Return Distributions

[sinner]

and lastly, to make sure OPs question isn't left unanswered I present

I recommend having a read of " A Value Investor’s Perspective on Tail Risk Protection:An Ode to the Joy of Cash" by GMOs James Montier, presented here http://trendfollowing.com/whitepaper/JM_TailRisk_611.pdf

as a well researched and thoughtful perspective on the topic (which interestingly enough also shows how simple rebalancing 75/25 stocks/cash can outperform in certain market regimes)

EDIT: even something for craft in there

As Warren Buffett has said, holding cash is uncomfortable, but not as uncomfortable as doing
something stupid

 

[DeepState]

The weaknesses highlighted by Hussman are specifically in relation to the "Fed Model" using Forward Operating Earnings to represent the cash flows of the entire lifetime of the security...


Hussman is a big fan of CAPE and P/Sales, anything that better describes the cashflows of the security over its lifetime.



EDIT: Also the multiples above are CAPE as in PE10, not PE1 and Q ratio which is more akin to Price/Book. You can see from the r-squared graph that explanatory power for PE1 is much lower than PE10 for all timeframes.

Yup. Thanks.

Basically stabilized earnings produces better outcomes than unstabilised ones. Part of this can be overcome by looking within economically coherent sectors or, otherwise, groupings. The cross section there applying Fwd whatever should be comparable to a higher degree than across the entire market. Two lemonade stands have earnings that a more comparable than lemonade vs iron ore. Hence cyclical peaks and troughs ought to match up more closely and the differences in the valuations should provide a better indication of future cross sectional return. Even then, the CAPE style measures should still be enhancing where firms have stable growth patterns.

I still have no good answer for firms coming into profitability after a build phase or which are moving into resource depletion. CAPE style measures would probably kick off misleading signals. Nothing is perfect I guess.

Thanks for the EDIT. What's what I needed to see.

 

[sinner]

I still have no good answer for firms coming into profitability after a build phase or which are moving into resource depletion. CAPE style measures would probably kick off misleading signals. Nothing is perfect I guess.

Not 100% certain but I think you can use Price/Sales, which is essentially Price/Earnings-adjusted-for-profit-margins (at least according to Hussmans latest missive) but as you noted this requires more due diligence to ensure not just idiosyncratic but also relative valuations.

 

[DeepState]

Not 100% certain but I think you can use Price/Sales, which is essentially Price/Earnings-adjusted-for-profit-margins (at least according to Hussmans latest missive) but as you noted this requires more due diligence to ensure not just idiosyncratic but also relative valuations.

Price to Sales, Price to EBITDA, Price to CF, trend as opposed to CPI adjusted of the lot are all used. All work. It's bloody amazing. Particularly at aggregate market levels where things are pretty much as stable as we are practically going to find and thus suitable for extrapolation via such means.

Even still, inside a market, the firms in those situations as first sale or depletion would have hockey stick up/down sales etc. and trend adjustments will lead us to come to the wrong conclusions. Guess we need to still lift our heads and think. Dammit.

Thanks

 

[craft]

and lastly, to make sure OPs question isn't left unanswered I present

I recommend having a read of " A Value Investor’s Perspective on Tail Risk Protection:An Ode to the Joy of Cash" by GMOs James Montier, presented here http://trendfollowing.com/whitepaper/JM_TailRisk_611.pdf

as a well researched and thoughtful perspective on the topic (which interestingly enough also shows how simple rebalancing 75/25 stocks/cash can outperform in certain market regimes)

EDIT: even something for craft in there

I find Montier excellent.

Liquidity value of cash has always been a perplexing question for me. KTP resparked the question below when he eluded to the liquidity value of short positions at opportune times to go long.

The way I think about it is to add an option value to what I could get from cash (or a short side approach etc).

So cash will get you about 4% now and I add a % to that figure for the extra I could make if I waited to invest as opposed to investing now. – That becomes the hurdle rate where I would stay in cash rather than deploy it.

Estimating the liquidity option value is pretty subjective – it depends on so many unknowable future events. For me and my skill/temperament set I think its somewhere around 3%. Mostly I can find some opportunity that meets a (current minimum 7%) hurdle rate and don’t tend to delay purchase. So my cash exposure is generally limited to only liquidity buffer for 5 years of living expenses.

