Nineteen Islamist terrorists murdered 2,979 people on September 11, 2001. The government’s counterterrorism (CT) response to 9/11, which was the deadliest terrorist attack in world history, included domestic security measures like the TSA, the PATRIOT Act, surveillance, and numerous wars overseas that we lost. The costs of 9/11 were horrific, and the response was gargantuan. When studying public policy, we must ask: Was the counterterrorism response to 9/11 worth it? Did it save enough lives to pay for itself or at least break even?

No, not even close. For CT spending to equal the value of lives saved, it would have to have saved between 390,000 and 442,000 lives, or between 15,600 and 17,680 per year since 9/11. From September 12, 2001, through September 10, 2026, I estimate that 347 total people were murdered on US soil by all terrorists of all ideological motivations and locations of birth. That includes some victims in attacks since that analysis was published. Thus, all CT spending would have to have saved 1,124 to 1,274 times as many lives as were taken by terrorists in attacks on US soil for the benefits of CT spending to equal the costs in terms of lives saved. 

Read the rest of this post →

September 11 marks one of the darkest days in US history. Terrorists murdered thousands of people on American soil, put countless people in fear for their safety, and left a scar in American history that can still be felt to this day. Yet what is less known is the financial surveillance that was expanded in the wake of the attack.

The Bank Secrecy Act already required banks and other financial institutions to report their customers to the government, but Congress decided that it wanted more. In the near blink of an eye, the USA PATRIOT Act was signed into law. At the time, the Cato Institute’s Robert A. Levy wrote that the law “zipped through in six weeks—gutting much of the Fourth Amendment in far less time than Congress typically expends on routine bills that raise no constitutional concerns.”

The damage to civil liberties caused by the PATRIOT Act is far-reaching. For just financial surveillance, the law mandated stricter customer identification (better known as know your customer, or KYC, requirements), expanded the list of financial institutions that must surveil customers, scrutinized foreign relations, created channels for more government officials to access this surveillance system, and made much of the system confidential from the broader public.

Twenty-five anniversaries in, let’s drop the “clear blue September sky” line and recalling where we were when the attack happened. Nearly 3,000 Americans were killed that day, and the suffering of their families dulls over time but will never end. 

Still, people use anniversaries to reflect on events both good and bad, and 9/11’s 25th has been and will be used to shape our consensus on what the terror attacks and the ensuing Global War on Terror were.

My piece in the Independent Review for the 20th anniversary says much of what I have to say about the world produced after 9/11.

In the absence of major international or domestic constraints, policy can become extravagant. The period from 2001 to the present represents a promiscuous waste of money, lives, and diplomatic attention, for which almost no one in charge of the policy has paid serious consequences…. Policies whose costs can be avoided, defrayed, or hidden are likely to be oversupplied.

I would add two additional points. As should any analyst, I think first about my own failures. Perhaps the biggest one is that I opposed the Iraq War in part because I thought at the time that radical Islamist terrorism was a huge threat to the United States and that invading would both distract us from fighting terrorism and make the threat worse. In 2002 and 2003, that was a widely held view among the few of us who opposed the war.

But there was one person who got the terror threat right while it was truly heretical to do so: my colleague John Mueller. John wrote an article in the fall of 2002 (!) arguing that 

a case can be made that rather than foreshadowing the future, the attacks may turn out to have been a statistical outlier, a kind of tragic blip in the experience of American national security.

To say that this was ridiculed in 2002 would be an understatement. But John was right.

In 2026, there are opportunities to bet—er, “trade”—on almost everything on sites like Kalshi and Polymarket. Many people use these as a way to gauge likely futures. John was willing to put his reputation on the line with a prediction shared by almost no one at the time. Had we had prediction markets back then, John would have a fleet of Bentleys in his driveway today, and the entire US foreign policy establishment would be rattling a tin cup. It is worth hearing out the heretics sometimes.

Second, twenty-five years later, the GWOT is both over and never-ending. The public has lost interest in large-scale counterterror projects, although it continues to say worries about terrorism are top of mind. But the institutions (DHS, TSA, ODNI) and protocols (presidential drone killings unbound by courts or Congress) set up in response to the attacks endure. As I closed the 2021 article,

ending the other legacies of the post‑9/​11 period will be difficult and many of these legacies may survive as nuts tightened onto the American ship of state by what Robert Higgs has famously termed the “ratchet effect.”

