Credit deflation and the reflation cycle to come (part 2)

[Popuplights]

5 hours ago, sancho panza said:

They're also looking at Sweden and realising the Marxists at Imperial have had them over.Imperial still havent released the code they used to predict 500,000 deaths in the UK if our govt didn't lock down. as far as Im aware.

I'll need to get in touch with our resident hospital informant and find out what the current thinking is,but I think a lot of Drs are beginning to question the govts method and hence we're getting the debate over 2m rule.Lock down opponents such as Swdeish Public Helath officals said some time back that getting out of a lock down was a lot easier than going into it.

I was chatting to and A&E consutlant and his GPwife at the park yesterday as our kids played together on the swings (well whats left of the play equipment) and he didn't seem too fussed.

Quick report from Winchester hospital. Daughter says there are ZERO Covid patients.

[Loki]

1 minute ago, Popuplights said:

Quick report from Winchester hospital. Daughter says there are ZERO Covid patients.

I can hear the doomers in the pandemic thread screaming already

 

[Cosmic]

3 hours ago, TheCountOfNowhere said:

I'll raise you...

While the subject of CRE is being bought up... @DurhamBorn You've previously mentioned Newriver, what would your thoughts be on them in a post covid world? I believe you favored them as they'd secured longer term debt at low rates and their centers were more B&M than Habitat (do they still exist?), which should do better than the high end fancy debt fuel'd more upmarket placees?

[Majorpain]

3 hours ago, Talking Monkey said:

Good point that, there may be a flurry of activity and lots of jobs to get the industrial capacity up, but if it is highly automated there may not be many long term jobs to run that capacity

You want to compete in the 21st century its got to be automated and/or maximum productivity.  UK trying to compete wage wise will get creamed.

There is a big advantage to starting from scratch, you get brand new modern facilities and no pesky long term pensions etc.  Its how Germany/Japan recovered from WW2 whilst UK was stuck with uncompetitive British Leyland and British Shipbuilders through the 1960-80's.

[sancho panza]

 

6 hours ago, Loki said:

Just checked Twitter and Dave Hunter's tweet today is encouraging.  He even calls it an 'unloved rally', I know what that expression means now 

https://twitter.com/DaveHcontrarian/status/1275009938927112195

Be interesting to see where the rally would be if you exclude the FAANGs

 

[sancho panza]

8 hours ago, jamtomorrow said:

Wondering the same. To extend the "pipes" analogy ... the system will only respond as quickly as the pipes can deliver the liquidity. So how big are the pipes these days?

Or to put it another way: to what extent is the "real economy" of pay packets, mortgage payments and food shopping sufficiently coupled to the financialised corporate economy of stocks, bonds, swap lines, ETFs, commodities, derivatives etc that we might expect liquidity to move between one and the other like it has done before?

The following is from the daily shot as well.Perhaps explains why the Govts will be looking to find other ways to grease the wheels of commerce.The recent sharp upturn in the UK savings rates woud put most govts interested in pursuing stimulus of finding other routes to market thn giving it to consumers.

Also worth noting that negative rates destroy bank margins,many banks are alreayd constrained by high leverage ratios  and have a poor history of delivering credit growth in such circumstances..If govts pursue them then that's possibly antoher reason why they'll channel the stimulus themselves.

 

[Loki]

15 minutes ago, sancho panza said:

 

Be interesting to see where the rally would be if you exclude the FAANGs

 

It will be even sweeter when that means Fresnillo Anglo-American New Gold

[JMD]

9 hours ago, sancho panza said:

They're also looking at Sweden and realising the Marxists at Imperial have had them over.Imperial still havent released the code they used to predict 500,000 deaths in the UK if our govt didn't lock down. as far as Im aware.

I thought the code had been released and was found to contain bugs (not thankfully no viruses!), one team that tested it found that the model gave different results when run on different computer platforms. I believe that the computer program was originally developed in the 90's and has been variously updated and patched-up since then. However my view is that the computer model was just a tool, and the real problems stemmed from the  deficiencies in the basic science... It seems the science of virology is far from 'settled', I wonder if we will be making the same critical judgements about the man-made global warming 'settled climate science' in 20 years time? 

[AWW]

The original code was given to Microsoft, who had a number of people try to refractor it into something that looked vaguely professional. That code was then released, still full of bugs. The model gave different results for the same inputs, which is a huge red flag. Rather than fix their thread sync issues, they would just run the model a number of times, take an average and wave away the issue as being due to the stochastic nature of the model.

Unfortunately for our economy, the average of wrong is still wrong.

