3 ms·

> How and to what degree has this impacted yields? Is it even possible to answer this question with certainty? To reduce its balance sheet, the Fed has rolled

> How and to what degree has this impacted yields? Is it even possible to answer this question with certainty? To reduce its balance sheet, the Fed has rolled down assets, allowing them to mature. By choosing to not reinvest the principal from these matured assets, they reduce the size of the balance sheet. My understanding is that the reinvestment occurs in longer dated securities, which would put flattening pressure on the curve. Estimates indicate a reduction of these purchases would raise 10y yields 40-60bps [0]. All other things being equal, this would lead to steepening of the yield curve. > I'd be leery of any predictions of recession in the US based solely on yield inversion. Not all yield curve inversions have preceded recessions, but almost all recent recessions have been preceded by yield curve inversions. Yield curve inversions lasting 10 days or longer tend to have more predictive power. The inversion in March did not last that long, for example. On average, a recession follows 14 months after an inversion [1]. [0] https://www.ft.com/content/828169e8-8902-3d5a-903d-586bf22cc35c https://www.ft.com/content/828169e8-8902-3d5a-903d-586bf22cc... [1] https://www.advisorperspectives.com/articles/2019/04/01/a-historical-perspective-on-inverted-yield-curves https://www.advisorperspectives.com/articles/2019/04/01/a-hi...

cs702 7y ago
They recently announced they would stop doing that. How has that affected expectations and behaviors?
- nickles 7y ago
  > They recently announced they would stop doing that. I'm aware. Markets are forward looking. As you can see, yields have dropped precipitously since the announcement, and the curve has flattened. This is consistent with expectations of increased purchases of securities in the tail end of the curve.
  - cs702 7y ago
    Yes...and therefore, this recent yield inversion may not be related to growing risk of recession in the US. Maybe it is, maybe it isn't. I don't know. That's why I'm leery of predictions made based solely on the inversion of yields. Things are different this time. That's my point :-)
    - nickles 7y ago
      > this recent yield inversion may not be related to growing risk of recession in the US It's been a reliable signal in the past, but perhaps this time is different. Interestingly, some economists postulate that such an inversion may be somewhat causative of recession, due to increased short term funding costs relative to long term funding costs.
      - cs702 7y ago
        > It's been a reliable signal in the past, but perhaps this time is different. Yes, we're in agreement :-) > Interestingly, some economists postulate that such an inversion may be somewhat causative of recession... That could be true... I don't know.
ianai 7y ago
It’s dangerous to reason about this without a white board, but I think that means the negative difference between 10y and 3 mo would be decreased by the fed not reinventing in long term securities.
- nickles 7y ago
  > I think that means the negative difference between 10y and 3 mo would be decreased by the fed not reinventing in long term securities. When the value of the spread between a long dated (e.g. 10Y) and a short dated (e.g. 3m) instrument increases, the yield curve is referred to as 'steepening'. When this spread decreases (even when becoming more negative), the curve is said to 'flatten'.
  - FabHK 7y ago
    / steep - flat \ even flatter
asmithmd1 7y ago
It looks like a 10Y-3M inversion has predicted each of the last 3 recessions by about a year with no false positives in the last 35 year. https://fred.stlouisfed.org/graph/?g=o4ro https://fred.stlouisfed.org/graph/?g=o4ro Of course maybe this time is different...
- mrep 7y ago
  Assuming you use 0 as your magic ball number. Move the magic ball number up just a tiny bit to 0.07 and you would have had a false positive in september 1998. You could also move it down to -0.32 and you would still a 100% success rate, albeit with a sample size of 3...

cs702 7y ago

They recently announced they would stop doing that. How has that affected expectations and behaviors?

nickles 7y ago
> They recently announced they would stop doing that. I'm aware. Markets are forward looking. As you can see, yields have dropped precipitously since the announcement, and the curve has flattened. This is consistent with expectations of increased purchases of securities in the tail end of the curve.
- cs702 7y ago
  Yes...and therefore, this recent yield inversion may not be related to growing risk of recession in the US. Maybe it is, maybe it isn't. I don't know. That's why I'm leery of predictions made based solely on the inversion of yields. Things are different this time. That's my point :-)
  - nickles 7y ago
    > this recent yield inversion may not be related to growing risk of recession in the US It's been a reliable signal in the past, but perhaps this time is different. Interestingly, some economists postulate that such an inversion may be somewhat causative of recession, due to increased short term funding costs relative to long term funding costs.
    - cs702 7y ago
      > It's been a reliable signal in the past, but perhaps this time is different. Yes, we're in agreement :-) > Interestingly, some economists postulate that such an inversion may be somewhat causative of recession... That could be true... I don't know.

ianai 7y ago

It’s dangerous to reason about this without a white board, but I think that means the negative difference between 10y and 3 mo would be decreased by the fed not reinventing in long term securities.

nickles 7y ago
> I think that means the negative difference between 10y and 3 mo would be decreased by the fed not reinventing in long term securities. When the value of the spread between a long dated (e.g. 10Y) and a short dated (e.g. 3m) instrument increases, the yield curve is referred to as 'steepening'. When this spread decreases (even when becoming more negative), the curve is said to 'flatten'.
- FabHK 7y ago
  / steep - flat \ even flatter

asmithmd1 7y ago

It looks like a 10Y-3M inversion has predicted each of the last 3 recessions by about a year with no false positives in the last 35 year. https://fred.stlouisfed.org/graph/?g=o4ro https://fred.stlouisfed.org/graph/?g=o4ro Of course maybe this time is different...

mrep 7y ago
Assuming you use 0 as your magic ball number. Move the magic ball number up just a tiny bit to 0.07 and you would have had a false positive in september 1998. You could also move it down to -0.32 and you would still a 100% success rate, albeit with a sample size of 3...