Gita Gopinath on Why Interest Rates Have Surged All Around the World

5/13/2026 3:12:17 AM	short term up	5 days later	+6.52%	+6.52%
4/29/2026 9:31:07 PM	short term sharp up	5 days later	-2.66%	-4.00%
4/3/2026 12:20:17 AM	medium term sharp up	21 days later	-14.27%	-21.40%

4/15/2026 4:14:06 PM	medium term up	20 days later	+1.09%	+1.09%
4/3/2026 12:20:17 AM	medium term up	21 days later	+0.02%	+0.02%
1/31/2026 3:00:44 PM	short term cautious down	7 days later	-1.80%	+0.90%
12/9/2025 1:32:13 AM	medium term cautious up	20 days later	-0.82%	-0.41%
11/1/2025 2:00:00 PM	medium term cautious down	21 days later	-1.66%	+0.83%
9/17/2025 1:00:49 PM	long term up	60 days later	+0.71%	+0.71%
9/5/2025 11:17:40 PM	short term cautious down	7 days later	-0.30%	+0.15%
9/4/2025 12:03:38 AM	short term cautious up	5 days later	-1.32%	-0.66%
8/16/2025 2:06:17 AM	short term cautious down	7 days later	-1.52%	+0.76%
8/8/2025 8:34:45 PM	short term cautious up	7 days later	+1.59%	+0.80%
8/4/2025 4:53:03 PM	medium term cautious down	20 days later	+1.88%	-0.94%

The bond market selloff is a global phenomenon, and it's connected to the AI boom. There's a repricing of something more secular than transitory, related to AI's capital consumption and the ability of developed world to finance itself.

Joe Weisenthal

The proportion of AI-related issuance in the corporate bond market is now 50% of all investment grade issuance year-to-date, and almost 40% in junk-rated debt. Is it reasonable to think investors might prefer big tech IG debt over US Treasuries?

Tracy Alloway

Gita Gopinath

The bond market selloff is most worrisome due to public debt levels. Yields are up due to higher inflation expectations, a higher real rate (R-star), risk premia from large fiscal deficits, and the AI boom's capital expenditure. We have moved away from 'low for long' and the end of secular stagnation.

R-star has drifted from 0.5% pre-pandemic to 1% now, and AI capex may push it even higher. The combination of inflation, AI boom, fiscal deficits, and high public debt is pushing yields up globally.

What changed between pre-COVID and now that made high public debt matter, when before rates kept going lower even in Japan?

Joe Weisenthal

Gita Gopinath

Three things changed: 1) The unexpected AI boom increased private sector demand for capital. 2) Fiscal deficits are projected to stay at 7% of GDP in the US. 3) The marginal buyer of debt shifted from central banks to more volatile investors like hedge funds and non-bank financial institutions.

Foreign central banks are no longer major buyers of US Treasuries; capital flows are now from more volatile non-bank financial institutions, increasing yield curve volatility.

Gita Gopinath

Both the financial markets version (demand for capital raising real rates) and the real economy version (AI buildout creating inflationary pressure via competition for inputs) of crowding out are in play. The real rate channel is more important currently, while inflation is more driven by energy prices.

The first channel raises real rates even without inflation moving. The second channel creates inflationary pressure needing higher nominal rates. Right now, the real rate piece is more important.

If real rates are structurally rising due to AI, what should policymakers do? Should the Fed look through the AI boom?

Tracy Alloway

Gita Gopinath

It matters whether R-star is rising due to good (higher productivity growth) or bad (fiscal deficits) reasons. Regardless, with a 2% inflation target and R-star at 1%, we are looking at 3% nominal rates. The key question is whether the productivity boom reduces the need for rates above R-star.

Good R-star from productivity helps bring debt-to-GDP down. Bad R-star from deficits is problematic. The Fed needs to tease apart whether inflation is from energy/immigration/supply chains or from AI.

Is a 'disinflationary boom' or 'deflationary boom' scenario conceivable, where AI makes everything cheap and raises living standards dramatically?

Joe Weisenthal

Gita Gopinath

That scenario is possible, but no one puts a significant probability on it. There is very high uncertainty. There is no evidence yet of a major productivity wave. It's curious that markets price in the rosy scenario but not the many other possible scenarios.

Some analysts believe productivity could go up by 2 percentage points a year above the current 2%, but there is no evidence of that yet. The technology is valuable, but uncertainty is high.

