Thought Bubbles from the Deep

[DeepState]

nice fall in Wall Street, rebound as expected has been quite short, find harder to justify the strength of the AUD vs US
I would have expected an actual worsening, even in front of a falling USD vs Euro etc

It's strange. The rise in Fed rates was supposed to trigger a sell-off in markets because it signaled the end of the era of cheap money. Now the delay in this activity is taken as a reason to sell-off. The delay being generated by global market volatility with associated downside risk to demand and inflation. What, and this wasn't already known by the market which has had a bit of a dip since the Yuan devaluation(s) and weak Flash PMI. That and the suspect six, I guess. Anyhow, as good a time as any to backfill a reason why the market fell today as opposed to yesterday when all the required information was released in minutes and did not provide anything terribly complex to analyse. I guess it took a day or two for the market to react to the Flash PMI release as well.

The USD weakened because the carry trade is reversed as interest rates turned out not of have been raised when some likelihood of this was being factored. As a result, the USD fell broadly against other currencies. Because we are looking at pushing back the curve only a few months, the impact is not very large, at least from this source.

 

[DeepState]

Always amazed at your reading.

Banks are hitting aroun 70% payout ratio which annoys me. APRA should be forcing them to use the good times to increase their buffers since all the majors are in the bottom half of their global peers when you look at the non weighted leverage ratio. The CET1 ratio will likely be discarded during the next crisis, just as it was durign the GFC. The banks had their risk weighting for home loans as low as 16 but have been forced to move it back towards 25%, though in the past it has been as high as 50%.

Banks maintain dividend yields for cosmetic purposes. They are raising capital via rights issues. I'm not sure I entirely believe it but this and retained future earnings is supposed to be sufficient to see the banks at the bottom end of first quartile in terms of capital adequacy - however this is harmonized.

Risk weights depend on aggregate LVRs. Lending in the past has been more aggressive in this respect than currently. However, if you look at places like Rabobank, you will see high risk weights.

Another headwind for the banks will be NSFR (Net Stable Funding Ration) that's still being developed. According to the Reserve Bank of Australia, short-term debt currently comprises about 20 per cent of bank funding, with about 60 per cent of that coming from offshore markets. That is the very type of funding that the financial system inquiry chaired by David Murray said was the most risky, because it could disappear in a crisis. Banks haven't reduced their reliance on this kind of short term funding since the GFC.

It will be interesting to see how the Turnbull Govt handles this issue as the banks will argue they can get 1 year paper at 26 bps + bbsw but 5 year is 90 bps + bbsw and will do their best to scare those in debt they'll face higher interest costs.

RBA would step in in the event of a liquidity crisis. That's its job. However, the less intervention the better I suppose. Deposit based funding has been steadily climbing. In the GFC, bank debt issued offshore was government guaranteed which prevented material disruption from funding.

 

[DeepState]

What we have here people is an example of what is wrong with the democratic system of government:



Trump could be the Commander in Chief of the US and the world's strongest military.



Is this the same phenomenon that saw Corbyn elected to lead Labour in the UK? Or Syriza to run Greece?

If the usual stuff doesn't work, I suppose it's time to get ridiculous. The Republican voters hate their own government representatives.

 

[skc]

What we have here people is an example of what is wrong with the democratic system of government:

This explains everything.

https://www.aussiestockforums.com/forums/showthread.php?t=4817&page=4&p=882863&viewfull=1#post882863

Sarah Palin was very close to becoming vice president. Pauline Hanson led One Nation managed 22.7% of primary votes in the 1998 QLD election. I really think voting shouldn't be compulsory... at least those who don't care won't cast random protest votes.

We might start to see a new trend here. Those with presidential ambitions would all start their own TV shows, a few years before entering presidential election, featuring them as intelligent leaders... at the same time, they'd shy away from having any actual political experience as it's a potential detriment.

 

[McLovin]

What we have here people is an example of what is wrong with the democratic system of government:

People across the Anglosphere seem to be sick of cookie cutter politicians who offer the exact same thing as the other side just wrapped differently. Corbyn, Trump and Bernie Sanders seem to be the end result. Tbh, I am more shocked by Corbyn's popularity than I am of Trump's.

