America's fiscal deficit and inflation are pushing the Federal Reserve into an intractable bind.
Jeffrey Gundlach, founder, CEO and CIO of DoubleLine Capital, warned that whether the Fed chooses to hike or cut rates, it will pay a heavy price—it is a "waterbed game": push inflation down and the interest expense problem bulges up; let inflation run and the debt burden spirals further out of control.
In a recent interview with Solita Marcelli of UBS, Gundlach expressed a broadly cautious stance on long-term US Treasuries, equity valuations and the dollar. He believes the path of least resistance for the 10-year US Treasury yield is toward 6%, and the 30-year would need to reach 6% to 6.5% to be worth buying; the S&P 500's Shiller CAPE ratio stands at 42, and historical data show real returns over the following decade are essentially negative; and once a recession arrives, the dollar faces depreciation of at least 20%.
Gundlach's judgments have direct implications for fixed income and multi-asset investors: he has already cut market-cap-weighted equity exposure to zero, shifting to equal-weight allocation to avoid the rich valuations of AI-related names; on the fixed income side he employs a barbell strategy, while raising real assets to 20% of the portfolio, and he is bullish on emerging market local currency debt and gold.
Bond Supply Overhang, Credit Market Begins to Stratify
Gundlach characterizes the recent rise in global rates as a systemic move. He notes that yields are rising across all major developed markets except Switzerland, and even Japan's 10-year government bond yield—once thought never to rise—has broken above 3%.
The core driver behind this move is the sharp expansion of bond supply. The US budget deficit continues to widen, with issuance over the coming months estimated at between $900 billion and $1.4 trillion, compounded by massive financing needs in AI-related sectors, which is clearly weighing on investors' capacity to absorb it all.
Gundlach observes that the credit market is already showing initial signs of stratification. In the investment grade market, spreads on non-AI bonds remain at historically tight levels of about 70-plus basis points, while AI-related bonds have widened by roughly 60 basis points over the past eight to twelve months. The divergence is even more pronounced in high yield: spreads on non-AI high yield bonds have changed little, while AI-related high yield spreads have widened by more than 150 basis points, trading like deep junk. He cites Oracle bonds as an example, noting they trade at spreads far wider than their official BB or BB- ratings.
Gundlach believes this divergence reflects market doubts about the credibility of the relevant ratings—some non-traditional rating agencies have rated a disproportionate number of bonds relative to their size. At the same time, AI-related issuers will continue to add supply, piling further pressure on top of Treasury issuance.
Stretched Valuations, Long-End Treasuries Lack Appeal
Against the backdrop of rising rates, pressure on equity valuations cannot be ignored either. Gundlach cites data showing that as of the end of July, the S&P 500's Shiller CAPE ratio was 42.04, one of the highest levels on record. Based on historical records, when the CAPE ratio is above 35—and especially around 42—the S&P 500's real return over the subsequent decade has essentially always been negative, with the best case close to -5% to -6%.
On rate judgments, Gundlach believes the reasonable anchoring range for the current 10-year US Treasury yield depends on the model used. If nominal GDP is used as the benchmark—with a seven-year average of about 6.1%—the 10-year yield could theoretically reach 6%; if a blended model of nominal GDP and the German 10-year yield is used, the current level of about 4.9% is already basically reasonable, with the existing level slightly high by about 25 basis points.
On that basis, he states clearly that he will not buy long-term Treasuries unless the 10-year yield reaches 6% and the 30-year reaches 6% to 6.5%. "The path of least resistance is clearly upward," he says, especially against a backdrop of persistently strong commodity prices and inflation that is hard to bring back to 2%. He also notes that the implied inflation expectations from TIPS versus nominal bonds may understate actual pressure—consumer confidence surveys from the University of Michigan, the Conference Board and others all show inflation expectations moving higher, and Costco's recent doubling of its private-label motor oil price along with rationing is, in his view, a classic signal of consumers buying ahead of expected price increases.
