PayEvidence 指数
被指数化的劳动力市场
三条可引用的序列,全部建立在标准统计设计之上——PMI 式招聘扩散指数、欧盟指令意图改变的薪酬透明度占比、以及成分受控的明示薪酬指数。每个点都经测量、带版本、可复现;样本不足的日子发布为空档,绝不估算。
HMI · 招聘动能指数
面板级 PMI 式周度扩散:扩张在架职位的雇主占比对收缩占比,50 为荣枯线。逐公司相对变化,面板本身的扩张无法推动指数。
53.4
截至 2026-10-11
方法论
| design | PMI-style weekly diffusion |
| reporter_gate | >= 5 live openings at both endpoints of a 7-day window |
| publish_gate | >= 50 reporters per point |
| formula | (% expanding + 0.5 x % unchanged) x 100 |
嵌入此指数
<iframe src="https://payevidence.com/embed/index/hmi" width="640" height="300"
style="border:1px solid #d9d4c5" title="HMI — PayEvidence"></iframe>
PTI · 薪酬透明指数
可测在架职位中雇主明示薪资的占比——全量与欧盟双序列。这正是欧盟薪酬透明指令意图改变的量,我们按日测量而非按年调查。
52.8%
截至 2026-10-11
方法论
| design | direct daily measurement over the deduplicated panel |
| denominator | open roles on platforms whose full text is observable (salary-blind platforms excluded) |
| eu_series | same measure restricted to EU-located roles |
嵌入此指数
<iframe src="https://payevidence.com/embed/index/pti" width="640" height="300"
style="border:1px solid #d9d4c5" title="PTI — PayEvidence"></iframe>
SPI · 明示薪酬指数
基于职类×国家单元格的 Laspeyres 固定篮子明示薪酬指数,权重冻结于基期——面板结构漂移无法伪装成薪酬变动。
83.12
截至 2026-10-11
方法论
| design | Laspeyres fixed basket over job_family x country cells |
| base_period | first 5 qualifying days, cell medians and weights averaged (a cell needs >= 3 base-window days) |
| cell_gate | >= 30 stated ranges per cell-day |
| publish_gate | >= 10 common cells per point |
| weights | frozen at base-period average stated-range counts; renormalized over the cells present on each day (coverage_pct reports the retained weight share) |
| note | composition-controlled: panel growth cannot masquerade as pay movement |
嵌入此指数
<iframe src="https://payevidence.com/embed/index/spi" width="640" height="300"
style="border:1px solid #d9d4c5" title="SPI — PayEvidence"></iframe>
如何引用
PayEvidence Indices (HMI/PTI/SPI), methodology v1, accessed YYYY-MM-DD, https://payevidence.com/indices
欢迎注明出处并附链接引用;CSV 与 JSON 端点稳定且带版本。 · 方法论