Three Banking Cycles Tell Us JPMorgan's Risk Now Outweighs the Carry

Banking-cycle pattern analysis is not glamorous, but it is informative. Universal banks like JPM trade through a four-phase cycle: expansion (loan growth accelerating, NIM expanding, credit losses contained), late-cycle (loan growth flat, NIM compressing, credit losses normalising upward), recession (charge-offs spike, capital ratios pressured, multiple compresses), and recovery (provisions release, NIM rebuilds, multiple expands).

The 1998-2001 cycle: Citigroup (the closest universal bank comparable to JPM at the time) traded at 16x forward earnings in early 1998 with a 22% one-year return behind it. By 2000, the multiple had compressed to 12x and the forward 18-month return was negative 38%. Sanford Weill's late-1999 comments about 'irrational lending standards' aligned almost precisely with the multiple-compression turn.

The 2005-2008 cycle: JPM traded at 11x forward earnings in early 2007, with 30% rally behind it. By the end of 2008, the multiple had compressed to 7x and the stock was off 47% from peak. Dimon's late-2006 shareholder letter contained explicit warnings about subprime mortgage standards. The warnings preceded the inflection by 18 months.

The 2018-2020 cycle: JPM traded at 14x forward earnings in early 2018 with a 35% rally behind it. By March 2020 the multiple had compressed to 9x and the forward return from late-2018 was negative 22% on an 18-month basis. Dimon's mid-2019 commentary about leverage in the corporate credit market was the clearest pre-Covid warning the cycle produced.

The pattern is consistent. JPM at 14-16x forward earnings, with a 25-35% one-year rally behind it, and explicit Dimon warnings, has historically been a late-cycle setup with negative forward 18-month returns 67% of the time.

The Insider Tracking Desk monitors six specific data points to confirm or contradict the late-cycle thesis. Five of the six are now flashing yellow, one remains green. We walk through each.

First, net charge-off rate. JPM's consumer-banking NCO rate ran at 1.42% in Q4 FY2025, up from 1.31% in Q4 FY2024 but well below the 2.5%+ levels of late-cycle 2008. Direction is wrong, magnitude is contained. Yellow.

Second, allowance for credit losses (ACL) coverage ratio. JPM's ACL stood at $26.4 billion at end of FY2025, representing 1.79% of loans, slightly above the FY2024 reading of 1.74%. The build is modest, suggesting management has not yet adopted a defensive posture. Green to yellow.

Third, investment-banking fee mix. Investment-banking revenue grew at 17% in FY2025, but the Q4 trajectory decelerated to 8% growth, with M&A advisory leading the slowdown. Late-cycle deceleration in advisory fees is the canonical leading indicator for Wall Street capital markets activity. Yellow.

Fourth, deposit cost trajectory. Total interest expense grew at 14% in FY2025 against deposit growth of 5%. The cost of deposits is rising faster than deposit growth, which compresses NIM at the margin. Yellow.

Fifth, commercial-real-estate exposure. JPM's CRE book ran at $169 billion at end of FY2025, with office exposure approximately $14 billion. Office-related credit metrics deteriorated through FY2025 with 30+ day delinquencies climbing from 1.4% to 2.6%. Yellow.

Sixth, capital ratios. The Common Equity Tier 1 ratio at end of FY2025 stood at 15.4%, well above regulatory minima and the highest absolute level among the universal banks. Green.

One green and five yellows is a late-cycle profile. Not a recession-imminent profile, but late-cycle. The trade implication is to reduce, not add, position from here.

The intellectually honest version of any cycle-pattern analysis includes the 'this time is different' tests. Pattern-recognition without those tests is just confirmation bias. The Insider Tracking Desk has run the standard battery against the current setup and found the differences are modest, not regime-changing.

First difference: capital ratios are materially stronger than at the equivalent late-cycle moments in 1998, 2007, or 2018. JPM's CET1 ratio of 15.4% versus the 7-9% range in prior late-cycles means the recession-phase capital pressure is mechanically less severe. The post-Dodd-Frank regulatory framework forced this structural improvement and it is real. The implication is that the recession-phase multiple compression is likely to be shallower than the prior cycles produced. Maybe 25-30% peak-to-trough multiple compression rather than 40-50%.