Barbell Strategy to Navigate Uncertainty, EM Local Currency Debt a Rare Bright Spot
Faced with this environment, Gundlach divides asset allocation into four dimensions: equities, conventional fixed income, real assets and dry powder, with an overall defensive stance.
On the fixed income side, he employs a barbell strategy: one end is "pure safety" assets including government-guaranteed securities, securitized products and MBS, while the other end is selected high-quality floating rate instruments. He is particularly bullish on non-agency MBS issued before 2020—since home prices corresponding to those mortgages have risen at least 50%, the underlying loan default risk is extremely low, while investors can earn a spread of about 120 to 130 basis points on top of the roughly 5% five-year Treasury, for a total yield of about 6.25% to 6.30%, which Gundlach characterizes as a "risk-free spread." AAA-rated CLOs, which also carry top-tier credit protection and benefit from their floating rate nature amid expectations of rising rates, are also among his holdings, though he has not allocated heavily due to poor convexity.
Emerging market local currency debt is one of the few positive-return categories he singles out. He notes it is the best-performing segment of traditional fixed income year-to-date, and if the dollar weakens, investors can earn additional currency gains.
On real assets, Gundlach allocates 20% of the portfolio, with 10% going to commodity funds, and believes gold, after its recent weakness, has entered a favorable window for long-term accumulation. In his view, the commodity complex is the best-performing traditional asset class year-to-date, beating all equity markets and all categories of US bonds.
Dollar Structurally Weaker, Recession Could Be the Trigger
On the dollar outlook, Gundlach holds a clearly structurally bearish stance. His core logic: the US fiscal deficit is currently near 6% to 7% of GDP, already widening while the economy performs decently in nominal terms; once a recession arrives, the deficit could surge to 12% of GDP, about $3.6 trillion, more than half of current tax revenue. At that point, the Fed will face an extreme choice between massive money printing or debt restructuring, both of which put severe pressure on the dollar.
He cites the 2025 "taper tantrum" as important corroboration for this judgment. In the past, every time the S&P 500 fell more than 10%, the dollar rose almost without exception by 8% to 10%; but the most recent time the S&P 500 fell about 18%, the dollar fell 8% to 10% in tandem—completely contrary to the dollar's traditional safe-haven property. Gundlach believes this marks a new regime in which rates switched from a long-term downtrend to a long-term uptrend after 2020, rendering many old rules obsolete. "Once a recession arrives, I think the DXY will fall at least 20%," he says. Under this scenario, EM local currency debt, non-US assets, gold and broad commodities are all expected to benefit.
The "Waterbed Dilemma": The Fed Cannot Solve Inflation and Interest Expense at the Same Time
Gundlach sums up the Fed's current predicament as an unsolvable problem. He believes the Fed has effectively acquired two new "dual mandates": suppressing inflation and controlling interest expense—and these two goals are inherently in conflict. Hiking rates can suppress inflation but pushes up interest expense; cutting rates can ease debt pressure but reignites inflation. "It's like a waterbed," he says. "Push one spot down and another bulges up. You're just moving the problem around."
He is equally skeptical of AI capital expenditure-driven growth, believing this cycle is unsustainable. When AI enthusiasm shifts from mania to doubt, a sharp decline in capital expenditure will become the main trigger of the next recession, and the recession will cause the aforementioned fiscal predicament to erupt fully. He calls this moment "a very, very heavy reckoning."
For the Fed, Gundlach suggests that standing pat may be the current second-best option—"at least they know where they stand"—but that does not mean the problem will disappear; it merely delays an unavoidable choice.
The following is the full transcript of the interview:
Host (Dan Cassidy, UBS): Hello everyone, I'm Dan Cassidy, and welcome back to the UBS Market Moves podcast. Today we bring you our annual "back-to-school market outlook" conversation, featuring Jeffrey Gundlach of DoubleLine Capital. Today's discussion is hosted by Solita Marcelli of UBS, who is Head of UBS Global Investment Management. Solita, over to you.
