Karen Karniol-Tambour - Macro Headwinds vs. Tech Tailwinds - [Invest Like the Best, EP.329]
What this covers
Recorded at the Sohn Conference, this is Patrick talking with Karen Karniol-Tambour, co-CIO of Bridgewater Associates, two years after her first appearance. It begins with AI: what she watches for, and why she doubts it can be used for macro investing. Macro markets have a short history, so a model trained on them risks Limited Sample Overfitting, unlike image recognition, where the data is vast. From there it moves to the state of capital markets and portfolio construction.
The middle of the conversation is about slow-moving change. She traces how globalization and automation lifted manufacturing productivity and corporate profits, held inflation down, and fed inequality and populism. She cites OpenAI and Goldman estimates that AI could touch a far larger share of the labor market, and faster. Against that, she argues that spending on supply-chain resilience and decarbonization is non-economic spending and so inflationary. She also covers the US-China conflict as a working part of policymaking, and why the bank failures should not have surprised anyone given the pace of tightening.
On markets, she describes Central Bank Tension: easing is no longer costless, and she says markets have not priced this. She notes that index investors already hold roughly 65 to 70 percent US assets, while China and emerging markets look cheap. She explains gold through its opportunity cost, then adds the effect of frozen Russian reserves. The portfolio discussion draws on All Weather Logic, including shocks with no diversifier but cash, and she raises governance risk and the US debt-ceiling standoff.
Karniol-Tambour argues that slow-moving structural forces (deglobalization, automation, AI, and new fiscal and geopolitical pressures) are reshaping inflation and asset correlations, so investors whose assumptions come from the low-inflation, central-bank-put era are mispricing both US equities and the value of diversifiers like gold.
- Central banks no longer have free rein to ease because inflation is volatile, creating real policy tension that markets have not priced
- US equities and US tech already embed continued US dominance, leaving asymmetric downside, while Japan, China, and emerging markets are cheaper
- Gold's opportunity cost fell as real rates were near zero and sanctions on central bank reserves raised the value of non-financial stores of value
This asset isn't compiled yet
You're seeing its claims, ranked. Compile it to build the argument threads, weight them, and check each claim against your library — the full view.
Asset prices move because someone must buy and sell, and those buyers and sellers act for reasons beyond fundamental value, such as balance sheet constraints, regulation, and forced flows. Understanding who holds an asset and why they transact is therefore more useful than debating fundamental value.
“for an asset to move, someone has to buy and sell it”
Macro markets have a short history, so learning from it can easily produce bad conclusions, especially when past relationships no longer hold. Unlike image recognition, where massive data gives very low error, macro AI would have little data and high risk of being wrong.
“there's a pretty short history of macro markets”
Holding gold has an opportunity cost equal to the real interest rate on financial assets, since gold pays no yield. Gold tends to fall when real rates rise, as in taper tantrums, and it rallied when real rates were near zero or negative.
“the real interest rate is a good approximation for, what are you getting by holding financial assets?”
For the last 40 years, deflation meant central banks could ease whenever growth weakened, with no tension. That era is over: if the economy weakens, a central bank now faces a real choice between growth and inflation, and markets have not yet priced this tension.
“There's actually tension for the first time.”
Capital markets adapt slowly to structural change because investors trade on their own lived experience and what they are used to, so prices lag when the regime changes.
“capital markets tend to be slow to adapt when things structurally change because people trade based on their own experience”
An all-weather portfolio is built by finding assets that respond oppositely to the same macro shock, such as growth or inflation. Some shocks, like a Fed real-rate hike that raises discount rates on all future cash flows, have no diversifier except cash.
“can I find assets that have the opposite responses to that thing happening?”
An All Weather portfolio should be built to be immune to the big macro forces, growth and inflation, by holding assets that perform in both rising and falling growth and inflation, and geographic spread adds resilience because different regions cycle differently.
“the big forces you can be immune from, and I believe growth and inflation are good examples where there are a lot of assets that do well when growth is rising and growth is falling”
A bank that gathered a flood of deposits from venture-capital-linked startups, which came from money printing, had few loan opportunities. To earn returns it took duration risk, and when rates moved against it, it took losses.
“And so that imbalance between lots of money printed and not many people needing to borrow existed.”
After Russia's invasion of Ukraine, Western governments froze Russian access to dollar and euro reserves, showing that dollar assets can be confiscated. This made gold's opportunity cost less important, which is why central banks have shifted into gold even while the Fed tightened.
“they in some sense weaponized the dollars that we have”
The surprise at bank failures is misplaced because the fastest rate tightening in decades was designed to slow the economy, and it works through credit tightening that makes the credit system more vulnerable, though it does not identify which specific bank will fail.
“You forgot that when you massively raise interest rates, there's a point to doing that.”
