@chasesfish this is a very interesting way to look at the choices. I've gathered you're asking "what happens to the value of the stock if the yield goes to X% in the future". The answers to this question illustrate how some of these preferreds (e.g. BAC-N) have relatively modest upside potential while others are more like moonshots (e.g. WAL-A).
Of course, WAL is much, much more likely to go out of business than BAC, so knowing the
potential future returns tells us little about our
expected future returns. Our expected future returns are probability-weighted, and tied to the odds of bank collapse.
There are hundreds (thousands?) of analysts hard at work to answer this "probability of collapse" question, and they're looking at all the information they can find about assets and deposit flightiness. Today's preferred stock prices and yields can be seen as their collective output. A preferred stock's yield spread over a similar-duration risk-free asset can be interpreted as the odds of a 100% loss.
For example, BAC-N has no stated maturity and yields 206 basis points over comparable-duration 30-year treasuries. The extra 2.06% in yield could be said to compensate the holder for a 2.06% risk of complete loss. Of course, there are at least 4 problems with this approach: (1) the actual probability math for odds calculation and avoidance of gambler's ruin is more complex, (2) default risk is not a one-time binary event and could happen after years of dividends, (3) there may be complex strategies with a lower risk of 100% loss, such as using options or shorting the common and going long the preferred, and (4) the price and yield are set not by weighing the estimates of a neutral panel of investors, but by the marginal willingness to pay by the subset of investors who are making trades at these prices - a critique of EMH. Regarding that last point, if the distribution of investors is not normal or is skewed, all the price-setting trades may be occurring amongst a tiny sliver of the investor population when the vast majority says "don't even touch it with a stick".
Despite these limitations, let's treat the spread over risk-free rates as a binary probability-of-total-loss estimate over the next 12 months. Using 30 year treasury yield as the risk-free rate, which may not be appropriate for preferreds with an earlier maturity date, one obtains:
Market-estimated risk of total loss:
BAC-N 2.06%
HBANP 3.80%
SNV-E 4.34%
RF-E 3.70%
OZKAP 4.50%
AUB-A 5.09%
UCBIO 6.27%
BOH-A 4.98%
FHN-F 4.96%
FHN-E 5.14%
PNFPP 4.30%
COF-N 3.24%
WAL-A 6.34%
COF-L 3.30%
CFR-B 2.53%
FITBO 2.47%
CADE-A 3.55%
FCNCO 3.26%
Seen in this light, WAL-A's or BOH-A's potential for returns up to 100%+ is a no-brainer if the odds of failure are only 6.34% or 4.98%. This is our alert that we need to slow down because there's something wrong with our process.
In particular, I'm looking at limitations #3 and #4 above. #4 seems untestable, so let's look at #3.
Is it possible to hedge BOH-A for example, at a lower cost than the yield?
The farthest-out duration for BOH options is currently 158 days or 0.433 years. The $35 put option for October 20 (slightly ITM) had an amazing $11.08 in time value yesterday. Thus it will decay by that much if BOH's stock price is the same on Oct. 20. Meanwhile, the preferred stock will pass through two ex-dividend dates in that timeframe, paying a total of $0.545. The common stock is selling at 274% of the price of the preferred stock, so if we assume both would go to zero together, one put option on 100 shares of BOH should hedge 274 shares of BOH-A. So our hypothetical trade would gain (274*0.545=) $149.33 in dividends while losing (100*-11.08=) $-1,108 in time value on the put. No bueno.
Of course, a trader might not ride the put to expiration because time decay quickens the closer you get to expiration. The option's theta is currently about 3 cents per day, so if the next BOH-A ex-dividend date is 7/13 or so, the highly optimistic level of time value loss (59 days*-.03 decay*100 shares =) $-177. That isn't at all compensated by the ($0.2725 dividend * 274 shares=) $74.67 dividend.
Maybe there are other tricks one could play, such as selling BOH-A short (can you do that?) to hedge a short put, but I don't see #3 as an adequate explanation for why we shouldn't trust these risk-of-failure estimates.