Host (Solita Marcelli, UBS Head of Global Investment Management): Good morning, Dan. Thank you very much, and hello everyone. I'm delighted to be joined today by Jeffrey Gundlach. He is the founder, CEO and CIO of DoubleLine. Given that he has been one of the industry's most influential thought leaders for many years, particularly in fixed income, I think Jeffrey needs little introduction. Jeffrey, thank you for joining. I think the timing of this conversation is especially good, given the latest developments in rates, inflation and the Fed. Let's dive right in.
Let's start with this week's hot topic: yesterday's sharp move in rates. The 10-year Treasury yield rose nearly 17 basis points to its highest level since 2007.
Market Volatility and Debt
Solita: The 5-year yield broke above 5% for the first time, also the first time since 2007; the 30-year is near its 2004 high. Although we received new data such as stronger PMIs, the overall feeling is that more uncertainty is embedded in the fixed income market. How would you characterize this recent move?
Jeffrey Gundlach: First, this is a global move. Rates are rising across all developed markets except Switzerland. Even in Japan, yields have, I believe, broken above 3%. People thought Japanese yields would never rise because they were so carefully controlled, but the weakening yen led to some degree of policy shift.
One of the factors driving all this is bond supply. Issuance right now is enormous because budget deficits keep widening, with no end in sight. For example, we're now going to build two more military bases in Greenland—I don't know how much that will cost, but I remember the Bagram base we gave up when we withdrew from Afghanistan cost about $300 billion. This is happening, on top of a lot of issuance in AI and AI-related areas. For issuance over the coming months, I've seen different estimates, roughly $900 billion to $1.4 trillion.
This is genuinely suppressing investor demand. For the first time we're starting to see some initial signs—I wouldn't call them very convincing—that credit is beginning to stratify. In particular, if you compare AI-related issuers with every other issuer in the corporate bond market, you see a completely different picture.
In the investment grade market, spreads on non-AI investment grade bonds are basically at their tightest levels, around 70-plus basis points over Treasuries; but AI-related bonds have widened by about 60 basis points over the past eight months to a year. From a credit analysis perspective, this may or may not constitute a credit problem. I think it's more that investors doubt the ratings some AI-related bonds have received. Note that when SpaceX issued debt, people were somewhat surprised: shortly after its IPO it borrowed about $85 billion and got a BBB- rating—which, as everyone knows, is the lowest rung of investment grade. As long as it retains an investment grade rating, insurance companies give it favorable capital treatment. And the market almost immediately rejected that BBB- rating; these bonds trade like single-B credits.
Looking at the high yield category, AI bonds have widened much more. Non-AI high yield spreads have also widened a bit, but not much, roughly still within the range they've occupied over the past year. But AI bonds had widened by 150 basis points as of last week—I've been traveling recently and don't have the latest quotes—but 150 basis points means they trade like deep junk. Take Oracle bonds, for example—I recall the rating is BB or BB-, and the actual trading spread is several hundred basis points wider than the rating.
I think this is more because these ratings are being questioned. These ratings come from some non-traditional rating agencies, some of which have rated an absurdly large number of bonds relative to their staffing. So the market knows these ratings may be inflated. But beyond that, the market also assumes—and I think this assumption is accurate—that there will be more issuance from these names, and that stacks on top of Treasury issuance.
I don't think it's a coincidence that Scott Bessent announced "Operation Twist" the day after the total Treasury debt on the official debt clock broke through $40 trillion. Earlier this year I talked on a podcast about $4 gasoline and how that's a psychological level. The past six months or so have proven that seems right: when oil is below $4, people don't seem to care much; now the national average is about $4.40, and California may already be at $7. I haven't been to California in a few weeks. It will be interesting to see where prices go.
So there's a lot of unease in the market right now. The $4 gasoline level really matters; the $40 trillion Treasury level—six months ago on this podcast I said we'd hit $40 trillion by year-end, and that this could prove to be a psychological level—now it seems that's true. Suddenly, "we broke through $40 trillion" became big news. Now the debt ceiling is set at $41.1 trillion, and we'll have to do something about the debt ceiling again—another fight, which is not good for the market.
