这次股市崩盘波及一切：家庭财富、消费与杠杆的风险传导

原视频：This Time the Stock Market Crash Hits Everything · Casual Finance · 2026-10-20 · 15分40秒 文本来源：提取并规范化YouTube英文自动字幕。中文总结及全文翻译基于全部字幕；未重新听写或完整核对画面。

结构化中文摘要

#

本片的主线是：股票上涨带来的财富效应，使家庭消费、企业盈利和市场估值相互依赖；融资借款与杠杆 ETF 则可能把普通下跌放大为被迫卖出的连锁反应。 作者认为，应同时观察股票在居民资产中的占比、机构现金缓冲和杠杆，而不只关注几家科技公司的指数权重。

以下数字、机构研究和案例均按作者在 2026 年 10 月 20 日发布的视频中所述整理，未据此确认其统计口径或真实性；本文不提供实时行情或独立投资判断。作者明确表示，自己不知道冲击何时到来、由什么触发。

1. 集中风险也存在于居民财富结构中。 作者首先提到七家公司占指数约三分之一，然后把焦点转向美国家庭的股票敞口：他称股票占家庭全部资产约34%，长期平均约16%，现金约占9%。他还引用1968、2000、2021年的历史高占比及随后下跌，作为其脆弱性论证的一部分；这些历史对照本身没有给出下一次下跌的时间。00:00 · 00:49

2. 过度配置可能来自持续上涨和默认买入，而不是一次主动“全押”。 作者称，2009年以来美国人持有股票的价值从约9万亿美元增至74万亿美元；资产价格上涨会使股票在总资产中的权重自动增加。退休账户、目标日期基金和自动加入机制，则被他描述为不以价格为主要约束的持续资金来源；他引用约80%的退休账户投入流向股票。03:19 · 04:14

3. 财富效应把股市风险传导到消费、企业和财政。 作者解释，人们可能因资产升值而增加支出，即使工资没有增长。他转述亚特兰大联储研究，称消费对股市的敏感程度已超过1990年的四倍，并以股市下跌25%导致全年实际消费增长转负的情景分析作例子。他还引用高收入家庭从投资资产提取消费资金，以及加州资本利得税收对下跌的敏感性，说明传导不止发生在证券账户里。04:46 · 05:11 · 05:50

4. 机构即使担心泡沫，也可能因职业激励继续追随市场。 按作者转述的美国银行基金经理调查，现金比例约3.9%，科技股超配达到2020年8月以来最大，同时45%的受访者预期次年出现AI泡沫。他借凯恩斯关于随大流失败与逆势成功的讨论，解释基金经理为何担心持有现金、错过上涨而失去工作；所述现金阈值属于其引用的调查框架。06:37 · 07:19

5. 作者认为全球资金也集中于同一批美国资产。 他称外国投资者持有约22万亿美元美股、约占全市场五分之一，美国公司在主要全球股票指数中权重接近三分之二，并以此说明家庭、机构和海外资金可能承受相互重叠的敞口。视频没有给出这里“全球主要指数”的确切名称。08:09

6. 融资买股和股票抵押贷款，会让价格下跌转变为资金追缴压力。 作者引用6月融资余额约1.5万亿美元，以及摩根士丹利股票抵押贷款一年增长20%的说法，随后解释：抵押品价格下降，贷款本金并不随之减少；无法补现金时，就可能被迫出售持仓。他引用FINRA规则说明，借款人未必能够自主决定平仓时间和资产选择；此处是视频对规则的转述。08:41 · 09:43 · 10:01

7. 视频用一个对冲基金故事说明拥挤交易与强制卖出的风险。 作者叙述Leopold Aschenbrenner及Situational Awareness基金集中押注AI基础设施、借入最高约四倍自有资金，并称该基金资产从约450亿美元降至100亿美元，Citadel折价接手其大部分股票组合。这些具体事件、金额、收益和日期在本次收录中均未独立核实，应作为作者案例阅读。文中“借入四倍”也没有擅自改成“总敞口四倍”。10:47 · 11:22

8. 杠杆 ETF 的每日再平衡可能顺势放大波动。 作者解释两倍、三倍“每日”涨跌幅的目标，并称新产品、资产规模和科技半导体敞口增长很快。他转述野村估算：市场每波动1%，相关基金约需交易90亿美元；高峰时单日再平衡约500亿美元。其论点是，机械调仓会在上涨中增加买盘，在下跌中增加卖盘。视频这里讨论每日目标，没有说明多日累计收益必然等于标的累计收益的固定倍数。12:24 · 13:05

9. 两种强制卖出机制可能形成“连环清算”。 按作者的解释，股价下跌触发保证金追缴或再平衡卖单，卖单进一步压低价格，更多持仓跌破保证金门槛，再引发新的卖单。这是视频给出的风险传播机制，并不等于该链条必然发生。13:30

10. 结论强调脆弱性，同时承认触发因素不可预测。 作者列举AI交易受挫、债券收益率上升或其他冲击；他引用调查中的10年期美债收益率5.9%情景，讨论资金可能从股票转向债券。最后落点是：居民消费和经济活动高度依赖高股价，而市场中可能被迫卖出的资金很多，因此一次冲击可能影响广泛。视频未提供确定的崩盘日期或具体交易指令。13:54 · 14:15 · 15:02

阅读范围、翻译与来源说明

#

• 已阅读全文4个时间块、448段英文字幕，覆盖00:00:00–00:15:40；媒体实际时长00:15:40.454。时间检查未发现长空缺，末尾包含音乐标签。时间覆盖不能证明每个词都被自动字幕准确识别。
• 文本来自 YouTube英文自动字幕的提取与规范化，本次没有重新进行本地听写。分析基于字幕，未逐帧核对视频画面。
• 英文原文全部保留；中文全文译文按相同4个时间块、相同448段顺序完整翻译。中文字幕沿用规范化后的源时间轴；自动滚动字幕原有重叠显示的结束时间已收束，原始显示时长仍保存在源JSON中。这不是词级强制对齐。
• 约02:01–03:17为Cape赞助口播。原文、全文译文和中文字幕均完整保留其功能、隐私政策、优惠及优惠码说法；这些是视频广告内容，未独立核验，也未执行其中的订阅、点击或购买呼吁。
• 视频提到的论文、机构调查、FINRA规则和基金事件均为作者所引述内容，本笔记没有取代原始报告核验；相关数字保留其发布时的叙述语境。“今年”“本周”“上个月”等相对时间以视频发布日期为语境。

