Why Were We Told to Brace for a "Detox Period" if the White House Won't Allow the Data Confirm It?

I was in graduate school at Cambridge University when I first heard this joke:

当我第一次听到这个笑话时，我正在剑桥大学读研究生：

I was working on a paper for a micro theory class, and there were some pretty big gaps (missing observations) in the time series data I had collected.

我正在为微观理论课程写一篇论文，我收集的时间序列数据中存在一些相当大的差距（缺少观察结果）。

My professor advised me to find a creative solution.

我的教授建议我找到一个创造性的解决方案。

I could use a proxy (unfortunately, there were no good ones) or do some kind of averaging to construct my own data points.

我可以使用代理（不幸的是，没有好的代理）或进行某种平均来构建我自己的数据点。

He told me that he didn’t really care how I resolved the issue as long as I explained my technique in a footnote.

他告诉我，他并不关心我如何解决问题，只要我在脚注中解释我的技术即可。

When I told some of my classmates that I was going to have to paper over the problem, one of them told me that joke.

当我告诉我的一些同学我必须掩盖这个问题时，其中一个人给我讲了这个笑话。

One way or another, I was expected to invent the numbers.

不管怎样，我被期望发明这些数字。

It felt like cheating.

感觉像是作弊。

I did what I had to do, but it nagged at me.

我做了我必须做的事，但这件事却困扰着我。

How was I supposed to feel good about the paper’s conclusions, knowing that I hadn’t been able to analyze all of the right data?

知道我无法分析所有正确的数据，我该如何对论文的结论感到满意呢？

Having access to data is important, but having access to quality data is essential.

访问数据很重要，但访问高质量数据也至关重要。

Last week, President Trump fired the Bureau of Labor Statistics (BLS) commissioner.

上周，特朗普总统解雇了劳工统计局（BLS）专员。

The move came just hours after the agency reported that total nonfarm payroll employment grew by 73,000 in July.

就在该机构报告 7 月份非农就业人数增加 73,000 人几个小时后，就采取了这一举措。

That’s a pretty weak number, but it was the downward revisions to the May and June numbers that caused the president to cry foul.

这是一个相当疲弱的数字，但正是五月和六月数据的向下修正引起了总统的不满。

The BLS had previously reported job gains of +144,000 in May and +147,000 in June.

美国劳工统计局此前报告称，5 月份就业岗位增加 144,000 人，6 月份就业岗位增加 147,000 人。

The July report erased more than a quarter-of-a-million jobs, revising May to +19,000 and June to +14,000.

7 月份的报告削减了超过 25 万个就业岗位，将 5 月份的就业岗位修正为+19,000 个，6 月份的就业岗位为+14,000 个。

Before firing the commission head, President Trump accused her of releasing “phony” numbers that had been “faked” for political reasons.

在解雇该委员会主席之前，特朗普总统指责她发布了出于政治原因“伪造”的“虚假”数据。

In the same way that Trump looks at U.S. trade deficits as de facto proof that other nation’s are “ripping us off”, he interprets large revisions in the data as evidence of malfeasance.

就像特朗普将美国的贸易逆差视为其他国家“欺骗我们”的事实上的证据一样，他将数据的大幅修改解释为渎职的证据。

When there were big downward revisions in August 2024, he accused the BLS of an extended coverup to help Biden ahead of the 2024 election.

当 2024 年 8 月出现大幅下调时，他指责劳工统计局长期掩盖事实，以在 2024 年大选前帮助拜登。

(Never mind that the same agency reported , just days before the November election, that a paltry 12,000 jobs were created in October 2024 and that Trump used that the weak number to argue that the Biden/Harris economy was “a catastrophe ”.) Now that Trump is in the White House, he’s convinced that the BLS is either incompetent or malicious.

（尽管该机构在 11 月大选前几天报道称，2024 年 10 月创造了微不足道的 12,000 个就业岗位，而特朗普利用这个疲软的数字来辩称拜登/哈里斯的经济是“一场灾难”。）现在特朗普入主白宫，他确信劳工统计局要么无能，要么恶意。

Revisions aren’t mistakes.

修改不是错误。

They don’t’ happen because the statisticians at the BLS counted up the numbers the wrong way in May and June, forcing them to make big adjustments in subsequent months.

这种情况之所以没有发生，是因为美国劳工统计局的统计人员在 5 月和 6 月统计数据的方式错误，迫使他们在接下来的几个月中做出重大调整。

Nor are the rank-and-file workers in the agency coordinating with one another to rig the data for political reasons.

该机构的普通工作人员也没有出于政治原因相互协调来操纵数据。

They’re just incorporating new information— as it becomes available— in order to produce more accurate data.

他们只是将可用的新信息纳入其中，以产生更准确的数据。

The revisions have nothing to do with politics and everything to do with the very real problem of receiving time-staggered survey responses from employers who don’t transmit their own data to the BLS as efficiently as they could/should.

