How rates, inflation and the dollar push on every asset we track.利率、通胀与美元如何推动我们跟踪的每一类资产。
The combined push of today's rates and inflation on each asset we track.当前利率与通胀对我们跟踪的每类资产的综合作用。
Showing the strongest movers on each side. Bar length = how hard the measured forces are pushing; hover a label for the exact score and its drivers.仅显示各方向最强的资产。柱长＝实测力量推动的强度；悬停标签查看精确评分与驱动因素。
The direct hit first, then the knock-on effects in order. Hover a step for the full mechanism.先是直接冲击，再按顺序层层传导。悬停某一步查看完整机制。
What each shock has historically meant for prices — illustrations from past episodes, not forecasts.各类冲击在历史上对价格的含义——取自过往情形的示意，并非预测。
Everything above, with the machinery showing — for readers who want the receipts.以上全部内容的完整技术细节——供想查看依据的读者。
Curve: Bear steepener. recession dashboard low (0/4 flags).曲线：熊市陡峭。衰退面板低（0/4 项亮起）。
The three classic factors span三个经典因子合计解释 96.3% of curve variance — the textbook result, measured on our data.的曲线方差——教科书结论，在我方数据上实测。
Historically the BEST equity regime overall (~+22% annualised across post-2000 growth-led episodes) — value, banks and real-asset cyclicals over long-duration growth/utilities/REITs. BUT a TERM-PREMIUM/fiscal-driven steepener (long end up with no growth) is not the bullish kind — check 5s30s + breakevens to disambiguate; rising term premium also breaks the 'duration = safe haven' assumption.历史上整体最佳的股票状态（2000年后增长主导的episode年化约+22%）——价值、银行、实物资产周期跑赢长久期成长/公用/REITs。但由期限溢价/财政驱动的变陡（长端上行而无增长）并非看多版本——用5-30年与盈亏平衡来区分；上升的期限溢价也会打破“久期＝避风港”的假设。
The near-term forward spread (3m rate ~18m forward − 3m bill) statistically dominates the 2s10s — Engstrom-Sharpe (Fed 2018-19) find it renders the 2s10s redundant in a joint probit. We read 3m10y, not 2s10s, as the recession slope (Estrella-Mishkin γ = −0.53, −0.66). The genuinely ominous configuration is the curve RE-STEEPENING out of a deep inversion: the un-inversion has historically COINCIDED WITH or slightly led recession onset (1990 ≈3m, 2001 ≈6m after the curve turned positive) — a fresh dis-inversion is the late-cycle handoff, not an all-clear.近端远期利差（约18个月远期3月利率−3月票据）在统计上优于2-10年——Engstrom-Sharpe（美联储2018-19）发现在联合probit中它使2-10年变得多余。我们以3月-10年（而非2-10年）作为衰退斜率（Estrella-Mishkin γ＝−0.53、−0.66）。真正危险的形态是深度倒挂后曲线重新变陡：历史上重新转正与衰退开始大致同步或略微领先（1990年约3个月、2001年约6个月），因此刚出现的重新变陡是周期晚段的交接，而非解除警报。 Funds 3.63% vs ~2.5% nominal-neutral = restrictive. Wright (2006): the same curve slope is more restrictive at a higher funds level — a flat curve at 5% funds is a stronger recession tell than at 2%.联邦基金3.63% vs 约2.5%名义中性＝偏紧。Wright（2006）：相同曲线斜率在更高基金利率下更具紧缩性——5%基金利率下的平坦曲线比2%下更强的衰退信号。
| Spread利差 | Level水平 | %ile分位 | 63d Δ63日变动 |
|---|---|---|---|
| 2s10s (cycle slope)2-10年（周期斜率） | +0.37 | 65 | -15bp |
| 3m10y (recession slope)3月-10年（衰退斜率） | +0.70 | 81 | +9bp |
| 5s30s (long-end / term-premium)5-30年（长端/期限溢价） | +0.81 | 74 | -17bp |
| 2s5s (front belly)2-5年（前段腹部） | +0.12 | 69 | -1bp |
| Real 5s10s (TIPS curve)实际5-10年（TIPS曲线） | +0.31 | 57 | -30bp |
