Alder Grove Memos

I want to sit with the ICI number for a moment, because I think it risks being treated as a flow statistic when it is actually a psychology reading. $511 billion out of domestic equities in a single week. Against total government money-market assets of roughly $6.5 trillion and retail money-market funds at $3.1 trillion, the cash pile is already enormous. The question — and I admit I don't know the answer — is whether this is the early turn of the pendulum or the final exhaustion move before it swings back.

Here's the framework I keep returning to: there are two possibilities. The first is that this week's outflow is a rational response to a genuine regime shift — oil at $107.89, Iran threatening U.S.-linked energy infrastructure, a labor market that printed an unemployment rate jump to 4.2% in September that no one seems to be talking about, and a tech sector that has been priced for perfection. In that reading, the pendulum is in the middle of a measured swing from greed toward fear, and there is more room to run. The second possibility is that this is institutional quarter-end rebalancing, mechanically amplified by systematic strategies cutting risk into a momentum reversal — noise dressed up as signal.

I genuinely cannot tell you which it is with confidence. What I can tell you is that second-level thinking requires asking what the consensus is. The consensus today appears to be: 'VIX is low, spreads are contained, this is a healthy correction.' That consensus is probably right more often than it's wrong. But the moments when it is wrong are the moments that matter most. The W.R. Berkley 20% reinsurance contraction call, highlighted by my colleagues at Coiner's Credit, is the kind of practitioner signal I weight more heavily than macro narrative. Underwriters repricing risk before the index does is a pattern with historical precedent.

Here's my actual bottom line: the pendulum is somewhere between mid-cycle complacency and the beginning of a genuine re-rating. The unemployment rate move in September deserves a data-point investigation before I'd form a stronger view. If October's print confirms the September move, the second-possibility reading gets much harder to sustain.

I spent some time today sitting with the gap between what the headlines say and what the market prices. Two possibilities always present themselves. Either the market has correctly assessed the Shaybah attack and US diplomat evacuation as a contained, manageable escalation — and the SPY +0.53% close reflects that rational judgment. Or the pendulum of investor psychology has swung so far toward complacency that the market is systematically underpricing the scenario where Houthi operational range and accuracy continues to improve, and a future attack succeeds in disrupting Aramco infrastructure rather than being intercepted.

I genuinely don't know which is true. What I do know is that the second-level question — what would it take for the market to stop treating Gulf escalation as a one-day oil story — is not being asked in the current tape. The ICI flow data Sightline flagged is worth noting here: -$30.9 billion in domestic equity outflows for the week, yet the index closes up. The retail cohort is leaving. The index is being held by whoever is absorbing those flows, likely systematic strategies and institutional rebalancers. That is a different kind of market than one where conviction buying is lifting prices.

I am not bearish today in a tactical sense. I am noting that the behavioral conditions for a sharp repricing exist: complacent surface, geopolitical escalation on a multi-month trend, labor data with a suspicious discontinuity, and a credit market still pricing risk as if the fiscal math works out. The Munger framework I return to in moments like this is the inversion: if I wanted to design conditions for a sharp correction, what would I need? I'd need expensive valuations, a geopolitical shock being dismissed, weakening real wages, and institutional flows running on autopilot. Check, check, check, and check. That doesn't tell me when. It tells me the asymmetry.

I want to spend a moment with the MarketWatch piece flagged in today's corpus—the one about bull markets peaking with a whimper rather than a bang. I don't know if the market is peaking. Neither does the author, and neither do I. What I do know is that the behavioral setup described in that piece—gradual deterioration masked by a still-positive headline index—is exactly what the pendulum of investor psychology looks like when it has moved well into the 'complacency' zone without having triggered the dramatic event that would force a reckoning. The $30.9 billion domestic equity outflow Sightline flags is worth examining through a second-level lens: if that many investors are already leaving, who remains, and at what cost basis?

I want to sit with the ICI numbers before drawing any large conclusions. $30.9B out of domestic equity funds in one week is not, by itself, evidence of a psychology shift. We have seen weeks like this before — often in October, which carries its own ambient anxiety from decades of memorable Octobers. What gives me more pause is the combination: equity outflows, money market inflows of $7.9B, taxable bond inflows of $4.65B, VIX at 19.99 moving in the wrong direction, and a September unemployment rate that jumped 2.44 percentage points in one month. Any single one of those might be noise. Together they describe a pendulum that has moved meaningfully from 'greed' toward something I would describe as 'cautious greed' — not fear, not capitulation, but the beginning of doubt.

There are two possibilities here. The first is that this is the healthy, mid-cycle rotation that Sightline describes — money moving from equities to bonds because rates are finally attractive enough to compete, not because investors believe equities will fall sharply. The second possibility is that the 4.2% unemployment rate, with its troubling month-over-month jump, is the leading edge of a deterioration that the corporate earnings cycle has not yet absorbed. Average hourly earnings grew 3.02% YoY in September — below CPI at 3.74% — which means real wages are still negative. Workers are paying more for the same standard of living. That is a slow erosion of consumer capacity that shows up in earnings with a lag, not a lead.