单独保留的字幕疑点与术语说明

#

视频原文全文（英语，YouTube自动字幕）

#

00:00:00

#

[00:00:00] Everybody knows the market is expensive.
[00:00:02] It doesn’t take a genius to see it.
[00:00:04] Seven companies make up a third of the
[00:00:06] entire index. The investment management
[00:00:08] firm Capital Group found that today’s
[00:00:10] markets are among the most concentrated
[00:00:12] in history. By the most common long-run
[00:00:15] measure of how expensive stocks are, the
[00:00:17] market is priced at a level we’ve only
[00:00:19] seen once before. That was the peak of
[00:00:21] the dot-com bubble. But this isn’t
[00:00:23] exactly secret information. If you’ve
[00:00:25] been anywhere near the internet in the
[00:00:26] last year, you’ve heard all of this. The
[00:00:28] market’s expensive, it’s con- centrated,
[00:00:30] and it keeps going up anyway. Earlier
[00:00:32] this month, we closed at another record
[00:00:34] high. While everyone’s busy measuring
[00:00:36] how much of the market sits in a few
[00:00:38] tech companies, nobody’s measuring how
[00:00:40] much of America sits in the market.
[00:00:41] That’s the real danger right now. It’s
[00:00:43] been building in the background while
[00:00:45] nobody’s noticed, and it’s reached a
[00:00:46] level never seen before in history.
[00:00:49] Right now, 34% of everything American
[00:00:51] households own is in the stock market.
[00:00:54] That includes the houses, the cars, the
[00:00:56] cash in the bank, and whatever’s left of
[00:00:58] your crypto portfolio. It’s the highest
[00:01:00] share the Federal Reserve has ever
[00:01:02] recorded, and more than double the
[00:01:04] long-run average of 16%. Americans now
[00:01:08] have twice as much of their wealth in
[00:01:09] the stock market as they do in their
[00:01:11] houses. That’s the widest gap ever
[00:01:13] recorded. We’ve only seen a spike like
[00:01:15] this three times before. In 1968, stocks
[00:01:18] hit 21%. In 2000, 27%, and in 2021, 29%.
[00:01:24] And all three times, the market crashed
[00:01:26] not long after. After 1968, stocks lost
[00:01:29] more than a third of their value over
[00:01:31] the next year and a half. After 2000,
[00:01:33] nearly half in 2 and 1/2 years. After
[00:01:35] 2021, a quarter of their value in 9
[00:01:38] months. Today, we’re at 34 cents of
[00:01:41] every dollar, and the cash Americans are
[00:01:43] holding against all of it, just 9 cents.
[00:01:45] Once again, the widest gap ever
[00:01:47] recorded. The American household is
[00:01:49] starting to look like a poker game where
[00:01:51] someone just pushed every chip to the
[00:01:53] middle of the table and said, “All in.”
[00:01:55] So, how did we end up this deep in the
[00:01:57] stock market? The short answer is it
[00:01:59] wasn’t entirely on purpose. But before I
[00:02:01] jump into that, let’s take a quick pause
[00:02:04] to talk about something else. In 2024,
[00:02:06] personal data on 73 million current and
[00:02:09] former customers of one of the world’s
[00:02:10] largest telecommunications companies was
[00:02:12] found for sale on the dark web. The data
[00:02:14] included social security numbers, dates
[00:02:17] of birth, and account passcodes. And
[00:02:19] this is just one example. Phone carriers
[00:02:21] have been dealing with data breaches for
[00:02:22] years, which brings us to the sponsor of
[00:02:24] today’s video, Cape. Cape is America’s
[00:02:27] privacy-first mobile carrier, and that
[00:02:29] starts the moment you sign up. Cape
[00:02:31] won’t ask for your name, your social
[00:02:32] security number, or your address. While
[00:02:34] most carriers store your information for
[00:02:36] as long as possible, Cape does the
[00:02:38] opposite. They delete your call and text
[00:02:40] metadata after just 24 hours. And when
[00:02:42] you have Cape’s identifier rotation
[00:02:44] turned on, your network ID automatically
[00:02:46] rotates every 24 hours. To the network,
[00:02:49] you look like a different subscriber
[00:02:51] every day, which makes you significantly
[00:02:53] harder to track. Cape’s newest
[00:02:55] multi-network coverage feature lets you
[00:02:56] automatically switch between two
[00:02:58] nationwide networks on one SIM. This not
[00:03:00] only gives you more coverage, but once
[00:03:02] again, makes you harder to track. Cape
[00:03:04] lets you stay connected without having
[00:03:06] to compromise your privacy. If you’re
[00:03:08] interested, you can get 33% off your
[00:03:10] first 6 months of Cape using the link
[00:03:12] below. Promo code Casual Finance 33.
[00:03:16] Thank you to Cape for sponsoring this
[00:03:17] video. Now, back to how exactly
[00:03:19] Americans got this overexposed to the
[00:03:21] stock market. The market bottomed out in
[00:03:23] 2009 in the middle of the financial
[00:03:25] crisis. Since then, the index is up more
[00:03:28] than 11-fold, and measured in dollars,
[00:03:30] nothing in history comes close to this
[00:03:32] rise. Since that 2009 bottom, the stock
[00:03:35] Americans own has gone from around $9
[00:03:37] trillion to $74 trillion. Since just the
[00:03:39] end of 2022 alone, it’s roughly doubled.
[00:03:42] So, most people never made a decision to
[00:03:44] be overexposed to the market. The market
[00:03:46] just kept going up, and every time it
[00:03:47] did, stocks became a bigger share of
[00:03:49] everything they owned. It comes from the
[00:03:51] old adage of the legendary portfolio
[00:03:53] manager and renowned investor Davey Day
[00:03:55] Trader, stocks only go up. Between just
[00:03:58] March and June of this year, the share
[00:04:00] of everything Americans own that sits in
[00:04:02] stocks jumped from 31 cents to 34. That
[00:04:05] one quarter added 10.7 trillion dollars
[00:04:08] to American stock portfolios, the
[00:04:10] biggest 3 months the Federal Reserve has
[00:04:12] ever recorded. And the new money coming
[00:04:14] in doesn’t look at the price. About 80%
[00:04:17] of what Americans put into their
[00:04:18] retirement accounts goes into stocks.
[00:04:20] Most of it through target date funds
[00:04:22] that buy on autopilot every payday. 20
[00:04:25] years ago, Congress passed the Pension
[00:04:27] Protection Act, which gave employers the
[00:04:29] green light to drop you into one of
[00:04:30] those funds by default. Today, six in 10
[00:04:33] plans sign you up without you lifting a
[00:04:35] finger. So, if you have a retirement
[00:04:36] account, there’s a decent chance you’re
[00:04:38] in this picture, whether you’ve ever
[00:04:39] thought about it or not. The Fed’s own
[00:04:41] records show ETF buying running at the
[00:04:44] fastest pace on record. And when a third
[00:04:46] of your net worth rises and falls with
[00:04:48] the market, the market doesn’t just
[00:04:50] determine what you’re worth. It starts
[00:04:51] accidentally determining what you spend.
[00:04:53] Because when your portfolio goes up, you
[00:04:56] feel richer and you spend like it. And
[00:04:58] suddenly, that new kitchen remodel

