Jim Rickards sat in a room in 2006, inside the machinery of American intelligence, and told the people who run it that the financial system was about to break. The full career arc is in the Jim Rickards dossier; what follows is the specific story of that warning.
He had the data. He had the methodology. He had a machine - a predictive analytics system funded by In-Q-Tel, the CIA’s venture capital arm - that was reading market signals the rest of the world was not seeing. And he had already proven it worked.
Five days before the London liquid bomb plot was foiled in August 2006, his system flagged American Airlines with a bright red indicator, and he escalated it while the CIA watched for days until Scotland Yard broke down doors across London and arrested 24 men who were planning to blow up 10 passenger jets over the Atlantic.
Rickards’ machine caught it at the signal level using market data - puts on airline stocks, volatility anomalies, patterns in the noise that the human eye could not track at scale.
The CIA took notice and wanted more of that, and what they did not expect was what the machine would find next.
The Warning
In 2007, the same system that flagged the terrorist plot started lighting up on something else. Real estate. Mortgage-backed securities. The big Wall Street banks that had layered debt on top of debt on top of debt until the whole thing looked stable only because nobody was counting the floors.
Rickards took the data to Treasury officials. He laid it out. The housing market was overvalued in ways that the credit structures underneath could not support. Not a cyclical downturn but a structural collapse - the kind that takes down institutions, not just prices.
The Treasury officials listened. Then they dismissed it.
That dismissal is worth sitting with because it reveals something about how the system works. In 2006 and 2007, the housing market was booming. Credit was flowing. The models that banks and regulators used said everything was fine. VAR - value at risk - showed manageable exposure. Bond ratings were AAA. The narrative was that risk had been distributed, diversified, and tamed through financial engineering.
Rickards saw the engineering itself as the risk.
He had spent the late 1990s inside the LTCM rescue, watching Nobel laureates discover that their models could not handle the real world. He had watched $4.6 billion evaporate from a fund that the models said could not lose more than $35 million in a day. He knew that the people running the system often did not understand it as well as they thought they did. The LTCM rescue made Jim Rickards the forecaster he became, and what he learned about the LTCM leverage lesson became the lens for every warning that followed.
Rickards was already advising the Director of National Intelligence on financial threats. He was already building the financial war games that the Pentagon would run in 2009. He was already seeing the connections between credit markets, derivatives, and systemic risk that most economists were not even looking for. He had been delivering analysis over months to people whose job was to think about worst cases.
He told them the crash was coming. They did not act.
Eighteen Months
Lehman Brothers filed for Chapter 11 bankruptcy on September 15, 2008.
The Dow dropped 504 points that day as the entire global financial system froze, money market funds broke the buck, and banks stopped lending to each other. The government had to inject $700 billion through TARP just to keep institutions from collapsing overnight.
Rickards was not surprised. He had been saying it since 2006.
The timeline matters because it separates Rickards from the people who called the crash after it started. There were plenty of those. In June 2007, Bear Stearns hedge funds blew up and suddenly everyone was an expert on subprime mortgages. By September 2008, every financial commentator on television had a story about how they saw it coming.
Rickards saw it coming when housing was still climbing, when the Dow was still hitting records, when the consensus was that the system had never been more stable. He saw it because he was looking at the same structural fault lines that LTCM had exposed a decade earlier. Hidden leverage. Interconnected counterparties. Models that did not account for the one thing that always breaks: human panic.
The war game he helped design for the Pentagon in 2009 simulated a scenario where countries use financial weapons instead of kinetic ones — Russia and China pool their gold, issue a new currency backed by gold, challenge the dollar’s reserve status, and the whole thing unravels.
That war game did not come out of nowhere. It came from the same analytical framework Rickards had been using since 1998: look at the plumbing, find the hidden leverage, and ask what happens when trust evaporates.
What the Machine Saw
The Project Prophecy system that Rickards built with his partners Chris Ray and Randy Tauss was a market intelligence tool - MARKINT, in intelligence jargon. It applied complex systems theory, Bayesian analysis, and network science to financial market data to detect patterns that preceded major events.
The system worked by reading the market’s own information. When a terrorist attack or financial crisis is coming, the people who know about it trade — options volume spikes, volatility smiles distort, correlations between unrelated assets shift, and the market knows before the news does. The question is whether anyone is reading that signal.
Project Prophecy read it.
It caught the liquid bomb plot. It caught the 2008 crash signal. The CIA funded it, used it, and then declined to scale it - partly because of political concerns about the appearance of the intelligence community monitoring financial markets. The project was classified, and much of the detail remains classified today.
But the core insight is not classified. It is the same insight that drives everything Rickards does: the market is a information-processing system that reveals its state through prices, and if you know how to read those prices, you can see the future before it arrives.
The Lesson
The 2006 CIA warning is the clearest demonstration of his methodology in action - applied in real time, with real stakes, and with a verifiable outcome. More than a credential on his resume, it shows how he took an analytical framework developed from the wreckage of LTCM, applied it through a machine designed to find patterns in market data, and produced a directional call that was early enough to act on but specific enough that it could be verified.
Eighteen months later, the crash came.
The people who ignored the warning in 2006 had their reasons, since the housing market, the models, and the broader economy all looked fine. But Rickards was not looking at the economy. He was looking at the plumbing. And the plumbing was corroded.
That is what the CIA warning teaches you about Jim Rickards. He does not predict crashes because he is naturally bearish. He predicts them because he has spent thirty years studying how financial systems actually break, and he has learned that the mechanism is always the same: hidden leverage the models cannot see, building until something triggers the unwind.
The specific mechanisms change while the pattern stays the same.
In 1998, it was LTCM. In 2006-2008, it was mortgage-backed securities and bank leverage. In 2020, it was COVID and the liquidity panic. In 2026, it is AI debt - off-balance-sheet structures that look a lot like the ones Enron and Lehman used. He extended the same framework to AI in MoneyGPT - the Guru Files entry for one of the most followed macro thinkers working today.
Rickards is still using the same lens. He is still finding the same pattern. And he is still early enough that most people do not listen until the break happens.