Friday, March 11, 2016

JW Mason — 3 on international trade and balance of payments

"Free trade" seems to be the hot topic now.

J. W. Mason's Blog

We’ve Always Had Free Trade with Eastasia

How to Think about the Balance of Payments

How to Think about the Balance of Payments: The US Position 2012-2013

JW Mason | Assistant Professor of Economics, John Jay College, City University of New York


Neil Wilson said...

The mistake here again is that there is an underlying assumption of infinite liquidity in the FX market.

And it is all brought about by analysing the system using a dividing line on the borders of the country rather than a dynamic dividing line based around who is holding dollar denominated assets at any point in time.

There is no operational difference between somebody holding dollar denominated assets in Birmingham, Alabama & Birmingham, England.

Once you switch to a dynamic dividing line then the processes of 'dollarisation' that make a currency zone smaller than its home territory are precisely the same as the 'reserve asset' ones that make a currency zone bigger than its home territory. You have a consistent mechanism that explains both.

Once you see it the other way, it immediately throws up the policy question as to why central banks bail out people who have made bad FX bets with their 'currency reserves' rather than simply forcing those entities into administration so that the loss is imposed on the FX creditor rather than distributed across the home country via excessive exchange rate moves.

Matt Franko said...

"an underlying assumption of infinite liquidity in the FX market. " yes good point...

But Neil what if the firm in the foreign nation decides to drop its price in USD terms?

Take Daimler and their US market... a C-class is the same automobile as a well equiped Honda Accord... 4 door 5 passenger FWD 4-cyl leather/AC/power automobile...

C-class $40K.... Accord $30k due to "status" or whatever ... additional $10k of pure rent...

What if Daimler decides to reposition the C-class and takes it down towards $30k to compete with the Honda for business reasons?

If a EZ bank is financing the inventory of C-class then EUR/USD would accordingly respond and goes down to reflect the new reduced terms of trade as there is NOT as you point out 'infinite liquidity' in the fx market....

JW Mason said...

One ting that bugs me about the world of heterodox economics discussions online -- the instinct so many people have to begin a conversation by saying "you're wrong!"

Neil, I'm not *assuming* infinite liquidity. I'm *asking the question*, how much liquidity there is in practice. Why don't you try engaging the question in a positive way, instead of rushing to draw a line between your ideas and everyone else's?

I actually agree with what you wrote here -- I think it's a good way to think about these issues. And I think your final question has an answer. In general we DON'T freely allow financial entities to fail. A balance of payments crisis, it seems to me,should be thought of as a banking crisis that the the central bank can't handle in the usual way (as lender of last resort) because of the prevalence of foreign-currency denominated liabilities. Does that seem like a useful way of thinking about it to you?

Matt Franko said...

JW if you wait until the accounting comes out you've missed what happened... and dont learn anything about the process... not saying accounting is not important (2 siblings accountants...) but accounting ex post is not empiricism...

Suggest try collapsing the delta T to smaller and smaller intervals of time a la the calculus to try to figure out what is going on...

As you collapse the delta T, try to think about IF the different parameters within the regulated system can respond within smaller and smaller time period...

Banks are regulated via capital to asset ratio.... Neil's point here wrt "no infinite liquidity" I assume comes from his knowledge that banks are limited by capital... the whole system is limited by the collective total of all bank's capital hence imo Neil's assertion that there is not "unlimited liquidity" but I would add OVER SHORT TIME PERIODS...

Asset values can vary with a high frequency while capital cannot it takes a long time to do a capital raising for a bank, meanwhile eg. financed oil inventories can collapse in price in hours...

Maybe what Neil is trying to point out is that all conventional ex post analysis (like you are doing in your post) EFFECTIVELY assumes infinite liquidity because it does not take into account the frequency response of the different regulatory parameters... if you have one parameter that exhibits a high frequency response while another one has a low frequency response, if you dont recognize that IT CAN APPEAR that you are assuming "infinite liquidity" by assuming ALL parameters have THE SAME frequency response... when in an empirical analysis using calculus techniques it can be shown that is manifestly not true...

John said...


What's all this stuff about dynamic dividing lines making delta T infinitesimally small and frequency responses? Any equations or links for this stuff? You guys have inadvertently picked up Warren Mosler's impenetrable insider speak. The priesthood speaking in Latin.

