I've posted this originally on the IVR board - with adjusted numbers the points apply here - happy to have comments.
Let me summarize what one of my IRR analyses reveals -
Buy IVR today at 21.50 and 1.00 quarterly dividend;
Hold IVR for 20 quarters (five years total), doing nothing;
Assume share price and dividend both decline 2% each and every quarter, resulting in an ending value five years hence of $14.35 for the shares and a then current dividend of $ .68 per quarter.
At the end of this period one would sell IVR at $14.35........
Here are the investment results (tax issues disregarded) assuming a 1000 shares position (for simplicity only):
The capital loss on the shares of $7,146 is offset by twenty dividends received (as impaired) of $16,620, resulting in a periodic IRR (quarterly) of 2.65%, or in excess of an 11% annual return.
This result arises from share price and dividend declining on a continuous and consistent basis by a total of 33%. Certainly there are worse cases to imagine - and many better - but I am trying to put a quantitative value on the actual impact of decling share and dividend values on a security starting with such a rich dividend.
I can play with an infinite array of variables and assumptions but you get my point: an investment in IVR today wherein the share price and dividend decline by 33% over the next five years results in an annualized gain of 11% and may well outperform most alternative investments.
That would be one possible reason not to be out of IVR even if one believed it were going to decline.
I've set the spreadsheet up so that I can input a "Coefficient of Decline" - the amount by which the current price and dividend are adjusted from the prior period (.98 would equal a 2% periodic decline) - and it's interesting to note that the five year investment achieves an IRR of 0 when the COD is .9535 -- meaning that shares and dividends could go down almost 5% per quarter for five years and the investment would break even.
That is, of course, unless my math is wrong - and obviously this is all theoretical stuff, intended, however, to shed insight on the investment decision process.
I too feel that your assumptions are too static. It would be better to model three scenarios and probability weight them based on your macro outlook. (Also, CYS and IVR are very different companies with very different portfolios, which you probably know.) Anyway.
Scenario #1 -- Japan type long term outlook where the yield curve flattens somewhat but fund costs remain low and stable.
Dividends decline to 16% ROE on average over the period (historically average for NLY is 13% and for levered financials generally is 12%). Share price is flat over this period. Personally I'd give this 40%.
Scenario #2 -- interest rates follow the 2003-2007 road map, i.e., the yield curve gradually goes flat. Dividends decline to an average of 13% ROE over this period. Share price gradually declines 20%. 40% likelihood.
Scenario #3 -- black swan type scenarios. Yield curve goes flat to inverted quickly and stays there, or unanticipated credit risks with agency products surface. Assume dividends average 6% ROE over the period and share price declines 50%. 20% likelihood.
That's the beauty of IRR analysis. Investors can enter quad's initial scenario and your three cases in less than 15 minutes. And different portfolios like AGNC, IVR, ARR and CYS would only take one hour to complete all four projections.
Thanks to all the nerds who create these web apps!
Great i guess you made things easy for everyone for the next 20 years.
see you in 20 years for the next suggestion.
bud this on dip with the dollar to support a 20% yielder.
or buy and cash in 20 years later.
Please identify any "guess" or "suggestion" in my post, which you have apparently misconstrued as soothsaying rather than analysis.
There are variables and results, if that's confusing you.
I like your math but your assumption is too static.
Prices of mortgage REITs will vary widely based on the real estate market, their own prudent/less prudent investment decisions, and the spread between short-long term rates.
I do think your idea is likely valid for the rest of this year. But I think as soon as the Fed does any tightening, that 2% decline in PPS may become 8% in a trading day. And when dividends are reduced, usually they are almost 'catastrophic' reductions, not incremental.
That said, I think IVR is not a bad choice. CYS is OK too. I do think there are more secure dividend payers, but they won't provide these yields - thus justifying why they are paying 20%.
Obviously the spreadsheet permits the insertion of any value one cares to model in any of the twenty quarterly periods.
I wasn't making a prediction about price or dividend outcomes - rather I was trying to quantify the impact of the same on investment return.
An 8% decline in a trading day is fine and certainly could be part of a 33% decline in five years.
Both CIM and NLY have reduced dividends recently and they were not "catastrophic" reductions. To some extent that expectation may be the result of the recent economic shock and represent a "black swan" discount not statistically warranted.
Obviously the dividend yield at 20% represents some kind of mismatch between earnings and share price. If yields on preferreds for decent REITs hover around 8%-9% there is a huge yield premium on the common dividend (perhaps justified, perhaps overcooked). Obviously they pay such a dividend because they have to distribute REIT earnings - the yield is elevated because investors are apprehensive about something - with or without quantifiable cause.
My point in this IRR exercise is to evaluate whether neutral quantification of outcomes reveals an investable thesis. It doesn't really matter, per se, whether dividends are stable or not or whether they go down 8% in one day. What matters to me is whether perception of such possibility or expectation of such events leads to share pricing that creates opportunity.
Thanks for your comments - I appreciate the perspective.