, it has barely made any meaningful changes.
Interestingly, as the dollar weakened, we saw non-US stocks begin to outperform the broader market , resulting in a severe imbalance in global price-to-book ratios. This is an argument against US stock valuations . The MSCI US Index has a price-to-book ratio of 5.72 , while the rest of the world (excluding the US) has a price-to-book ratio of only 2.49 .
You might think the US is the best investment in history, but you'll notice that at certain times, especially during market corrections , the brown and light blue lines in the chart tend to converge, causing the Morgan Stanley US Index to significantly underperform the Morgan Stanley Global Index.
Looking at the comparison between the S&P 500 and the MSCI Emerging Markets Index , the underperformance is quite significant. The US stock market's outperformance is expected to stop at the end of 2024 , and it has already underperformed by about 20% , a considerable margin.
I further believe that this gap will continue to widen in the future . Here is the relative performance of the S&P 500 against the MSCI Emerging Markets Index, represented by the red line. A rising red line means the S&P 500 is outperforming the emerging markets index; a falling red line means emerging markets are outperforming the S&P 500. The blue line is the Federal Reserve's trade-weighted nominal broad dollar index . You can see that the red and blue lines have very similar trends. Therefore, if the blue line (i.e., the trade-weighted nominal broad dollar index) falls, the S&P 500 is likely to underperform emerging markets .
I'm also very sensitive to seasonal factors now. It's early September, and September and October are typically tougher months for risk assets , which is one of the reasons that influences my upcoming recommendations.
Next, let's look at the US domestic bond market . This is the situation of the bond market, comparing the total return of US corporate bonds with the JPMorgan Emerging Markets Local Currency Index . The brown line represents the performance of the local currency index relative to the Bloomberg Total Return.
Here I have a dark line representing the US Dollar Index (shown inverted) , so when the blue line rises, it means the dollar is falling. Similarly, the brown line and the blue line show very similar patterns. Therefore, if the dollar falls (which is exactly what I expect), we anticipate emerging market local currency bonds will outperform US corporate bonds .
Alright, that's all for now, let's dive into the holiday season. Thank you for participating in this conference call, and thank you for your support of Double Line. Goodbye!
Risk Warning and DisclaimerInvesting involves risk; please exercise caution. This article does not constitute personal investment advice and does not take into account the specific investment objectives, financial situation, or needs of individual users. Users should consider whether any opinions, views, or conclusions in this article are suitable for their specific circumstances. Any investment decisions made based on this information are at your own risk.