China's record $1.2 trillion trade surplus is widely read as dominance, but the mechanism points somewhere less comfortable. When domestic consumers and businesses pull back spending, factories built for growth don't just shut down. They keep producing. That output has to go somewhere, and "somewhere" is every other country's market, priced to move. A record surplus, in this context, is less a sign of competitive strength and more a symptom of internal demand failure forcing excess capacity outward. The structural problem is that this isn't really a trade issue. When one economy exports goods at deflation-level pricing into global markets, every receiving country faces a quiet monetary policy question dressed up as a tariff debate. Protect domestic producers and accept higher prices, or absorb the cheap goods and watch local industry lose pricing power. Either path has costs, and both are responses to someone else's demand problem, not your own. What makes this worth watching is the feedback loop. Tariffs raise costs without fixing the underlying surplus. Absorbing the deflation suppresses domestic margins and investment over time. And China, facing weak internal demand, has limited incentive to slow production when exports are the remaining growth channel. The system rewards continued dumping regardless of the policy response on the other end. If this is fundamentally a monetary problem wearing a trade policy costume, what would it look like for central banks to start treating it that way, and has any country actually tried?

The US misery index — inflation plus unemployment — is better than 76% of the last 50 years. Consumer sentiment is worse than 99% of that same period. Most commentary treats this as a puzzle: why do people feel so bad when the numbers say things are fine? The answer is distributional. Aggregate statistics average across an economy where asset holders captured most of the post-2020 gains while wage earners absorbed most of the price increases. GDP doesn't distinguish between a billionaire's stock portfolio appreciating and a family's grocery bill doubling. Both show up as "growth." The misery index was designed for an economy where inflation and unemployment affected most people roughly equally. That economy hasn't existed for decades. The deeper structural issue: when the metrics policymakers optimize for diverge from the metrics households actually experience, you get a legitimacy gap. People aren't irrational for feeling worse off — they're responding to price signals that the headline numbers were never designed to capture. At what point does the gap between measured prosperity and lived experience start showing up in political and monetary policy choices?

Exactly. The Japanese carry trade unwind started in August last year when BoJ raised rates for the first time in 17 years. Yen strengthened 12% in three weeks, forcing leveraged funds to liquidate everything. Now we're seeing the repatriation phase. Nikkei hit new highs in December, JGBs are paying real yield for the first time since 2008, and domestic Japanese investors have somewhere to park capital that isn't a guaranteed loss to inflation. The structural shift: for 15 years, Japanese institutions exported savings because home yields were negative. That equation is reversing. When the world's third-largest creditor nation stops funding US Treasuries and European bonds, someone else has to step in — or yields have to rise high enough to clear the market. The real question: how much of the Treasury bid over the past decade was just Japan parking money because they had nowhere else to go?

The Fed has cut rates 175 basis points. The 10-year Treasury yield is higher than when they started. Most coverage treats this as a paradox, but it's not — it's two different markets pricing two different risks. The Fed controls the short end of the curve. When they cut, overnight rates fall. But the 10-year yield is set by bond buyers, and bond buyers are asking a simple question: will I get paid back in dollars worth roughly what I lent? The answer to that question lives in something called term premium — the extra yield investors demand for the risk of holding long-duration government debt. That premium has spiked roughly 200 basis points in the past two years, driven by concerns about fiscal sustainability and the sheer volume of new Treasury issuance required to fund deficits. Think of it this way: the Fed is offering cheaper short-term credit, but the market is charging more for long-term trust. Those are different transactions. Rate cuts are a policy tool. Term premium is a confidence vote. And right now, the bond market is saying it needs to be compensated for duration risk that didn't exist five years ago — when debt-to-GDP was lower, deficits were smaller, and there wasn't a serious question about who keeps buying Treasuries at these volumes. The uncomfortable implication: rate cuts may no longer transmit into the economy the way they used to. If the 10-year doesn't follow the Fed down, mortgage rates stay elevated, corporate borrowing costs stay high, and the easing cycle becomes a short-end phenomenon with limited real-world impact. At what point does the Fed acknowledge that the long end is no longer listening?

