Economic Forecasting宏观经济指标预测与研究Macro Indicators, Forecasts & Research
更新于 2026-09-12 09:30:45Updated 2026-09-12 09:30:45
周频劳动力信号Weekly Labor Signals

初请失业金人数Initial Jobless Claims

初请失业金人数是观察劳动力边际走弱最敏感的周频信号之一，通常早于月频就业数据给出变化。Initial claims are among the most sensitive weekly labor signals and often move before monthly employment releases confirm the trend.

返回终端总览Back to overview terminal查看本次变化Open latest batch查看发布时间日历Open release calendar查看发布历史Open release history

当前周期读数Current-Cycle Readout

把当前可用值、上一期、下一预测和时间语义放在同一页，方便从总览切到单指标深读。Keep the current reading, prior observation, next forecast, and timing semantics on one page for single-indicator deep reading.

中性Balanced
最近已实现模型目标Latest realized model target
205,250

首发冻结四周均值（模型目标）Frozen first-release four-week average (model target)

上一期模型目标Previous model target
206,000

观测 2026-08-29Observed 2026-08-29

基线现时估计Baseline nowcast
205,250

目标 2026-09-12 · 上一期持平基线Target 2026-09-12 · Last-observation baseline

目标期已结束，等待官方首发值后结算。The target period has ended and awaits the official first release for settlement.

模型状态Model status
基线保护Baseline protected

研究候选尚未在外层回测中稳定胜过基线。Research candidates have not consistently beaten the benchmark in outer backtests.

数据状态Data status
实时抓取Fresh fetch

当前运行使用了本轮实时抓数结果。This indicator used a live fetch result in the current run.

线上冻结结算Frozen live settlement
4 个已结算实盘样本4 settled live samples

已冻结预测 22 条；待结算 18 条。回测样本不计入实盘。22 frozen forecasts; 18 awaiting settlement. Backtests are excluded from live evidence.

当前 2026-09-12 的 205,250 是透明基线估计，不是复杂模型已证明的优势。研究候选 Baseline[last] 继续保留在研究轨道。The 205,250 estimate for 2026-09-12 is a transparent benchmark, not demonstrated complex-model edge. Baseline[last] remains on the research track.

本次变化摘要This Cycle Change

这一层比较相邻模型目标实现值，并单独标明本周期是否已有新的官方发布。This layer compares adjacent realized model targets and separately marks whether a new official release has arrived this cycle.

变化方向Direction
下行Down

初请失业金人数 本周期尚无新的模型目标实现值，当前仍按官方发布节奏等待。Initial Jobless Claims has no new realized model target this cycle and is waiting on the official release cadence.

最新源序列值Latest source-series value
206,000

上一来源值 207,000Previous source reading 207,000

本周期状态Cycle status
按官方节奏待发布Awaiting scheduled release

当前决策月内尚无新正式值，但这符合该指标的正常发布时间。No new official value has been released in the current decision month yet, which is normal for this indicator.

下一次正式发布Next official release
2026-09-17

5 天后In 5 days

初请失业金人数 当前沿用 205,250 作为最近模型目标实现值；本周期尚无新值可得，上一期为 206,000。Initial Jobless Claims is still carrying 205,250 as the latest realized model target; no new value is available and the prior period was 206,000.

时间语义Time semantics

最新观测、官方发布、站点更新和预测目标分别回答的是不同问题，不能彼此替代。Observed, official release, site update, and forecast target answer different questions and should not be treated as interchangeable.

原始最新观测Latest source observation2026-09-05
官方发布日期Official release date2026-09-10
站点更新Site update2026-09-12 09:30
预测目标日期（非发布时间）Forecast target date (not release time)2026-09-12

来自首发版本可得日期From recorded first-release availability

初请失业金目标合同Initial claims target contract原始周值、当前版本四周均值、首发冻结目标和下一预测不可混用。Raw weekly claims, current-vintage average, frozen first-release target, and next forecast are distinct.
最新原始周值Latest raw weekly claims206,0002026-09-05
当前版本四周均值Current-vintage four-week average206,000可随修订变化May change with revisions
首发冻结四周均值Frozen first-release four-week average205,2502026-09-05
下一周首发目标预测Next first-release target forecast205,2502026-09-12

图表阅读Chart view

先看焦点窗口确认最新拐点和预测边界，再看长历史窗口理解该指标在更长周期中的位置。Use the focus window for the latest turn and forecast boundary, then the long-history window for broader cycle context.

