The Bond Vigilantes’ Delusion The bond vigilantes are back, or so Wall Street would have investors believe. Their claim is that Treasury yields must remain high because America has too much debt and artificial intelligence is creating enormous new demand for credit. It is a seductive story. It is also wrong. It’s a shame even Fed officials use it. AI borrowing may lift the term premium at the margin. Deficits may make investors demand more compensation for duration. Neither explains a 5% 10-year Treasury yield, nor makes it a permanent condition. The real wager is monetary. The 10-year, near 5.2%, reflects a market assumption that the Fed funds rate will average roughly 3.8% over the next decade, plus a positive term premium. The San Francisco Fed’s model decomposition makes the point plainly. It attributes roughly 3.8 percentage points of the 10-year yield to expected short rates over the coming decade and about 1.5 points to the term premium. The New York Fed’s estimate of the premium is lower, closer to 0.9 points. The split is debatable. The message is not. That 3.8% assumption is a remarkable forecast. It is more hawkish than the Federal Reserve’s own 3.2% longer-run estimate. Markets are not merely assuming that rates settle above the zero-rate era. They are assuming that the Fed will run policy about 60 bps tighter than its own idea of normal for an entire decade. That is the bond vigilantes’ delusion. Yet who can blame them? The Fed’s recent conduct has encouraged it. It has shown a readiness to tighten into a negative growth shock, to treat tariff-related price changes as persistent inflation, and to mistake supply-side expansion for excess demand. That is bad theory. Tariffs may change relative prices and impose real costs, but they do not automatically produce a sustained increase in the inflation rate. Nor does faster growth automatically require tighter policy. When investment, technology and productivity increase the economy’s productive capacity, they permit stronger real growth without a corresponding acceleration in underlying inflation. A central bank that ignores this distinction makes itself the source of the restrictive policy that the bond market now expects. The debt thesis does not survive contact with history. US non-financial debt rose from about 136% of GDP in 1980 to roughly 250% by 2010. The 10-year yield fell from the low teens to around 3%. Debt then stayed near 250% of GDP through much of the 2010s as yields fell toward 1%. It peaked near 293% in 2020, when long yields were near their lows. Since then, the debt ratio has declined as nominal GDP rose faster than debt, even as the Fed lifted rates above 5% and the 10-year rose above 4%. The largest move in rates occurred while economy-wide leverage was flat or falling. Japan and China reinforce the lesson. Both accumulated vast debt burdens without producing a mechanical surge in long yields. Debt can matter at the margin. It does not dictate the cycle. My estimate of nominal r-star is 2.75%, 2% inflation plus a 0.75% equilibrium real policy rate. Technology is deflationary. AI, capital investment and productivity growth expand supply, lower costs and permit faster growth without persistent inflation. Yes, a bull steepener awaits. As the Fed’s reaction function catches up, short rates should fall faster than long yields. If the average funds rate converges to 2.75% and the term premium remains positive, the 10-year should settle near 4%, with a normal 100 bps spread over the two-year. The long end is not going parabolic because of debt. It is high because markets expect the Fed to mistake supply-led growth for inflation and keep policy too tight. That error will eventually fade, likely after the midterms.