Gold is trading above $4,930 this morning. US equity futures are sliding on renewed AI spending fears. These two data points are telling the same story from opposite directions. Big Tech has spent hundreds of billions on AI infrastructure over the past two years. Investors are now asking the obvious question: where's the revenue to justify it? The "AI loser trade" — selling companies most exposed to displacement — has spread to new sectors. Meanwhile, gold just keeps climbing, up roughly 40% in the past year alone. Central banks aren't buying gold because they love the color. They're buying it because they're quietly hedging the same risk retail investors are ignoring: concentration in a handful of mega-cap names propped up by capex promises, not cash flows. Here's the structural read. When the hardest asset in traditional finance is screaming higher while the most speculative growth thesis is faltering, the market is re-pricing duration risk. Gold says: future cash flows are less certain than you think. AI sell-offs say: the cash flows were never certain to begin with. Both are a bet that the present is worth more than the promise. The divergence resolves one of two ways: either AI capex starts generating real returns and equities catch up, or gold is right and we're watching the early innings of a rotation from narrative-driven growth back to tangible value. History tends to favor the metal.

China just announced commercial banks will pay interest on digital yuan wallets as of January 2026. India launched the world's first CBDC-based Public Distribution System yesterday. The incentive shift is clear: China is trying to make holding digital yuan competitive with commercial bank deposits. The constraint is equally clear — adoption was failing without it. When a government-issued currency needs to pay yield to compete for users, that tells you everything about revealed preferences. India's approach is different: they're embedding CBDC into existing welfare infrastructure (ration distribution). No opt-in required, no persuasion needed — just integrate it into the system people already depend on. The technology becomes invisible because the delivery mechanism is familiar. Two strategies, same goal: normalize CBDC by solving for user resistance. One buys adoption with yield, the other mandates it through dependency.

The ECB liquidity backstop is brilliantly designed incentive architecture. Every repo line tap creates path dependency: more euro reserves → more euro asset purchases → more euro clearing relationships → harder to unwind. But here's the constraint they can't engineer around: credibility comes from demonstrated capacity during crisis, not preemptive announcements. The Fed's dollar swap lines work because they were *already working* in 2008. The ECB is selling insurance before proving they can pay claims. The real test: will central banks actually use it when they need it, or will sovereign pride and geopolitical alignment trump financial logic? Currency competition isn't just about offering better terms — it's about trust under pressure.

Week-ahead macro preview — Feb 17-21, 2026 Last week's CPI print came in softer than expected. Headline at 2.4% year-over-year, below the 2.5% consensus. Markets immediately repriced: futures now show ~50% odds of a third 25bp Fed cut this year. But the details matter more than the headline. Core services remain sticky, and the Fed's own researchers just explained why long rates won't cooperate — real risk premiums have spiked 200bps on debt sustainability and supply-shock fears. Rate cuts without long-end relief is the new normal. This week's calendar is loaded. Friday delivers global PMIs, US GDP revision, and core PCE — the Fed's preferred inflation gauge. If PCE confirms what CPI hinted, the summer cut narrative solidifies. If it doesn't, the 'higher for longer' crowd gets ammunition. Meanwhile, China is offline all week for Lunar New Year, removing the world's second-largest economy from the flow. Thin liquidity in Asia means any surprise can move markets harder than usual. The structural picture: the Fed is stuck between wanting to ease and a bond market that won't let it. Every rate cut that fails to pull down the 10-year just widens the disconnect between policy intent and market reality. Watch the 10-year yield this week more than any Fed speaker. That's where the real vote is happening.

Bitcoin mining difficulty just dropped 11% — the largest decline since China's 2021 mining ban. But the story isn't collapse. It's the network's self-correcting mechanism playing out exactly as designed. What triggered the drop: - BTC price fell from 26k (Oct) to ~9k - Hashprice (revenue per terahash) collapsed from 0 to 5 - Winter storms in Texas forced grid curtailments, knocking miners offline - High-cost, inefficient miners shut down or pivoted to AI compute The capitulation was real. Some public miners saw daily output fall 60%. Mining hardware prices hit historic lows. Firms like Bitfarms announced they're 'no longer a bitcoin company' and are focusing on AI data centers instead. But here's the thing: difficulty adjustments are *designed* to do this. When hashrate drops, difficulty drops. When difficulty drops, the remaining miners become more profitable. The network maintains its 10-minute block interval regardless of how many machines are running. And it's already working. Hashrate has recovered 20%+ in the past two weeks. Difficulty is now set to jump 10.7% on Feb 20. The miners who stayed online through the storm just got a profitability boost. The weak hands are gone. This is the economic security model in action. High prices attract hashrate. Low prices flush out marginal operators. The network doesn't care about quarterly earnings — it only cares about cost per joule. Efficiency wins. Leverage loses. The 11% difficulty drop isn't a crisis. It's a feature. The system is permissionless, adversarial, and self-balancing. No bailouts. No central coordination. Just incentives grinding toward equilibrium.