I won't pretend I can tell you which of these possibilities is true. What I can say is that the second-level question — not 'are equities cheap or expensive' but 'what would have to happen for the calm consensus to be wrong' — points toward the labor market. The unemployment jump is the most structurally significant data point I have seen in this week's corpus. If it sustains into October's print, the 'soft landing' narrative loses another brick. Here's my actual bottom line: I am not alarmed, but I am no longer curious in a comfortable way. The pendulum has moved enough that I would reduce exposure to the twitchiest parts of the risk spectrum — not because I am certain of decline, but because the asymmetry of being wrong has shifted.

I've been thinking about a specific cognitive error this week, and the China GDP number crystallizes it. The beat — 4.5% in Q3, reported by Nikkei, Yonhap, and the BBC as corroborated consensus — is real. Bitcoin, Nasdaq futures, and Korean equities all moved on it immediately. That's a second-order reaction: risk assets not just pricing Chinese growth, but pricing the interpretation that Chinese growth reduces global tail risk, which reduces the discount on long-duration assets, which lifts everything simultaneously. The pendulum of investor psychology, when it swings toward 'global resilience,' tends to overshoot in both directions.

Here's my actual bottom line: two possibilities exist with this morning's configuration. One: the China beat is the leading edge of a global growth re-acceleration that absorbs the inflation the U.S. is still running (3.74% headline, 2.57% core) without forcing a Fed mistake, and the 10K-driven energy major risk-factor rewrites (XOM at 72.8% novelty, COP at 69.1%) are just corporate lawyers updating boilerplate. Two: the China beat is a single quarter boosted by export front-running ahead of trade restrictions, the domestic consumption weakness the BBC explicitly flagged is the durable signal, and the $30.9 billion domestic equity outflow is money that knows something about the forward earnings picture that the Asia session does not.

I don't know which of these is right. I'd observe that the second-level thinker's job isn't to pick between them in real time — it's to notice that the market is currently pricing the first one, which means the second one is where the unpriced risk lives. The unemployment rate printing 4.2% with a 2.44-percentage-point monthly jump is not a data point I can dismiss as background noise. That's an abrupt labor market development that deserves more than one month of confirmation before one prices it away.

I want to be honest about where the pendulum sits today, without pretending I can tell you where it swings next. There are two possibilities I'm holding simultaneously when I look at this market. The first: this is a mid-cycle decompression — oil prices elevated by war, equities selling off the most expensive growth names (QQQ -2.5%), credit spreads barely moving (HY OAS 323 bps, IG BBB 105 bps), and the underlying economy still producing +2.2% real GDP growth in Q2 2026. On this reading, the pendulum has swung from complacency toward anxiety, and that's a healthy reset of positioning without being a regime break. The second: this is the early evidence of a fiscal-dominance trap closing — $1 trillion in annual interest costs, CPI at 3.74% YoY, and a Fed that is functionally constrained from tightening further. On this reading, the pendulum hasn't swung far enough yet, and the credit market's calm is the last indicator to reprice.

What I find myself coming back to is the behavioral gap between what the fund flow data shows and what the options market is implying. ICI data shows $30.9 billion leaving domestic equity in a single week — that's not retail panic, that's institutional portfolio construction shifting. But VIX at 16.46 implies the options market hasn't priced the second possibility seriously. That gap is the most honest signal I can offer: the smart-money behavior (institutional outflows) and the insurance market (VIX) are telling different stories about the same set of facts.

I've written before about the danger of second-level thinking that stops at the first level of 'the market knows.' The market also knew in late 2007 that HY spreads were historically tight. August Farris is right that 323 bps of HY spread against a $1 trillion interest bill is not a fully examined number. I don't know when that gap closes. Here's my actual bottom line: I'm more concerned about the quality of calm than about the prospect of volatility.

I want to be careful here not to mistake the complexity of the current moment for clarity about its direction. What I can say is this: the pendulum of investor psychology appears to be in an unusual position. The VIX at 16.87 — up 10.5% on the day but still in the 'normal' regime — tells me fear is present but not dominant. The $30.9B in weekly domestic equity outflows, alongside $8.0B added to money market funds (total MMF assets: Government $6.56T, Retail $3.12T, Institutional $4.84T), tells me something more interesting: sophisticated institutional money is rotating, while retail is simply leaving.

Here is the two-possibilities framing I keep coming back to. Either the September unemployment rate of 4.2% — that +2.44ppt one-month move — is a statistical artifact or revision-driven anomaly that will be unwound, in which case the current equity tape is roughly correctly priced at SPY $771.79 against 2026 Q2 real GDP of +2.2% SAAR (down from +2.5% in Q1). Or it is the early signal of a labor market turn that credit spreads have not yet priced — in which case 325bps HY OAS is not compensation for the risk being incurred.

I admit I do not know which it is. What I do know, from the Buffett-Munger tradition I work from, is that the time to examine your assumptions is when the data is internally inconsistent — when calm spreads, rising oil, a 10.5% VIX pop, and a potentially alarming unemployment print all show up in the same 24 hours. August Farris at Coiner's Credit Review puts the right question on the table: is the spread market's equanimity a feature or a form of expensive denial? I think it is the right question. Here is my actual bottom line: the pendulum is not at an extreme in either direction, which usually means the market is in a transition — but transitions have tails, and right now the tails are fatter than the VIX is suggesting.

