Yesterday Today Tomorrow

Inseong J Kim: Hello, this is Inseong J Kim from Yesterday, Today, and Tomorrow. We are going to talk about panic, and before we get into it, I want to ask questions: why does law exist? The freedom without law is a jungle, but freedom with too much law is like a prison. Let's just chew on this very important statement about where we live in America. Freedom without the law is a jungle; freedom with too much law becomes a prison.

Every panic in history is a search for the line between. We want to understand a few panics in a historical background and look at the history of financial bubbles and what the AI boom is telling us now. We will look from the 1930s with railroads to the 2026 memory chip crash in Seoul, Korea, which was a recent event, and what is happening in China as well. We will cover the five historical panics around the mid-1800s, 1907, 1929, the dot-com crash, and the 2008 housing market crash to study and find a common denominator for why it keeps happening.

The common factors we are going to talk about are the four forces behind every panic and one hidden clock. In the third part, we will discuss the AI boom today and what's happening in South Korea, China, and the United States. Finally, we will talk about whether we can prevent the next one and what history says actually works. Why are we studying history? We study it to learn why and how it happened and to avoid problems for the future.

First, what is panic? A correction is when prices fall to reflect new information, sellers act on judgment, and the decline is gradual and roughly proportional to the bad news. A panic is different. In a panic, selling is forced, not chosen. Margin calls, leverage, and fear trigger the next round of selling in a domino effect, and the decline outruns the news that started it.

The pattern we see here involves five patterns. Number one: a real transformative technology or trend arrives. This is a new technology or new industry. Second: cheap capital sales occur as excitement passes reason. Number three: a story excuses skipping normal valuation discipline. Number four: hidden interconnected exposure builds the system. Number five: a trigger forces reality back in fast, not slow. A trigger is actually an underlying long-term accumulation of a problem. It doesn't happen suddenly when everything was good; it's actually brewing until it pops. We can prevent that brewing.

We can see five historical panics from 1837 to 2008 in different eras with the same underlying machinery. In 1873, there was railroad mania. The railroads were first. This changed everything. It was a brand-new concept: you could go long distances fast. It was not just a single individual moving fast, but a group of people moving together. Capital poured into new tracks based on optimistic traffic forecasts. It was financed heavily with bonds. Competing lines were built between the same cities chasing the same freight.

When the actual revenue fell short, railroads defaulted. Because the banks across the country held those bonds, the failure spread into a full banking panic. The railroad financier Jay Cooke and Company collapsed, igniting the panic of 1873. A five-year-long depression followed. The railroads paid off eventually, but the first wave of financiers mostly did not. That was the first one I found that had a similar story. This happens every generation.

Then came the Panic of 1907. What actually happened was a failed attempt to corner the market in United Copper stocks, which triggered runs on the trust companies that had financed it. Trusts held thinner reserves than banks and, crucially, no central bank existed to act as a lender of last resort at the time. Panic spread through interlinked institutions until financier J.P. Morgan personally organized a private rescue. There was an approximately 50% peak-to-trough decline in United States stocks during the 1907 panic. It exposed how fragile banking was with no backstop and directly led to the creation of the Federal Reserve in 1913.

Another big one was the Great Depression. The real boom of electrification, automobiles, radio, and mass production was layered with heavy margin lending. Investors could buy stocks with as little as 10% down. That leverage meant even a moderate decline triggered margin calls, forcing sales and driving prices down in a self-reinforcing spiral. We can see the pattern here: overproduction and overinvestment with no balance.

The Federal Reserve then tightened the money supply during the unwind instead of easing. The interpretation of whether this was accidental or intentional depends on the audience. Stocks declined 89% from the 1929 peak to the 1932 low. The crash became a decade-long depression and produced the SEC and the Glass-Steagall Act, the first real U.S. securities laws. We tried to prevent similar problems, but it takes everyone participating in healthy economic behavior.

Every time we go through a phase of new inventions, new technology, or new industries, a crash happens. During the dot-com crash, the internet was real. Access was financing unproven business models on narrative alone, often called "Get Big Fast"—eyeballs over earnings. Cheap capital and a wide-open IPO window funneled money into companies with no path to profit. When the capital tightened, investors started demanding actual earnings.

Valuation collapsed within months. This is another pattern. The Nasdaq declined 78% from March 2000 to the October 2002 low. It affected that era dramatically. The internet itself built the next 20 years of growth, but most mania-phase capital didn't survive to see it.

The financial crisis is the most recent one we can remember. It was not a new technology bubble but a credit leverage bubble in housing. Lending standards collapsed. Risky mortgages were bundled into securities rated as safe. Banks held enormous leverage against them. When home prices stopped rising, defaults cascaded through the financial system. Derivatives exposure was far more interconnected than anyone had modeled. Factories moved to China and the housing market collapsed because people could not pay the mortgage.

