Dem
52
Ties
1
GOP
47
Treasury Secretary Scott Bessent is actively trying to reduce interest rates, because that is what Donald Trump wants. Unfortunately, Bessent has no direct control over interest rates, so he is trying to do it indirectly. He is ordering the Treasury to buy back existing bonds on the secondary market. By reducing the supply of bonds on the market, investors will bid up the prices for the remaining bonds since many investors believe U.S. government bonds are the safest investment possible. Since the semiannual coupon payout on existing bonds is fixed, investors buying bonds at a higher price than it used to be are effectively getting a lower interest rate than investors who bought them at the old and lower price. When bonds are paying a lower interest rate than they used to, that tends to force all interest rates down. At least that is the idea.
Not all investors and financial experts like the idea of the treasury trying to fight the bond market. Some financial types think it is pointless for the government to fight the fundamentals. They believe that if bond prices are going down and yields are going up, the bond market is speaking to you and you should listen. It knows what it is doing better than you do.
Legendary billionaire investor Stanley Druckenmiller, who worked with Bessent at George Soros' fund management firm in the 1990s, and who is seen as Bessent's mentor, is one of those who strongly opposes Bessent buying up bonds. He published a scathing op-ed in The Wall Street Journal attacking his former pupil. The op-ed was called "Let the Bond Market Speak." The op-ed argues that with inflation at 3-4% and the federal deficit at 6% of GDP, investors are demanding a higher return, and trying to force down bond yields artificially is a pointless and foolish endeavor. The op-ed says the markets combine the wisdom of thousands of investors, much more than any one person, even a treasury secretary, has. It is a long and detailed argument that Bessent is foolish and will fail. Since Druckenmiller is a well-known billionaire with decades of market experience, it made a huge wave on Wall Street.
Only Druckenmiller didn't write the essay. An AI bot wrote it. Someone used Pangram, an AI detection tool, to unmask the op-ed as fake. When confronted with this observation, Druckenmiller admitted it and said he uses all the tools at his disposal, and that includes AI.
But there is a larger lesson here. AI has gotten so good that it can fool the editors at the WSJ and people on Wall Street. If that is so, it can certainly go the other way as well. Next time Bessent or any other federal official puts out a statement that can move markets or otherwise affect the real world, can people trust it? Can people believe the "author" even read the whole thing or agrees with it? Maybe the "author" just instructed an AI bot to "write a press release about the bond market or about energy prices or about vaccines or about the war in Iran, or about just about anything." With a few vague hints, a bot could churn out something plausible. What about a politician instructing an AI bot: "Figure out what polling says the people want about [X] and then write me a press release supporting [X]." Or instead of issuing a press release, the instruction could be to write a speech supporting [X], even though the politician knows nothing about [X] and just wants to say what the people want to hear. Or even better, tell the bot "Find the three issues people care most about and write me a speech supporting all three." This is probably not a good development for a democracy, but here we are. (V)