[JMD]

3 hours ago, Majorpain said:

You want to compete in the 21st century its got to be automated and/or maximum productivity.  UK trying to compete wage wise will get creamed.

There is a big advantage to starting from scratch, you get brand new modern facilities and no pesky long term pensions etc.  Its how Germany/Japan recovered from WW2 whilst UK was stuck with uncompetitive British Leyland and British Shipbuilders through the 1960-80's.

I agree. Companies will be encouraged to invest big in the UK, in return for allowing them to make large profits from automation. I think the political/commercial compromise will be that company profits and more importantly the taxes must remain in the country.                                                                                                                                                                      As you state, modernisation/automation is not all bad, you mention Germany and Japan, which did this admittedly from the ground up, post war. But I would also add the mooted 'Singapore on Thames' idea of our own Boris Johnson... looks like he might still get his 'Boris island' idea in some shape or form, one way or the other, but maybe on a far grander scale than originally conceived!

[sancho panza]

17 minutes ago, AWW said:

The original code was given to Microsoft, who had a number of people try to refractor it into something that looked vaguely professional. That code was then released, still full of bugs. The model gave different results for the same inputs, which is a huge red flag. Rather than fix their thread sync issues, they would just run the model a number of times, take an average and wave away the issue as being due to the stochastic nature of the model.

Unfortunately for our economy, the average of wrong is still wrong.

AWW you might understand some of the nuances in this Lockdown sceptcis review.They state the story as you do,about microsoft tarting it up,but I they say say only certain bits of the code have been released or has it all been released now?

https://lockdownsceptics.org/code-review-of-fergusons-model/

 

In your opinion,how bad was the Imerial model?Have you looekd at it much?Sue Denim states there were several things that would have given professionals in teh industry warnings.Would you agree with that?Have you any idea why the govt advisors didn't tell Bozza?

[Loki]

There's a part 2 of the review

[JMD]

1 hour ago, sancho panza said:

 

Be interesting to see where the rally would be if you exclude the FAANGs

 

 

 

SP, Interesting graph. Made me wonder if it would be a vote winner to announce radical new regulations - effectively breaking up the faangs? This would probably 'crash' the faang stocks whilst protecting the wider stock market. I dont know if this would be a win-win in political terms for someone like Trump? But he is there to 'disrupt the system' after all. And the US has a history of breaking up its own oil and telecom industries, which at the time must have been seen as very radical and 'evan self harming'.

[DurhamBorn]

5 hours ago, Cosmic Apple said:

While the subject of CRE is being bought up... @DurhamBorn You've previously mentioned Newriver, what would your thoughts be on them in a post covid world? I believe you favored them as they'd secured longer term debt at low rates and their centers were more B&M than Habitat (do they still exist?), which should do better than the high end fancy debt fuel'd more upmarket placees?

They have the best managers in the industry and iv bought a few for that reason.The problem though is debt to equity if values keep falling.Their retail park assets are good,their pub assets are good and a lot of their shopping centres have very low rents already.Their balance sheet is unsecured,thats what you want in times like this as people cant force you to sell anything etc.Loan to value is the only real worry as that could cause the banks to call the loan in.If it got close they would have to raise equity.I think they will be free cash positive over the year easily and with no divi that should help the loan to value.I would expect they will return to a 9p divi maybe next year,maybe 2.5p a quarter once they start paying again.Im a contrarian and retail property is about as hated as hated could be so i myself would buy a few and have,but if they went under it would be a mild inconvenience and any affect above that and people should avoid the sector.Intu is fubar mainly because you cant re-position massive centres,small ones you can because there is a core of discounters etc.NewRiver for instance have just singed a lease with Aldi on a unit Next have and they are extending it on the land at the side.They have also just signed B+M for a big empty unit in the park they bought from Intu for half price.Intu was trying to get the likes of Waitrose.NewRiver no doubt had B+M lined up before buying and likely budget hotels etc will follow.

NewRivers management are experts at buying up places when dislocation is going on,the problem they have though is due to the nature of the crash they cant raise equity to do it without diluting shareholders who are underwater.I hope they make it because provincial towns need some form of retail and entertainment in their areas.

[jamtomorrow]

8 hours ago, sancho panza said:

The following is from the daily shot as well.Perhaps explains why the Govts will be looking to find other ways to grease the wheels of commerce.The recent sharp upturn in the UK savings rates woud put most govts interested in pursuing stimulus of finding other routes to market thn giving it to consumers.

