Heres a photo of the 10yr-3mo term spread in Australia:
Source: worldgovernmentbonds.com.
What inversion there is 3yr-2yr; the curve is indeed much flatter than 6 months ago
.
Figure 1: 10yr-3mo Australian Treasury spread, % (blue). ECRI peak-to-trough recession dates shaded gray. Source: OECD Main Economic Indicators, ECRI, and authors calculations.
A probit model estimated over 1968-2021 yields (presumes no recession occurred as of December 2022):.
Prob( recessiont +12) = -1.585 — 0.253 spreadt + ut +12.
McFadden R2 = 0.13, NObs = 630. Coefficients significant at 5% msl strong. The spread remains in percentage points.
While the spread is statistically considerable, it wouldve totally missed out on the 2019-2020 economic downturn. Indeed, this model wouldve missed the 2008 recession, even utilizing as low as a 20% limit. This outcome partially verifies Not Trampiss view; however, an increasing 3 month Treasury rate doesnt seem to predict recessions well either.
Keep In Mind that Karunaratne (2002) does show the spread does anticipate development and economic downturns well, for data through 1997. His definition of recession is based on the two-consecutive-quarter criterion, and so differs from the approach that uses a NBER or ECRI sign.
The Australian yield curve as of today:.
Figure 1: 10yr-3mo Australian Treasury spread, % (blue). The spread is in percentage points.
While the spread is statistically considerable, it wouldve completely missed out on the 2019-2020 economic downturn.
This entry was published on December 7, 2022 by Menzie Chinn.
