> I’m not convinced this is a recession signal, as much as it’s a reflection of the new normal for worldwide central banks.
3m/10y curve inversion has proven to be a reasonable signal for upcoming recession in the USA for the post WW2 period [0]. If the Federal Reserve were to factor this signal into its decision making processes (as ECB and BoJ do), its predictive power would likely decline (can't recall the episode of the podcast Macro Musings at present for citation).
> Put the 10 year in context: in Japan and Germany and other stable countries yields are negative. So you have to pay to lend those countries money, because the central banks are pushing yields negative.
You can earn positive (nominal) yields on JGBs [1] and Bunds [2], so not all yields are negative.
> But in the US you can actually get a modest (but real) return on the ten year, so it’s quite popular. This popularity has pushed the price lower.
You're comparing outright duration to curve risk. These are distinct. When discussing curve inversions, you are comparing the spread between two points on a curve, precisely to eliminate any parallel shift component. In this case, the comparison is between a 3m investing period and a 10y investing period.
> And while recession may not be in the cards for the US in the next few years, ten years is a whole different story
On what basis do you assert that recession isn't possible in the US over the next few years? Fed funds futures currently imply a 63% chance of easing by the end of 2019 [3], indicating expectations of deteriorating economic conditions.
[0] https://www.frbsf.org/economic-research/publications/economi...
[1] https://www.bloomberg.com/markets/rates-bonds/government-bon...
[2] https://www.bloomberg.com/markets/rates-bonds/government-bon...
[3] https://www.cmegroup.com/trading/interest-rates/countdown-to...