The Temperature Gap: When 4.7% GDP Meets "It Feels Like Everything Collapsed"

温差：当 4.7% 的 GDP 撞上"感觉天塌了"

2026 H1, a Chengdu noodle shop owner, Lao Zhang, closes at 9pm instead of midnight. His revenue is down 30% YoY. Across town, a battery-cell exporter is running three shifts, shipping containers to Rotterdam. Both are "the Chinese economy." Only one of them is in your feed.

2026 上半年，成都开面馆的老张把打烊时间从半夜改到晚上 9 点，营业额同比掉了三成。城市另一端，一家动力电池出口商三班倒，集装箱发往鹿特丹。两者都是"中国经济"。但只有前一个，会出现在你的朋友圈。

The headline says GDP +4.7% YoY, disposable income +5.2% nominal / +4.2% real. The group chat says "钱难挣、活难找、房子跌穿、不敢花". Neither side is lying. They are just standing on different floors of the same building.

宏观口径里，上半年 GDP 同比 +4.7%，居民人均可支配收入 名义 +5.2% / 实际 +4.2%。群聊里却是"钱难挣、活难找、房子跌穿、不敢花"。两边都没撒谎，只是站在同一栋楼的不同楼层。

1. Macro counts output; households count cash flow

一、宏观算的是产出，老百姓算的是现金流

GDP is a production tally​ — factories shipped, ports cleared, silicon wafers etched, EVs rolled out. In H1 2026, industrial value-added rose 5.3%, high-tech manufacturing 13.8%, exports +13.4%. That is the "hot" China.

GDP 是生产台账——工厂出货、港口清关、晶圆刻蚀、电车下线。2026 上半年规上工业增加值 +5.3%，高技术制造 +13.8%，出口 +13.4%，这是"热"的那面中国。

A household balance sheet is a different object. It has four entries that actually move mood:

家庭资产负债表是另一本账，真正撬动情绪的只有四项：

• Wage stickiness​ — average wage income +5.3%, but the median household earned 19,036 yuan H1​ (≈3,173/month), only 82.8% of the mean. The median grew 4.7%, slower than the average. Translation: the top tail is pulling the number, the middle class is treading water.
• 工资黏住——工资性收入均值 +5.3%，但中位数家庭上半年只挣 19,036 元（月均约 3,173），是平均数的 82.8%，且中位数增速 4.7% 跑输平均数。翻译过来：高收入尾巴在拉均值，中间阶层原地踩水。
• Property, the silent killer​ — housing is ~52–65% of household assets. National home prices are off ~30%​ from peak; Goldman sees another ~10% before bottoming ~2027. A 30% drop on a 30%-down-payment home wipes out the equity and then some. The "I feel richer" illusion of 2015–2021 is now its mirror image.
• 房产，沉默的杀手——房子占家庭资产 52%–65%。全国房价较峰值回撤约 30%，高盛判断到 2027 触底前还有约 10% 下行空间。30% 的跌幅套在首付 30% 的房子上，净值归零还倒挂。2015–2021 年"我觉得自己有钱"的幻觉，现在是它的镜像。
• Property income stalled​ — per-capita property income (rent, dividends, interest) grew just 1.1%​ in H1, the weakest line in the income statement. The "passive wealth" channel is clogged.
• 财产性收入停转——上半年人均财产净收入（房租、股息、利息）只增 1.1%，是所有收入项里最瘸的一条。"被动致富"的管子堵了。
• Precautionary hoarding​ — households added 7.58 trillion yuan​ to deposits in H1 2026 while net loans fell 366.8 billion. People are not broke; they are de-risking. Saving rate effectively back to 32–34%.
• 预防性囤钱——2026 上半年住户存款新增 7.58 万亿，同时贷款净减 3668 亿。人不是没钱，是在去风险。储蓄率实质回到 32%–34%。

Macro GDP = how much we made.

Micro sentiment = how much we kept, how much our house pretends to be worth, and whether next year's paycheck is certain.

宏观 GDP = 我们造了多少。

微观体感 = 到手留了多少、房子还装多少面子、明年工资条还在不在。

2. The growth engine shifted, but the job engine didn't follow

二、增长引擎换了，就业引擎没跟上

The 4.7% is carried by EVs, chips, AI racks, solar, batteries, shipbuilding​ — capital-heavy, labor-light. A 50-billion-yuan EV plant hires thousands of engineers, not twenty thousand assembly hands. The old job multiplier — one real-estate project → builders, agents, furniture, appliances, decorators, lenders​ — is broken: H1 real-estate development investment −18.0% YoY, land-sale revenue −30%.

4.7% 靠 电车、芯片、AI 机柜、光伏、电池、造船​ 扛着——重资本、轻人力。一个五百亿产值的新能源厂招几千名工程师，不是两万装配工。旧乘数——一个楼盘→建筑、中介、家具、家电、装修、放贷——断了：上半年房地产开发投资 同比 −18.0%，土地出让金 −30%。

So the economy adds "advanced manufacturing value" while the restaurant owner, the real-estate agent, the interior designer, the 22-year-old liberal-arts graduate​ all sit on the wrong side of the transition. Youth unemployment stays sticky; flexible work now covers ~44% of employed people​ — no monthly certainty, no bonus horizon.