We're currently in a seasonal phase where, for some unknown reason, September and October seem seasonally weak. This September has indeed been very bad—there are still a few trading days left, but the 10-year and long bond yields have risen about 45 basis points over the past month. So we've seen Bloomberg Aggregate bond returns go from decent, near zero, maybe around negative 0.5%, to about -3.5% for investment grade bonds. This unsurprisingly puts pressure on equity valuations.
The S&P 500's Shiller CAPE ratio—as of July 31, the latest data I have—is 42.04, one of the highest on record.
Valuations and Rates
Jeffrey: And in a rising rate environment, such a high CAPE ratio becomes increasingly difficult to sustain. Looking back over a long historical period, when the CAPE ratio is at this level—in fact, whenever it's above 35—the S&P 500's subsequent 10-year real return has basically never been positive, always negative. And around our current level of 42, there are several data points: after recording a CAPE ratio of around 42, the S&P 500's 10-year forward real return is about -6%, with the best data point around -5%. That's the environment we're in.
At least bonds are no longer as absurdly overvalued as they were five years ago—when real rates were deeply negative. Now with bond yields at 5%, real rates are at least positive. But historically, when people question how debt is being managed—as in the 1980s and early 1990s—investors typically demand at least 200 basis points of real return. And right now inflation—there are many different series out there, let's just take 3%, and I think real inflation may be slightly higher than that number. That means the 10-year Treasury needs to yield 5%. And we're right there, at 5.14%.
But if the deficit continues to plague people's judgment about financial market pricing, my view is: to get me interested in 30-year Treasuries, I need at least 250 basis points, maybe even 300 basis points of real return. If inflation is 3%, that means the 30-year Treasury needs to yield 6%. If you'd told people that a few years ago, they'd have thought it impossible. But now we're already at 5.44%, the highest in about 20 years. So the path of least resistance is clearly upward, especially with the commodity complex continuing to boom.
The best-performing asset class year-to-date—if measured only by traditional indices—is the Bloomberg Commodity Index, which has beaten all equity markets and all US bond sectors. So the commodity market is indeed booming. That's good for commodity investors, but less friendly for gold holders, because gold stalled after a big rally last year into the first quarter of this year. And oil prices just won't come down. We keep expecting something magical to happen: the war ends, oil prices fall—maybe. But we've already drawn down a large chunk of the US Strategic Petroleum Reserve, and we can't draw it down much further because you can't drain all the oil from the reserve—below a certain level the oil degrades. So you have to refill the strategic reserve.
And it's not just the US. Look at global oil reserves—they're at their lowest level in decades in terms of barrels; compare that with a growing population over decades, and oil usage that grows with the population—that's a very low reserve level. So although oil prices might fall if there's good news on the war, I don't think they'll drop to $50 or $60. I think the government will want to start refilling the oil reserve. I don't know the exact number, I'm not a policymaker, but I'd guess around $70. Oil prices may fall, but they won't fall back to pre-war levels. And we haven't yet seen the full shock radius of these high oil prices truly unfold, and it will unfold over the coming weeks and months. Fertilizer prices will likely rise, shipping rates are surging, with some shipping corridors seeing increases of 400%.
So I see no realistic basis for inflation falling to 2%. I pointed this out last week at the event before the Fed press conference: the dot plot and the Summary of Economic Projections (SEP) show the committee forecasting PCE at 3.7% by year-end, that's the committee's median forecast; and by the end of 2027, they say 2.3%—that's a 1.4 percentage point decline, and I don't know the logic behind that decline. It won't happen by itself, and inflation won't fall by itself when the commodity price complex is at current levels. So to get to 2.3%, I think you'd have to hike rates.
