Quantitative Tightening and Mortgages

The Fed went from reinvesting all of the proceeds from maturing securities to reinvesting all but $10 billion every month for the first three months, with further reductions of monthly reinvestment by another $10 billion every three months after that. So, by Q4 2018 the amount non-reinvested will have climbed to $50 billion a month. The treasury/mortgage split within those amounts will be 60/40. All told, if things go to plan the Fed’s balance sheet will have been shrunk by about a half a trillion dollars by this time next year.

There was a lot of fear that the market would start discounting what people were calling Quantitative Tightening and that yields would start rising–perhaps disruptively–putting pressure on risky assets and economic activity.

It is against this backdrop that mortgage spreads jumped out at me. Granted, reducing your holdings by $4 billion a month should not have a huge mechanistic effect. But, allegedly, markets are forward looking and the Fed has a lot of mortgages hanging over the market.

Here is the mortgage spread on one and five year horizons:

Here I’ve taken the 30 year FNMA current coupon and subtracted the yield on the five-year treasury. I use the five year because the five to seven year area of the Treasury and swaps curves is the part of the curve that institutional investors use to hedge the duration of their mortgage books.

The collapse in mortgage spread since September speaks for itself. What we’ve seen is five-year yields increase by about 50bps while mortgage yields have stayed flat.

Some might be tempted to argue that the 50bps rise in the 5yr yield is the sign that QT is having an effect. But if the unwind of the Fed’s balance sheet were having the supply effect so many feared, you’d expect to see the yields on both rise with only a modest change in spread. Could it be that effects from the Fed’s balance sheet at this point–to the extent there are any at all–are mostly about signalling and psychology?