这些修订与政治无关，而是与接收雇主的时间错开的调查答复这一非常现实的问题有关，这些雇主没有像他们能够/应该那样有效地将自己的数据传输到美国劳工统计局。

1

1

All data must be generated.

必须生成所有数据。

Sometimes it’s relatively easy to produce good data, and sometimes you have to work—or wait—for it.

有时产生好的数据相对容易，有时您必须努力或等待。

Some of the data I rely on to get a sense of what’s happening in the economy comes from outside the government (e.g. ADP’s Employment Report or The Conference Board ) and some of it comes from federal agencies like the Bureau of Economic Analysis (BEA) or the Bureau of Labor Statistics (BLS).

我用来了解经济状况的一些数据来自政府外部（例如 ADP 的就业报告或世界大型企业联合会），还有一些来自经济分析局 (BEA) 或劳工统计局 (BLS) 等联邦机构。

In addition to the nonfarm payrolls report, the BLS is responsible for publishing the most widely-cited measure of inflation, the Consumer Price Index (CPI).

除了非农就业报告外，美国劳工统计局还负责发布最广泛引用的通胀指标——消费者价格指数（CPI）。

Generating this series is a heroic undertaking that literally involves scores of data collectors calling or visiting thousands of retail stores, service establishments, doctor’s offices, websites, etc. across the United States to keep track of the prices that consumers are paying for ~80,000 goods and services each month.

生成这个系列是一项艰巨的任务，实际上需要数十名数据收集者致电或访问美国各地数千家零售店、服务机构、医生办公室、网站等，以跟踪消费者每月为约 80,000 种商品和服务支付的价格。

If you’re interested in a behind-the-scenes look at the important work the agency does, then I highly recommend this podcast conversation with Bill Beach, who was named Commissioner of the BLS by President Trump in 2017.

如果您有兴趣了解该机构所做的重要工作的幕后花絮，那么我强烈推荐与比尔·比奇 (Bill Beach) 的播客对话，他于 2017 年被特朗普总统任命为美国劳工统计局 (BLS) 专员。

(Before moving into that role, Bill was the chief economist for the republicans on the U.S. Senate Budget Committee.

（在担任该职务之前，比尔是美国参议院预算委员会共和党的首席经济学家。

I was his counterpart on the democratic side.

我是他在民主方面的对手。

I viewed him as kind, smart, and hardworking.) As Bill explains, the Bureau has worked hard to maintain data integrity in the face of acute budgeting challenges.

我认为他善良、聪明、勤奋。）正如比尔解释的那样，面对严峻的预算挑战，该局一直努力保持数据完整性。

When it comes to collecting the data that’s used to construct the CPI, one of the big challenges is that it costs a lot of money to pay all of those data collectors to monitor what’s actually happening on the ground.

在收集用于构建消费物价指数的数据时，最大的挑战之一是需要花费大量资金来向所有这些数据收集者支付大量费用来监控实地实际发生的情况。

Back in June, the BLS posted a public notice on its website, explaining that it had stopped collecting data in several cities (Lincoln, Nebraska; Buffalo, New York; and Provo, Utah) and that it would rely more heavily on “imputations” to configure the prices of certain items (basically, that means inferring rather than directly observing certain prices).

早在 6 月，劳工统计局就在其网站上发布公告，解释其已停止收集多个城市（内布拉斯加州林肯、纽约州布法罗和犹他州普罗沃）的数据，并将更多地依赖“估算”来配置某些商品的价格（基本上，这意味着推断而不是直接观察某些价格）。

That started to raise some alarm bells as CPI is a vital statistic for all sorts of reasons .

这开始敲响了一些警钟，因为出于各种原因，CPI 是一项至关重要的统计数据。

The Trump administration’s drastic cutbacks of government spending and the federal workforce have economists, researchers and statisticians sounding the alarm that the reliability and accuracy of economic data could become a casualty to those efforts.

特朗普政府大幅削减政府支出和联邦劳动力，经济学家、研究人员和统计学家敲响了警钟，经济数据的可靠性和准确性可能会成为这些努力的牺牲品。

The BLS pretty much gets one shot at producing good inflation data.

美国劳工统计局几乎有机会产生良好的通胀数据。

The data team scours the consumer landscape and reports the prices that people are paying to acquire ~80,000 different goods and services.

数据团队仔细研究了消费者格局，并报告了人们为购买约 80,000 种不同商品和服务而支付的价格。

They don’t look back at prior months in order to fill in gaps or otherwise modify the price data they report each month.

他们不会回顾前几个月来填补空白或以其他方式修改他们每月报告的价格数据。

What you get in each monthly CPI report is what you get.