Of course, now after this big September move in the bond market—remember, in January the bond market expected the Fed to cut rates twice this year. I said after the Fed press conference: "If you're betting on that, you're betting on the wrong horse, because there will be no rate cuts in 2026." And now, what we're seeing is two rate hikes this year—I always slip and say "cuts," but I mean hikes—one in October, maybe one in December. Some people dismiss the idea of an October hike because it's near the midterm elections, but I think that argument is absurd. Early voting is already underway; people are voting. I know it may not look good politically, but I don't think that's the real factor affecting the election. So the Fed may hike once or twice more. The 2-year Treasury is at 4.9%, which implies the fed funds rate should be around 4.5%, while it's actually only at the upper end of the range at 4%.
So I started thinking: based on some useful indicators from the past, where should the 10-year Treasury yield be? By "past" I mean the past 40 years, going back to 1986. It's strange—in the past we used to use US nominal GDP as a starting point for the 10-year Treasury yield. Now the 10-year yield is at 5.15%. Current nominal GDP—using a seven-year average—is 6.1%; and this quarter's GDPNow real growth looks like it's in the 5% range. So 6% is not unreasonable for the US 10-year.
Ten years ago we also built a model that incorporated the fact that when the other $19 trillion of bonds had yields manipulated to negative levels, US rates were positive. We started averaging US nominal GDP with the German 10-year yield. Using that model, the 10-year Treasury should now be at 4.9%. So under the GDP model, the 10-year is currently slightly high, but not by much, maybe about 25 basis points; while using pure nominal GDP it could reach 6%.
I won't buy long-term Treasuries unless yields reach about 6%, and the 30-year may even need to reach 6.5%. We've seen that as the Fed moves into a tighter mode, the yield curve has flattened significantly. The 2-year/30-year spread was once at 130 basis points, now it's about 55 basis points—a big flattening, which makes the long end even less attractive.
That's my basic assessment of the current situation. Since the end of July, actually mid-August, we've taken a defensive posture across all markets, mainly because of valuations—I talked about the CAPE ratio earlier; before this bond yield move, valuations already looked quite unattractive. Also, as I mentioned earlier, seasonality is just bad. Negative returns in September and October wouldn't be surprising. I'll stop there and see where you want to take it.
Portfolio Strategy
Solita: Thank you very much, Jeffrey. You touched on a lot. To tie it together: where are the best opportunities in fixed income right now? We know you're cautious on the long end and like the short end. But looking across the entire fixed income universe—securitized credit, emerging markets, etc.—where do you see the best opportunities?
Jeffrey: I divide asset allocation into four buckets: equities; conventional fixed income, i.e., Bloomberg index-oriented thinking; real assets; and finally "dry powder," i.e., cash-like reserve ammunition.
On equities, what I really want to do is avoid the epicenter of overvaluation. Since about two weeks ago, I've cut market-cap-weighted exposure to zero. Instead of concentrating 40% to 50% in AI and AI-related names—which are the drivers of overvaluation—I've shifted to equal weight, so I hold almost no AI and AI-related names. So I'm not conservative in this category.
On fixed income, I use a barbell strategy, which is a bit unusual—this is the first time in my career I've really used it: one end is truly safe, corporate-bond-free, consisting only of government securities, securitized products and mortgage-backed securities (MBS). What's risk-free right now? Non-agency MBS issued five years ago or earlier, because although home prices are now falling modestly, the homes underlying these pools' mortgages have risen at least 50%. Suppose someone bought a house for $1 million with an $800,000 mortgage, and now the house is worth about $1.5 million—the loan-to-value ratio (LTV) is close to 200%. No one would default—house worth 1.8, mortgage 800, only a complete fool would default. So you get the spread, about 130 basis points in good times, about 120, 115 when spreads are particularly tight. But I believe this is a truly risk-free spread over Treasuries. The five-year Treasury area is 5%, so you can get about 6.25%, 6.30%, and I consider that a risk-free asset. This is one of our important holdings.
Another risk-free asset is AAA-rated CLOs—we don't hold them heavily, mainly because of poor convexity. Recently people have started thinking about the Fed hiking, and it's not surprising that AAA CLOs have become popular, because everyone wants floating rate assets at the top of the capital structure. If AAA CLOs ever took a loss, that would be a big problem—that's not rigorous analysis, but I do think AAA CLOs won't lose money; they didn't during the global financial crisis. There you can get a spread of about 100, 110 basis points. Again, I consider these risk-free spreads.
