Dennis DeBusschere: The Economy Must Respect The 'Speed Limit'

Dennis DeBusschere · 22V ResearchBullish

Software names using AI to expand margins are a long; he holds them via IGV.

Consensus is underappreciating service-sector names. He has focused for months on software companies that benefit from using AI tools to raise margins and profitability, and that is finally working. He suggests owning the generic software ETF, IGV, to monetize the theme.

“that software as an example, much more focused and we have been for a couple of months, but it started now finally starting to work, much more focused on companies that are benefiting from using these AI tools to increase margins and profitability.”at 20:36

Dennis DeBusschere · 22V Research

AI capex alone should add about 1% to US GDP growth for three to four years.

Citing SemiAnalysis, he says cumulative AI-related capex through 2030 will be about $11.5 trillion, with $7 trillion debt-financed. That adds duration supply that competes with Treasuries, and it signals a large investment pipeline ahead. Even if the numbers are 20-30% too high, AI capex alone should keep GDP growth about 1 point higher for three to four years, which helps explain why yields are rising.

“So you can just say, I'm gonna get 1% at least to GDP growth for the next three to four years from AI CapEx alone. If you believe those numbers.”at 9:35

Dennis DeBusschere · 22V Research

AI demand baskets are a long, since their earnings growth is unlikely to change soon.

He had been negative on AI demand baskets but is less so now. They are a long because their earnings growth is cyclical yet can grow outside the economic cycle. He says the capex chains are unlikely to change anytime soon, so earnings should hold up even as the broader economy slows.

“So we had been a little bit negative on the price momentum basket, or your AI demand baskets. We are less so now. We wanna be long that because those earnings growth are unlikely to change.”at 20:15

Dennis DeBusschere · 22V Research

Data-center backlash is a pretend hatred that will clear after the election.

Opposition to data centers is politically popular, but it is performative. He says DC conversations show municipalities will receive large sums from data-center projects, and there is too much money involved for the backlash to last. He expects a clearing event after the midterms that reverses the commentary.

“So I think there will be a clearing event post the election as it relates to data center commentary. But it's a nice thing for everybody to push back on right now.”at 24:28

Dennis DeBusschere · 22V Research

US economy is running above its roughly 2% speed limit at full employment, creating inflation risk.

Nominal GDP is about 6.5% and not slowing, and unemployment is 4.1% versus the Fed's projected 4.3% by year-end. Real consumer spending tracks 3-3.5%, and AI capex adds 70-100bp to GDP. He argues potential growth is about 2% because productivity growth is 2% and labor force growth is negligible. Growing above that at full employment with above-target inflation means tighter financial conditions and higher yields, not higher equity prices.

“So kind of related to that is the economy can only really grow at 2%. And that's your speed limit because productivity growth is 2%...”at 11:31

Dennis DeBusschere · 22V Research

A 15-20% stock decline would slow growth to about 1%, but not cause a recession.

A 15-20% equity decline would absolutely slow consumer spending. Growth would land around 0.5-1%, but not in recession, since real consumption at about 3.5% is far above its post-GFC norm of about 1.25%. Only a combined AI capex slowdown and 20% market decline would change that assessment.

“So consumer spending would absolutely slow, and I think you would end up with an economy that grows somewhere in the 50 basis points to a 100 basis point range, so 1%.”at 22:36

Dennis DeBusschere · 22V Research

Fundamentals will overwhelm Treasury efforts to hold down long yields over the cycle.

Ten-year yields are rising because markets price strong growth for longer: consumer spending accelerated and AI capex is heavy. Treasury or Fed efforts to keep yields low can have a small effect in low-liquidity periods, but they cannot hold yields below fundamentals over six to twelve months. The Treasury market is far larger than any program. He also argues the administration understands that a lower term premium is desirable, and that hawkish restraint that slows growth achieves it.

“So I just think it's very hard over a long duration to actually keep rates at some level that is not justified by the fundamentals, especially given the size of US Treasury market, which dwarfs any program that he's willing to use.”at 14:00

Dennis DeBusschere · 22V Research

Expects the 10-year Treasury yield to fall back to 4% by Christmas Eve.

Investor surveys show the highest share since 2022 saying financial conditions must tighten to meet the Fed's goal. If those surveys are a guide, the next three months will bring dovish data: higher unemployment and soft core inflation. That should pull the 10-year yield back to 4% by year-end. This is a near-term call that cuts against his structural view that yields keep rising unless inflation falls.

“we're gonna have dovish data galore for the next three months... Ten year yields back at 4% by Christmas Eve.”at 32:10

Dennis DeBusschere · 22V Research

Rising unemployment with positive payrolls would be good for markets by easing Fed hikes.

If unemployment rises because labor supply grows while payrolls stay positive, the economy's speed limit rises and the Fed has less reason to hike. He says that would help markets more than strong labor would. He dismisses the K-shaped pain at the bottom end as not mattering for growth or markets.