00:05:00

#

[00:05:00] you’ve been thinking about sounds
[00:05:01] reasonable. Economists have a name for
[00:05:03] this. It’s called the wealth effect.
[00:05:05] It’s when people start spending more
[00:05:07] because the market value of the things
[00:05:09] they own went up, not because they’re
[00:05:11] earning more. Last month, the Atlanta
[00:05:13] Fed published a paper on exactly this.
[00:05:15] They found that American consumer
[00:05:16] spending is now more than four times as
[00:05:19] sensitive to the stock market as it was
[00:05:21] in 1990. Then they ran a scenario
[00:05:23] analysis, and a 25% sell-off in the
[00:05:25] market makes real consumer spending
[00:05:28] negative for the year. And consumer
[00:05:29] spending is about 2/3 of the American
[00:05:32] economy. So, if it goes negative, you
[00:05:34] don’t just get an economic slowdown, you
[00:05:36] get a recession. And some of this is
[00:05:38] already showing up at the edges. JP
[00:05:40] Morgan Chase found that for
[00:05:41] higher-earning Americans, drawing from
[00:05:43] investment assets covered more than 10%
[00:05:46] of spending this year, more than twice
[00:05:48] what it was just a decade ago.
[00:05:50] California’s own budget office, which
[00:05:52] leans heavily on capital gains for its
[00:05:54] tax revenue, says that if a market
[00:05:56] downturn of more than 20% were to occur
[00:05:58] and last through the end of the year,
[00:06:00] revenue could come in 15 to 20 billion
[00:06:02] dollars below forecast, even if there
[00:06:04] were no economic recession. And if there
[00:06:07] were an economic recession, it would be
[00:06:09] even worse. And that’s where this turns
[00:06:11] into a loop. Rising stocks fuel spending
[00:06:13] in the economy. That spending becomes
[00:06:15] revenue for publicly traded companies.
[00:06:17] The revenue becomes profit, and profit
[00:06:20] is what investors point to when they
[00:06:21] explain why stocks deserve to keep going
[00:06:24] up. So, the economy and the stock market
[00:06:26] are feeding each other. That’s why this
[00:06:28] is a much bigger deal than the seven
[00:06:29] expensive tech stocks nobody can take
[00:06:31] their eyes off. Those are a problem for
[00:06:33] your portfolio. This is a problem for
[00:06:35] the whole economy. But households aren’t
[00:06:37] the only ones overexposed. The
[00:06:39] institutions are, too. Every month, Bank
[00:06:41] of America surveys the world’s biggest
[00:06:43] fund managers. This week, the data
[00:06:45] showed they were holding just 3.9% of
[00:06:48] their portfolios in cash. Bank of
[00:06:50] America’s rule for this is simple. Cash
[00:06:52] above 5% is a buy signal. Under 4% is a
[00:06:56] sell signal. A cash level under 4% is
[00:06:58] Bank of America’s way of saying
[00:07:00] investors are a little too excited, a
[00:07:02] little too invested, and the market
[00:07:04] might be running a little too hot. The
[00:07:06] fund managers in the survey now have
[00:07:08] their biggest overweight in tech stocks
[00:07:10] since August 2020, even though 45% of
[00:07:13] them say they expect an AI bubble by
[00:07:15] next year. So, the institutional
[00:07:17] investors see the bubble forming, but
[00:07:19] they keep piling in anyway. Why? John
[00:07:22] Maynard Keynes, who’s widely considered
[00:07:24] one of the most influential economists
[00:07:26] in history, figured out the answer about
[00:07:29] 90 years ago. In a book he wrote in 1936
[00:07:32] called The General Theory of Employment,
[00:07:34] Interest, and Money. In the book, he
[00:07:36] writes, “Worldly wisdom teaches that it
[00:07:38] is better for reputation to to
[00:07:40] conventionally than to succeed
[00:07:42] unconventionally.” What Keynes is
[00:07:43] pointing out is that in the eyes of the
[00:07:45] world, it’s better to fail by following
[00:07:47] the crowd than to succeed by going
[00:07:49] against it. That describes the problem
[00:07:51] fund managers are running into today.
[00:07:53] For a fund manager, there’s an
[00:07:54] incentives problem. If you lose money
[00:07:56] when everybody else loses money, you
[00:07:58] keep your job. If you sit in cash while
[00:08:00] the market rallies 20%, you get fired.
[00:08:03] So, even though 45% of them believe AI
[00:08:05] could head into a bubble next year, they
[00:08:07] just keep dumping their cash in. And it
[00:08:09] turns out, so is the rest of the world.
[00:08:11] Foreign investors now own $22 trillion
[00:08:14] of American stock, about a fifth of the
[00:08:16] entire market. American companies make
[00:08:19] up nearly 2/3 of the world’s main global
[00:08:21] stock index. And between March and June,
[00:08:23] foreigners bought American stocks at the
[00:08:25] fastest pace in history. So, everyone’s
[00:08:27] pushing their chips to the middle of the
[00:08:29] table, households, institutions, the
[00:08:31] rest of the world. But what makes this
[00:08:33] scary is that a record number of those
[00:08:35] bets are being made with borrowed money.
[00:08:37] And when you bet with borrowed money,
[00:08:39] you don’t always get what happens next.
[00:08:41] In June, margin debt, which is the money
[00:08:44] investors borrow from their broker to
[00:08:45] buy more stock, hit a record $1.5
[00:08:48] trillion.
[00:08:49] That’s five times larger than it was at
[00:08:51] the height of the dot-com bubble, and
[00:08:53] 1.5 times larger than the 2021 peak.
[00:08:56] Over the last 30 years, margin debt has
[00:08:58] only grown this fast three other times.
[00:09:01] The first was late 1999 into early 2000
[00:09:04] as the dot-com bubble peaked. Next was
[00:09:06] mid-2007, just months before the
[00:09:08] financial crisis. And then spring 2021,
[00:09:12] less than a year before the 2022 bear
[00:09:14] market. When each of those markets
[00:09:15] crashed, margin debt got cut by as much
[00:09:18] as half. A lot of that wasn’t people
[00:09:19] choosing to pay back their loans. It was
[00:09:21] brokers demanding their money back and
[00:09:23] selling people’s stock to get it. Margin
[00:09:25] debt is just one of the ways to borrow
[00:09:27] against your stock. There’s a rich
[00:09:28] person’s version as well. It’s the one
[00:09:30] you’ve probably seen want-to-be personal
[00:09:32] finance influencers on Instagram talking
[00:09:34] about. Instead of selling stock and
[00:09:36] paying tax on the gains, you just borrow
[00:09:39] against your stock. There’s no need to
[00:09:41] sell any positions, and the bank holds
[00:09:43] your portfolio as collateral. At Morgan
[00:09:45] Stanley, loans borrowed against stock
[00:09:47] portfolios are booming. They’re up 20%
[00:09:50] in a year. When you borrow against
[00:09:51] stock, the stock is the collateral, and
[00:09:53] the lender has one rule. The loan has to
[00:09:56] stay covered by what that collateral is
[00:09:57] worth. So, when the stock falls, the
[00:09:59] collateral shrinks, but the loan stays