Another one is emerging market local currency debt, which I hadn't done for a long time. It has a double benefit: it's actually the best-performing traditional bond sector year-to-date, with a modest gain of about 4%, which after the September correction is probably only 1% to 1.5%. But if the dollar weakens, you also get currency gains. Either way, you have a positive return—and almost nothing in the bond market is positive right now. As I said, most categories are down 3.5%, some down 4.5%. So you have to be defensive.
Real assets make up 20% of the portfolio. I think 10% of that should be in commodity funds. After gold's recent weakness, I think now is the time to add to gold for long-term accumulation purposes. We're looking for things that hedge against higher rates.
For six years I've held one judgment: 2020 was the cyclical bottom for rates. Now rates have risen 500 basis points, and that statement is no longer controversial. But I think the fundamental mistake people make comes from their experience, especially for those who entered the business 20 to 25 years ago—they really think current rates are high. People call this a "great buying opportunity" to get 5% on Treasuries; of course that's much better than 3%, and although inflation isn't at 2%, it's at least relatively stable, and inflation probably won't really start falling until the high year-over-year base effect fades in the second quarter of next year. But 5% rates really aren't high.
I think people have been spoiled by the past—roughly the 20 years before 2020—of zero rates, thinking zero rates and long bonds below 2% were some natural norm. They weren't. I think what really confuses investors is that people in private credit and private equity keep saying: "Once rates come down, we can exit, and we won't have to keep doing extensions and continuation funds. Clients want their money back, and we can't liquidate—there are trillions of dollars trapped in private companies." But rates won't come down like that. That was an anomaly. I think 5% rates are completely reasonable based on nominal GDP and the nominal GDP-German 10-year yield model, and GDP still looks like it's rising—possibly even driven by inventory restocking, which worries me about inflation reigniting. When I was a kid, I remember people, including my family, would buy ahead once they felt prices were about to rise. There was a great example last week—just one example, but interesting: Costco raised the price of Kirkland private-label motor oil by about 100%, literally doubling it, while telling customers to ration supply, limiting purchases to a certain amount per month. I think Costco's calculation—and I think they're smart to do this—is: when you double the price of something, people think, "I keep hearing gold is going to $100, they've already doubled motor oil, maybe it'll double again." So people start hoarding and buying ahead.
I've noticed that in the recent GDPNow upgrades, both consumer spending and inventory restocking are increasing, suggesting consumers may be buying ahead in anticipation of higher prices, and retailers may be stockpiling inventory, thinking "buy now, maybe sell higher at year-end." So I don't buy the argument that "this is all about real rates, inflation expectations haven't really changed"—sure, comparing TIPS with nominal bonds you can reach that conclusion, but I don't buy it. Look at consumer confidence indices—University of Michigan, Conference Board and so on—I think we're starting to see inflation expectations move higher.
So this is a period where equities have done well, commodities have done very well, and bonds have done absolutely nothing good, except EM local currency debt. I'm still on the side of "the path of least resistance for long-end rates is upward," until Operation Twist actually kicks in—and that hasn't happened yet. $6 billion of buybacks is equivalent to one day's deficit, nothing. Of course, they do have the ability to control rates: we saw it in the 1940s to the mid-1950s, and in Japan for decades—then they spent so much effort maintaining yields and ultimately couldn't do it, and now they've joined the higher-rate camp too. So I remain conservative on credit and conservative on duration.
Solita: Thank you, Jeffrey. You mentioned EM local currency debt, and we've talked a lot about deficits and high rates. So what do the current fiscal deficit and high rate environment mean for the dollar?