“It's probably a good thing if the unemployment rate goes up rather than down for markets.”at 11:09

Dennis DeBusschere · 22V Research

Wants AI to displace jobs; it's needed to lift productivity and ease a too-tight labor market.

Labor force growth is essentially zero in the US, so economic growth is capped at about 2% by productivity. Higher productivity raises that speed limit. He says AI taking more jobs, possibly lifting unemployment, would help. Without it, the labor market stays too tight and growth stays capped.

“Another way to do that is increase productivity growth meaningfully. That could be associated with a higher unemployment rate. I'm actually rooting for AI to take more jobs. You need it to happen. Otherwise, we're gonna be in a way too tight of a labor market.”at 12:01

Dennis DeBusschere · 22V Research

Agrees with the Fed's hike; growth is too strong, so the Fed must keep restraining inflation.

He agrees with the hike. Economic strength is far stronger than forecast and the supply shock has lasted longer. Growth running well above the supply side raises inflation risk, so the Fed needs mild restraint. He says Warsh knows inflation must be quelled, and that a somewhat higher unemployment rate is desirable to offset deficit concerns, though officials would never say so publicly.

“this is why do not underestimate the Fed's willingness to raise rates. They know this. Warsh knows this. You need to quell inflation.”at 17:37

Dennis DeBusschere · 22V Research

Consumer spending must slow toward income growth; consumer areas are relatively negative.

Real personal consumption expenditure is tracking 3-3.5%, while real income growth is only 1-1.5%. Nominal net worth rose $12 trillion in Q2, letting consumers save less to maintain spending. Spending should converge toward income growth over time, and the Fed's short-rate hikes are meant to speed that up. He positions to be relatively negative on consumer areas.

“So it seems very obvious to us that consumer needs to slow. ... And that consumption growth should come down closer to income growth over time.”at 19:36

Dennis DeBusschere · 22V Research

Financials will not benefit from the yield curve flattening as the Fed hikes.

The Fed's short-rate hikes are flattening the yield curve, and he says financials will not benefit from that flattening as it happens. He also expects consumer and transport names to suffer on a relative basis during the slowdown, and positions defensively against them.

“Financials will not [be] benefiting from this yield curve flattening as it takes place, but also your consumer, your transports should also suffer on a relative basis.”at 19:53

Dennis DeBusschere · 22V Research

Stock market is comfortable with a Democratic sweep; a GOP sweep would raise yields.

The market is comfortable with a Democratic sweep, and there is little difference between a sweep and one chamber flipping. A Republican sweep would be the worry, because it could bring more inflationary, less budget-constrained policy. That would push the ten-year yield and term premium significantly higher.

“So I think the stock market is very comfortable with a democratic sweep at this point.”at 23:47

Dennis DeBusschere · 22V Research

Debt ceiling fights only matter for a few weeks when they become a real problem.

The debt ceiling has never mattered beyond the few weeks when it becomes a real problem. He also sees no meaningful change to consumer spending, AI capex, earnings, margins, or Fed policy from a government shutdown, so such episodes are likely only headline noise.

“the debt ceiling has never really mattered beyond the time when it becomes a real problem for the few weeks.”at 26:40

Dennis DeBusschere · 22V Research

Brazil's equity rerating potential is large if Flavio Bolsonaro wins; market prices about 20% of it.

Brazil has a very low multiple, high interest rates and heavy fiscal spending. Local investors dislike Lula and have under-invested. With inflation stabilized and a good growth outlook, Flavio Bolsonaro gaining in polls means significant rerating potential. Asked whether the market prices about 20% of the move, he agreed, and said normalizing the multiple would be roughly in line.

“multiples that are very, very low relative to anything else, a very good growth outlook. So there looks like there's a significant rerating potential, if we actually get now Flavio to win.”at 29:06

Dennis DeBusschere · 22V Research

Deeper cyclicals are a bad overweight into Fed hikes; tech now leads relative to its history.

In a sector survey taken before Labor Day, materials, industrials, and energy ranked best. Industrials have since been the worst performer, while tech, which rarely ranks low, is now the best relative to its own history. Investors were overweight deeper cyclicals into the teeth of a Fed hiking cycle, which he says is never good for them.

“So that's the one I thought was most interesting as people were very interested in being overweight the deeper cyclicals right into the teeth of a fed raising rates, which is never good for deeper cyclicals.”at 30:58

Dennis DeBusschere · 22V Research

Expects core PCE to print around 0.1% over the next three months.

Investor surveys show the highest share since 2022 saying financial conditions need to tighten to meet the Fed's goal. If the surveys are any guide, the next three months bring dovish data, including a higher unemployment rate and core PCE prints around 10 basis points. This is a short-term call.

“Expect core PC[E] coming out like 10 basis points, something like that.”at 32:18