I guess it's also possible we're at an inflexion point, pax Americana is coming to a close, the economic boom of the 80's, 90's, 00's now seems a long time ago, and people are no longer accepting the status quo.

 

[banco]

People across the Anglosphere seem to be sick of cookie cutter politicians who offer the exact same thing as the other side just wrapped differently. Corbyn, Trump and Bernie Sanders seem to be the end result. Tbh, I am more shocked by Corbyn's popularity than I am of Trump's.

I guess it's also possible we're at an inflexion point, pax Americana is coming to a close, the economic boom of the 80's, 90's, 00's now seems a long time ago, and people are no longer accepting the status quo.

Corbyn's an interesting example of why left wing parties need to limit the role their members have in choosing a leader if they want to stay electorally viable.

 

[Atari rose]

``America still has plenty of problems, starting with a political class motivated by self-gain above national interest. Trump, Democratic presidential candidate Bernie Sanders and other candidates have clearly tapped into disgust with political elites who seem worlds away from the citizens they’re supposed to represent. And compared with its own glory days from 1960 to 2000 or so, the U.S. is underperforming, with growth slowing, wages stagnating and national leaders failing to devise solutions.``


Full article here:

http://finance.yahoo.com/news/memo-to-trump--america-still-wins--here-s-how-173550320.html

 

[craft]

Imagine that the yield is the bond markets "implied annualised growth forecast" for the duration of the bond. It isn't, but imagine it is. So if the 10y risk free rate is 2% then imagine the bond market is forecasting annualised growth of 2% per annum for 10 years.

Now consider Intrinsic Valuation of a listed company. On the one hand, because the interest rate is low, your bar for earnings yield is lower. But the other edge of that sword is that you can no longer realistically forecast high levels of growth far out into the future.

Assuming S&P500 EPS growth of 6.3% and current Cyclically Adjusted P/E (10) of 26.48 and the long term historical average of CAPE as 16, dividend of 2% we approximate the annualised return for the next 10 years as:

100 * (1.063 * (16/26.48)^(1/10) - 1 + 0.02) = 3.07%

Currently the US 10Y yield is 2.05% ...

So the (much more realistic) equity risk premium is only 1%. Basically, you have to be willing to take on double the volatility for an extra 1% per annum return, and that is assuming you believe EPS can grow at 6.3% per annum with profit margins at record highs, with no cuts to dividends.

Hi sinner sorry for dragging up an old post but I only just got my login re-established.

(16/26.48)^(1/10) adjusts for the normalisation of the earnings multiple to historical average.

Do you think this normalisation of earnings multiples will take effect in isolation to changes in inflation/interest rates? If the normalisation doesn’t occur because we stay in a low rate environment then the adjustment factor could/should be zero (negligible) Implying a much higher equity premium.

If inflation/interest rates do change then the other variables (ie growth and div) will also vary, so it’s not a static equation. If interest rate jump to 6% bonds get smacked big time in accordance with their duration in a way that is clearly understandable but stocks it depends on how the margin holds and how asset utilisation is managed under increasing inflation. Some companies with strong competitive advantage and good business economics will thrive with some inflation, others will be hurt, you can’t make blanket assumption like you can bonds, but I guess you can average for the market overall.

I guess I’m saying normalisation of the valuation multiple independent of inflation/interest rate consideration doesn’t seem entirely likely to me so I’m not sure there’s a lot of usefulness in the equation.

Thoughts?

 

[CanOz]

Craft, good to see you back posting

 

[sinner]

I guess I’m saying normalisation of the valuation multiple independent of inflation/interest rate consideration doesn’t seem entirely likely to me so I’m not sure there’s a lot of usefulness in the equation.

Thoughts?

Heya,

Yep. Firstly just to clarify the point: the equation posted is a shorthand model for the market overall based on a bunch of assumption using long term historical averages. However these simple models do seem to have a pretty high correlation (>80%) with long term nominal returns and adding another piece for profit margin normalisation increases the correlation even further.

No mean feat, if you ask me! I dispute the equation is not useful, but the problem is that unless you're willing to take historical data for proof, we'd have to wait 10 years or so to find out empirically. However you can go back and look at Hussmans weekly comments from just before the tech bubble burst, 2003 bottom, pre GFC or 2008 bottom to see that they did already provide what might be otherwise considered an uncanny "walk forward" utility.

