NLY/mREIT Vs Yield Spread Regression Update 12/31/20

I seem to have forgotten the update on 9/30/20.  The regression at that time indicated that NLY should be held or bought.

The new regression equation for NLY, which is calculated quarterly, as of December 30, 2020 is:

PredNLY = 5.683 - 1.623 (FedFundsrate) + 2.693(TYXMA10).

The equation r-sq is .478, which has been gradually increasing the last 3 years. The standard error is 2.78. The predicted price for NLY using the 30-year T-bond yield and the fed funds rate is $8.16. The last price of NLY on June 30 was $8.45, which is 0.09 standard errors above the predicted mean using the equation. The stock, or mREIT ETFs, therefore, at current interest rates, are close to fair value, and the stock price has advanced a lot since June 30, when I wrote that it was a screaming buy. A price of $12.59 would be 1.5 standard errors above the mean, and that would be a sell, unless the spread between 30-yr T-bonds and 30-yr T-bills widens.

NLY had sold off because of fear of a credit freeze, and fear of mortgage defaults, but agency mREITs look like they can get financing from the Fed, and so there should be no problem. NLY has some risk from the default of CMBS, but most of its assets are agency MBS. The drop in short-term rates was good for mREITs. Although mortgage rates are low, people are no longer likely to refinance because they can't, or have done so, and so mREITs are unlikely to have a problem with prepayment of mortgages. 

I had sold REM because of the price crash in the first quarter, but I bought it again in the second quarter much cheaper than where I sold it. Rem has had good gains since then, and it may pause for a while unless the yield spread widens.