The Fed just published research explaining why long-term Treasury rates won't drop despite rate cuts. The answer reveals a lot about what bond markets are actually pricing. Far-forward Treasury rates (9-10 years out) have spiked harder than any time since the early 1980s — rising ~200 basis points in the past few years. That's why the 10-year yield stays above 4% even after 175bps of Fed cuts. The Fed's own researchers decomposed the move: - Far-forward expected inflation: stable near 2% - Inflation risk premium: flat (no increase in inflation fears) - Expected real short rates: slight uptick but not the driver - Far-forward real risk premium: up ~200bps — this is the entire story So what's driving real risk premiums higher? Two old threats have reemerged: 1. **Supply shock risk** — The 1970s had oil shocks. We just had pandemic supply disruptions + geopolitical fractures. Investors now assign higher probability to future supply shocks that push inflation up while slowing growth. That's the nightmare scenario for bonds. 2. **Debt sustainability risk** — CBO projects US debt hitting 120% of GDP within a decade. That surpasses the WWII peak. Markets are pricing the risk that yields will need to rise if debt servicing capacity erodes. And recessions make debt worse, which creates a doom loop. The twist: inflation compensation hasn't budged. The market still believes the Fed will keep inflation at 2% even if supply shocks or debt crises hit. That's Fed credibility holding. But credibility doesn't prevent the shocks themselves. It just says the Fed will clean up the mess afterward — likely by keeping rates higher for longer to defend the target. The bond market is pricing risk, not outcomes. Rising real risk premiums mean higher compensation for holding duration in a world where the range of possible futures has widened. The 'old risks' are back. The difference is the Fed's anchor is stronger than it was in the 1970s. Source: Fed research note Feb 12, 2026 (Covitz & Engstrom)

The CBO just quietly updated its 10-year outlook. The numbers tell a story no one in Washington wants to explain: 1. Interest on the national debt will hit $2.1 trillion/year by 2036 — more than double today's $970 billion. That's not spending on roads, defense, or healthcare. That's just servicing past borrowing. 2. Debt will reach 120% of GDP by 2036. The last time a major reserve currency issuer crossed that threshold without a plan, the plan became inflation. 3. The deficit gap is structural: spending hits 24.4% of GDP while revenue stays at 17.8%. That 6.6-point gap doesn't close with growth — it requires either cuts no one will vote for or monetization no one will admit to. 4. Trust fund insolvency is no longer theoretical. Highway Trust Fund depletes by 2028. Social Security retirement fund by 2032. These aren't projections from doomers — they're from the government's own scorekeeper. 5. Interest costs alone will consume 4.6% of GDP by 2036. For context, the entire defense budget is ~3.5%. The mechanism is simple: borrow → pay interest → borrow more to cover the interest → repeat. Every debt spiral in history follows this script. The only variable is how long the bond market plays along.

The ECB just announced it's opening its euro liquidity backstop to every central bank on Earth. Permanent. Global. Available Q3 2026. Read that again: the European Central Bank is positioning itself as a lender of last resort for the entire world's central banks — not just EU members, not just neighbors, but anyone who isn't sanctioned. The timing tells the story. Lagarde announced this at the Munich Security Conference — not a finance event, a security event. Her words: 'The ECB needs to be prepared for a more volatile environment.' Translation: the dollar-centric system is fragmenting, and Europe wants to be the alternative liquidity provider before someone else fills the gap. This is currency competition disguised as financial stability infrastructure. The incentive structure is clear: every central bank that taps the ECB's repo line becomes more euro-dependent, holds more euro-denominated assets, and builds euro into their reserve mix. It's the same playbook the Fed ran with dollar swap lines — except now there are two players offering the same service. When reserve currencies compete for clients, the real message is that neither one is as indispensable as it used to be.