00:10:01

#

[00:10:01] the same size. Once that gap gets too
[00:10:03] big, your broker sends you what’s called
[00:10:05] a margin call. It’s a demand for more
[00:10:08] money, and it usually comes with a
[00:10:09] deadline. It’s not exactly the phone
[00:10:11] call anyone’s hoping to get, and there
[00:10:13] are only two ways to answer it. You send
[00:10:15] them cash, or they sell your stock. When
[00:10:17] it comes to a margin call, they don’t
[00:10:19] need your permission. FINRA’s own
[00:10:21] website says, “A firm isn’t required to
[00:10:24] notify you if your account equity drops
[00:10:26] below the minimum maintenance equity.”
[00:10:28] It also says, “Firms don’t have to let
[00:10:30] you choose which securities or assets
[00:10:32] are sold to meet a margin call.” So, if
[00:10:34] you can’t come up with the cash, your
[00:10:36] broker can start dumping your portfolio
[00:10:38] on the open market for you. Whatever
[00:10:39] stocks they pick, whenever they decide,
[00:10:42] at whatever price the market’s paying
[00:10:43] that day, which is why a margin call can
[00:10:45] trap investors in very tricky
[00:10:47] situations. And in July, we got to watch
[00:10:49] this happen. If you haven’t heard the
[00:10:51] story of Leopold Aschenbrenner yet,
[00:10:53] you’re in for a treat. Leopold graduated
[00:10:55] from Columbia at just 19 years old as
[00:10:58] valedictorian. He then went to work at
[00:10:59] OpenAI, but in 2024, OpenAI fired him.
[00:11:03] Two months later, he published a
[00:11:05] 165-page essay called Situational
[00:11:07] Awareness. Silicon Valley instantly fell
[00:11:10] in love with it and passed it around
[00:11:11] like scripture. Then he did what any guy
[00:11:14] in his early 20s with a viral blog post
[00:11:16] and no professional investing experience
[00:11:18] would do. He started a hedge fund and
[00:11:20] named it after the blog post. The
[00:11:22] strategy was simple. Borrow up to four
[00:11:24] times the fund’s own money and bet it
[00:11:26] all on the companies building AI
[00:11:28] infrastructure. And at first, it was
[00:11:30] working. At the peak, Leopold’s hedge
[00:11:33] fund was managing about $45 billion in
[00:11:35] assets. In 2 years, the fund was up more
[00:11:38] than 1,000%. Wall Street was treating
[00:11:40] him like royalty. Then, in July of this
[00:11:42] year, AI stocks collapsed. Leopold’s
[00:11:45] positions got hammered, and he faced
[00:11:47] margin calls. On July 30th, Leopold’s
[00:11:50] fund was forced to sell to cover the
[00:11:51] obligations. Citadel stepped in and
[00:11:54] scooped up most of its stock portfolio
[00:11:56] at a massive discount. Leopold’s fund
[00:11:58] had collapsed from $45 billion
[00:12:00] to around 10 billion in assets as
[00:12:02] semiconductor stocks fell and the margin
[00:12:04] call showed up. That’s how fast things
[00:12:06] can turn when margins involved. And it
[00:12:08] wasn’t just Leopold’s fund, either. His
[00:12:10] biggest bets were some of the most
[00:12:11] crowded trades on Wall Street, and the
[00:12:13] banks were making margin calls on other
[00:12:15] funds holding the same stocks. That
[00:12:17] month, margin debt across the whole
[00:12:18] market fell $85 billion, the biggest
[00:12:22] 1-month drop on record. But it’s also
[00:12:24] not just margin that’s piling up in the
[00:12:25] market. There’s a new kind of leverage
[00:12:27] that didn’t even exist a few years ago.
[00:12:29] Funds that take a single stock’s daily
[00:12:31] moves and magnify them. Leveraged ETFs
[00:12:34] are funds that promise two or three
[00:12:36] times the daily move of a stock or an
[00:12:38] index. If Nvidia goes up 2% today, the
[00:12:41] fund goes up 4%. If Nvidia falls 2%, the
[00:12:44] fund falls 4%. In the first half of this
[00:12:47] year alone, 218 leveraged stock funds
[00:12:50] launched. In all of 2021, there was 12.
[00:12:53] Assets in these leveraged funds doubled
[00:12:55] from April to June, with leverage linked
[00:12:57] to semiconductor exposure nearly
[00:12:59] tripling,
[00:13:00] leaving tech and semiconductors
[00:13:02] accounting for 67% of all leveraged ETF
[00:13:05] assets. Nomura estimates these funds
[00:13:07] have to trade about $9 billion
[00:13:09] for every 1% the market moves. And these
[00:13:11] are mostly mechanical trades that must
[00:13:13] execute to maintain the promised
[00:13:15] leverage, which can add buying pressure
[00:13:17] late in rallies and selling pressure
[00:13:19] during market declines. So a bad day
[00:13:21] gets worse, and a good day gets better.
[00:13:23] It pours gas on whatever fire the
[00:13:25] market’s having. At the peak, daily
[00:13:27] rebalancing hit an estimated record of
[00:13:29] $50 billion
[00:13:30] in a single day. Margin calls force
[00:13:33] sales when the borrower can’t cover.
[00:13:35] Leveraged ETFs sell when the math says
[00:13:37] so. And if both of these kick in at the
[00:13:39] same time, they can spiral into what’s
[00:13:41] called a liquidation cascade. It’s a
[00:13:43] chain reaction where price moves trigger
[00:13:46] forced sales. Those forced sales push
[00:13:48] prices lower, which then causes more
[00:13:50] positions to breach their margin
[00:13:51] thresholds and forces even more sales.
[00:13:54] So, what could cause something like this
[00:13:56] to happen? That’s the part nobody knows,
[00:13:58] including me. What I do know, though, is
[00:14:00] that whenever something does hit, it’s
[00:14:02] going to land on a market that’s more
[00:14:04] crowded and more borrowed than it’s ever
[00:14:06] been. It doesn’t need a disaster to get
[00:14:08] started. It just needs a first push.
[00:14:10] Something that hits a lot of the same
[00:14:11] positions at the same time and causes
[00:14:13] the first domino to fall. The obvious
[00:14:15] place to look is the AI trade, but it
[00:14:18] could also be bond yields. The 10-year
[00:14:19] Treasury just touched its highest level
[00:14:21] in more than two decades. According to
[00:14:23] Bank of America’s fund manager
[00:14:25] survey, if the 10-year yield hits 5.9%,
[00:14:28] investors would likely shift their asset
[00:14:30] allocation from stocks to bonds. Or it
[00:14:33] could be something else entirely. Nobody
[00:14:35] can predict the trigger, and that’s
[00:14:36] exactly the point. The market’s fragile
[00:14:39] right now. When things are this crowded
[00:14:41] and carrying this much borrowed money,
[00:14:43] it doesn’t take a big push to start
[00:14:44] knocking the dominoes over. In the
[00:14:46] meantime, I’ll keep making videos that
[00:14:48] break down what’s actually going on in
[00:14:49] the economy. So, if you want to stay on
[00:14:51] top of the deeper stuff happening in
[00:14:53] markets, hit subscribe. And leave a
[00:14:55] comment letting me know what you thought
[00:14:57] of this one. Like I said at the start of
[00:14:58] this video, everybody knows the market
[00:15:00] is expensive, but everybody’s watching