| Breakeven 5s10s (inflation curve)盈亏平衡5-10年（通胀曲线） | -0.03 | 67 | +19bp |
| TP-adjusted 2s10s期限溢价调整2-10年 | +1.15 | 89 | +0bp |
Curve speed (63d)曲线速度（63日）: real 10y实际10年 +43bp · front 2y前端2年 +38bp · |real-speed| pctile|实际速度|分位 74. Rate-of-change beats level — the SPEED of a real-yield move breaks equities more than its level.变动速度胜过水平——实际收益率移动的速度比其水平更能冲击股票。
Term-spread trend (2y smooth)期限利差趋势（2年平滑）: +0.09 · rising上行. The literature flags the smoothed, business-cycle-frequency component of the 3m10y spread as the curve read with genuine equity-premium content (Faria-Verona 2020) — but OUR calibration finds it regime-dependent (the IC sign flips post-2015), so we carry it as context, never a timing signal.文献认为3月-10年利差经平滑的商业周期频率成分是收益率曲线中具有真正股权溢价含义的读数（Faria-Verona 2020）——但我方校准发现其依赖状态（IC符号在2015年后翻转），故仅作背景，绝非择时信号。
Forward rates are the market-implied future short rates embedded in today's curve; carry+roll-down is the total return a duration holder earns if the curve is unchanged in a year (the cushion before price loss).远期利率是今日曲线隐含的市场预期未来短端利率；carry＋滚动收益是久期持有者在曲线一年不变时获得的总回报（价格亏损前的缓冲）。
Equity duration: growth stocks discount cash flows further out, so they behave as long-duration assets and de-rate when real yields rise (~2× the rate-sensitivity of value, BIS 2022); value/financials are short-duration. HONEST CAVEAT: the long-run correlation of the value-minus-growth factor with Δ10y is only ~0.10 (AQR/Asness) — the link is concentrated in episodes (the +25% growth→value rotation of 2022) and is largely NOT pure duration. Small caps are driven by the FRONT end (floating-rate debt cost), a distinct channel from value/growth duration.股票久期：成长股贴现更远期的现金流，因而表现为长久期资产，实际收益率上行时估值下修（利率敏感度约为价值的2倍，BIS 2022）；价值/金融为短久期。诚实提醒：价值减成长因子与10年期变动的长期相关性仅约0.10（AQR/Asness）——该联系集中于特定时期（如2022年+25%的成长→价值轮动），且大部分并非纯久期效应。小盘由前端驱动（浮动利率债务成本），与价值/成长久期是不同的渠道。
No fast real-rate repricing — the curve is not flagging near-term equity stress.无快速实际利率重定价——曲线未对近端股票压力发出警示。
VISIBILITY ONLY — a falling breakeven is a thermometer, not a forecast. Across 2008 / 2014–15 / 2018 / 2020 / 2022 / 2023 it led the risk cascade in ~0 clean cases; breakeven velocity has ~0 forward-drawdown edge and the credit/vol co-state adds nothing beyond VIX. Use this to disambiguate an in-progress move (which kind of fall?), never to front-run one.仅供观察——下行的盈亏平衡是温度计，不是预测。在2008/2014-15/2018/2020/2022/2023中，它几乎没有一次干净地领先风险资产的瀑布式下跌；盈亏平衡速度对前瞻回撤几乎没有预测力，信用/波动率联动相对VIX无增量。用它来辨别正在发生的变动（属于哪一类下跌），而非抢跑。
We tested every rate and inflation driver on this page for real predictive edge. None passed the bar to be scored — so everything here is shown as context to explain moves, never as a signal to trade on. The table below is the full honest record.我们对页面上每个利率与通胀驱动都做了预测优势检验，无一通过计分门槛——因此本页内容仅为解释行情的背景，绝非交易信号。下表为完整的诚实记录。
| Candidate leg候选腿 | Verdict判定 | IC dd (f/pre/post)IC回撤(全/前/后) | hi-tercile edge高三分位超额 | scored?计分? |
|---|---|---|---|---|