I do absolutely agree that some companies will thrive in inflationary circumstances more than others, but generally from looking at the data we can see that investors almost always misprice inflation risk so the price of most stocks (even those benefiting from inflation) suffer an initial inflation shock and only then go on to thrive. This can even be witnessed in the data for Argentina stocks which you might know has still huge inflationary pressure and has suffered multiple inflationary periods.

I'm careful not to equate rates with inflation. As I mentioned in another thread recently, the short end of the curve is controlled by the economies CB and the long end of the curve is driven more by marginal expectations for both growth and inflation. You can run a spread against the 10Y and 10Y TIPS to try and see which bit is inflation and which isn't. So it depends of which you speak.

Having said that, none of the factors are in isolation, the 6.3% assumption for growth is based on the long term peak-to-peak rate in S&P500 EPS growth. Of which I believe about 3% is implicitly due to inflation, so if it made you feel any better you could break it out to be (1 + 0.33 + 0.3) rather than 1.063 and say the model assumes 3% inflation . The forecast itself provides its most apparent utility when plugged into a Sharpe ratio (as I attempted to show in the post you quoted), i.e. in comparison to the known future (nominal) risk free rate, with the idea being that you don't want to invest while that Sharpe ratio is negative (i.e. the signal is not impacted by the volatility portion of the ratio).

On that note here is an awesome little comic representation of one of Eugene Fama's papers on the Fed and interest rates:
http://www.chicagobooth.edu/capideas/magazine/fall-2015/whos-really-in-charge

Interesting topic, happy to discuss more, so if you feel that wasn't adequate I'm happy to try and convince you some more

EDIT: I do advise just reading the Hussman Weekly Market Comments directly if you're interested, because I am sure he does a much better job of explaining things than I can. When it comes to stuff like this I'm just poorly translating what I read there in my own words.

 

[systematic]

Craft, good to see you back posting

Hear, hear!

 

[fraa]

Hello Craft

I am a lurker who is learning from reading and have nothing to contribute here other then to thank you for all your previous posts that you have here. Glad that you are back and posting again.

Back to lurking for me.

Hi sinner sorry for dragging up an old post but I only just got my login re-established.

(16/26.48)^(1/10) adjusts for the normalisation of the earnings multiple to historical average.

Do you think this normalisation of earnings multiples will take effect in isolation to changes in inflation/interest rates? If the normalisation doesn’t occur because we stay in a low rate environment then the adjustment factor could/should be zero (negligible) Implying a much higher equity premium.

If inflation/interest rates do change then the other variables (ie growth and div) will also vary, so it’s not a static equation. If interest rate jump to 6% bonds get smacked big time in accordance with their duration in a way that is clearly understandable but stocks it depends on how the margin holds and how asset utilisation is managed under increasing inflation. Some companies with strong competitive advantage and good business economics will thrive with some inflation, others will be hurt, you can’t make blanket assumption like you can bonds, but I guess you can average for the market overall.

I guess I’m saying normalisation of the valuation multiple independent of inflation/interest rate consideration doesn’t seem entirely likely to me so I’m not sure there’s a lot of usefulness in the equation.

Thoughts?

 

[sinner]

I must concur, craft, good to have you back.

Your keen insight and analysis skills are both an inspiration and an aspiration that has definitely been missed by me.

But it's not just your absence, because I have seen that other smart people play off your posts and you play off theirs, a whole greater than the sum of parts. This has been sorely lacking.

I only learned this recently, did you know that ASF has an ignore function?

 

[craft]

Heya,

Interesting topic, happy to discuss more, so if you feel that wasn't adequate I'm happy to try and convince you some more

EDIT: I do advise just reading the Hussman Weekly Market Comments directly if you're interested, because I am sure he does a much better job of explaining things than I can. When it comes to stuff like this I'm just poorly translating what I read there in my own words.

Thanks Sinner

I really should read Hussman more regularly, thought provoking. Normalising the growth rate by peak to peak does make sense in light of the equation. So with everything normalised you get the 3.07% expected return against a 2.05% 10 year bond rate – but the normalised inflation figure embedded is 3% which doesn’t equate to a 2.05% bond rate for comparison. You are going to get a capital loss on the bond as it embeds a 3% inflation number. Maybe the comparison should be against zero duration (cash)

I don’t know, something just isn’t gelling here for me and I can’t put my finger on it exactly.

