Friday, March 27, 2015

Ramanan — Respect For Identities

Stock flow consistency versus econometrics.

David Glasner doesn't seem to accept that accounting identities are tautologies that determine what is logically necessary in that system of expression. They are true by definition (rule). That which contradicts a tautology by violating a definition (rule) is logically impossible. 

Tautology and contradiction are formal boundary conditions of expression — unless one is writing nonsense like Lewis Carroll, who happened to be a logician and mathematician in addition to being a very clever writer. He is best known for writing nonsense verse — Alice's Adventures in Wonderland and Through the Looking Glass, which continue to delight children and adults alike.

Mathematics is the language of science, and accounting is the language of business. In so far as economic is a study of production, distribution and consumption, in a modern economy this involves business and is conducted in terms of business rules, institutional arrangements in which double entry bookkeeping and accounting procedure (rules) are foundational.

Does anyone actually question this? How could accounting identities that follow from the institutional arrangements (rules) not be foundational as constraints (boundary conditions)? Otherwise, it's possible to just make stuff up.

Let's be clear. In a modern monetary economy, accounting identities are the starting point of doing economics. Period. Anyone who doesn't get this and hew to it is not playing by the rules, which amounts to talking nonsense.

The Case for Concerted Action
Respect For Identities


Brian Romanchuk said...

I looked at Glasner's articles, and asked him about it. I think the situation is that he wants to use a different definition of savings and investment. I am thinking about other things, and so I did not have the time to convert his terminology into a form that I would use. But as long as you are careful in stating your definitions, you can define terms as you wish within mathematics, up to the point of creating a new usage of an existing term. Therefore, what he is doing appears defensible.

The thing is that you would normally only start defining your own terminology if there is an advantage of doing so. I personally do not see what advantages he gains.

Ramanan said...


Good points Brian.

However if you read carefully, he insists that even the traditionally defined identities are problematic:

Glasner: "First, if we define savings and investment (or income and expenditure) to be identically equal, we can’t solve, either algebraically or graphically, the system of equations describing the model for a unique equilibrium."

SFC models are proofs that one can construct models perfectly consistent with the system of national accounts and flow of funds.

So although superficially he might sound defensible, when you look in detail, he's not.

How else does one explain his quote? He is simply saying that there cannot be a right model consistent with standard definitions.

Brian Romanchuk said...

He may be thinking about the need to drop one constraint in order to avoid over-determining the system. But yes, I had a hard time following his logic. He is referring to older models that I never bothered looking at carefully.

Ramanan said...


Good point. Hadn't thought of it.

Perhaps that thing has led him to conclude something different than just dropping one constraint.

I liked this comment by Jamie. I think he has followed all the posts and gone to the root of the matter.

Tom Hickey said...

I like Ramanan's.

To be even clearer, your posts simply read:

“Accounting is wrong”.

But that’s not the end. You attempt to hedge your posts by also claiming “accounting is right”, so that nobody can accuse you of claiming “accounting is wrong”.

But all your posts simply read: “accounting is wrong”

You have to start with accounting symbols and rules, just like in logic and math you have to start with symbols, (formation rules) and then add operations transformation rules and transformation rules.

When working within an established system, and accounting is based on clearly defined institutional arrangements that hold in law, you can't just make stuff up

The whole series of posts seems to be based on the ongoing debate in economics over the meaning of terms like saving and investment, which are based on accounting records of flows that show up on accounting reports

Problems of ambiguity arise when the same term is used analogously rather than univocally. This is one reason that formal expression is used, so that every term is defined precisely and used the same way throughout.

Saving is generally used to mean the household income residual after consumption. Firms don't "save" in this sense. They "retain earnings." One say loosely that firms' retained earnings are (like) saving. But that is using the term analogously.

But it is also possible to a consolidate sector, e.g., domestic private sector relative to other sectors, government, external sector, or even aggregate further in terms of government and non-government, and talk about the "saving" and "dissaving" of the sectors.

No problem. But one has to be clear about the use of terminology and stick to it throughout.

The advantage of keeping the accounting in mind is so that one doesn't make errors or fall into stock-flow inconsistency. This is exactly what double entry is for. It also prevents cooking the books and not getting caught in an audit. Sometimes I think that some economists avoid accounting in order to make the point they want to make even though it would not pass an audit.

I got tired reading Glasner's posts on this when I concluded (with Ramanan) that he was claiming the accounting is wrong.

NeilW said...

It's vitally important to remember that accounting is not facts. It is opinion.

The national accounts are a consistent *opinion* from a particular point of view (largely the 'reporting currency').

However that constrains you to that point of view which may or may not be helpful.

For example is it helpful to have non-convertible currency elements automatically converted into the reporting currency? I would suggest not.

Is it helpful to fail to classify human capital and treat training, etc as 'consumption'. I would suggest not.

We have to be careful with the national accounts. They are simply a snapshot from the point of view of the borders of a nation and do not capture the time dynamics or the currency area interactions (intra or inter) very well at all.

Tom Hickey said...

It's vitally important to remember that accounting is not facts. It is opinion.

Right. While accounting begins with "facts" through recording transactions in journals ("bookkeeping"), accounting is more than bookkeeping and involves decisions about application and interpretation of accounting rules and procedure. Reports are already abstractions and aggregating reports into national accounts involves estimation rather than actually adding up individual reports that are traceable to journal entries as "facts."

But using accounting as a basis for doing economics is probably more accurate and faithful to the facts than microfoundations in economic theory based on theoretical assumptions.