Does the demand for AI mean we are seeing capital scarcity instead of a global savings glut?

Tracy Alloway

Gita Gopinath

We no longer have a global savings glut; proof is real rates going up. Instead, there is a 'gluttenous' demand for US equities, with foreign holdings at a historic high. The world has never been this invested in US equity markets.

Foreign holdings of US equities are at $40 trillion, about twice as high as a share of rest-of-world GDP compared to the peak of the dot-com boom.

How much of the upward pressure on rates and inflation is due to national resource hoarding and the decline of free trade, with every country duplicating capital investment?

Joe Weisenthal

Gita Gopinath

We have moved decisively away from the pure efficiency model to one where every country is building its own capacity. Energy security, defense expenditure, semiconductor chips, rare earths—there is so little trust that countries are spending a lot more. The list of sources of demand for capital is very long, while the supply of capital is only from aging demographics.

The only source of capital supply is aging demographics, which is why interest rates are not even higher.

How can policymakers navigate the tension of needing to reduce entitlement spending in elected democracies?

Tracy Alloway

Gita Gopinath

Policymakers are not really worried about debt-to-GDP except where bond markets force them. Historically, debt-to-GDP has been reduced through growth spurts (often from commodity luck or productivity), inflation, or default/restructuring. Developed economies are now moving into a space where borrowing costs are more volatile and sensitive to market conditions.

The US is still an exception with some privilege, but even Japan and Germany have seen their 10-year rates move up.

What does a debt crisis look like in a country that borrows in its own currency, like the US, UK, or Japan?

Joe Weisenthal

Gita Gopinath

A crisis in a developed economy would look like a credit crunch leading to a financial crisis: a sharp increase in borrowing costs, a slump in investment, and a debt overhang that slows growth and dynamism.

Despite all shocks, we haven't had a financial crisis in developed or major emerging markets recently, which has helped recovery. But if debt levels stay high, costs will eventually slow economies or trigger a crisis.

What would happen if a central bank explicitly capped long-term rates by buying government debt?

Joe Weisenthal

Gita Gopinath

If a central bank does that outside of a zero-lower-bound scenario, inflation expectations would drift up, nominal and real rates would rise due to inflation risk premium, and it would be the end of central bank independence. It only works for a short period before markets price it in.

Countries that have tried this usually end up in crisis. You get a tiny period of help, then much higher interest rates with no benefits.

Why has the global economy, especially the US, been so resilient despite multiple shocks? Are economists underestimating something?

Tracy Alloway

Gita Gopinath

Resilience came from massive government support (25% of GDP in advanced economies during the pandemic), which left households and businesses with stronger balance sheets, and from the AI boom holding up demand and trade. But the next crisis will find countries without fiscal space, so we should not assume this resilience is structural.

If we didn't have AI, growth would be much lower. Trade is being held up by AI inputs. But the next crisis will see much less resilience.

You wrote about the 'bliss trade'—the assumption of state support and a backstop. Is this mistaken?

Joe Weisenthal

Gita Gopinath

Yes, there is a mindset that the state will fix problems, seen in countries capping fuel prices and cutting energy taxes. This 'blitz' (big lasting state support) has helped resilience, but given high debt levels, it's increasingly questionable. Governments may move to unorthodox approaches like price controls and financial repression.

The expectation that state support will continue is not sustainable given current debt levels.

How much of the stress on economies is due to the fact that China can make almost any physical good cheaper and better?

Joe Weisenthal

Gita Gopinath

China runs a surplus on goods but a deficit on services (e.g., tourism). If China were so successful and productive, it would run deficits, not surpluses. The surplus is due to consumption suppression and a crash in investment (property market). This is not sustainable; other countries will put tariffs on China.

China's current account surplus is around 3% of GDP, down from 10% before the GFC, because they are big consumers of services. The weakness in the property market and declining investment explain the trade surplus.

What could be a catalyst for a real-life debt crisis?

Tracy Alloway

Gita Gopinath

What happens with AI and the productivity boom is very important. If AI fails to deliver productivity gains, or if there is a setback (e.g., hallucination problems), the pricing of debt could drop significantly, raising concerns about governments' ability to repay. Growth needs to come from good places.

If the AI story goes away, there will be a problem with fiscal positions around the world.

SignalTube