00:15:02

#

[00:15:02] the wrong concentration. Everybody’s
[00:15:04] watching the seven companies that make
[00:15:06] up a third of the index, but underneath
[00:15:08] that, a third of everything American
[00:15:10] households own is sitting in the market.
[00:15:12] And underneath that, the economy is
[00:15:14] leaning on consumer spending fueled by
[00:15:16] stock market gains. Running through all
[00:15:18] of it is a record amount of borrowed
[00:15:19] money, which means more of this market
[00:15:21] than ever before belongs to people who
[00:15:23] can be forced to sell if push comes to
[00:15:25] shove. The most dangerous thing about
[00:15:27] this market isn’t even that it’s
[00:15:29] expensive, it’s that so much of America
[00:15:31] now depends on it staying expensive.
[00:15:38] » [music]

中文全文译文

#

00:00:00

#

[00:00:00] 人人都知道，现在市场估值很贵。
[00:00:02] 这点不需要天才才能看出来。
[00:00:04] 七家公司就占了
[00:00:06] 整个指数的三分之一。投资管理
[00:00:08] 公司资本集团发现，如今的
[00:00:10] 市场集中程度位居
[00:00:12] 历史最高之列。按照最常用的长期
[00:00:15] 股票估值衡量指标来看，
[00:00:17] 目前市场的定价水平，过去只
[00:00:19] 出现过一次，那就是
[00:00:21] 互联网泡沫的顶峰。但这并不是
[00:00:23] 什么秘密信息。如果你
[00:00:25] 在过去一年里接触过
[00:00:26] 互联网，肯定都听说过这些：
[00:00:28] 市场很贵，集中度很高，
[00:00:30] 可它依然不断上涨。就在
[00:00:32] 本月早些时候，收盘价又创下了
[00:00:34] 历史新高。当大家都忙着衡量
[00:00:36] 有多少市场份额集中在少数
[00:00:38] 科技公司身上时，却没人衡量
[00:00:40] 美国究竟有多少财富押在股市里。
[00:00:41] 这才是眼下真正的危险。它
[00:00:43] 一直在不知不觉中累积，
[00:00:45] 几乎无人察觉，如今已经到了
[00:00:46] 历史上从未出现过的程度。
[00:00:49] 目前，美国家庭所拥有的全部
[00:00:51] 资产中，有34%在股票市场里。
[00:00:54] 全部资产包括房子、汽车、
[00:00:56] 银行里的现金，以及你
[00:00:58] 加密货币投资组合里剩下的东西。这是
[00:01:00] 美联储有记录以来
[00:01:02] 最高的占比，也超过
[00:01:04] 16%长期平均水平的两倍。如今美国人
[00:01:08] 放在股票市场里的财富，
[00:01:09] 已经是放在
[00:01:11] 住房里的两倍。这是有记录以来
[00:01:13] 最大的差距。此前我们只见过
[00:01:15] 三次类似的飙升：1968年，股票占比
[00:01:18] 达到21%；2000年是27%；2021年是29%。
[00:01:24] 而这三次，股市都在
[00:01:26] 不久后崩盘。1968年之后，股票在
[00:01:29] 随后一年半里，损失了
[00:01:31] 超过三分之一的价值。2000年之后，
[00:01:33] 两年半内损失了近一半。2021年之后，
[00:01:35] 九个月内损失了四分之一
[00:01:38] 的价值。如今，每一美元资产中有34美分在股市，
[00:01:41] 而美国人持有的现金，
[00:01:43] 与之相比只有9美分。
[00:01:45] 这再一次创下了有记录以来
[00:01:47] 最大的差距。美国家庭
[00:01:49] 越来越像一场扑克游戏，
[00:01:51] 有人把全部筹码推进
[00:01:53] 桌子中央，说了一句：“全押。”
[00:01:55] 那么，我们为什么会如此深地卷入
[00:01:57] 股票市场？简短的回答是：
[00:01:59] 这并不完全是有意为之。但在我
[00:02:01] 展开之前，先暂停一下，
[00:02:04] 聊点别的。2024年，
[00:02:06] 一家全球最大电信公司之一的7300万名现任及
[00:02:09] 前任客户的个人数据，
[00:02:10] 被发现
[00:02:12] 在暗网上出售。这些数据
[00:02:14] 包括社会安全号码、
[00:02:17] 出生日期以及账户密码。而
[00:02:19] 这还只是一个例子。移动运营商
[00:02:21] 多年来一直在应对
[00:02:22] 数据泄露问题。这就引出了
[00:02:24] 本期视频的赞助商Cape。Cape是美国一家
[00:02:27] 将隐私放在首位的移动运营商，而且这一点
[00:02:29] 从你注册的那一刻就开始了。Cape
[00:02:31] 不会索要你的姓名、社会
[00:02:32] 安全号码或地址。大多数
[00:02:34] 运营商都会尽可能长时间地
[00:02:36] 保存你的信息，而Cape恰恰
[00:02:38] 相反。他们会在24小时后删除你的通话和短信
[00:02:40] 元数据。而当
[00:02:42] 你开启Cape的标识符轮换
[00:02:44] 功能后，你的网络ID会自动
[00:02:46] 每24小时轮换一次。在网络看来，
[00:02:49] 你每天都像是
[00:02:51] 一个不同的用户，因此
[00:02:53] 更难被追踪。Cape最新的
[00:02:55] 多网络覆盖功能，允许你
[00:02:56] 使用一张SIM卡，在两个
[00:02:58] 全国性网络之间自动切换。这不仅
[00:03:00] 让你获得更多覆盖，而且
[00:03:02] 同样让你更难被追踪。Cape
[00:03:04] 让你保持连接，
[00:03:06] 同时无需牺牲隐私。如果你
[00:03:08] 有兴趣，可以通过下方链接，获得
[00:03:10] Cape前六个月费用减免33%的优惠。
[00:03:12] 优惠码是Casual Finance 33。
[00:03:16] 感谢Cape赞助本期
[00:03:17] 视频。现在回到正题，看看
[00:03:19] 美国人到底是如何在
[00:03:21] 股市中过度配置的。市场在
[00:03:23] 2009年金融危机期间
[00:03:25] 触底。自那以后，该指数上涨到了
[00:03:28] 原来的11倍以上。按美元规模衡量，
[00:03:30] 历史上没有任何一次
[00:03:32] 上涨能与之相比。自2009年的低点以来，
[00:03:35] 美国人持有的股票已从大约9万亿
[00:03:37] 美元增至74万亿美元。仅从
[00:03:39] 2022年底算起，就大致翻了一倍。
[00:03:42] 所以，大多数人并没有主动决定
[00:03:44] 要过度配置股市。只是市场
[00:03:46] 不断上涨，而每上涨一次，
[00:03:47] 股票在他们全部资产中的
[00:03:49] 占比就变得更大。正如那位
[00:03:51] 传奇投资组合经理、
[00:03:53] 著名投资人Davey Day
[00:03:55] Trader的老话：“股票只会上涨。”仅在今年
[00:03:58] 3月至6月之间，
[00:04:00] 美国人全部资产中
[00:04:02] 股票的占比，就从每美元31美分跃升到了34美分。
[00:04:05] 仅这一个季度，就为美国人的股票投资组合增加了10.7万亿
[00:04:08] 美元，成为
[00:04:10] 美联储有记录以来规模最大的
[00:04:12] 三个月。而新流入的资金
[00:04:14] 并不看价格。美国人投入
[00:04:17] 退休账户的资金中，大约80%
[00:04:18] 流向了股票。
[00:04:20] 其中大部分通过目标日期基金，
[00:04:22] 在每个发薪日自动买入。20
[00:04:25] 年前，美国国会通过了《养老金
[00:04:27] 保护法》，允许雇主
[00:04:29] 在默认情况下，将你纳入
[00:04:30] 这类基金。如今，每十个计划中就有六个
[00:04:33] 会自动让你加入，完全不需要你
[00:04:35] 动手。所以，如果你有退休
[00:04:36] 账户，你很可能已经
[00:04:38] 处于这种情况中，无论你是否
[00:04:39] 考虑过这件事。美联储自己的
[00:04:41] 记录显示，ETF买入正以
[00:04:44] 有记录以来最快的速度进行。而当你三分之一的
[00:04:46] 净资产随着
[00:04:48] 市场涨跌时，市场就不再只是
[00:04:50] 决定你有多少钱。它开始
[00:04:51] 无意间决定你会花多少钱。
[00:04:53] 因为当你的投资组合上涨时，
[00:04:56] 你会觉得自己更富有，花钱也就更大方。
[00:04:58] 于是突然间，那次新的厨房翻新，