The S&P 500 declined 57% from its October 2007 peak to its March 2009 low. The leverage sat inside the banking system, so the damage hit the real economy directly, not just portfolios. This produced the Dodd-Frank Act. To summarize: the panic of 1837 was 32% down; 1907 was 50% down; 1929 was 89% down; and we had subsequent recessions of 52%, 78% for dot-com, and 57% for 2008. These are the major American economic panics that caused economic collapse and damaged people's lives.

As we are facing a world with new technology and AI booms, we have to reconsider how we can avoid panic again. The AI is developing much faster than other industries did before. Railroads were invented and took a long time to be established. Automobiles were invented, and we still use them. AI is growing so fast we cannot even catch up. The speed of change is incredible. About five years ago, I started to hear about Large Language Models. Google started using AI for a front desk person to make appointments and communicate with actual people.

At that time, they were just answering the phone and scheduling appointments. Today, when you call an AI, you are actually having a conversation. Sometimes you forget you're talking to an AI rather than a human. They have accents or special voice traits. One place I went used AI to make appointments. The AI picked up the accent from the caller and switched to the corresponding language right away. That's how fast AI is moving forward.

We are attracted to this new development of AI, but at the same time, are we going to face bubbles or panics like we experienced before? Can we prevent it or analyze if we are intentionally creating it? It's not my role to make a decision for the listeners, but every listener has to decide: is it intentional or is it accidental?

The common factors that created these past issues: ethics. Every panic on this list involves someone with better information extracting value while pushing the risk onto others. This includes inflated traffic forecasts, market cornering, insider selling at the top, and loans written knowing they would fail. The 1929 crash is a partial exception—it was more collective delusion and leverage than outright fraud, though insider pump-and-dump was common then too.

Ethical issues are not something we can ignore. They show up as railroad promoters overstating traffic projections to sell bonds, the 1907 copper corner attempt, dot-com insider selling during IPO peaks while burning investor cash, and "liar loans" where mortgage security banks bet against them.

Another factor is the quest for quick fortune. Ordinary people, not just professional investors, see neighbors getting rich fast and pile in with money they can't afford to lose. This turns professional speculative excess into mass systemic madness as millions of small, leveraged bets act in the same direction at the same time. The real issue is that there is a real economy, and then there is a different set of the industry working through stocks. Stocks are basically public lending, run by supply and demand based on information. It is not always connected between the real economy and the stock market. This is where the danger comes in.

Public railroad bond speculation during the 1830s and 1870s didn't happen overnight; it built up gradually. Retail margin trading in the 1920s and dot-com day traders quitting jobs to trade full-time happened because they saw the attraction. Before the dot-com era started, there was no actual business built. Even Amazon was in its infant stage, yet people were heavily trading. In 2008, house flippers with no money down sought quick fortunes. Most recently, in South Korea in 2026, retail investors used personal loans to fund leveraged ETFs. People borrowed money to buy stocks because the lending interest was lower than the money they could earn. That's a very risky business, and it actually failed.

Overfunding is another issue. Without excess capital chasing a trend, mania cannot get big enough to matter systemically. Cheap financing via bonds, margin, venture capital, and mortgage leverage turns a good idea into an overpriced one at ten times the scale. What we are experiencing today is a reward without sweat—expecting a reward without proper labor or industrial backing. This is the problem. Cheap bond financing for parallel duplicate rail lines in the 1800s, 10% margin stock buying in the 1920s, wide-open IPO and VC windows in the late '90s, and mortgage leverage bundled and releveraged in 2008 are all examples.

Another problem is fragmented, hidden exposure. Danger arises when exposure is spread across many layers in a transparent and silhouetted system. The danger is exposure that is fragmented in ownership but hidden and correlated underneath. Nobody, including regulators, knows where the risk sits until it breaks.

Actually, how it happened was banks nationwide held railroad bonds without realizing shared exposure. In 1907, it was interlinked trust companies. In 1929, it was millions of anonymous margin accounts unwinding together. In 2008, derivative webs spread the risk everywhere and concentrated it where no one could see. These are the underlying patterns of how the panic happens.

At the end of the day, it takes good stewardship in finance to avoid reaching the point of panic. We're in a psychological state where everyone needs to have a clear mind. Panic is a psychological issue. We cannot make a good decision when we are in a panic. We have to learn to step back, calm down, analyze, and then make a decision rather than a reaction.

The investment-earning gap must be calculated. Timing is also an issue: after people invest, how long does it take for earnings to begin? How long does it take for an industry to be actually implemented and bring a profit? When that gap gets longer and longer, it has its own problems. These are the problems we can see from history. How can we prevent them? Are they preventable, accidental, or intentional? These are the things that we have to watch.

The money that people are invested in the stock market or pension plans represents their hope for the future. Everyone needs to be educated on what causes those panics and how we can avoid them in the near future. Next week, I'm going to continue this issue. We are going to dive deep into what is happening today in AI. Is it a bubble, a hype, or a real economy? What is the timeline of actual earnings as a return on investment? We will look at South Korea and how their economy is evolving from a solid, strong economy to the bubble they just experienced recently. Thank you for listening to Yesterday, Today, and Tomorrow. We'll be back next week. Thank you.