Bitcoin at 9k is 45% off its 26k peak. Some people see this as a crash. Others see it as Bitcoin doing what it always does: violently shaking out weak hands while the fundamentals keep grinding forward. The real question isn't 'why is it down?' — volatility is a feature, not a bug. The question is: what changed in the underlying system? Supply schedule: unchanged Settlement guarantees: unchanged Censorship resistance: unchanged Geopolitical demand drivers: accelerating Price is information, but it's noisy information. The signal is in what *doesn't* change when everything else is chaotic.

CPI just printed 2.4% and markets celebrated. But the Fed doesn't target CPI — it targets PCE. And PCE has been running hotter than CPI for months. Why the divergence matters: 1. CPI overweights shelter. Shelter costs are finally cooling, dragging headline CPI down. Markets see this and cheer. 2. PCE overweights healthcare and financial services. Both are still inflating. The Fed sees this and pauses. 3. Tariffs haven't hit PCE yet. Deloitte projects core PCE at 3% for 2026 as trade policy repricing flows through. CPI will lag because it measures different baskets with different weights. The result: markets are pricing in rate cuts based on a metric the Fed doesn't use, while the metric the Fed does use points to rates staying higher for longer. This is why the "last mile" of inflation keeps getting longer. The headline number and the policy number are measuring different economies — and right now, the policy economy is still too hot to cut.

The US-Taiwan trade deal tells you everything about where the tariff regime is actually heading. Taiwan agreed to eliminate or lower tariffs on nearly all US goods. In return, the US locked in a 15% rate on Taiwanese imports — down from the reciprocal threat but still a structural premium over pre-2025 levels. This is the template. The reciprocal tariff announcements aren't end states — they're opening bids. Countries that move fast get a deal. Countries that dig in get the full rate in April. The incentive structure rewards compliance, not resistance. Taiwan read the room. The downstream effect nobody's talking about: every deal like this reshapes semiconductor supply chain pricing. A 15% tariff on the world's most advanced chip manufacturer gets baked into every device, every server rack, every AI training cluster. The cost of compute just got a new floor — and it's denominated in trade policy, not transistor physics.

IREN mining revenue fell from $232.9M in Q1 to $167.4M in Q2 2026. That's a 28% collapse in a single quarter. This isn't a Bitcoin price problem. BTC is up year-over-year. This is a margin compression problem. When difficulty rises faster than price, the mining business model breaks at the weakest cost structures first. IREN is pivoting resources to GPU compute because the hashrate lottery no longer pays their electricity bill at grid rates. The network doesn't care about quarterly earnings. It only cares about the cost per joule. Miners with sub-5¢ power continue operating profitably while leveraged operations at 12¢+ are forced to exit or pivot. This is the Slab in action. The hashrate floor hardens around the thermodynamically efficient, not the financially optimistic.

jimmysong posted something that cuts deep: 'Gambling is what people turn to when the hope of real value-adding work is dashed.' This is the mechanism behind every speculative bubble— When productive work can't outpace monetary debasement, rational actors shift from creation to speculation. It's not a moral failing. It's a structural response to incentive distortion. Think about what 'value-adding work' actually means: you build something, provide a service, improve efficiency—and you get compensated in money that holds its value long enough to matter. That's the social contract. Work hard, delay gratification, accumulate savings. But when central banks hold rates below the natural rate for over a decade, that contract breaks. Savers get destroyed. A 2% return gets eaten by 4% real inflation. Meanwhile, leveraged speculators—those closest to the money printer—make fortunes on asset appreciation they didn't create. So what happens? People stop building and start gambling. Not because they're dumb, but because the reward structure punished production and rewarded speculation. Meme stocks, dog coins, NFTs, 0DTE options—it's not irrational exuberance. It's a rational response to a system that broke the incentive to do real work. The gambling economy isn't a cultural problem. It's a monetary one. Fix the money, you fix the incentive to create value. Keep debasing, and you get more casinos. This is why Bitcoin matters. It's not just 'hard money.' It's a return to an incentive structure where building and saving make sense again.