每个月度 CPI 报告中所得到的就是你所得到的。

That’s not how it works with certain other data series, for example Gross Domestic Product (GDP) and the Nonfarm Payrolls report (NFP).

这并不是它与某些其他数据系列的配合方式，例如国内生产总值（GDP）和非农就业报告（NFP）。

In each of these cases, the BEA and BLS, respectively, report preliminary data.

在每种情况下，BEA 和 BLS 分别报告初步数据。

Everyone knows it’s just an initial glimpse—i.e. an early estimate that will be revised as more data comes in.

每个人都知道这只是初步的一瞥——即早期的估计将随着更多数据的出现而进行修改。

That’s why you often hear Fed officials (and others) explaining that they don’t want to put too much weight on a single report, especially when the data is subject to multiple revisions.

这就是为什么你经常听到美联储官员（和其他人）解释说，他们不想过分重视一份报告，尤其是当数据经过多次修订时。

The president’s top economic advisor (and potential next Fed chair ) has defended the firing of the BLS commissioner, adding that President Trump’s “top priority” is to make the economic data more reliable.

总统的最高经济顾问（以及潜在的下一任美联储主席）为解雇劳工统计局专员进行了辩护，并补充说特朗普总统的“首要任务”是让经济数据更加可靠。

But it’s hard to see how installing a partisan loyalist will foster greater confidence in the integrity of the nation’s economic data.

但很难看出任命一名党派忠诚者将如何增强人们对国家经济数据完整性的信心。

It’s also difficult to imagine how the administration’s own budget proposal , which calls for an 8% cut in staff and funding for the BLS, will help the agency modernize and improve its data collection.

同样很难想象政府自己的预算提案（要求 BLS 的人员和资金削减 8%）将如何帮助该机构实现现代化并改进其数据收集。

But here’s a question no one is asking: Why did top officials in the Trump administration spend months telling us that DOGE was going to help wean the economy off of government spending and that the “detox period” would be painful but “worth it” even if it tipped the economy into recession ?

但有一个问题没有人问：为什么特朗普政府的高级官员花了几个月的时间告诉我们，DOGE 将帮助经济摆脱对政府支出的依赖，并且“排毒期”将是痛苦的，但“值得”，即使它会让经济陷入衰退？

I mean, this is precisely what Lutnick, Bessent, Trump, and other top officials considered the downside risks (albeit transitory) of the tariffs, trade wars, DOGE layoffs, and mass deportations.

我的意思是，这正是卢特尼克、贝森特、特朗普和其他高级官员所认为的关税、贸易战、DOGE 裁员和大规模驱逐的下行风险（尽管是暂时的）。

David asks, “What did he think would happen?” but Trump told us that his policies could cause substantial upheaval.

大卫问：“他认为会发生什么？”但特朗普告诉我们，他的政策可能会导致重大动荡。

When the unemployment rate edged higher in February, some wondered whether the “detox period” was already underway.

当二月份失业率小幅上升时，一些人怀疑“排毒期”是否已经开始。

Consumer confidence had plunged and economic growth was slowing .

消费者信心大幅下降，经济增长放缓。

In response, Treasury Secretary Scott Bessent said it was reasonable to think the economy would lose momentum as we begin to rely more heavily on the private sector to support output and employment.

对此，财政部长斯科特·贝森特表示，随着我们开始更加依赖私营部门来支持产出和就业，有理由认为经济将失去动力。

So what was the point of telling everyone to brace for a rough ride if they were going to shoot the messenger for publishing the data that confirms that we just took a bad bounce?

那么，如果他们要因为发布证实我们刚刚经历了糟糕反弹的数据而开枪打死每个人，那么告诉每个人做好准备迎接艰难的旅程又有什么意义呢？

As always, thank you for reading The Lens .

一如既往，感谢您阅读《镜头》。

Here’s Trump’s former BLS Commissioner (Bill Beach) in a newly-released podcast , explaining that if anyone is politicizing the jobs data, it is the president himself.

特朗普前劳工统计局专员（比尔·比奇）在新发布的播客中解释说，如果有人将就业数据政治化，那就是总统本人。

As David Rosenberg observed “The problem lies with the depleted business response rate in the first payroll survey (58% - was over 80% pre-Covid) which then rises to the normal 94% by the final revision.

正如大卫·罗森伯格（David Rosenberg）所观察到的那样，“问题在于第一次薪资调查中的企业响应率下降（58% - 新冠疫情爆发前超过 80%），而到了最终修订时，该响应率上升至正常的 94%。

That’s the story.

这就是故事。

The data are only as reliable as the sample size, and that has nothing to do with the BLS but rather the tardiness of the U.S. business sector in getting its updated staffing levels to the Bureau on time.”

数据的可靠性取决于样本量，这与劳工统计局无关，而与美国商业部门迟迟未能及时向劳工统计局提供最新的人员配置水平有关。”