 

[sinner]

Even still, inside a market, the firms in those situations as first sale or depletion would have hockey stick up/down sales etc. and trend adjustments will lead us to come to the wrong conclusions. Guess we need to still lift our heads and think. Dammit.

Thanks

Make an arbitrary decision like "all stocks in universe must have 5Y earnings history" and then rank on earnings stability (e.g. rolling standard deviation of QoQ earnings change - or similar measure)...

 

[craft]

Price to Sales, Price to EBITDA, Price to CF, trend as opposed to CPI adjusted of the lot are all used. All work. It's bloody amazing. Particularly at aggregate market levels where things are pretty much as stable as we are practically going to find and thus suitable for extrapolation via such means.

Even still, inside a market, the firms in those situations as first sale or depletion would have hockey stick up/down sales etc. and trend adjustments will lead us to come to the wrong conclusions. Guess we need to still lift our heads and think. Dammit.

Thanks

Once you found the spot to go fishing nothing beats bottom up business analysis

 

[DeepState]

And another from Crestmont research based on USA data.

20 year rolling market return based on starting P/E.

View attachment 57815

The P/E should really be adjusted for discount rates prevailing at the time. A stock at p/e 10x is much cheaper when bonds are at 2% than when they are at 10% all else equal. If you look at charts showing P/E vs interest rates it shows a relationship between P/E and interest rates as might be expected. So I'm puzzled as to why CAPE is always shown in absolute terms vs subsequent equity return (which really should be equity risk premium).

This is all theoretical BS, but a lot of the people producing this stuff often go to the trouble of making heaps of different adjustments to make the underlying concept more accurate. But not here. Shiller uses these straight up....and everyone else follows. It also has strong empirical outcomes.

I raise this because CAPE shows US as being in the drop zone. US 10yr is 2.62%...If that's what you can buy bonds for, a fair P/E should be higher than usual. And, in a recent interview, Buffett, who uses a cash rate type discount rate indicated that the market is in the fair value zone. This is entirely different conclusion to CAPE.

 

[DeepState]

Once you found the spot to go fishing nothing beats bottom up business analysis

I'll pay $1 for that.

 

[craft]

The P/E should really be adjusted for discount rates prevailing at the time. A stock at p/e 10x is much cheaper when bonds are at 2% than when they are at 10% all else equal. If you look at charts showing P/E vs interest rates it shows a relationship between P/E and interest rates as might be expected. So I'm puzzled as to why CAPE is always shown in absolute terms vs subsequent equity return (which really should be equity risk premium).

This is all theoretical BS, but a lot of the people producing this stuff often go to the trouble of making heaps of different adjustments to make the underlying concept more accurate. But not here. Shiller uses these straight up....and everyone else follows. It also has strong empirical outcomes.

I raise this because CAPE shows US as being in the drop zone. US 10yr is 2.62%...If that's what you can buy bonds for, a fair P/E should be higher than usual. And, in a recent interview, Buffett, who uses a cash rate type discount rate indicated that the market is in the fair value zone. This is entirely different conclusion to CAPE.

Some more on inflation (discount rate driver) in relation to PE from Crestmont.



Crestmont do (and I agree) consider discount rate in their assessments. From their latest P/E report.

CONCLUSION
Today’s P/E is approximately 26; the stock market remains in secular bear market territory””
close to the mid-range of fair value assuming a relatively low inflation and low interest rate
environment.

 

[KnowThePast]

and lastly, to make sure OPs question isn't left unanswered I present

I recommend having a read of " A Value Investor’s Perspective on Tail Risk Protection:An Ode to the Joy of Cash" by GMOs James Montier, presented here http://trendfollowing.com/whitepaper/JM_TailRisk_611.pdf

as a well researched and thoughtful perspective on the topic (which interestingly enough also shows how simple rebalancing 75/25 stocks/cash can outperform in certain market regimes)

EDIT: even something for craft in there

Thanks sinner!

It was his book that initially got me thinking of shorts as a hedging strategy.

DeepState, thanks for the answer once again.

 

[craft]

The P/E should really be adjusted for discount rates prevailing at the time. A stock at p/e 10x is much cheaper when bonds are at 2% than when they are at 10% all else equal. If you look at charts showing P/E vs interest rates it shows a relationship between P/E and interest rates as might be expected. So I'm puzzled as to why CAPE is always shown in absolute terms vs subsequent equity return (which really should be equity risk premium).