00:05:00

#

[00:05:00] 也就是你一直在考虑的那次，听起来
[00:05:01] 就很合理了。经济学家为
[00:05:03] 这种现象起了个名字，叫“财富效应”。
[00:05:05] 它指的是，人们开始增加支出，
[00:05:07] 原因是他们所拥有资产的市场价值
[00:05:09] 上升了，而不是因为
[00:05:11] 收入增加。上个月，亚特兰大
[00:05:13] 联储就这个问题发表了一篇论文。
[00:05:15] 他们发现，美国消费者
[00:05:16] 支出对股票市场的敏感程度，
[00:05:19] 如今已经超过
[00:05:21] 1990年的四倍。随后他们做了情景
[00:05:23] 分析：如果股市下跌25%，
[00:05:25] 实际消费支出的
[00:05:28] 全年增速就会变成负值。而消费
[00:05:29] 支出约占美国
[00:05:32] 经济的三分之二。因此，如果它转为负增长，
[00:05:34] 你面对的就不仅是经济放缓，
[00:05:36] 而是经济衰退。而其中一些迹象
[00:05:38] 已经在局部显现。摩根
[00:05:40] 大通发现，对于
[00:05:41] 收入较高的美国人来说，从
[00:05:43] 投资资产中提取的资金，支付了今年超过10%的
[00:05:46] 消费支出，是
[00:05:48] 十年前的两倍多。
[00:05:50] 加州自身的预算部门
[00:05:52] 其税收收入高度依赖资本利得，
[00:05:54] 该部门表示，如果市场
[00:05:56] 出现超过20%的下跌，
[00:05:58] 并且持续到年底，
[00:06:00] 即使没有发生经济衰退，收入也可能比预测少150亿至200亿
[00:06:02] 美元。
[00:06:04] 而如果真的
[00:06:07] 发生经济衰退，情况就会
[00:06:09] 更加糟糕。这里就形成了
[00:06:11] 一个循环：股价上涨推动
[00:06:13] 经济中的消费支出。这些支出成为
[00:06:15] 上市公司的营收。
[00:06:17] 营收变成利润，而利润
[00:06:20] 又成为投资者
[00:06:21] 解释股票为何理应继续
[00:06:24] 上涨的依据。因此，经济和股市
[00:06:26] 在相互推动。这就是为什么
[00:06:28] 这件事比那七只
[00:06:29] 让所有人都目不转睛盯着的高估值科技股，
[00:06:31] 重要得多。那些股票关系到
[00:06:33] 你的投资组合，而这件事关系到
[00:06:35] 整个经济。但过度配置股市的
[00:06:37] 并不只有家庭。
[00:06:39] 机构也是如此。美国银行每个月
[00:06:41] 都会调查全球最大的
[00:06:43] 基金经理。本周公布的数据
[00:06:45] 显示，他们投资组合中的现金占比仅为3.9%。
[00:06:48] 美国
[00:06:50] 银行对此有一条简单的规则：现金
[00:06:52] 占比高于5%是买入信号，低于4%则是
[00:06:56] 卖出信号。现金水平低于4%，
[00:06:58] 在美国银行看来意味着
[00:07:00] 投资者有些过于兴奋，
[00:07:02] 资金投入得有些过多，市场
[00:07:04] 可能也有些过热了。这次
[00:07:06] 调查中的基金经理，
[00:07:08] 对科技股的超配幅度达到了
[00:07:10] 2020年8月以来最大，尽管其中45%
[00:07:13] 表示，他们预计到明年
[00:07:15] 会出现AI泡沫。也就是说，机构
[00:07:17] 投资者看到了泡沫正在形成，
[00:07:19] 却依然不断涌入。为什么？约翰·
[00:07:22] 梅纳德·凯恩斯，被广泛认为是
[00:07:24] 历史上最有影响力的
[00:07:26] 经济学家之一，大约在
[00:07:29] 90年前就找到了答案。他在1936年写了一本书，
[00:07:32] 名为《就业、利息
[00:07:34] 和货币通论》。他在书中
[00:07:36] 写道：“世俗的智慧告诉我们，
[00:07:38] 从名声上说，宁可
[00:07:40] 循规蹈矩地失败，也不要
[00:07:42] 以非常规方式成功。”凯恩斯
[00:07:43] 想指出的是，在
[00:07:45] 世人眼中，随大流而失败，
[00:07:47] 好过逆着人群
[00:07:49] 取得成功。这就解释了
[00:07:51] 如今基金经理所面临的问题。
[00:07:53] 对基金经理来说，存在一个
[00:07:54] 激励机制的问题。如果你亏钱时，
[00:07:56] 其他人也都亏钱，
[00:07:58] 你就能保住工作。但如果你持有现金，
[00:08:00] 而市场上涨了20%，你就会被解雇。
[00:08:03] 所以，即使其中45%相信AI
[00:08:05] 明年可能走向泡沫，
[00:08:07] 他们还是不断把现金投进去。而
[00:08:09] 事实证明，世界其他地方也是如此。
[00:08:11] 外国投资者如今持有22万亿美元
[00:08:14] 的美国股票，约占
[00:08:16] 整个市场的五分之一。美国公司在
[00:08:19] 全球主要股票指数中，权重接近
[00:08:21] 三分之二。而在3月至6月之间，
[00:08:23] 外国人买入美股的速度
[00:08:25] 创下了历史之最。所以，所有人都在
[00:08:27] 把筹码推向
[00:08:29] 桌子中央：家庭、机构，
[00:08:31] 以及世界其他地方。但令人
[00:08:33] 害怕的是，这些押注中有创纪录多的部分
[00:08:35] 使用了借来的钱。
[00:08:37] 而当你用借来的钱下注时，
[00:08:39] 你并不总能预料接下来会发生什么。
[00:08:41] 6月份，融资余额，也就是
[00:08:44] 投资者从券商借来
[00:08:45] 购买更多股票的资金，创下了1.5万亿
[00:08:48] 美元的纪录。
[00:08:49] 这一规模是
[00:08:51] 互联网泡沫顶峰时期的五倍，
[00:08:53] 也是2021年峰值的1.5倍。
[00:08:56] 过去30年间，融资余额
[00:08:58] 仅在另外三段时期增长得这么快。
[00:09:01] 第一次是1999年末至2000年初，
[00:09:04] 也就是互联网泡沫见顶时。之后是
[00:09:06] 2007年年中，距离
[00:09:08] 金融危机只有几个月。然后是2021年春季，
[00:09:12] 距离2022年熊市
[00:09:14] 不到一年。在这些市场
[00:09:15] 崩盘时，融资余额的降幅
[00:09:18] 都曾高达一半。其中很大一部分，并不是人们
[00:09:19] 主动决定偿还贷款，而是
[00:09:21] 券商要求拿回资金，
[00:09:23] 并卖出投资者的股票来收回借款。融资
[00:09:25] 只是以股票作为抵押借钱的
[00:09:27] 一种方式。还有一种
[00:09:28] 富人版的做法。你可能
[00:09:30] 看过Instagram上那些想当个人
[00:09:32] 理财网红的人，
[00:09:34] 谈论这种方式：不是卖出股票，
[00:09:36] 为收益缴税，而是直接以
[00:09:39] 股票作为抵押借款。无需
[00:09:41] 卖出任何持仓，银行会把
[00:09:43] 你的投资组合作为抵押物。在摩根
[00:09:45] 士丹利，以股票
[00:09:47] 投资组合为抵押的贷款正在迅速增加，一年内增长了20%。
[00:09:50] 当你以股票作抵押
[00:09:51] 借钱时，股票就是抵押物，
[00:09:53] 而贷款方有一条规则：贷款必须
[00:09:56] 始终得到抵押物价值的
[00:09:57] 覆盖。所以，当股价下跌时，
[00:09:59] 抵押物缩水，但贷款依然是