Globalization shifted US manufacturing jobs to cheaper countries, mostly China, while automation made remaining workers far more productive. This produced a manufacturing productivity explosion, higher corporate profits, deflationary pressure that let the Fed keep raising rates, rising inequality, and populist politics.
“move a lot of those jobs to cheaper countries, mostly China”
The US-China conflict has moved from a topic of discussion to a core part of policymaking in both countries and in Europe and Japan, driving decisions about competitiveness and reducing reliance on the other side, which creates winners and losers.
“that conflict between the U.S. is very much heating up. It's in some sense, I have to say, metastasizing”
Over the last 40 years, deflation meant central banks could always ease when anything went wrong with no inflation tension, so risky assets were easy to profit from because every downturn was met with easing.
“every time anything went wrong, central banks would just solve it because there was no tension in what they were experiencing.”
Estimates from OpenAI and Goldman, using different methodologies, suggest AI could affect a much larger share of the labor market than globalization and automation affected manufacturing, and the effect could be faster and bigger.
“it's a much bigger bite in theory”
Companies are being told to spend on supply chain resilience and decarbonization, which is non-economic spending that does not reduce costs. Unlike the earlier cost-cutting capex wave, it is structurally inflationary and will slow the disinflationary effect of technology.
“it's the biggest wave of what I would call noneconomic spending companies have ever had to do”
Investors assume a baseline of functioning governance, but governance quality deteriorates and is hard to measure. Emerging markets that lost access to capital markets because of bad governance show it matters at extremes, and the US debt-ceiling standoff suggests the US is approaching that line.
“the quality of governance matters a lot. It underpins a lot what we're doing or kind of investing in markets.”
Long-term variables like demographics and government debt move so slowly that they define what an asset is and how it behaves, and they seep into portfolio construction. Investors should not trade them on three-month horizons but should revisit their assumptions when they shift.
“those longer-term variables, they go so slowly that they form your assumptions of what assets even are”
Being a good investor is mainly about curiosity and knowing what you don't know. As breadth grows, there are more topics where others have better expertise, so the job is to identify which areas of expertise the decision requires.
“you got to know what you don't know”
Emerging markets used to be a higher-risk version of the S&P 500 with high correlation, but that has changed because many emerging central banks now act independently and tighten aggressively against inflation, so correlations between stocks and bonds and across markets are less stable.
“a lot of emerging markets have gotten to beating their own drum, they're doing their own monetary policy”
For ordinary investors, more volatile inflation raises the probability of a debasement event that erodes real purchasing power, so the opportunity cost of gold matters less as a driver of the decision to hold it.
“the fact that inflation is more volatile certainly raises the probability that you're going to get some version of a debasement event”
Global investors hold about 65 to 70 percent US assets in index portfolios, so expectations of continued US winning are already embedded in prices and must be exceeded to earn returns. China, by contrast, is priced cheaply, so the pricing compensates for its risks.
“expectations of continued U.S. winning against everyone are already in your portfolio”
Market pricing assumes the Fed will always bring inflation back to 2 percent. Five years ago, almost no one would have expected markets to believe that inflation would simply go away, which reflects how well central banks anchored expectations for decades.
“markets are basically telling you, don't worry, the Fed is always right”
Emerging market stocks are underrated because the wave of post-pandemic money, including stimulus checks, went into crypto and tech rather than into the risky emerging markets where it usually goes, so the capital build-up that drives EM bubbles did not happen.
“They bought crypto. They bought tech. They used to buy emerging markets stock.”
Slow-moving secular forces like globalization and automation affect macro variables gradually over decades, so their effects only become visible in aggregate after long periods, and the economic, social, and political aftereffects are still being felt.
“when you have secular forces, they tend to be slow moving and affect a lot of things that matter for macro but affect them over time and over decades”
When social, environmental, and economic outcomes diverge, large cleavages form and political systems become hard to hold together. The globalization and automation era created such cleavages, and further divergence could make governance worse.
“it's hard to keep a political system together when you get too much polarization”
Equity prices imply only a modest economic slowdown with little recession risk, and that the Fed will cut rates from 5 to 3 percent quickly even though inflation remains high. Bond pricing implies inflation will be fine and that easing will come without a significant slowdown.
“the Federal Reserve is going to find that sufficient to go ease from 5% to 3% extremely quickly”
Because easing is already priced in, a strong economy with no recession leaves the Fed unable to cut and disappointing markets, while a weak economy also brings disappointment, since a recession would hurt stocks. Either outcome creates room for disappointment, so the risk is asymmetric to the downside.
“that's a lot of room for disappointment that can happen whether the economy is strong or weak”
Japan is highly underrated because investors have written it off after its past problems, yet it trades at much better valuations than the US, and Japanese policy is not fighting an inflation problem.
“I think Japan is highly underrated.”