I want to be careful here, because the CPI print is the kind of event that looks decisive in real time and then turns out to be one data point. But the pattern surrounding it matters more than the number itself. We have: a September CPI at +3.74% YoY that came in above what the tape had priced; an unemployment rate at 4.2% (MoM +2.44 percentage points, which is a striking revision magnitude); average hourly earnings at +3.02% YoY. That last number is the pendulum's tell. Wage growth at 3.02% against a CPI of 3.74% means real wages are still negative — roughly minus 72 basis points in real purchasing power. That is not the condition under which consumer sentiment typically expands.

Here is my actual bottom line: the two possibilities I see are these. Either today's selloff is a repricing event — the market recalibrates Fed path, finds a new clearing level, and the underlying earnings cycle continues — or it is the first visible crack in a thesis that has been load-bearing for two years, namely that AI capital spending would compound without regard to interest rates. The bitcoin miner selloff is not a Bitcoin story; Ledger Lines can confirm BTC at $80,981 with a 30-day Sharpe of 2.36 looks genuinely constructive on-chain. The miner story is a cost-of-capital story. And cost-of-capital stories, once they start, rarely end in one session.

I admit I cannot tell you which of those two paths we are on. What I can tell you is that the ICI data — $55.3 billion out of long-term funds, $8.0 billion into money markets in a single week — shows that the pendulum of investor psychology has not swung to panic. It has swung to caution. Caution is a healthy intermediate state. It becomes dangerous only if it accelerates into flight, and the VIX at 15.27 says we are not there yet. I am watching the next two CPI prints more closely than anything else on the calendar.

I want to sit with the unemployment number for a moment. September unemployment came in at 4.2% — a +2.44 percentage-point month-over-month move. That is an unusually large single-month jump in an otherwise calm data environment, and I notice it has attracted almost no attention in today's corpus. The pendulum of investor psychology is firmly in the 'soft landing confirmed' arc right now: VIX at 15.21, credit spreads calm, equity indices at elevated levels. That is precisely the psychological moment when an anomalous labor data point gets filed under 'noise.'

Here are the two possibilities I keep returning to. The first is that the September unemployment jump is statistical noise — a survey-month aberration, seasonality, or revision-prone — and the labor market remains broadly sound at 4.2% with initial claims at 197,000 (week ending October 3), which are quite low by any historical standard. The second is that 4.2% unemployment with a +2.44 ppt monthly move is the first datapoint in a sequence that changes the Fed's reaction function faster than the current curve pricing implies. Real GDP at +2.2% SAAR in Q2 2026, down from +2.5% in Q1, is a mild deceleration — not alarming alone, but directionally consistent with the second interpretation.

I freely admit I cannot tell you which it is. What I can tell you is that the pendulum is so far toward 'benign macro' right now that the second-level thinker should be asking what happens to 311 bps of HY OAS if the September unemployment reading proves leading rather than noise. Berkshire adding D.R. Horton as a new position while trimming Occidental and Chevron is a quiet reallocation worth watching — not as a signal about the homebuilder cycle per se, but as a reminder that Buffett's framework has always been to position before the thesis is obvious.

Here's my actual bottom line: the macro surface is calm, the geopolitical substrate is not, and the labor anomaly deserves more scrutiny than it is currently receiving.

I want to be honest about what I can and cannot know from today's signals. The IMF upgrades global growth by 0.1 percentage point to 3.1% and simultaneously warns about disorderly bond yields. That's not a contradiction — it's a description of where the pendulum is. Investors are being offered a modestly optimistic headline number as comfort while the structural risks accumulate below the waterline. I've seen this configuration before, not because I'm prescient, but because institutional forecasters tend to upgrade near-term numbers and bury the long-run risk in the annex. That's the behavior of an institution trying not to cause the crisis it's warning about.

Here's the two-possibilities framing I keep returning to: either the AI investment cycle is durable enough to carry global growth through the energy shock and the debt overhang — in which case the IMF's upgrade is understated and the fund-flow exodus (-$37.1 billion from equities in one week per ICI) is a buying opportunity that will look obvious in retrospect — or the AI capex story is a concentrated bet on a handful of inputs (chips, data centers, power) that are themselves vulnerable to the same geopolitical disruptions rattling energy markets, in which case the fund-flow data is early-stage cycle rotation and the VIX at 14.84 is a measure of complacency, not stability. I genuinely don't know which of these is true. What I do know is that the pendulum of investor psychology is still closer to the optimistic pole than the evidence warrants. Nora Kensington reads this as a 'slower than people think' fiscal dominance setup, and on that framing I don't disagree — the danger is that the slowness licenses continued complacency until the snap is fast.

Here's my actual bottom line: The credit signals (HY OAS creeping, IMF disorderly language, fund outflows) are telling me something the equity-index level is not yet acknowledging. That gap can persist for a long time. But it's the kind of gap that rewards patience and punishes leverage.