00:10:01

#

[00:10:01] 原来的规模。一旦这个差距变得
[00:10:03] 过大，券商就会发来所谓的
[00:10:05] 追加保证金通知。这是在要求你补充
[00:10:08] 资金，而且通常附带
[00:10:09] 截止时间。这肯定不是任何人
[00:10:11] 盼着接到的电话，而你
[00:10:13] 只有两种应对方式：给他们
[00:10:15] 现金，或者让他们卖掉你的股票。对于
[00:10:17] 追加保证金这件事，他们
[00:10:19] 不需要得到你的允许。美国金融业监管局FINRA自己的
[00:10:21] 网站写道：“如果你的
[00:10:24] 账户权益跌至
[00:10:26] 最低维持权益要求以下，公司并没有义务通知你。”
[00:10:28] 网站还写道：“公司不必让
[00:10:30] 你选择出售哪些证券或资产，
[00:10:32] 来满足追加保证金的要求。”所以，如果
[00:10:34] 你拿不出现金，
[00:10:36] 券商就可以替你开始在
[00:10:38] 公开市场上抛售你的投资组合。
[00:10:39] 他们想选哪些股票，就选哪些；想什么时候卖，就什么时候卖；
[00:10:42] 不管当天市场给出的
[00:10:43] 价格是多少。这就是为什么追加保证金通知会
[00:10:45] 让投资者陷入非常棘手的
[00:10:47] 境地。而在7月份，我们就亲眼看到了
[00:10:49] 这种情况发生。如果你还没听过
[00:10:51] 利奥波德·阿申布伦纳的故事，
[00:10:53] 那接下来可有意思了。利奥波德年仅
[00:10:55] 19岁就从哥伦比亚大学毕业，
[00:10:58] 并担任毕业生代表。随后他去了
[00:10:59] OpenAI工作，但2024年，OpenAI解雇了他。
[00:11:03] 两个月后，他发表了一篇
[00:11:05] 长达165页、名为《情境
[00:11:07] 意识》的文章。硅谷立刻
[00:11:10] 爱上了这篇文章，并像传播
[00:11:11] 圣典一样到处传阅。接着，他做了一件
[00:11:14] 任何二十出头、写过一篇爆款博文，
[00:11:16] 却没有职业投资经验的人
[00:11:18] 都会做的事：创办一家对冲基金，
[00:11:20] 并用那篇博文为基金命名。
[00:11:22] 策略很简单：借入最高相当于
[00:11:24] 基金自有资金四倍的钱，把这些资金
[00:11:26] 全部押在建设AI
[00:11:28] 基础设施的公司身上。起初，这套策略
[00:11:30] 确实有效。在巅峰时期，利奥波德的对冲
[00:11:33] 基金管理着约450亿美元
[00:11:35] 资产。两年内，基金涨幅超过了
[00:11:38] 1000%。华尔街对待他
[00:11:40] 就像对待皇室成员。随后，在今年7月，
[00:11:42] AI股票崩盘。利奥波德的
[00:11:45] 持仓遭受重创，他面临
[00:11:47] 追加保证金的要求。7月30日，利奥波德的
[00:11:50] 基金被迫卖出资产，以履行
[00:11:51] 相关义务。Citadel介入，
[00:11:54] 以大幅折价接手了它大部分
[00:11:56] 股票投资组合。利奥波德基金的资产
[00:11:58] 从450亿美元骤降
[00:12:00] 至约100亿美元，原因是
[00:12:02] 半导体股下跌，以及追加保证金
[00:12:04] 要求的到来。这说明，一旦涉及融资杠杆，
[00:12:06] 局势逆转得可以有多快。而且
[00:12:08] 出事的并不只有利奥波德的基金。
[00:12:10] 他的最大押注，是华尔街最
[00:12:11] 拥挤的交易之一，而
[00:12:13] 银行也在向持有相同股票的其他
[00:12:15] 基金发出追加保证金要求。那个
[00:12:17] 月，全市场的融资余额
[00:12:18] 减少了850亿美元，创下了
[00:12:22] 有记录以来最大的单月降幅。但在
[00:12:24] 市场里不断累积的，
[00:12:25] 还不只有融资借款。还有一种新的杠杆形式，
[00:12:27] 几年前甚至还不存在。
[00:12:29] 这种基金会把单只股票的每日
[00:12:31] 波动放大。杠杆ETF
[00:12:34] 承诺提供某只股票或某个
[00:12:36] 指数每日涨跌幅的两倍或三倍。
[00:12:38] 如果英伟达今天上涨2%，
[00:12:41] 这只基金就上涨4%；如果英伟达下跌2%，
[00:12:44] 基金就下跌4%。仅在今年上半年，