The Dollar and Fiscal Policy
Jeffrey: I think the dollar's problem is this: when—not if, but when—the US enters a recession, the dollar will weaken severely. I don't think it will be a safe haven at all. The big rally in gold over the past four years or so already shows this. I think the dollar will fall because of the deficit problem. The deficit is already running at about $2 trillion a year. There's one more month of fiscal data to be released, so it could change, but the deficit as a share of GDP for fiscal 2026 (ending at the end of this month) will be the highest in five years, while this quarter's GDPNow nominal or real growth is around 5%. That is, when the economy is supposedly performing well in nominal terms, the deficit is widening. And once the situation reverses, the deficit will increase substantially. We're at about 6% of GDP now; if you include all war spending and disaster relief spending not counted in the official deficit, it's closer to 7%. Once a recession arrives, it could easily hit 12% of GDP—of course that would be a horrifying deficit, about $3.6 trillion, more than half of current tax revenue.
So we have to do something about this fiscal problem, especially when the next recession comes. Actually, we should do it now, before the recession arrives, but unfortunately politics doesn't work that way. In the next recession, I think the dollar will fall. The DXY index has been absurdly stable, almost as if manipulated—it's heavily influenced by the euro—with a year-to-date low near 98 and today near a high of about 101.5, but a range of less than 4%. So when the DXY starts falling in the next recession, I think it falls at least 20%. A sharply weaker dollar means EM local currency debt could perform well, any non-dollar investment could perform well, and gold and broad commodities could perform well too.
I think the biggest confirmation of my structurally bearish dollar view came during last year's "taper tantrum": the S&P 500 fell into correction—that was the 13th correction since 2020, defined as a decline of more than 10%. In the previous 12 corrections, every time the S&P fell more than 10%, the dollar rose every time, with the DXY rising about 8% to 10% each time. But the most recent correction—the 2025 taper tantrum—the S&P fell about 18%, yet the dollar fell about 8% to 10%. Completely opposite to the dollar's typical behavior during risk-off periods for assets like the S&P 500.
This confirms my view: when we suddenly switch from the 40-year long-term rate downtrend regime from 1980 to 2020 to a long-term rate uptrend regime—which is what I call the present—everything changes. One very telling example: one of the longest-running, most economically driven, most academic analysts is Lacy Hunt of Hoisington. They deserve credit for turning bullish on Treasuries when yields were in the double digits, even mid-double digits 35 to 40 years ago, and remained bullish until early this year—they really stayed too long. They were firmly bullish on Treasuries going into the big bear market of 2021 into 2022, and remained bullish throughout the bear market. But after decades of being bullish on bonds, they turned bearish on bonds for the first time in the first or early second quarter of 2026.
I think this is very telling: he built a framework he used for 40 years, a framework that worked for 35 to 40 years, he kept using it, but when he applied it to the bond market after 2021, and certainly after 2020, it completely failed. To his credit again, he concluded the framework no longer works. The framework was useful and beneficial in a long-term bull market, but now the mirror shows: it's no longer a long-term bull market, but a long-term bear market. That's the conclusion he reached. Of course, his bearish call also made some money, but if he'd figured it out in 2020 or 2021, he'd have made far more.
I think we're in "opposite land," and the most important thing people need to consider is: the old rules will become increasingly inapplicable. Think about how good friends we used to be with Canada—such a big trading partner, similar cultural values, not very different government forms. When I was a kid, crossing the border you didn't even need to show ID, you just said "I'm going to Toronto" and that was it. Now they're talking about becoming an associate member of the EU, and they've labeled us as some kind of "economic enemy." What more do you need to see to understand that the foundations of this framework are changing? All of this is part of my thinking.
So I think the last straw will be when a recession arrives and the deficit rises to unimaginable numbers like $3 trillion or even $4 trillion. You either print money—bad for rates and inflation—or restructure debt: lower coupon rates, extend maturities, or other extreme aggressive measures certain to face strong opposition. So that's what you should be thinking about.