Hope this kind of stuff is explained in the new MMT textbook. Or is it too esoteric and advanced for a macro textbook? It's all very interesting. Or is this just something you two have been thinking about?

Matt Franko said...

JW here:

Just think that the horizontal axis is time in this case ... deltaX is always deltaT...

John said...

I know what calculus is! It's a piece of piss, to use an English expression. I was referring to the replies you gave about time, currency zones and all the rest of it.

Matt Franko said...


Neil (I assume) is some sort of software systems analyst... My degree is in Electrical Engineering ... so we are coming at the economic systems from those contexts...

In the curriculum I studied you get a lot of general systems theory out of it which employs the calculus...

You have to do what Neil is doing FIRST, you have to identify the 'agents' in the system and the relationships, THEN you have to try to create an empirical illustration of those relationships via the calculus...

iow take Newton's 'apple and the ground'.... Neil might look at that as TWO agents in a system ... 1 is the apple and 2 is the ground in the earth system ... once you have the agents identified, then you attempt to define the relationship between them in order to make PREDICTIVE statements that have to turn out to be accurate...

"apple releases and accelerates towards the ground at 9.8m/s^2... hits with force = (mass of apple) * 9.8m/s^2..... kinetic energy at impact = 1/2mv^2...."

Dont know if Bill and Randy have been trained this way.... Bill uses a lot of correlation studies for sure... Randy seems more rationalist than empiricist imo...

imo Neil exhibits some of the best understanding of the system boundaries and perhaps 'agents' in this system and functional relationships between these agents.... now it would normally be turned over to perhaps systems theorists to start to quantify these relationships in order to demonstrate predictive ability ... then we can move on to proper regulation of the system....

Matt Franko said...

ha I dont know what that "piss" figure of speech means over there.. (good or bad? lol)

in any case its important to have technocrats on the team who understand the calculus and how it is appropriately used in proper operation/regulation of systems...

John said...


Translation: "piece of piss" is the same as "piece of cake". Use it and see how it works out for you! Major news channels will be reporting on why all of a sudden Maryland is using a weird English expression.

I was always impressed by the "More is Different" argument by Anderson. I'd say much of economics should be thought of as some sort of emergent phenomena - why macro is not the sum of micro phenomena. However, a systems approach seems a very convincing way of addressing macro phenomena.

I'll just have to trawl through Neil's blog, the bloody clever bugger! My only consolation is that I've come to this stuff about five or six years after you guys. In five years time I'll be as knowledgeable as all of you!

Neil Wilson said...

"Neil, I'm not *assuming* infinite liquidity."

Yes you are. That is the underlying assumption *in the FX market*. Not in any other lending market. *The FX market*.

The one where you have to settle at 10pm GMT *every day* or default.

That is the fundamental problem with all economic analysis using primitive mathematical techniques and 'closure'. They miss the day to day rhythm of the way financial systems work.

"In general we DON'T freely allow financial entities to fail."

Which shows that you didn't read it properly. I wasn't talking about financial entities failing. I was talking about normal firms that borrow in a foreign currency.

Once the standard FX system runs out of liquidity the central bank should then only allow discounting of invoices for needed real goods and services *not* loans or fripperies. Those firms still in debt then go bust and the FX losses are sheeted home to the creditors. The banking system could help with pre-pack administration and refinancing in a local currency.

By sending the losses to the FX creditor, the FX creditor may panic and call on its own central bank for assistance, which can then quell the storm by creating the FX and buying up the spare other currency - providing the liquidity necessary to unblock the system. Or alternatively the wave of bankruptcies will destroy sufficient FX loans to clear the demand/supply imbalance.

A balance of payments crisis is caused by trying to drive an airplane as though it was a car. It is bad policy - brought about by bad analysis from mainstream economists who just can't seem to handle the dynamics between the currency zones - let alone multi-national banks.

Bad analysis that is caused by them using the wrong tools to analyse a system - primitive mathematical models.

You guys need to upgrade your toolkit. I never use mathematics in analysis because it is too easy to make a mistake and it is crap for communication with the actual domain experts.

Neil Wilson said...

"If a EZ bank is financing the inventory of C-class then EUR/USD would accordingly respond and goes down to reflect the new reduced terms of trade as there is NOT as you point out 'infinite liquidity' in the fx market...."