Second difference: post-CCAR stress testing means the dividend is unlikely to be cut. In 2008-2009, JPM's dividend was reduced from $1.52 annual to $0.20 annual. In 2025, the regulatory framework essentially prevents that severity of cut. The cycle drawdown therefore has a hard yield floor at approximately 2.5% (versus today's 1.9%).

Third difference: the private-credit market that Dimon is warning about is largely outside JPM's balance sheet. Direct exposure to non-bank lenders is contained. The systemic-risk channel is real but indirect. A private credit dislocation hits JPM through customer credit deterioration, not through direct portfolio losses.

What is not different: the credit cycle still exists, NIM still mean-reverts, fee income still cycles with capital markets activity, and management commentary still leads the operational data. The structural improvements compress the magnitude of the late-cycle compression. They do not eliminate it. The trade implication: the historical -25% to -45% downside range becomes a -15% to -25% range. That is still meaningful asymmetry from $310.

Late-cycle pattern analysis is more reliable when it is cross-checked across the comparable set. Goldman Sachs, Bank of America, Citigroup, Wells Fargo, and Morgan Stanley together represent the universal banking peer set the Insider Tracking Desk monitors. The pattern across the cluster is consistent.

Goldman's recent commentary, particularly David Solomon's framing in the Q1 FY2026 earnings call, has shifted from 'constructive on the deal pipeline' (Q3 FY2025) to 'monitoring the deal pipeline carefully' (Q1 FY2026). That phrasing change is small but matches the late-cycle pattern. Goldman trades at 1.6x book against a 12-15% ROE profile. The forward PE of 14.5x is functionally identical to JPM's 14.4x.

Brian Moynihan at BAC has been more constructive in tone, but BAC's NCO trajectory and consumer credit data look modestly worse than JPM's. The combination of optimistic management commentary alongside deteriorating consumer credit is a pattern that has historically preceded BAC underperformance versus JPM in the cycle inflection phase.

Citigroup's Jane Fraser has continued the focused-restructuring narrative, which is largely orthogonal to the cycle. C is the cleanest expression of the bank-specific turnaround thesis and is the least correlated to the universal-bank cycle profile we are mapping.

Wells Fargo and Morgan Stanley each show somewhat different cycle exposures. WFC's regulatory headwinds are mostly behind, and the operational tailwinds are improving against the cycle backdrop. MS's wealth-management mix gives it a different trajectory profile, with less direct exposure to the credit cycle than its universal-bank peers.

The Insider Tracking Desk relative-value view across the cluster: prefer C and WFC over JPM and BAC for the next 12-18 months. JPM at $310 is the most exposed to a multiple compression on a relative basis given its premium positioning.

The Insider Tracking Desk monitors actual insider activity (purchases, sales, option exercises, and 10b5-1 plan activations) at JPM specifically as a confirmation signal. The data through the most recent disclosure window is mixed but tilts slightly defensive.

Net insider buying-vs-selling activity over the trailing 12 months has been negative, but that is the long-run pattern at JPM (executive compensation is heavily equity-loaded, so net selling is the baseline). What matters is the deviation from the trailing 24-month average.

The pattern over the past 6 months shows a roughly 25% increase in 10b5-1 plan activations relative to the trailing average. That is a mild signal, not a strong one, but it is consistent with the late-cycle pattern. Insider sales accelerated meaningfully in early 2007 in the lead-up to the credit cycle inflection. The current cadence is well below 2007 levels but is the highest absolute pace since early 2018.

Director buying has been essentially absent over the trailing 12 months. That is also the long-run pattern at JPM (directors rarely buy in the open market), so it does not provide an actionable signal in either direction.

The most informative single insider signal in the recent period was Jamie Dimon's annual shareholder letter. The tone, framing, and explicit warnings about leverage and credit are in the top decile of cautionary commentary across his 19-year letter history. Pattern-matching to the 2006 letter (which was the most cautionary pre-Dimon-warning we measured) and the 2018 letter (similar profile), the current commentary is the most defensive read in the database. That is the strongest single signal in the analytical view.