The other thing that doesn’t seem to gell is a 6.3%growth rate and 2% div with a normalised ROE for the US of ~ 12%

Any rate thanks for the reply – I’ll consult Hussman and keep trying to fill in some blanks for myself.

Cheers

Ps
Thanks to those who offered kind words of support but it is not my intention to post regularly anymore. I only really got my login re-stated to access some of the links posted and for occasional topics like this when I’m trying to clear a roadblock in my thinking.

 

[McLovin]

Ps
Thanks to those who offered kind words of support but it is not my intention to post regularly anymore. I only really got my login re-stated to access some of the links posted and for occasional topics like this when I’m trying to clear a roadblock in my thinking.

I hope you have more roadblocks in your thinking then.

 

[DeepState]

Welcome back Craft. The quality of debate and exchange improves for your presence. In line with Sinner, consider exercising the Ignore feature more than what clearly has been past practice. Those who know about this stuff can see your quality.

Shall revert to your question later and hopefully disentangle it with you. I am conflicted, though. If I answer your question, you'll go dark. Hmmmm...what to do, what to do.

 

[sinner]

Thanks Sinner

I really should read Hussman more regularly, thought provoking. Normalising the growth rate by peak to peak does make sense in light of the equation. So with everything normalised you get the 3.07% expected return against a 2.05% 10 year bond rate – but the normalised inflation figure embedded is 3% which doesn’t equate to a 2.05% bond rate for comparison. You are going to get a capital loss on the bond as it embeds a 3% inflation number. Maybe the comparison should be against zero duration (cash)

You should use whichever risk free duration that matches your investment horizon, keeping in mind that these sort of valuation measures are not really effective in the short term.

The rest of the above is a little confusing for me, not sure I understood it.

The other thing that doesn’t seem to gell is a 6.3%growth rate and 2% div with a normalised ROE for the US of ~ 12%

You know that aside from the valuation mean reversion bit, you can plug in whichever numbers you like, right? The shorthand model uses peak-peak EPS growth and the current dividend yield. More accurate measures will likely lead to more accurate future forecasts

We are moving into a new house soon and I was over there tonight setting up my new desk and chair (fancy). On the train home I was thinking about this topic and something that came to mind is the Tobins Q. You can read about this on wikipedia as AFAIK the Tobin dude won a Nobel prize for it like Shiller (what's with these valuation ratio guys being Nobel Laureates?). In both cases as it turns out the mean reversion component of the valuation ratio is predictive of long term returns regardless of any other factors, used completely in isolation. Both CAPE and Tobins Q were far more predictive than any other measure previously analysed.

There's a good paper by Spitznagel of Empirica (Taleb protege) called (IIRC) "The Dao of Corporate Finance" which covers this (including some out of sample since it was produced well after Q was publicly released) and keep in mind this guy uses valuations as a major input to decide when to buy volatility!

 

[sinner]

I have posted this chart and blogpost before I think, but I'll post it again because it's apt


http://gestaltu.com/2014/07/valuation-based-equity-market-forecasts-q2-2014.html

including this small disclaimer, obviously written solely for craft himself ...after rereading I actually think I remember something similar being said in one of the Hussman posts about this.

Moreover, we are acutely aware that interest rates are a discounting mechanism and thus low interest rates (especially rates which are expected to remain low for a long time) may justify higher than average equity valuations. This may be a normal condition of asset markets, but it doesn’t alter forecasts about future returns. While markets might be ‘fairly priced’ at high valuations relative to exceedingly low long-term interest rates, this does not change the fact that future returns are likely to be well below average. Again, a market can be ‘fairly priced’ relative to long-term rates, yet still exhibit high valuations implying lower than average future returns. We wouldn’t argue with the assertion that current conditions exhibit these very qualities. However, this fact does not change ANY of the conclusions from the analysis below.

The "Considerations & Next Steps" section at the bottom has some juicy stuff to think about.

 

[DeepState]

CAPE

In theory, 'Equilibrium' CAPE should be impacted by the level of interest rates and inflation prevailing at the time. That and also the nature of the companies in the index at the time.