严格外样本策略焦点图Strict OOS policy focus当前服务：Baseline[last]Current serving: Baseline[last]
长历史图Long-history chart全窗口Full window

模型与诊断Model & diagnostics

近期 MAERecent MAE2379.81
近期 MASERecent MASE1.000
近期相对基线技能Recent skill0.0%
当前95%区间宽度Current 95% width33000.0
全期 MAEFull MAE8567.96
全期 MASEFull MASE1.006
全期相对基线技能Full skill-0.6%
2022后 MASEPost-2022 MASE1.000
中位数缩放误差Median scaled error1.000
方向辅助诊断（平衡）Direction diagnostic (balanced)53.4%
校准期95%覆盖Calibrated 95% coverage94.6%

公开服务模型：Baseline[last]（上一期持平基线）Serving model: Baseline[last] (Last-observation baseline)

目标发布与历史Target release & history

最近一次官方发布Latest official release2026-09-10
下一次官方发布Next official release2026-09-17

发布时间有早有后通常来自指标自身的官方节奏差异，而不是站点异常。Differences across release timing usually come from the indicator's official cadence rather than a site anomaly.

2026-09-10观测 2026-09-05Observed 2026-09-05205,250
2026-09-03观测 2026-08-29Observed 2026-08-29206,000
2026-08-27观测 2026-08-22Observed 2026-08-22204,250
2026-08-20观测 2026-08-15Observed 2026-08-15202,750
2026-08-13观测 2026-08-08Observed 2026-08-08198,000
2026-08-06观测 2026-08-01Observed 2026-08-01197,750
正式发布时间Official release观测期Observed at模型目标实现值Realized model target本周期标记Cycle flag
2026-09-102026-09-05205,250历史目标实现值Historical target
2026-09-032026-08-29206,000历史目标实现值Historical target
2026-08-272026-08-22204,250历史目标实现值Historical target
2026-08-202026-08-15202,750历史目标实现值Historical target
2026-08-132026-08-08198,000历史目标实现值Historical target
2026-08-062026-08-01197,750历史目标实现值Historical target
2026-07-302026-07-25201,750历史目标实现值Historical target
2026-07-232026-07-18206,250历史目标实现值Historical target
2026-07-162026-07-11213,250历史目标实现值Historical target
2026-07-092026-07-04217,750历史目标实现值Historical target
2026-07-022026-06-27221,250历史目标实现值Historical target
2026-06-252026-06-20223,750历史目标实现值Historical target

方法与边界Method & boundary

周度波动较大，适合看趋势和拐点，不适合对单周读数做过度解释。Weekly noise can be large, so it is best used for trend and turning-point analysis rather than over-reading a single print.

这一详情页把月度配置摘要、混频观察面和发布时间语义汇总到单一指标视角，方便从总览切到深读。This detail page condenses the monthly summary, mixed-cadence surface, and release semantics into a single-indicator view for deeper reading.

初请以官方首发周值构造四周移动平均；组成式基线假设下一周原始初请等于最近一周，并据此滚动四周窗口。疫情结构突变保护仅使用预测时点之前的 52 周信息。Initial claims use first-release weekly values to form the four-week average. The component baseline carries the latest raw claim into the next rolling window, while the structural-break guard uses only the prior 52 weeks available at the forecast origin.

相关研究Related research

优先展示与当前指标直接相关的正式研究和模板化更新摘要，帮助从单指标延伸到主题判断。Prioritize anchor research and templated update briefs linked to this indicator so the user can move from a single series into thematic reading.

主研究Anchor Research

Economic Forecasting 研究组Economic Forecasting Research Desk正式研究Anchor research

七个公开信号如何拼成同一张周期记分卡How the seven public signals fit into one cycle scorecard

这七个公开指标不是七个平行栏目，而是一张分层记分卡。通胀、增长、劳动力和利率分别提供不同的确认职责，只有把它们按发布时间节奏和信息功能放到同一框架里，平台才会真正像终端，而不是一组孤立图表。The seven public indicators are not seven parallel columns but a layered scorecard. Inflation, growth, labor, and rates each play a different confirmation role, and only a shared cadence-and-function framework makes the platform feel like a terminal rather than a set of disconnected charts.

通胀Inflation增长Growth劳动力Labor利率Rates主研究Anchor
核心PCE物价指数Core PCE Price Index非农就业人数Nonfarm Payrolls工业生产Industrial Production初请失业金人数Initial Jobless Claims10年期美债收益率10Y Treasury Yield2年期美债收益率2Y Treasury Yield10Y-2Y美债利差10Y-2Y Treasury Spread
发布时间 2026-04-03Published 2026-04-03

月度简报Monthly Briefs

Economic ForecastingEconomic Forecasting月度简报Monthly brief

2026年9月月度配置简报Sep 2026 monthly brief

本月主线围绕 10Y-2Y美债利差、10年期美债收益率、2年期美债收益率 展开，并统一交代已更新与按节奏待发布的公开指标。This month's brief centers on 10Y-2Y Treasury Spread, 10Y Treasury Yield, 2Y Treasury Yield and aligns updated versus still-pending published indicators.

月度配置Monthly allocation公开指标Published indicators
核心PCE物价指数Core PCE Price Index非农就业人数Nonfarm Payrolls工业生产Industrial Production10年期美债收益率10Y Treasury Yield2年期美债收益率2Y Treasury Yield10Y-2Y美债利差10Y-2Y Treasury Spread初请失业金人数Initial Jobless Claims
发布时间 2026-09-12Published 2026-09-12

最新指标更新Latest Release Notes