Bond Yields, Debt Traps, and AI Bubbles: What the 2024 Macro Outlook Got Right — and Where It Fell Short

EDITOR'S NOTE (March 2026): This podcast was recorded in March 2024. In the two years since, the 10-year U.S. Treasury yield did not reach 3% — it finished 2024 near 4.6% and remained elevated through 2025 even as the Federal Reserve cut rates three times. There was no U.S. recession in 2024, though growth moderated from the exceptional pace of 2023. The AI theme accelerated rather than peaked. David Abramson’s structural framework — fiscal dominance, the anatomy of a debt trap, and the four conditions for a speculative bubble — has grown more relevant with every passing quarter.  One specific prediction has resolved with unusual precision. At the 36-minute mark David stated: “You’re going to need lots of oil two to three years from now. It’s not being incentivised.” Two years from March 2024 is March 2026. On February 28, 2026, the Strait of Hormuz was closed to commercial shipping following coordinated U.S. and Israeli strikes against Iran, removing approximately 20 per cent of global oil supply from the market overnight and sending Brent crude above $100 per barrel. The structural supply vulnerabilities David described — underinvestment, constrained refinery capacity, and concentrated chokepoint risk — are precisely what the March 2026 crisis exposed. Northland’s CEO Arthur Salzer has published a full analysis of the portfolio implications in The Artisan: “The Energy Hedge of a Decade: What the 2026 Oil Crisis Means for Canadian Investors.” 

On energy positioning: Joseph’s comment in this episode that “our clients have benefited” from early energy exposure reflected the firm’s 2022–2023 positioning. 2025 was a difficult year for oil prices, with Brent trading in the low $60s through much of the year — testing the patience of energy bulls and delivering a period of underperformance relative to broader equities. The March 2026 supply shock has substantially reversed that. The full investment thesis, including the difficult 2025 chapter, is documented in the 2026 article referenced above. This piece revisits the conversation to separate the durable frameworks from the point-in-time forecasts.

When Alpine Macro's David Abramson sat down with Northland's Co-CIO Joseph Abramson in early 2024, the conversation landed on a question that looked tactical but was actually structural: how far would bond yields fall, and what would force them there?

The surface answer — U.S. 10-year yields toward 3% on the back of a mild recession — has not materialized on the timeline David described. What has materialized is the analytical machinery underneath it. Fiscal dominance is real. Private sector resilience has surprised, partly because upward revisions to household savings data extended the runway longer than the consensus expected. The Kindleberger bubble framework David applied to the Magnificent Seven has become a reference point that serious macro analysts continue to invoke as the AI investment cycle matures.

This conversation is worth revisiting not as a scorecard — forecasting specific price levels across a one-to-two year horizon is a low-percentage game — but as a case study in how rigorous macro analysis separates durable frameworks from point-in-time targets.

The Fiscal Policy Puzzle: Why the Change in Deficit Matters More Than Its Size

David's central argument rested on a distinction that most commentary missed: it is the change in the cyclically adjusted deficit that provides economic stimulus, not the deficit's absolute level. A large deficit that stays large is not stimulative — it is baseline. What drove U.S. growth in the second half of 2023 was a widening of the structural deficit at a point in the cycle when such widening almost never happens outside of recession. Once that impulse faded, he argued, the economy would face a meaningful drag.

This reasoning has proven more durable than the specific forecast it supported. The deficit did not dramatically widen further in 2024, growth did soften from its exceptional 2023 pace, and the Fed did pivot to rate cuts. What David did not fully anticipate was the persistence of consumer spending even as delinquency rates on bank credit ticked up — partly because revised savings data showed households had more buffer than appeared in real time.

For Canadian UHNW families managing globally diversified portfolios, this distinction matters practically. When the U.S. fiscal impulse fades but doesn't collapse, bond duration functions as partial portfolio insurance rather than as an aggressive return generator. That is a very different sizing and risk conversation than 'bonds will rally in a recession.' A multi-family office approach requires being precise about which scenario is being hedged and at what cost.

The Debt Trap Thesis: Structural Risk, Not an Imminent Crisis

The debt trap discussion is the most structurally important section of this conversation. The arithmetic is straightforward: when the real interest rate on government debt persistently exceeds the economy's real growth rate, and the primary deficit does not shrink, debt as a share of GDP compounds upward. The U.S. was not in a formal debt trap by David's 2024 definition — real interest rates were roughly equal to real growth in most quarters. The margin has not improved since.

As of early 2026, U.S. federal debt exceeds $36 trillion. Interest payments now consume a larger fraction of the federal budget than at any point in recent decades. The 10-year TIPS yield, which David used as a proxy for the government's real borrowing cost, has remained elevated. This is not a crisis. It is the precondition for one if the political will to address the primary deficit does not materialize.

The bond vigilante scenario David described — where investors refuse to accept low real yields on sovereign debt because they no longer trust the fiscal trajectory — remains a tail risk rather than a base case. It is also the risk that matters most if it arrives, because the adjustment would be sudden and disorderly. Long-duration nominal government bonds carry this risk in an unreflective way. A UHNW portfolio built around liability matching — the pension model — is deliberately exposed to this duration risk because it matches a liability. A family office portfolio without corresponding long-duration liabilities carries that same risk without the structural justification for it.

The Japan comparison David raised deserves continued attention. Japan spent three decades demonstrating that a debt trap can be tolerated longer than bond vigilantes predict — provided the domestic savings base remains willing to absorb the bonds at low yields. The U.S. does not have the same domestic savings structure as Japan, which makes the path-dependence of this risk genuinely uncertain.

The Kindleberger Framework and the AI Bubble

The most intellectually useful section of this conversation is David's application of Charles Kindleberger's bubble anatomy to the Magnificent Seven technology stocks. Kindleberger identified four conditions for a speculative mania. David applied each with precision.