Giving governments extraordinary powers in times of crisis seems irresistible. But it is often worth resisting, because this ratchet only tightens.

The Treasury Department has learned nothing from markets’ muted reactions to its bond price engineering schemes. Having doubled its long-end buybacks in August to at least $4 billion per operation, this week the Treasury unveiled a $6 billion operation in the 10- to 20-year sector, triple the normal size. Instead of lowering bond yields, as it was supposed to, the 10-year yield climbed to 4.85 percent, a three-year high, and the 30-year yield pushed back above 5.3 percent.

I wrote last month that unless the root causes of high yields are addressed, many of which originate in Washington, no amount of government financial engineering would help. Bond price manipulation cannot persuade markets to stop pricing government debt for excessive spending and inflation risks stemming from failed macroeconomic policies such as tariffs and the war in Iran. 

And that’s what happened again, only faster than before. In August, yields dipped for at least a day before recovering. This time they rose straight through Treasury Secretary Scott Bessent’s announcement. Around $6 billion per operation is no better than $4 billion when weighed against a $31.5 trillion Treasury market, and the market is no longer pretending otherwise.

The AI boom and resulting data center backlash are taking several forms. One dimension is the growing concern over subsidies. The latest estimates from Good Jobs First show that data center tax exemptions in Georgia, Texas, Virginia, and Ohio could reduce revenue by more than $1 billion a year in each state. The National Conference of State Legislatures counts 38 states with data center tax incentives. 

But not every deviation from the revenue-maximizing tax base is created equal.

Data centers should be treated like every other industry. Government policy should neither penalize nor favor specific sectors or types of investment. However, as is the case with tax expenditure analysis more generally, the term “tax subsidy” often describes two fundamentally different things.

For the past 14 years, 83-year-old Ken Jouppi has been trying to prevent the State of Alaska from seizing his airplane. Mr. Jouppi is an Air Force veteran and has been flying charter planes in Alaska since the 1970s. Alaska has a number of villages and municipalities that are “dry,” where the sale, importation, and even possession of alcohol are illegal. Mr. Jouppi agreed to fly a customer to one of these locales. But the plane was stopped by police before takeoff—the passenger had packed alcohol in her luggage and carried three cases of Budweiser and Bud Light onboard to share with her husband on his birthday. Mr. Jouppi was convicted of a misdemeanor, sentenced to the minimum three days in jail, and ordered to pay a $1,500 fine. But Alaska also mandates forfeiture of any airplane used in connection with transporting alcohol to dry villages. Hence Mr. Jouppi has been fighting for the past decade and a half to save his $95,000 plane. The Supreme Court has agreed to hear his case.

Cato, joined by the Fines and Fees Justice Center, filed an amicus brief arguing that this forfeiture is an unconstitutional excessive fine. The Eighth Amendment was designed to prevent this kind of abuse. Long before the United States was founded, the common law protected people from extreme monetary penalties. The Framers understood the dangers of such abuse. But the Alaska Supreme Court’s view is that challenges to excessive fines “should rarely succeed.” This dismissive view led the court to conclude that there is nothing excessive about the forfeiture of an airplane worth “only 9.5 times the maximum fine”—and over 60 times the fine actually imposed.

Many innovators find the growing uncertainty, frequent changes, and secret frameworks in the AI regulatory space as barriers to investment in the research, development, and deployment of new technology. At the same time, there’s broad recognition that laws such as the EU AI Act risk being out of date by their effective date, representing a different constraint on the AI ecosystem. The frustrating reality is that there’s no “right” way to govern AI. Existing laws, including consumer protection statutes, already address many of the most common harms associated with AI. Yet frontier issues and edge cases will require new paradigms and regulatory certainty to allow innovators to fully pursue possibilities while resolving the public’s and policymakers’ concerns about legitimate risks, such as cybersecurity threats.

New governance efforts should be tailored to address cases where clear market failures underpinning broader AI adoption and ongoing AI research occur or where the lack of regulatory certainty prevents important development of the technology. Three examples of such failures include: Inadequate information about model capabilities and values is hindering the ability of consumers to make knowledgeable decisions; insufficient supply of independent AI auditors threatens to prevent the timely and thorough investigation of company practices; and ineffective initiatives to increase the talent pipeline of AI experts may hinder the nation’s long-term tech ambitions.