The AI boom in historic perspective

Several investment booms have occurred throughout economic history. These booms have not always been part of a bubble or led to a bust, yet periods of investment exhaustion have occurred in which marginal firms have failed and early high returns have reversed or, at a minimum, significantly declined. Yet the current infrastructure boom is orders of magnitude higher than what we have seen in the past.

The AI investment boom exceeds the huge investment in railroads post-Civil War and is much larger than other economically changing technological investments. This is cause for concern because today's euphoria may not match reality tomorrow.

The short-term market is telling us rates will be higher

A good indicator of rate direction is the spread between 3-month Treasury bills and 1-year rates versus the EFFR. Market rates for slightly longer maturities show that the market is willing to clear at higher rates, and implied forward rates will, by definition, be higher. The switch from falling to rising rates occurred earlier this year, with the 1-year vs. EFFR spread rising sharply in the last month. Nothing suggests rates are headed lower.

AI impacting the arts employment

 

An interesting chart from Apricitas Economics on the employment changes in the arts since the introduction of ChatGPT. Live arts are showing job gains, but digital arts are showing significant employment declines. AI’s impact may come in unexpected places and shift job dynamics, with some winners and others losing out. 

Bonds at fair value - perhaps opportunity

We can look at real yields and say that real rates may be high, but there may be a view that real potential GDP is likely to fall and the natural rate of interest is lower.  If we assume growth will be higher, and the natural rate is closer to a longer-term average, we can say real rates have risen sharply, but their current level is not excessive. 

If we think nominal rates may be near fair value, we have to ask what the impact is of a 100 bps increase versus a 100 bps decrease. In that case, the risk-to-reward suggests that bonds may be attractive. If you assume the probabilities are equal, you may want to consider buying bonds.