The first condition is a real and compelling displacement — a genuine technological or structural shift that justifies initial re-rating. AI qualifies unambiguously. The productivity potential of large language models, the data infrastructure buildout behind them, and the concentration of this capability in a small number of well-capitalized companies are not fictitious.

The second is valuation difficulty — the new paradigm is hard to value by conventional metrics, which licenses investors to abandon those metrics entirely. In the late 1990s internet cycle, the license was 'eyeballs' and website traffic. In the early 2020s SaaS cycle, it was ARR multiples. In the current AI cycle, it is infrastructure spend as a proxy for future market share. Each license has been partially valid; each has also enabled extreme overvaluation when combined with the other conditions.

The third is broadening participation — retail investors and benchmark-constrained institutional managers both chase the story. The former from fear of missing out; the latter from fear of tracking error against an S&P 500 where seven stocks represent an outsized fraction of market capitalization. David's observation that money managers without large Magnificent Seven exposure faced massive tracking error risk in 2023-2024 was precisely accurate and remains relevant.

The fourth condition — and the one David identified as missing in early 2024 — is plentiful liquidity. The Federal Reserve's shift to rate cuts in September 2024 began to restore this ingredient. The precise timing and magnitude of any resulting mania in the most extended AI-adjacent names is not forecastable with useful specificity. The framework is useful not for calling the top but for identifying when all four conditions are simultaneously present — which is the moment maximum caution is warranted.

For Canadian investors, the AI theme creates a specific portfolio construction challenge. Canadian equity allocations are structurally resource-heavy by default. Adding large U.S. technology exposure through index products creates a dual concentration: sector concentration in tech and resource energy simultaneously, combined with geographic concentration in North American equities. A multi-family office manages this explicitly through factor analysis and scenario stress-testing rather than relying on broad market beta to provide diversification.

The Energy Investment Thesis

The exchange on energy illustrates the difference between cyclical caution and structural conviction, and it remains instructive. David and Joseph agreed that a U.S. recession would pressure oil prices — the historical record on demand-driven recessions is consistent on this point, with oil falling in every instance. They also agreed that this represents a cyclical interruption to a structural bull case rather than a reversal of it.

The structural argument is grounded in supply, not demand. Years of underinvestment in new production, ESG constraints on institutional capital allocation to exploration, and the diversion of capital to green energy projects with policy-dependent returns have all combined to limit the supply response to elevated prices. The comparison to tobacco equities in the 1980s — a sector with structurally declining demand that produced exceptional equity returns because supply declined faster and capital discipline improved dramatically — is both apt and underused in investment discourse.

The energy transition thesis raises a related observation on green energy investments: when government subsidy is the primary driver of project economics, capital allocation becomes decoupled from return-on-capital discipline. Projects that cannot clear an unsubsidized hurdle rate accumulate on a foundation of political continuity risk. Both speakers noted this risk in early 2024; subsequent policy shifts in several major markets have confirmed it as a live concern rather than a hypothetical.

Fixed Income Positioning for Canadian UHNW Portfolios

Synthesizing the macro framework from this conversation into portfolio guidance for Canadian UHNW families suggests three durable conclusions.

First, bond duration is not a passive strategic position. The fiscal dynamics David described — persistent deficits, rising interest cost as a share of GDP, and a bond vigilante tail risk that is structural rather than imminent — argue for managed, regularly reassessed duration exposure. The endowment model and the multi-family office model handle this differently from a pension model precisely because the liability structure is different. Positioning duration to match a long-dated liability is rational; holding long-duration government bonds as a speculative bet on a deflationary recession requires a correspondingly high conviction on that specific scenario.

Second, real assets and hard assets provide a hedge the bond allocation cannot. In the debt trap scenario — or even in a moderate fiscal deterioration scenario — real estate, infrastructure, commodity exposure, and real-yield instruments provide a hedge against the purchasing power erosion that nominal fixed income cannot. This argues for maintaining real asset exposure even when near-term return expectations favour equities or credit on a risk-adjusted basis.

Third, the AI theme warrants a framework rather than a binary view. The Kindleberger lens suggests neither 'AI is a bubble, avoid tech' nor 'AI is transformative, buy everything.' It suggests: monitor all four conditions, size positions to survive a severe drawdown in the most extended names without forcing a distressed rebalance, and distinguish between the companies with genuine durable competitive advantage in AI infrastructure and the many more that are benefiting from narrative association rather than fundamental economics.

Frequently Asked Questions

The following questions and answers reflect content discussed in the podcast and editorial commentary above. These answers must match visible body text for structured data compliance.

About This Episode

This episode of The Artisan Podcast was recorded in March 2024. The Artisan Podcast features conversations between Northland's investment team and leading investors, strategists, and thinkers. Episodes are available on Northland Wealth's YouTube channel and through the firm's website at northlandwealth.com.

About David Abramson

David Abramson is the Chief U.S. Strategist and Director of Research at Alpine Macro, an independent global macro research firm. He teaches MBA at McGill University and spent more than 20 years as a Macro Strategist at BCA Research prior to joining Alpine Macro.

About the Host: Joseph Abramson, CFA, MBA

Joseph Abramson is Co-Chief Investment Officer and Portfolio Manager at Northland Wealth Management Inc. He holds the CFA designation and an MBA from McGill, and brings extensive experience in global macro research from his prior roles at BCA Research and Senvest. At Northland, Joseph leads the investment strategy process, overseeing both traditional and alternative asset allocation for ultra-high-net-worth Canadian families.

Make sure to check Northland Wealth’s YouTube Channel for more episodes.

Full Transcript

SPEAKERS: Joseph Abramson, David Abramson