This is all theoretical BS, but a lot of the people producing this stuff often go to the trouble of making heaps of different adjustments to make the underlying concept more accurate. But not here. Shiller uses these straight up....and everyone else follows. It also has strong empirical outcomes.

I raise this because CAPE shows US as being in the drop zone. US 10yr is 2.62%...If that's what you can buy bonds for, a fair P/E should be higher than usual. And, in a recent interview, Buffett, who uses a cash rate type discount rate indicated that the market is in the fair value zone. This is entirely different conclusion to CAPE.

Hi RY

This might interest you in relation to ERP and the Aust Market.



It was a pretty good indication in June last year when last discussed.

https://www.aussiestockforums.com/forums/showthread.php?t=23385&p=710550&viewfull=1#post710550

 

[DeepState]

Hi RY

This might interest you in relation to ERP and the Aust Market.

View attachment 57848

It was a pretty good indication in June last year when last discussed.

https://www.aussiestockforums.com/forums/showthread.php?t=23385&p=710550&viewfull=1#post710550

Thanks Craft. Appreciate it. Nice chart and mostly in line with what I expected to see - in terms of the inverse relationship. Interesting that the time series was broken into the pre-float regime and post-float. Pre-float ERP showing much more elasticity to 10 year yield and much more poorly fitting than the post-float era. Guess that the loss of that source of economic flexibility meant that the markets absorbed more of a hit than they do today, under the current monetary regime. Also, the composition of the index then was much more heavily tilted towards resources than they are today. So I guess it all makes sense.

However, the steepness of the relationship in the post float era also suggests that the change in ERP is largely driven by changes in the bond yield with little left for PE rerating. In other words ERP is less sensitive to the level of 10 yr bond rate than before. To the point that PE doesn't move all that much for different interest rate levels. This is a correlation, not causational statement. That's puzzling. It suggests a very weak relationship between PE and bond yields. And, whilst moving to AUD float can expect to absorb some economic shocks, I did not expect this.

This is a kind of "Fed Model".

Despite the theoretical attraction (I think what we do should match some underlying theory or at least rough economic understanding of a stable relationship if no formal theory has been developed), we then have the fly in the ointment of Hussman who said in his latest note ("The Future in Now"):

"The Federal Reserve has stomped on the gas pedal for years, inadvertently taking price/earnings ratios at face value, while attending to “equity risk premium” models that have a demonstrably poor relationship with subsequent returns. As a result, the Fed has produced what is now the most generalized equity valuation bubble that investors are likely to observe in their lifetimes."

If our theory does not match the reality in terms of the chart and then Hussman....then, what the heck?

This may also be tangentially relevant to you in terms of using long term valuation measures to determine intrinsic long term valuation - not taking anything at all away from the concept you are pursing and which I am as well in terms of market aggregates.

Thanks for the provocation. Shall look into this a bit further. It doesn't make immediate sense to me but the results are the results.

Thanks again for all your contributions.

 

[craft]

Thanks Craft. Appreciate it. Nice chart and mostly in line with what I expected to see - in terms of the inverse relationship. Interesting that the time series was broken into the pre-float regime and post-float. Pre-float ERP showing much more elasticity to 10 year yield and much more poorly fitting than the post-float era. Guess that the loss of that source of economic flexibility meant that the markets absorbed more of a hit than they do today, under the current monetary regime. Also, the composition of the index then was much more heavily tilted towards resources than they are today. So I guess it all makes sense. .

I split the time series based on the logic that the a floating currency does a fair bit of the economic lifting - the fit of the data seemed to agree with the logic.

However, the steepness of the relationship in the post float era also suggests that the change in ERP is largely driven by changes in the bond yield with little left for PE rerating. In other words ERP is less sensitive to the level of 10 yr bond rate than before. To the point that PE doesn't move all that much for different interest rate levels. This is a correlation, not causational statement. That's puzzling. It suggests a very weak relationship between PE and bond yields. And, whilst moving to AUD float can expect to absorb some economic shocks, I did not expect this.
.

The steepness of the relationship surprised me a little bit as well, but I guess that's why we look at data. One important thing to remember (and may not be obvious if you didn't look back at the linked thread) is that the Earnings used to calculate ERP is "Trend Earnings" Both trend earnings and regressed GDP earnings are above current earnings in Aus at the moment.