Also worth noting that negative rates destroy bank margins,many banks are alreayd constrained by high leverage ratios  and have a poor history of delivering credit growth in such circumstances..If govts pursue them then that's possibly antoher reason why they'll channel the stimulus themselves.

 

Cracking data SP!

So returning to the main theme of this thread - anticipating events, looking after our nearest and dearest - I think the only relevant consideration is how long Government sticks with the programme of whacking stimulus direct into the heart of the economy *if* it comes to pass that the worker (i.e. consumer i.e. voter) isn't invited to the party?

My best guess: long enough that it doesn't matter - industrial investment typically has a very long fuse indeed, and by the time it's apparent that the consumer/voter only benefits from the *creation* of the capacity, we'll be at the end of the decade anyway, bracing for The Really Big Kahuna.

[PaulParanoia]

9 hours ago, AWW said:

The original code was given to Microsoft, who had a number of people try to refractor it into something that looked vaguely professional. That code was then released, still full of bugs. The model gave different results for the same inputs, which is a huge red flag. Rather than fix their thread sync issues, they would just run the model a number of times, take an average and wave away the issue as being due to the stochastic nature of the model.

Unfortunately for our economy, the average of wrong is still wrong.

There's two main issues with the ICL model. 

1. Garbage in, garbage out.  A couple of examples.  a) They started with an unrealistically high IFR.  If i remember correctly they set it to ~3.  The CDC now say it's closer to 0.25.  b) They didn't account for increasing levels of NHS capacity (e.g. Nightingale hospitals).

2. The fact that the code never gives the same output twice makes producing quality code impossible.  Stochastic mathematics may not require reproducible results, but software development does (in order to prove that bugs are not introduced). Modelling programs use seeds to initiate the randomness of the model.  The same seeds can be used repeatedly to ensure the model hasn't been ‘broken’ by introducing bugs which cause unexpected outputs.  Then for a production run truly random seeds can be used to run thousands of iterations before averaging the results.

You only have to look at ICLs predictions for Sweden to see how broken their model was.  At the end of April, ICL predicted 96,000 dead by the end of June if Sweden didn't implement a 'proper' lock-down.  Actual number of deaths ... 5,122.

[AWW]

9 hours ago, sancho panza said:

AWW you might understand some of the nuances in this Lockdown sceptcis review.They state the story as you do,about microsoft tarting it up,but I they say say only certain bits of the code have been released or has it all been released now?

https://lockdownsceptics.org/code-review-of-fergusons-model/

 

In your opinion,how bad was the Imerial model?Have you looekd at it much?Sue Denim states there were several things that would have given professionals in teh industry warnings.Would you agree with that?Have you any idea why the govt advisors didn't tell Bozza?

AIUI, the original 15 year old C code (some of which was source-to-source compiled from even older Fortran code) has never seen the light of day and likely never will, but someone did unearth a short history of code changes going back a few weeks, which revealed that the "tarting up" undertaken was actually a major effort to make the code look like something other than a bug-riddled mess.

So, while I can't comment on the code itself, I can state categorically that the approach taken is negligent, particularly when you consider that the decisions informed by this model are life-changing for millions of people.  While I wish I could simply tut and say "bloody academics", I have seen this sort of thing going on in professional spheres, albeit in teams and organisations that know and admit they have some issues that need to be addressed, which Imperial don't seem able to do.

I have a theory that IT (as in computer engineering, not as in desktop support) isn't really like other engineering disciplines; computer engineers aren't chartered and there's no need to maintain a professional registration with an approved body.  It's also very accessible, with tons of learning material online, and a culture of sharing valuable resources that you don't really see in other industries.  This means that It's relatively easy to get something (broadly) "working" without having much understanding of what's actually going on under the bonnet.  However, the skills and knowledge required to make something work perfectly in all scenarios, particularly multi-threaded code like the Imperial model, are much harder to acquire, as they require lots of reading and lots of experience. Many in IT (myself included) are hobbyists turned pro, which isn't necessarily a bad thing, however it does mean that you're always at risk of having a hobbyist mentality, as is clearly the case with the Imperial "team".

Pretty much everything they did would be a red flag in a professional setting.  There are many things that wouldn't survive even the most cursory audit - lack of documentation for the source-to-source compiled code (which by its nature is unreadable) and zero testing capability would get you on the naughty list at most places, with stern instructions to sort it out.

If you're interested in best practice, check out the book "Clean Code".

[Don Coglione]

The Share That Shall Not Be Named has dropped out of the FTSE100.

Shocking performance.