| Real-rate SPEED (63d rise) — 'speed breaks equities' | DIRECTIONAL | 0.137/0.052/0.218 | 7.1pp | — |
| Real-rate LEVEL (high real yields) | CONTEXT | 0.042/0.053/-0.007 | -0.9pp | — |
| Core-PCE-vs-target gap (sticky inflation) | DIRECTIONAL | 0.073/0.058/0.122 | 0.6pp | — |
| Inflation re-acceleration (3m>12m) | DIRECTIONAL | 0.065/0.066/0.064 | 4.3pp | — |
| Expectations unanchoring (market>model) | INVERTED | -0.046/-0.05/-0.09 | -6.1pp | — |
| TP-adjusted curve inversion (flip: low=stress) | CONTEXT | 0.005/0.01/0.04 | -0.7pp | — |
| Nominal-rate SPEED (63d rise) | DIRECTIONAL | 0.116/0.097/0.203 | -0.1pp | — |
| Near-term forward spread inversion (flip: low=stress; Engstrom-Sharpe beats 2s10s) | CONTEXT | -0.02/0.048/-0.21 | -1.4pp | — |
| Curve curvature (2s5s10s butterfly — humped = late-cycle) | CONTEXT | 0.035/-0.013/0.181 | -0.2pp | — |
| Real-rate move VIOLENCE (|63d speed|, either direction) | DIRECTIONAL | 0.116/0.161/0.065 | 3.0pp | — |
| Curve flattening impulse (flip: − = flattening = stress; INVERTED if post-inversion steepening is the tell) | CONTEXT | -0.033/-0.085/0.161 | -2.0pp | — |
| 3m10y TREND inversion (flip: low trend = stress) — Faria-Verona OOS equity-premium claim, tested on the return-forecast bar | CONTEXT | 0.14/0.195/-0.039 | 6.7pp | — |
• Display-only — never scored. The repricing of rates/inflation is REACTIVE; the curve moves after the data and the Fed, it does not lead them.仅供展示，从不计入评分。利率/通胀的重新定价是被动的；曲线在数据与美联储之后变动，而非领先。
• Rate-of-change beats level: the SPEED of a real-yield move breaks equities more than its level. Coefficients are the measured forward IC, not a promise.变动速度比水平更重要：实际收益率移动的速度比其水平更能冲击股票。系数为已测得的前瞻IC，并非承诺。
• Scenario moves apply a historical beta to a hypothetical shock — they are regime-dependent illustrations, not predictions, and assume the shock persists one quarter.情景中的涨跌幅是将历史beta应用于假设冲击——属于依赖状态的示意而非预测，且假设冲击持续一个季度。
• The real / nominal / breakeven 63d-change drivers are mechanically collinear (real = nominal − breakeven); do not treat them as independent confirmations.实际／名义／盈亏平衡的63日变动驱动在机制上共线（实际＝名义−盈亏平衡）；不应视为相互独立的印证。
Data.数据。 Free / keyless FRED: full Treasury curve, real yields (TIPS), 5y/10y/5y5y breakevens, term premium; headline & core CPI, headline & core PCE (the Fed’s 2% target gauge), PPI, ECI wage growth, shelter & core-services CPI; Cleveland model + UMich survey inflation expectations. All point-in-time vintaged for honest backtests. Dollar read: broad trade-weighted dollar + the Dollar Desk on the Forex page.免费/无密钥的FRED：完整国债曲线、实际利率(TIPS)、5/10年与5年5年盈亏平衡、期限溢价；整体与核心CPI、整体与核心PCE（美联储2%目标）、PPI、ECI工资增长、住房与核心服务CPI；克利夫兰模型与密歇根调查通胀预期。均按时点存档以便诚实回测。美元读数：广义贸易加权美元及外汇页的美元台。
Everything on this page is a measured relationship shown to explain market moves — none of it is a prediction or a trade signal. When a relationship failed our tests, we say so in the receipts above.本页所有内容均为用于解释行情的实测关联——皆非预测或交易信号。未通过检验的关联，已在上方依据中如实注明。