Quick answer: It really depends on what your long term interest rate forecast. You could take out a 5/1 ARM and refi into a new 5/1 every 5 years, or every dip in interest rate. There are also periodic and lifetime caps on the rate. The risks are that your home may depreciate or there may be a large shock in interest rate, preventing you from refi-ing. Sound familiar?

Long answer: Construct a mortgage cashflow and hack some scenarios.

Clearly this is an indication that you need to be 500% long

The logic makes sense to me, but I've had a fixed mortgage that was the best rate available over the fix period.

I'll never understand why people post "P/E 10" stats from decades in the past with zero context or explanation. You appear to think it makes some kind of point, but it has less than zero relevance in today's markets.

The implication that earnings from 5-10 years ago are as relevant as current earnings (and more relevant than forward earnings) is just idiotic, yet people quote the Schiller PE as if we can just pretend losses during the mortgage/credit crisis have some kind of predictive power for equity returns over the next 20 years.

I've seen several institutional asset managers use that chart or similar - you really think it has "less than zero relevance"? I get the point on the method being imperfect, but if P/E 10 had zero relavance then the chart wouldn't have a significant fit line. Are you saying it's been relevant in the past, but is irrelevant now?

Here's a link to the nobel prize winning economists book who does think it has more than zero relevance and the up to date data set.

https://en.wikipedia.org/wiki/Irrati...uberance_(book)

http://www.econ.yale.edu/~shiller/data/ie_data.xls

If you take any period in time, as Schiller did around the turn of the century, and look back over 100 years, it's a mathematical fact that returns will be great if you buy at the lowest P/Es and trash if you buy at the highest P/Es.

Stocks are currently trading at trailing p/e in the low 20s or forward p/e in the high teens and there's no reason to think this will be a significantly higher than average p/e level going forward.

Historical p/e ratios in times where interest rates were in the double digits are meaningless today when short rates are around 1% and 30 year t-bond yields are below 3%. We also have every reason to believe equity risk premiums will be lower going forward than in the 1900s.

Shiller is a scam, but there's enough people who believe in similar concepts to keep prices low for those of us smart enough to accumulate equities at the current very reasonable prices. Those of us keeping significant equity exposure will be laughing in 20 years when p/e ratios are close to 30 and the indexes have returned 10%++.

But with that said, I'm not currently overexposed since I think there's a fairly high likelihood we'll be able to buy significantly cheaper within the next 5 years or so since investors are stupid and think we're currently trading at high valuations.

What do you do for a living?

It doesn't sound like you're saying "that analysis is irrelavant", which was what I thought you were saying more. You're saying that the bar has moved due to low interest rates and now the definition of "cheap" has changed?

I invest for a living.

You're misunderstanding what I'm saying.

Im saying P/E 10 is irrelevant. Current P/E numbers (with some basic context) are far more valuable.

I asked if you're saying it's always been irrelevant, or whether it's just irrelevant now? Also, are you saying that chart looks different for current P/E?

As in you're a fund manager or you invest your own money?

It hasn't even existed for 20 years yet, but once it has been forward tested we'll see it's been a poor predictor of returns. Obviously anyone can come up with a decent model to fit past data.

The chart would look similar for current P/E. P/E 10 sucks because it complicates things without adding predictive information. But that's not to say that P/E or earnings yield isn't a relevant number.

That looks pretty reasonable fit to me. And is more than 20 years.

So for that you're saying that the comparison needs context because the definition of cheap has changed due to low yields?

Edit Warren Buffett agrees with you, provided interest rates don't go up too much.

Recently, the famed investor Warren Buffett was on CNBC with Becky Quick and was asked about his thoughts about the valuation of the stock market.
"Measured against interest rates, stocks actually are on the cheap side compared to historic valuations. But the risk always is, is that interest rates go up a lot, and that brings stocks down. But I would say this, if the ten-year stays at 230, and they would stay there for ten years, you would regret very much not having bought stocks now."

The point is he "fitted" the model on the same data you're using to confirm its validity. Also, as I mentioned, it's a mathematical fact that if you look at any time period in a reasonably stable market your best returns will follow a low p/e and vice versa. The problem is we have no idea if 20ish will turn out to be a low or high or average p/e over the next 100 years. I'd bet it'll be around the average. So the fact that it would have been on the high side in the 1900s is irrelevant.

Anchoring your investment philosophies to numbers from the past century is more likely to hurt than help.

It's hardly like he's taken 6.457 year p/e and used it to predict 5.6 year returns. Presumably it was a theory and some analysis to test it. Also that's not strictly true, you can see the relationship is consistent across the five time periods chosen.


Exactly, so the definition of "cheap" has changed. I'm not disagreeing with you on that, I was taking exception to you saying before that the chart is less than meaningless which it clearly isn't.

Separately, if you bet "it'll be around the average", why are equities cheap now? Because interest rates now are lower than your expected average over the next 100 years?

The whole point of the shiller pe is to determine when stocks are cheap and when they aren't. And it's saying equities are overvalued right now. If you invest based on it you might never own equities for the next 30 years. That's why it's potentially detrimental, ie. worse than just ignoring it and investing in a balanced portfolio.

Saying that it's still relevant but the definition of cheap has changed is silly, because the whole point of it is to predict what's cheap and what's expensive.

I don't recall what I said exactly, but I think they're cheap compared to what Shiller predicts, ie. That we can expect decent returns. I also expect decently strong earnings growth in the major indices so they're quite cheap relative to other assets (bonds). I'm not saying that I think they're cheap compared to future p/e valuations, but it honestly wouldn't even shock me if index p/e ratios trended toward the 30s or even 40ish in the long run.

So what you are saying is we should f.e. discount P/E 10 by normalized interest rates since they dictated the cost of borrowing to predict if something is cheap or not?

As pointed out subsampling the P/E 10 on different time periods has been proven to be consistent, absent any cross-validation procedure.

Maybe I'm reading it wrong...


https://www.etftrends.com/2017/02/do...high-pe-ratio/

Based on that measure, we're currently at a fair valuation... I'm not claiming equities are historically cheap, I'm just saying the fearmongering isn't justified and I expect to see decent returns for the next 10-20 years.

...

https://www.hussmanfunds.com/wmc/wmc070521.htm

I would say the odds are not in the favor of the longs

That's a reasonable opinion. But there's more money in spotting secular changes early than betting on the status quo.

Where are you putting your money? Selling at the money straddles?

meh..I'll skip my strat. But yea inversely long equity. 60% in cash atm

That is wise. It is a well-known market phenomenon that if you mention what strategy you are using, it will immediately stop working. It is like how washing your car causes it to rain.

You believe in one factor that has reasonable theoretical and pretty good empirical support (value), but not the other factor (momentum) that has the same sort of theoretical and empirical support?

Mr. Market, I strongly suspect that, Rikers is all about the butterflies.