[jamtomorrow]

50 minutes ago, PaulParanoia said:

The fact that the code never gives the same output twice makes producing quality code impossible.  Stochastic mathematics may not require reproducible results, but software development does (in order to prove that bugs are not introduced). Modelling programs use seeds to initiate the randomness of the model.  The same seeds can be used repeatedly to ensure the model hasn't been ‘broken’ by introducing bugs which cause unexpected outputs.  Then for a production run truly random seeds can be used to run thousands of iterations before averaging the results.

This, this, and this again.

If your software is *incapable* of being regression-tested, you might as well be jabbing away at the keyboard wearing a blindfold.

The other thing I *really* didn't like about the Imperial modelling: they presented a single prediction for each set of assumptions. That just seems unscientific - I want to know about sensitivity to uncertainty. A confidence interval would have been a huge improvement.

All that said ... did any of their predictions attempt to model what the Government ended up doing? I thought all of their assumption sets turned out to be counterfactuals?

[DurhamBorn]

52 minutes ago, Knickerless Turgid said:

The Share That Shall Not Be Named has dropped out of the FTSE100.

Shocking performance.

“When the hurly-burly’s done, when the battle’s lost and won.” The Scottish Play

[AWW]

47 minutes ago, jamtomorrow said:

If your software is *incapable* of being regression-tested, you might as well be jabbing away at the keyboard wearing a blindfold.

The other thing I *really* didn't like about the Imperial modelling: they presented a single prediction for each set of assumptions. That just seems unscientific - I want to know about sensitivity to uncertainty. A confidence interval would have been a huge improvement.

It's breathtaking, isn't it?  Not helped by the deification of "scientists" and perpetuation of the term "The Science", as though the scientific process always produces single incontrovertible truths.  We're all victims of this hype which, in my opinion, has risen alongside the debate around climate change.  Powerful people fund scientists whose research concludes that money should be spent doing something that benefits those funding them.  At the same time, the same powerful people fund media outlets that deify scientists, putting them beyond reproach and shielding their work from close scrutiny.

For every Einstein and Bohr, there's a Piero Anversa or a Phil Jones.

[AWW]

Apols to OP for going OT... back on topic:

Did those posters who were nervously eyeing last week's price action actually place sell orders?

[DurhamBorn]

10 minutes ago, AWW said:

Apols to OP for going OT... back on topic:

Did those posters who were nervously eyeing last week's price action actually place sell orders?

Iv picked up a few more bits last week,mainly some more Telia as i was light on them,they are my 4th biggest Telco holding,but i want them up to 35% of my Vod and TEF holdings that are both roughly equal,VOD slightly bigger.

On Telefonica iv noticed the dividend scrip rights have appeared on HL under my holding,but doesnt show any amounts yet.Repsol should follow soon.I usually take all dividends and was going to invest all dividends coming in into silver,but im going to keep the extra shares on Telefonica and Repsol while they are priced where they are.They would both have to be 25% higher to take the cash.

The interesting thing is to see how the tax works.Repsol are doing a capital increase,so i would guess there would be no tax on the divi,though not certain on that.I think Telefonica is a straight scrip issue,so probably taxed.I should know all that of course before buying something,but my Spanish on my Club 18-30 holidays  only got as far as "Tienes una hermosa sonrisa" ,

[Castlevania]

26 minutes ago, AWW said:

Apols to OP for going OT... back on topic:

Did those posters who were nervously eyeing last week's price action actually place sell orders?

Who was thinking about selling?

[MrXxxx]

3 hours ago, PaulParanoia said:

There's two main issues with the ICL model. 

1. Garbage in, garbage out.  A couple of examples.  a) They started with an unrealistically high IFR.  If i remember correctly they set it to ~3.  The CDC now say it's closer to 0.25.  b) They didn't account for increasing levels of NHS capacity (e.g. Nightingale hospitals).

2. The fact that the code never gives the same output twice makes producing quality code impossible.  Stochastic mathematics may not require reproducible results, but software development does (in order to prove that bugs are not introduced). Modelling programs use seeds to initiate the randomness of the model.  The same seeds can be used repeatedly to ensure the model hasn't been ‘broken’ by introducing bugs which cause unexpected outputs.  Then for a production run truly random seeds can be used to run thousands of iterations before averaging the results.

You only have to look at ICLs predictions for Sweden to see how broken their model was.  At the end of April, ICL predicted 96,000 dead by the end of June if Sweden didn't implement a 'proper' lock-down.  Actual number of deaths ... 5,122.

That must have been the `outlier` they deleted before publishing it! :-) :-):-)