The biggest worry is a more volatile environment with more volatile inflation and sharper policy shifts, possibly much faster than during globalization and automation, which changes what assets look like.
“the biggest thing I worry is just a more volatile environment”
Gold is underrated because many global actors see dollar risk, there are few good alternatives, and the geopolitical and inflation-volatility pressures are slow-moving secular forces that will not reverse quickly, so gold has a long way to run.
“I think it's got a long way to run.”
US stocks are way overrated because a recession is possible, inflation volatility is not priced in, Fed easing is already priced in, and therefore there is room for downside. The AI beneficiaries may still be underrated, but the broad market is overrated, especially outside a few mega-cap companies.
“Way overrated.”
If the economy stays strong, inflation will not fall without recession, so the Fed cannot ease as priced, and every day it fails to pivot is effectively tightening relative to what is priced in.
“every day, once we hit summer, the Federal Reserve doesn't pivot an ease, that's effectively a tightening relative to what's priced in.”
Investors should diversify geographically toward places facing less central bank tension, such as Japan and China, where policy is not constrained by an inflation problem. Japan's central bank has wanted higher inflation for a long time.
“they can look at geographies that have less of this tension, places like Japan or China”
China is priced terribly with attractive valuations that compensate investors for its many risks, though it is not a slam dunk for a strong decade ahead.
“Then on the flip side, China is priced terribly. You have pretty attractive valuations there”
For AI to matter for portfolio positioning, its disinflationary impact must be large enough to offset the massive inflationary forces currently pushing the other way, including reshoring and decarbonization spending.
“the big thing is it has to be big enough to offset the massive inflationary forces that are going the other way.”
Most high-quality modern capital markets data comes from a period when the US was the global superpower with little geopolitical instability since World War II, so the rise of China is an unusual variable relative to the historical data investors rely on.
“the vast majority of the high-quality data comes during a period where effectively America was the global superpower”
Nominal bonds pay a fixed rate and lose purchasing power when inflation is high, while inflation-linked bonds pay CPI and protect against inflation. Holding both creates a position that does not depend on which inflation outcome occurs.
“those two bonds are set up to have opposite structural fundamental biases”
Although investors were worried that the Biden administration would not get anything done, it passed major legislation and produced one of the largest industrial policy pushes in a long time, showing that policy shapes the economy.
“Biden gets elected, we were really nervous, it's going to be really impossible to get anything done. And then actually Biden got a lot done legislatively”
Governments in the US and elsewhere are now comfortable choosing winners and subsidizing favored industries such as semiconductors and green energy, a sharp reversal from the earlier stance that governments should not pick winners.
“So much more comfort saying, let's subsidize the things we want. Whether it's semiconductors or green energy”
Bridgewater is famously systematic and an enormous consumer and user of data in its investment process.
“Bridgewater is famously systematic and an enormous consumer and user of data in its investment process.”
Potential deflation is underrated because the people closest to a possible technology-driven deflationary boom rarely discuss inflation and deflation, though deflation is a real possibility depending on how fast technology moves.
“most of the people who are really close to what's likely going to be a tech miracle are not the same ones that talk about inflation and deflation all day”
Pandemic-era money printing and stimulus checks flowed into houses, crypto, and tech, and that flow is now being sucked out and causing bubbles in those places to deflate.
“all this money got printed, and everybody got checks into their house. ... the capital gets sucked out and causing bubbles to deflate in the places we all know well.”
Almost no one questions governance, rule of law, or institutional function in the US, so these factors are effectively priced for perfection, which makes them an underappreciated risk.
“It's sort of like priced for perfection in that sense”
Stan Druckenmiller described near-term market events as a 30-foot tidal wave and slow-moving demographic and debt forces as a 200-foot wave farther out, meaning the slower forces are larger and more consequential than the near-term events.
“some 200-foot tidal wave that's coming in the form of things like much longer-term concerns”
Sam Altman suggested that some investor will figure out how to use AI to earn fantastic returns, and Karniol-Tambour responds that nobody knows yet, but Bridgewater and others must invest effort to find out, since the technology is evolving too fast to judge.
“we'd be idiots not to go invest energy trying to figure that out”
Risk assets look about as attractive as they have been in a very long time, yet investing is hard because rallies are not automatic, and diversification needs to be rethought.
“risk assets are about as attractive as we've seen in a very long time”
Intercom's new AI customer service product reports that a staggering share of customer interactions can be handled end-to-end by LLMs without people, which Patrick treats as early evidence that AI will affect labor.
“a company called Intercom, which does customer service chat interfaces, announced a new AI-powered service, and the stats that they report on the amount of things that can be handled end-to-end by LLMs”
She declines to rate Bitcoin as overrated or underrated because public comments on crypto attract headlines that make the speaker seem like an authority on the topic.
“I never really want to say the word because every time you say the word, you get headlines”