This is a kind of "Fed Model".

Despite the theoretical attraction (I think what we do should match some underlying theory or at least rough economic understanding of a stable relationship if no formal theory has been developed), we then have the fly in the ointment of Hussman who said in his latest note ("The Future in Now"):

"The Federal Reserve has stomped on the gas pedal for years, inadvertently taking price/earnings ratios at face value, while attending to “equity risk premium” models that have a demonstrably poor relationship with subsequent returns. As a result, the Fed has produced what is now the most generalized equity valuation bubble that investors are likely to observe in their lifetimes."

If our theory does not match the reality in terms of the chart and then Hussman....then, what the heck?

This may also be tangentially relevant to you in terms of using long term valuation measures to determine intrinsic long term valuation - not taking anything at all away from the concept you are pursing and which I am as well in terms of market aggregates.

Thanks for the provocation. Shall look into this a bit further. It doesn't make immediate sense to me but the results are the results.

Thanks again for all your contributions.

I don't have much trouble reconciling what I am seeing in Aus data to what Hussman is saying about US. We are a different market, we have a stronger currency at the moment, Current earnings are below trend. We have higher interest rates. The updated chart has moved away from indicating the type of potential it did in June last year, but it's more indicative of fair (considering inflation/interest rates) than overvalued.

 

[DeepState]

1. The steepness of the relationship surprised me a little bit as well, but I guess that's why we look at data. One important thing to remember (and may not be obvious if you didn't look back at the linked thread) is that the Earnings used to calculate ERP is "Trend Earnings" Both trend earnings and regressed GDP earnings are above current earnings in Aus at the moment.


2. I don't have much trouble reconciling what I am seeing in Aus data to what Hussman is saying about US. We are a different market, we have a stronger currency at the moment, Current earnings are below trend. We have higher interest rates. The updated chart has moved away from indicating the type of potential it did in June last year, but it's more indicative of fair (considering inflation/interest rates) than overvalued.

Thanks again Craft, truly inventive.

1. I caught the link and see you are using trend (and have calculated GDP inflated). So you have produced a type of Earnings as per Shiller except his standard measure inflates earnings by CPI/PCE or its equivalent and divided it by current market. These show that Australia is under-earning relative to trend. I agree, based on my own stuff, via chart below which is prepared on exponential trend over the last 10 years (the Shiller period and long enough to detrend, yet short enough that comparative universe issues don't become overwhelming).



The steepness of the line indicates that trend-adjusted PE doesn't move much. Almost all of the Equity Premium as you measure it is driven by variation in bond yields. So the signal could readily be used for bond yield forecasting rather than equity return forecasting. How ironic! Here we are hunting for cheap/rich equities and what we find instead is a bloody strong forecast variable for 10 year bonds! Very Newton-like. Most discoveries are accidential, I guess.

2. I think we are talking about different things here. On the grounds of this type of analysis and its close variants, I agree that the US is expensive to some extent and Australia screens up as in the fair value range. The issue is that Hussman claims this sort of equity risk premium model, roughly what the Fed uses, has poor forecasting power. He just uses straight up Shiller measure and this shows to have a lot of predictive power, presumably more than these models. That cheeses me off!

Part of the answer to that quandary may lie in the fact that the effort to adjust for interest rates actually detracts from forecasting power. There are several good reasons why this might be in practice. These defy standard theoretical valuation approaches which do not take into account the forecasting power inherent in bond yields. Hence they represent a challenge to what we are attempting here. I like the basis of the approach and I use it as a stepping stone for the calculation of a discounted cashflow-based valuation metric which takes a few other things into account, but, according to Hussman at least, we are barking up the wrong tree. I think he's cleverer than me...so I guess I'm screwed. Not.

 

[craft]

Daily ASX200 EPS –You have access to good data! – supplier?

The reconciling with Hussman was only in regards to his conclusion that US is expensive whilst I think the data shows Aust is fair. I haven’t followed him closely enough to comment on the correctness or otherwise of what he thinks does or does not have predictive power. (and besides as you might guess - I don't care either)

Cheers

 

[DeepState]

Daily ASX200 EPS –You have access to good data! – supplier?

EPS data supplied from pretty much all the active brokers, as adjusted. So these are 'above the line' figures. Rolling 12 months are calculated by splicing these together as required via an appropriate weighting scheme. Data aggregator is FactSet.