However, the stats pumped out suggest that the fit of CAPE vs future returns has worked better without any such adjustments. Hence, standard practice is to utilise absolute CAPE and regress it against future returns.

Most often quoted is US data since about 1929. Look at the long period returns and you will see only two major cycles in returns. This, ultimately, is where the regressions are pulled from. It is two major peaks and three(?) major troughs. For a measure like CAPE and the application to long term returns only, you will appreciate that this is not a heck of a lot of data points. Claims of high degrees of fit misuse the statistical method of regression and results in 'spurious regression' or otherwise not allowing for the lag structure in both CAPE and the subsequent returns being regressed. It's just not correct and overstates the fit.

The 6.3% long term real returns achieved do not gel with RoE of 12% and 2% divs. However, the 6.3% real return is measured over the long term. RoE of 12% is also a long term measure. However, dividend yields of 2% are a more recent phenomenon. Dividend payouts used to be much higher. In one of those head fakes, dividend yields today are much lower, but the inclusion of buy-backs has meant that the payout has been nearly 100% since 2004. EPS real growth has been anaemic since about 2007.

The 10 year bond yield is taken to be an equilibrium interest rate. It supposedly embeds the future expected path of short rates and adds a little bit for taking term premium risk. On a buy-hold forever style analysis you would compare the IRR on 10 year bonds with that expected from equities. The equity hold forever return is basically a Dividend Yield with is compounded by a magic number usually expressed as a sum of: expected future inflation, population growth, productivity growth less capital expenditure required to finance that growth. That's it. No fade path for CAPE to some average equilibrium figure. Unlike interest rates, there is no equilibrium value for CAPE. CAPE is used as a departure point for what essentially becomes a DDM. You can compare the two. There is also no magic number that says the gap needs to equilibrate to X% per annum.

Once again, it is claimed that CAPE works better without such adjustment anyway based on historical observation. Hence, you could just use current cash, expected inflation, or pretty much any reasonable sounding comparison to your CAPE based DDM. Just using CAPE on its own in absolute measure on the basis that it mean reverts only makes sense if the expected future growth rate of dividends doesn't foul it all up. As the variation in CAPE is much greater than the variation in expected/actual dividends, the CAPE reversion effect has dominated. Turns out Shiller wrote a paper on this too.

However, all that will not prevent observers or other market professionals from making allowance for reversion in CAPE and assuming hold to maturity for bonds when compared. I think this is broadly reasonable, but it is not the result of a bedrock theory. It is a guess at the market's behaviour based on the observation of the measure over two full cycles of equity return history. Quite possibly, for a ton of reasons, the right CAPE today is higher than the average of the past (of which there are only 9 independent observations over the data). Quite possibly, high CAPE will move back to average by means of a reinvigoration of earnings growth and on it goes.

----

I do believe in the use of cyclically adjusted earnings as one basis for long term valuation. I use it myself. However, I do not generally compare the measure today to the past and assume reversion in the belief that the figure is stationary. It is not. However, it has apparently behaved like it has been over the last century....when observing non-stationary data.

Hope this helps.

 

[sinner]

Some more sample has been conducted by these guys by trying to examine as many different markets as possible

http://papers.ssrn.com/sol3/papers.cfm?abstract_id=2470935

and they made this pretty site

http://www.starcapital.de/research/stockmarketvaluation

But I don't think CAPE is the be-all-and-end-all of what we are discussing, I think we discussed "sufficient statistic" before, lots of measures do just as well or better.

http://www.hussmanfunds.com/wmc/wmc140414.htm

A corollary to the Iron Law of Valuation is that one can only reliably use a “price/X” multiple to value stocks if “X” is a sufficient statistic for the very long-term stream of cash flows that stocks are likely to deliver into the hands of investors for decades to come. Not just next year, not just 10 years from now, but as long as the security is likely to exist. Now, X doesn’t have to be equal to those long-term cash flows – only proportional to them over time (every constant-growth rate valuation model relies on that quality). If X is a sufficient statistic for the stream of future cash flows, then the price/X ratio becomes informative about future returns. A good way to test a valuation measure is to check whether variations in the price/X multiple are closely related to actual subsequent returns in the security over a horizon of 7-10 years.