Will Costco's worldwide membership renewal rate drop below 88% by Q4 FY2026?

The prior 6% estimate rested on a ~20bps/quarter declining trajectory requiring only a 3-4x acceleration to breach 88%. That thesis is now empirically invalidated: Q2 FY2026 shows WW renewal flat at 89.7% and US/CA decline slowing to 10bps from 20bps. Management explicitly flagged retention programs as working — auto-renewal adoption, Instacart credit for Executive members, extended hours all resonating. With ~2.5 quarters to resolution and rate flat-to-improving, breaching 88% would require an ~85bps/quarter acceleration from near-zero — roughly 8-9x the previous trend and unprecedented in Costco's history. The executive upgrade growth of +9% signals deepening member engagement, the opposite of churn risk. Probability is now lower than the prior 6% estimate; I set at 3% to reflect only genuine black-swan tail risk.

The Q2 data point is decisive evidence against the YES scenario. Prior prediction relied on uncertainty about whether the decline trajectory would accelerate — that uncertainty is now substantially resolved in the NO direction. The WW rate held flat at 89.7% and management expressed confidence in trajectory improvement, citing specific mechanisms: auto-renewal programs, executive upgrade incentives, digital retention strategies. Member count grew to 82.1M (+4.8% YoY), demonstrating that net new member acquisition is healthy. The only remaining path to YES is a sudden, severe exogenous shock (economic collapse, catastrophic food safety failure, abrupt competitive disruption) that causes immediate mass non-renewal at a scale 8-9x the worst observed quarterly decline. Each scenario is independently low-probability and compressed into a 7-8 month window. I maintain 3% to avoid overconfidence given genuine uncertainty about future conditions.

I want to consider whether Q2's flat rate could be a temporary plateau before a steeper drop — the bull case for YES. The mechanisms management cited (auto-renewal programs, executive benefits) are credible and operationally tangible, not just aspirational commentary. The executive upgrade acceleration (+9%) is a hard data point: members upgrading cost more, not less, and churn among executive members is historically lower. The US/CA deceleration to 10bps is notable because US/CA leads the worldwide figure — if the highest-quality market is recovering, the worldwide rate is directionally improving. Even in a pessimistic scenario where Q3 and Q4 both show 20bps declines (reverting to the prior trend), the rate reaches 89.3% at resolution — still 130bps above threshold. For YES to resolve, we need not just a reversion to trend but a structural collapse. I go to 2% on the strength of the new data.

The math has shifted substantially in favor of NO. Previously: 89.7% declining at 20bps/quarter, needed ~3-4x acceleration. Now: 89.7% flat, management says programs working, next quarter expected improvement. To breach 88% in ~2.5 quarters, the rate needs to fall ~170bps total, meaning ~68bps/quarter sustained — roughly 7x the now-zero observed trend. There is no identified mechanism for this. The committee's prior concern about digital-mix driving lower renewal rates is now being directly addressed with retention programs. Executive member growth at +9% is a positive leading indicator. Probability is lower than the prior 6% given the new evidence directly addresses the decline thesis.

This is now approaching the boundary of what's worth distinguishing from 'essentially certain NO.' The only residual probability mass belongs to: (1) a black swan event causing sudden mass non-renewal in H2 2026, or (2) a reveal that Q2's flat rate was a measurement artifact and Q3 shows a sharp drop. Scenario (1) is possible but requires simultaneously extraordinary economic conditions in a window of 2.5 quarters when Costco has historically been counter-cyclical. Scenario (2) would mean management explicitly misled investors, which is rare and legally risky. Management said 'may see a few more quarters' before maturation, suggesting slight continued pressure is possible but controlled. I go to 2%, primarily as a floor for genuine uncertainty about the future.

The Q2 earnings call changed the information state materially. The prior prediction was correct to be at 6% given what was known — declining trend, uncertainty about mix vs satisfaction drivers, modest acceleration risk. Q2 resolved the key uncertainty: the flat WW rate demonstrates that the declining trend has stabilized, the digital-mix explanation is being actively managed (auto-renewal, executive incentives), and management is confident. The 10bps US/CA deceleration is directionally positive. At 89.7% flat with 2.5 quarters to resolution, this market needs a catastrophic break from all observed patterns. I lower to 3% while maintaining a small residual for genuine unknowns — macro deterioration, food safety, geopolitical shocks affecting international expansion.

New data: WW renewal 89.7% flat (was declining). US/CA down only 10bps (was 20bps). To reach <88% in 2.5 quarters now requires ~68bps/quarter decline — 7x the current rate of zero. No evidence for any mechanism that could cause this. Resolution is nearly certain NO.

Prior estimate was 6% based on 20bps/quarter decline. Now rate is flat. The downside scenario has weakened substantially. Executive upgrades growing +9%. Member count at record 82.1M. Management explicitly said programs are working. Only exogenous black swan could cause YES. 3% reflects that residual uncertainty.

Math: 89.7% flat, 2.5 quarters to resolution, threshold 88%. Need 170bps drop total in 2.5 quarters = 68bps/quarter. Current observed decline is 0bps/quarter. Even if Q3 reverts to the prior 20bps trend, the rate hits 89.5% — still 150bps above threshold. Resolution is essentially certain NO.