Are there any EZ banks that don't also have a USD branch? Really?

ISTM that multinational banks exist so that the wobbles in FX don't affect them on the group balance sheet. If one side goes down, the other side goes up.

And you are going to have to describe in detail what you think the financing of the inventory is, and how it works. Because I'm really not clear what you mean by that and how that helps capital efficiency in the business.

ISTM that a multi-national company will be serviced by a multi-national bank. And that means refinancing in an alternative currency is just moving a number up in one country and down in another within the same bank. The FX exchange never gets touched.

John said...

Neil: "I never use mathematics in analysis because it is too easy to make a mistake and it is crap for communication with the actual domain experts."

What do you use then? Everything needs mathematics! Analysis of any system requires mathematics of some description.

By "domain experts" do you simply mean currency experts? Do currency experts find mathematical knowledge an irrelevance? How on earth do they make any decisions? What in hell are all those computer screens blinking in the background doing?

I don't, and can't, dispute what you're saying. If you say so, I believe you. But it just doesn't seem to add up!

Neil Wilson said...

"Neil's point here wrt "no infinite liquidity" I assume comes from his knowledge that banks are limited by capital."

It's actually the intra-day v.s inter day differences. They are two separate processes.

Intra-day everything blows out as balance sheets expand. Then everything has to settle for the end of day process to reach the end of day positions. The same process then repeats tomorrow.

Intra-day and inter-day are not the same beasts at all. So the delta T cannot really go smaller than a day.

It's like in physics where you can go only so small before you start getting quantum effects overwhelming the 'normal' stuff.

The capital regulations don't bite very quickly. Capital buffer breaches in the ECB system, for example, only cause sanctions about dividend distribution not instant resolution. So there is time to sort out capital raising. And of course capital is always available at a price in an endogenous system.

So I don't see banks being limited by capital or regulation. It's more the price at which they can do business profitably that limits anything.

And I don't see an EZ bank lending against dollars (which is essentially what an open FX position actually is) being much different from lending against barrels of oil or an inventory of Italian ice cream. They are all bank balance sheet expansions.

The difference is the ease at which the dollars can be swapped out for some other asset or currency. Although with securitised loan assets I wonder how much difference that is these days. Shuffling the bank asset pack seems to be pretty easy for all classes of bank assets.

Tom Hickey said...

From what I understand, on an intraday basis the market participants progressively converge toward balancing out at the end of the day. Perhaps Neil can say something about this gradual covering, if this understanding is correct.

Neil Wilson said...

"What do you use then?"

I use systems languages. The mathematics is embedded in the type system and structure or those languages. Lots of clever information system experts did all the ugly mathematics and embedded them in a higher level tool so I don't have to bother with it.

Computers run entirely on 0s and 1s. Unsurprisingly I don't construct computer systems using 0s and 1s because I'm a human (allegedly) not a computer. The computers translate what I use into the 0s and 1s they understand.

A good systems language is readable so that it looks much like a natural language, but is computer translatable. Some are visual in nature instead - depending upon what domain you're analysing.

IT essentially abandoned what is known as 'formal systems methodology' (trying to do everything with pure maths and logic) decades ago. In agile terms it has 'cost above value'. In the vernacular - it simply doesn't work.

"By "domain experts" do you simply mean currency experts?"

A domain expert is somebody who actually does the job you are analysing. And that means that your models have to be expressed in a form the domain expert understands if you're to work out the essence of what they are actually doing rather than what they think they are doing.

I know I'm not clever enough to do complex modelling with maths alone. I want what I'm doing constantly checked by machine translators and domain experts to make sure I'm on the right path to a solution.

John said...


I have the reverse problem. Mathematics is almost trivial, even a lot of the so-called higher abstract stuff. It's the programming that is painful. Total mental block! Just don't get it. Any of it. The programmers I know find it hilarious that what they find so easy, I find literally impossible!

If you don't mind the intrusion, what is it you use?

Apparently, Python is a lifesaver. It makes programming possible for dummies like me. That's my 2016 resolution: learn Python, and use it to model the things I know must be true but can't prove mathematically (proofs are too highly regarded anyway). I think there was a John Brockman book by a similar title.