于是经济在加"先进制造增加值"，而饭馆老板、房产中介、室内设计师、22 岁文科毕业生全卡在转型的错误一侧。青年失业黏在高位；灵活就业已占就业人口 约 44%——没有月度确定性，没有年终奖地平线。

Fitch calls it a two-speed economy: external demand hot, domestic demand cold, deflator negative (−0.7% in Q4 2025, near flat 2026). When prices fall, holding cash beats holding inventory; when the house falls, spending the dividend feels immoral.

惠誉管这叫双速经济：外需热、内需冷，平减指数为负（2025Q4 为 −0.7%，2026 接近零）。价格下行时持币优于持货；房子下行时，花"纸面分红"像在败家。

3. "Collapse" is a balance-sheet word, not a GDP word

三、"崩盘"是资产负债表用语，不是 GDP 用语

Ordinary speech conflates three different events:

日常口语把三件不同的事揉成一团：

1. Output recession​ — GDP contracts YoY. Not happening. 4.7% is not collapse.
2. 产出衰退——GDP 同比收缩。没发生。4.7% 不是崩盘。
3. Balance-sheet recession​ — assets fall, debt stays, net worth shrinks, households repair for years. This is the Japanese 1990 template, and partially ours since 2021. A family that bought a 2021 suburban flat with 30% down is not "in a GDP recession"; they are in a 15-year wealth reset.
4. 资产负债表衰退——资产跌、负债留、净值缩、家庭多年修表。这是日本 1990 模板，也是我们从 2021 起的部分剧本。2021 年首付三成买郊区房的家庭，不在"GDP 衰退"里，他们在15 年财富重置里。
5. Sentiment crash​ — confidence index stuck below 50, every purchase deferred. This is where we are. Not Armageddon, but a long soft rain.
6. 信心崩——信心指数卡在 50 荣枯线下方，每笔开销都往后推。这是我们所处的层。不是末日，是一场漫长的毛毛雨。

The pain is real because the transmission belt from macro to micro is severed in three joints: property wealth effect (negative), labor intensity of new growth (low), and income distribution (median < mean). Growth lands in corporate retained earnings and export containers, not in the wage account of the person scrolling at midnight.

痛感是真的，因为宏观到微观的传送带在三节断了：房产财富效应（负）、新增长的劳动密度（低）、收入分配（中位数<平均数）。增长落在企业留存收益和出口集装箱里，不落在半夜刷手机那人的工资户头里。

4. What recovery actually means for a household

四、对一个家庭来说，"恢复"到底指什么

Officials will tell you recovery is GDP back to trend. Households translate it differently:

官方说的恢复是 GDP 回趋势线。家庭自动翻译成：

• not "factory output +5%"，
• but "my home price stops falling, my kid's internship converts, my bonus returns, my deposit rate doesn't have to be my only return."
• 不是"工厂产出 +5%"，
• 是"我房子别再跌了、孩子实习能转正、年终奖回来、存款利息不必是我唯一收益"。

On current trajectory, the nominal GDP V-shape​ is irrelevant to that list. The relevant clocks are:

按当前轨迹，GDP 名义 V 型和上面这张清单无关。真正相关的钟是：

• Property bottom: maybe 2027 (Goldman), then a flat plateau, not a rebound.
• 房价触底：大概 2027（高盛），之后是平台，不是回弹。
• Net-worth reset for leveraged buyers: 5–8 years to nominal, 8–15 to real purchasing power; Japan worst-case 20–30.
• 杠杆买家净值重置：名义 5–8 年，真实购买力 8–15 年；日本最差情形 20–30 年。
• Sentiment normalization: needs wage certainty + social safety net + a new asset anchor beyond housing. No calendar date yet.
• 信心正常化：要工资确定性+社保网+房产之外的资产锚。目前没有日历日期。

5. Closing: the gap is the story

五、温差本身就是正文

The Chinese economy in 2026 is not collapsing. It is re-rating​ — from a leverage-property-growth fund to a hard-tech-export-income-stagnation fund. The 4.7% is true. The "I feel poorer" is also true.

2026 年的中国经济没有崩。它在重定价——从"杠杆×房产×增长"的基金，重估成"硬科技×出口×收入停滞"的基金。4.7% 是真的，"我觉得变穷了"也是真的。

The distance between those two sentences is not a measurement error. It is the structural tax of transition, paid by the people who cannot wait out a decade.

这两句话之间的距离不是统计误差。它是转型的结构性税单，由那些等不起十年的人缴纳。

Beijing sees the temperature on the thermometer.

Lao Zhang sees the temperature on his skin.

北京看温度计上的温度。

老张看皮肤上的温度。

Both are 2026.

两者都是 2026。

Data anchors: NBS H1 2026 (GDP +4.7%, disposal income 22,981 yuan, median 19,036, property income +1.1%, consumption +2.7% real); Fitch Jan-2026 (4.1% forecast, deflator −0.7%); Goldman property −30% from peak, −10% to 2027 bottom; H1 property dev investment −18.0%; H1 household deposits +7.58 tn, loans −366.8 bn.

数据锚点：统计局 2026H1（GDP+4.7%、可支配收入 22981 元、中位数 19036、财产收入+1.1%、消费实际+2.7%）；惠誉 2026-01（预测 4.1%、平减−0.7%）；高盛房价峰值−30%、至 2027 底再−10%；上半年地产开发投资−18.0%；上半年住户存款+7.58 万亿、贷款−3668 亿。

继续阅读：GDP   中国   家庭资产负债表   崩盘

此文由 怡心湖 编辑，若您觉得有益，欢迎分享转发！：首页 > 会·生活 » 温差：当 4.7% 的 GDP 撞上"感觉天塌了"