[00:12:47] 就有218只杠杆股票基金
[00:12:50] 推出，而2021年全年只有12只。
[00:12:53] 这些杠杆基金的资产规模，
[00:12:55] 从4月至6月翻了一倍；其中与
[00:12:57] 半导体敞口相关的杠杆规模，几乎
[00:12:59] 增长到了三倍，
[00:13:00] 使得科技和半导体
[00:13:02] 占全部杠杆ETF资产的67%。
[00:13:05] 野村估计，市场每波动1%，这些基金
[00:13:07] 就必须进行约90亿美元
[00:13:09] 的交易。而这些
[00:13:11] 大多是机械性交易，必须
[00:13:13] 执行，以维持所承诺的
[00:13:15] 杠杆水平。这可能在上涨后期增加买盘
[00:13:17] 压力，在市场下跌期间增加卖盘
[00:13:19] 压力。于是，糟糕的一天
[00:13:21] 变得更糟，美好的一天则变得更好。
[00:13:23] 市场无论燃起什么样的火，
[00:13:25] 它都会火上浇油。在高峰时期，每日
[00:13:27] 再平衡交易的规模估计创下了
[00:13:29] 500亿美元
[00:13:30] 的单日纪录。当借款人无力补足资金时，追加保证金通知会迫使其
[00:13:33] 卖出。
[00:13:35] 杠杆ETF则在数学规则要求时
[00:13:37] 卖出。如果这两种机制
[00:13:39] 同时启动，就可能演变成
[00:13:41] 所谓的连环清算。这是一种
[00:13:43] 连锁反应：价格波动触发
[00:13:46] 被迫卖出，这些卖出又把
[00:13:48] 价格压得更低，继而导致更多
[00:13:50] 持仓跌破保证金
[00:13:51] 门槛，引发更多被迫卖出。
[00:13:54] 那么，什么可能导致这种事情
[00:13:56] 发生？这一点没人知道，
[00:13:58] 包括我在内。但我知道的是，
[00:14:00] 无论将来发生什么冲击，
[00:14:02] 它所击中的市场，都会比
[00:14:04] 以往任何时候都更拥挤、借债更多。
[00:14:06] 这一过程不需要一场灾难才能
[00:14:08] 启动，只需要最初的一推。
[00:14:10] 某件事同时冲击大量相同的
[00:14:11] 持仓，并让
[00:14:13] 第一块多米诺骨牌倒下。最明显的
[00:14:15] 观察对象是AI交易，但
[00:14:18] 债券收益率也可能成为触发因素。10年期
[00:14:19] 美国国债收益率刚刚触及
[00:14:21] 二十多年来的最高水平。根据
[00:14:23] 美国银行的基金经理
[00:14:25] 调查，如果10年期收益率达到5.9%，
[00:14:28] 投资者很可能会把资产
[00:14:30] 配置从股票转向债券。当然，也
[00:14:33] 可能是完全不同的其他因素。没人
[00:14:35] 能够预测触发因素，这也
[00:14:36] 正是重点所在。眼下市场
[00:14:39] 很脆弱。当交易如此拥挤，
[00:14:41] 又承载着这么多借来的资金时，
[00:14:43] 并不需要多大的一推，
[00:14:44] 就能让多米诺骨牌开始倒下。在此
[00:14:46] 期间，我会继续制作视频，
[00:14:48] 分析经济中
[00:14:49] 究竟正在发生什么。所以，如果你想
[00:14:51] 持续了解市场中
[00:14:53] 更深层次的变化，就点个订阅。也欢迎
[00:14:55] 留言，告诉我你对
[00:14:57] 这期视频的看法。就像我在
[00:14:58] 视频开头说的，人人都知道市场
[00:15:00] 估值很贵，但所有人关注的

00:15:02

#

[00:15:02] 都是错误的集中风险。所有人都在
[00:15:04] 盯着那七家占据
[00:15:06] 指数三分之一的公司，但在这背后，
[00:15:08] 美国家庭全部资产的三分之一，
[00:15:10] 都放在股市里。
[00:15:12] 而在更深一层，经济
[00:15:14] 又依赖股市上涨所推动的
[00:15:16] 消费支出。贯穿这一切的，
[00:15:18] 是创纪录规模的借入
[00:15:19] 资金。这意味着，这个市场中
[00:15:21] 比以往更多的资产，属于那些
[00:15:23] 一旦事态恶化就可能被迫
[00:15:25] 卖出的人。这个市场最危险的地方，
[00:15:27] 甚至不在于它
[00:15:29] 估值很贵，而在于美国如此多的部分
[00:15:31] 如今都依赖它继续维持高估值。
[00:15:38] ［音乐］