I used Mathematica a while back. Now that is one truly amazing piece of software! The time it saves. Take any damn bastard equation, throw it into Mathematica and...with in seconds...a fu**ing answer! I'll have to add that to learning Python!

Neil: "I know I'm not clever enough to do complex modelling with maths alone. "

Two things. First, no one is clever enough, not even John von Neumann, and that's saying something. It is not humanly possible. For example, no one can accurately model the weather using mathematics alone. Although the mathematics is simple, even for many other nonlinear dynamical systems, you'll need a supercomputer to give you some useful result. End of the matter. So don't beat yourself up about it.

Second, I've come across more truly clever people here at MNE and its fellow travelling blogs and sites than I've ever met, although clever bugger Matt can be infuriating at times, and I've met many of what are considered the cleverest people in the world. These people are usually brilliant in extremely narrow and arcane areas. They don't know anything else. Almost literally nothing else.

Anyway, 2015 was a revelation. For the life of me, I can't remember how I stumbled across you. I think Steve Keen, who I inadvertently came across when I was channel surfing, mentioned your name somewhere about accounting. That led me to your blog and then MNE and then Bill Mitchell and then to a realization of how little I really knew. So don't beat yourself up over that either. You're definitely one of the cleverest and most knowledgeable people I've ever come across. In The Matrix, Morpheus has that brilliant line about having a "splinter in your mind". Well, that splinter is slowly being removed by people like you. A piece of advice, however. Add a few extra lines to your blogs to explain things for the dummies out there. You'll turn into Warren Mosler at this rate, with his insider finance speak that even insiders find impenetrable. There's a happy medium to be found between spoon feeding and Warren Mosler's hyperefficient verbal actions (I refuse to call what he does "talking" or "writing"). Actually, that happy medium is probably found in Brian Romanchuk, another terrifyingly clever bastard.

Neil Wilson said...

"Apparently, Python is a lifesaver."

I can see why you're having problems :)

Python is a very weird and quite low level language with relatively poor expressiveness. Never quite understood why people like it.

The new kid on the block is Julia which I'm having a little bit of fun with because it's got a lot of geeky advanced type tricks in it that make it very expressive. But primarily it is a Scientific computing language that tries to combine the best bits of R, MATLAB and python that can also be used to do general purpose programming.

Since it has advanced support for mathematics, but also a lot of type structure in it I'm wondering whether it could be used to bridge the communications gap.

My main concern is whether the plumbing and output side is good enough to build the simulator scaffolding architecture we need. We don't need a single computer simulation with tabular output. We need an interactive experience with multiple inputs.

And I'm not sure some of the operator twiddles are that easy to read.

In development terms I spend a lot of time with Ruby - because in Ruby it is incredible easy to create Domain Specific Languages. And they aid communication no end. For example this is executable code:

publishing agreement dated '9/20/2005'
with_author 'Joe W. Author', social('555-493-3920')
for_title 'DSLs for Dummies'

report do
calculate 'Royalties', as net_retail_sales.during(last_six_months) * 20.percent

"A piece of advice, however. Add a few extra lines to your blogs to explain things for the dummies out there."

That's what the comment section is there for and the Google Group. Ask the question. Tell me what you don't understand and I'll try and come up with a way of explaining it.

The only daft question is the one not asked. Don't be shy!

Matt Franko said...

"Are there any EZ banks that don't also have a USD branch? Really?"

Well yes I'd assume they have a US branch but what happens when the US branch loan asset gets degraded by the auto firms dropping their prices in USD terms?

The US branch loan asset gets marked down so they breach the regulated capital:asset ratio in the US branch ... then they have to respond to this to bring the ratio back into the regulated ratio...

So they have to either adjust capital or assets in response to put the ratio back to where it has to be .... I posit that the frequency response characteristic of the capital is way lower than the assets... so they have to adjust assets... reserves are assets... so they adjust reserves while the individual US branch bank's capital AND the total USD system capital remains fixed (over short time periods)....

iow if Daimler cuts the c-class auto price in USD terms then the US branch of the bank financing the autos in the US has to obtain USD reserves in response... if the bank offers to exchange EUR reserves for USD reserves, the counterparty would want MORE EUR for a USD so MORE EUR per USD means the exchange rate goes DOWN and relflects the new reduced terms of trade for Europe (Europe based C-class manufacturer accepting LESS USDs for a c-class)....