I've pointed out that the Fed faces a real dilemma: if it fights inflation and hikes rates, it worsens the interest expense problem, which is a big problem; but if it doesn't hike, there's also a problem—the inflation problem. You have an inflation problem, but if you fight the inflation problem, you have a debt interest expense problem. Both choices are bad. Maybe the Fed wants to stand pat, because at least they know where they stand. But this is the Fed's dilemma: hiking is a problem, and cutting is also a problem.
I've always said the Fed has two mandates: maximum employment and stable 2% inflation. I say they still have two mandates, but clearly employment is no longer the mandate they focus on. Kevin Warsh—no, I mean the current Fed Chair—in his remarks maybe only 5% talked about the unemployment rate, just saying "employment is stable and strong, in balance, no problems at all." So at least in my view, that's not the mandate they're focused on. I think their dual mandate now is interest expense and the inflation rate.
Fed Mandate and Outlook
Jeffrey: Unfortunately, these two will forever be in tension. Jay Powell used to always say the dual mandate of employment and inflation were currently compatible; by last year, his final year, the two began to conflict more, and he started talking about the tension between the dual mandates, but after the unemployment rate stopped rising and fell back to 4.1%, that conflict eased. But now the dual mandate is forever in tension, because it's inflation and interest expense, and you can't solve both at the same time. You push one down, like a waterbed: push one spot down, another bulges up. You're just moving the problem around. It's a difficult situation, but we brought it on ourselves.
For 20 consecutive years I've been criticizing this ignored deficit and interest expense problem. People ask me what should be done, and I say: have Elon Musk invent a time machine, and we go back 20 years and start being a fiscally responsible country. Of course that's impossible, I'm joking. But people have indeed asked me: if you were in power, what would you do? I'd say: tell Congress you have to cut $1 trillion in spending, and you have to find a way to cut $1 trillion in spending by the end of fiscal 2027. That's how I would tackle this problem. It won't look good economically, but unfortunately the window for less painful measures has passed.
Solita: Jeffrey, you mentioned the impact of a recession, but you don't think a recession will come soon, do you?
Jeffrey: I think when the AI industry contracts—and I feel we've already passed peak AI—a recession will come. I've had this feeling since around June or July, when the narrative suddenly shifted from "AI is the best thing in human history, we can all retire to Tahiti and live in luxury, AI will do everything for us" to "but you'll lose your job first, AI will kill jobs." That wasn't scary enough, so starting a few weeks ago, the narrative shifted again to "AI might kill us all before the end of this decade." This is the most terrifying scare campaign I've ever seen. I think it's meant to create some kind of government intervention, to build a moat around existing AI companies. But government intervention won't help.
So when AI mania turns to AI doubt—which I think is underway—we'll see a sharp decline in capital expenditure. I don't know when it will happen, whether in the next six months or eighteen months, I don't know, nobody knows. So a recession will most likely arrive within the next few years. I think the current form of AI is unsustainable anyway. Our economy is entirely driven by AI capital expenditure, the entire economy depends on it, an unprecedented large-scale capital expenditure cycle. Based on this cycle, S&P 500 earnings expectations are being revised up, and I don't even know what this year's forecast gain is now, 30% or 36%. And people are still extrapolating, thinking earnings growth will continue, not as high, but still abnormally high. I would aggressively short that idea, because these things all look unstoppable until suddenly they're not.
This dynamic is different from 2000—at least then some companies were making money, but now many companies are bleeding. So there are some similarities. The trigger for the recession is this. When it happens, as I said, the way we run the government will face a very, very heavy "reckoning."
Solita: Okay, thank you very much, Jeffrey. You've left us with a lot to chew on. It's always a pleasure to hear you share. From rates to inflation, the Fed, global markets, geopolitics, portfolio construction, equities—today's perspective was very comprehensive. Thank you so much.
Jeffrey: Thanks for having me. I really enjoy doing this show, it feels like we've been working together for years. Thank you very much for the collaboration between our two organizations, it's been very pleasant and very satisfying. Thanks for having me. Finally, everyone cheer for the Bills—the Bills are 2-0, and I'll be at the next three games, which will be very exciting. I wish everyone all the best and good luck through the end of the year.
Comments