Reserve Adequacy and Moat Decay Scores Are Both Measuring the Past

What the June FOMC minutes actually say

The minutes of the June 16–17, 2026 meeting, released July 8, describe balance sheet policy as reserve management, not stimulus. The Committee rolled over Treasury principal at auction, reinvested agency proceeds into bills, and authorized outright bill purchases specifically to maintain what it called an ample level of reserves. That is plumbing, distinct from quantitative easing aimed at financial conditions.

The detail worth sitting with is one sentence: repo rates dropped meaningfully below the interest rate on reserve balances in May, a classic early marker of reserve scarcity outrunning the balance sheet runoff schedule, structurally the same precursor that preceded the September 2019 repo spike. The Standing Repo Facility held the floor at a 3.5 percent offering rate with a $160 billion per-counterparty daily limit, and staff called it effective. Effective for one dealer under isolated May stress is not the same claim as effective under correlated, system-wide stress, and the minutes do not address aggregate SRF capacity if several large dealers draw on it simultaneously.

The rate path language carries more information than the headline. The Committee dropped its prior easing-bias language. Participant views on where rates should sit by year end then split into two camps: some at or slightly below the current target range, others above it. That is a bimodal distribution being handed to the market, not a converged forecast.

“Ample reserves” is a staff assessment issued after the stress had already passed. It is a description of last month, not a forecast of the next one.

Where the reserve-adequacy read breaks down

• Tail risk not addressed

  The SRF ceiling described in the minutes is a per-counterparty limit. Nothing in the document speaks to system-wide capacity under correlated demand, the failure mode a genuine liquidity event produces.

• Forward versus backward looking

  Ample reserves is a staff assessment issued after May's stress had already passed. It is not a model of when the next scarcity episode arrives. Bill issuance calendar effects, debt ceiling dynamics, and the pace of runoff are the actual triggers, and none of them are quantified in this document.

• Structural or temporary

  The SRF and bill-purchase reserve-management approach is structural, permanent post-2019 plumbing. The specific rate-path split described in these minutes is not structural. It reflects this cycle's inflation data producing genuine committee disagreement, and it will resolve or shift within a few meetings. Positioning that requires a single consensus cut path to work is exposed to a wider actual distribution of outcomes than that path implies.

A parallel problem in moat measurement

The AI Transformation Gap Index (Dean Barr, arXiv, submitted March 17, 2026, so several months old rather than brand new) scores firms and industries on three axes: exposure to AI-driven productivity gains, vulnerability to AI-enabled disruption, and net value creation or erosion. It draws on O*NET and BLS labor-task data cross-referenced against SEC EDGAR capital allocation disclosures. The underlying claim is sound: moat erosion often shows up first in a firm's own labor and capex mix shifting toward AI-substitutable tasks, measurable before it shows up in market narrative or the multiple.

The construction has the same lag problem as the Fed's reserve read, for a different reason. O*NET and BLS classifications update on a multi-year cycle. AI capability is moving faster than that. The index likely lags real disruption by twelve to twenty-four months, understating risk exactly at the inflection points where it matters most. It is also filing-dependent, which structurally favors large, well documented public companies and leaves the framework close to blind to the private-market disruptors most likely to actually erode an incumbent's position. A firm can carry a stable score and still be repriced violently on narrative alone, because the index is a business-quality signal on a quarterly-to-annual cadence, not a timing tool.

What this means for underwriting

Neither document changes how we approach position sizing on its own. What both point to is a discipline worth stating plainly: treat any risk-adequacy claim, whether it is a central bank calling reserves ample or a quantitative model calling a moat intact, as a description of the recent past rather than a forecast. The question to ask is not whether the measurement looks comfortable today, but how far behind the underlying risk that measurement structurally has to run.

For repo-market-adjacent exposure, that means underwriting against correlated SRF demand, not single-dealer capacity. For rate-sensitive positions, it means stress testing against both camps the Committee itself is debating, not the market's single modal path. For moat-decay scoring, it means using frameworks like AITG as one input into business-quality underwriting, not a substitute for it, and remembering that the fastest-moving threats to a moat are often the ones a filing-based index cannot see at all.