Monday, November 16, 2009

Models vs. Accounting

I concluded a very enjoyable exchange with Nick Rowe from Worthwhile Canadian Initiative. You can check it out here. The crux of the discussion was the failure of monetary policy: we're at ZIRP and yet the economy keeps getting worse. How can we reduce interest rates below zero? What is the mechanism by which interest rates impact the real economy, and what is going wrong with that mechanism, &c.?

This mechanism does not exist, but that did not stop Nick from coming up with this--this!--to explain the framing (or "social construction") of monetarism:
Phillips Curve: p = 0.25(y-y*) + 0.5p(t-1) + 0.5E(p)

IS curve: y-y* = n-r

p = 0.25(n-r) + 0.5p(t-1) + 0.5E(p)




1. P=1 (or p=0 and E(p)=0).

2. B=1/(1-n) (or i=n)

3. R=1/(1-n) (or r=n)

I ask a very simple question: if the private sector wants to increase its net financial assets, can that be achieved by interest rates or can that only be done via fiscal policy (aka deficit spending)? The point of that question was for Nick to realize what monetary policy can and cannot do, and then we could begin a discussion of what cause of this current crises: ie. is the private sector overly leveraged? This is all d'uh stuff to me, and to most people I think, but d'uh never got you a PhD in Economics. Anyway, this is what Nick came back with:

Yep. In a closed economy:

S-I = G-T

That's just an accounting identity. It's a way of using words in a consistent way; it doesn't tell us how the world works.
Isn't that just fantastic?! Someone who spewed "p = 0.25(y-y*) + 0.5p(t-1) + 0.5E(p); IS curve: y-y* = n-r; p = 0.25(n-r) + 0.5p(t-1) + 0.5E(p); B=1/(1+i); R=P(1+E(p))/(1+r); (1+i)=(1+r)(1+E(p))" dismissing accounting because "it's a way of using words in a consistent way; it doesn't tell us how the world works"?! I want everyone to be totally clear on this point -- someone who is trying to understand finance is dismissing accounting. Finance is accounting. If you cannot track the accounting consequences of a financial act, then there was no financial act. I particularly love all the coefficients in Nick's model, 0.25, 0.5, 0.5 etc. Is he sure they are not 0.26, 0.49, and 0.51? Inquiring minds want to see the regressions.

I am more sympathetic towards economic models, and more suspicious towards accounting than most -- I learned both at U Chicago. But the disdain that economists hold accounting in, and the inability to see beyond their models blinds them to understanding the world around them. Dismissing accounting with "It's a way of using words in a consistent way; it doesn't tell us how the world works" by someone who models shows profound... primitiveness. Not sure if there's a better word.

Also, Nick got the pop quiz wrong:
If desired savings increases (i.e. desired consumption falls), then the rate of interest falls until desired savings falls back to where it started (or desired investment increases, or some mixture of the two).
Wrong! Lower interest makes it even more difficult to reach the desired savings level, so lower rates make things better, not worse. Moreover it is impossible to increase net assets within a sector, the solution can only lie out of the sector.


Blogger Nick Rowe said...

Hi Winterspeak:

Certainly (or almost certainly) my parameters (0.25, 0.5, etc) will be wrong. The best I can hope for is that they are "approximately" right. That's what makes my model a model: it says something about how the world works, and what it says cannot be right unless it could be wrong.

Assume a closed economy, with no government, and no investment. Then we can describe the national income accounting identity as c=y. Or equivalently, if we define s as y-c, as s=0.

If we are talking about actual c, y, and s, then those accounting identities must be true. They are true by definition (provided we are using the words c,y, and s consistently.

But if we are talking about desired c and s, they are equations, not identities, and are true only in equilibrium.

Suppose we start in equilibrium, then desired c falls (i.e. desired s rises). We are now out of equilibrium. It is logically impossible for people to do what they in aggregate desire to do. Something has to change to get us back to equilibrium.

In a "classical" model, the real rate of interest will fall until desired c increases enough to once again equal y (i.e. desired s falls back to 0).

(In a keynesian model it is y that falls until desired s=0).

I must do a post on this on WCI. It's a good conversation we're having.

5:21 AM  
Blogger STF said...


The distinction is between the BEHAVIORAL model of, say, C, and the ACCOUNTING IMPLICATIONS for individual and aggregate financial statements of this change in C.

Modelers must get the accounting right or they won't get the implications right.

Scott Fullwiler

11:43 AM  
Blogger Nick Rowe said...

Scott: I can agree with that. If you get the accounting right, you can still get the behaviour right or wrong. But if you get the accounting wrong, the behaviour will make no sense whatsoever.

1:53 PM  
Blogger STF said...

Hi Nick . . . We agree!!!!

2:28 PM  
Anonymous Anonymous said...

I posted over at Nick's an illustration of how the basic accounting identities are equivalent to basic macro GDP equation:

C + I + G + X - M = Y

Y = C + I (eliminate government & external)

In accounting, debit accounts = credit accounts:

Assets + Expenses = Liabilities + Equity + Revenue

A + E = L + Eq + R

Rearrange to:

R = E + (A - L - Eq)

Revenue = Expenses + net(Assets - Liabilities - Equity)

Which is basically the same as:

Income = Consumption + Investment

So, its nice to know accounting is consistent with macro - in other words accounting DOES describe how the world works - at least as well as macro.

I certainly hope that economics strives to use words in a consistent manner - otherwise I would have doubt as to the consistency of any equations developed.

What I think the issue is with accounting and economics is that classical economics takes a "gods" eye view of things - a position that no economic actor has.

Accounting is more point-of-view oriented - one person's debit is another's credit. This is why economists get frustrated with accounting, they are required to state from whose position they are speaking.

You bring up a great point - without government and external - the economy is zero sum. You can't move beyond the posited starting point (which is always equilibrium).

I don't know how you can even discuss raising interest rates (or even HAVING interest rates) without the government and external sectors - all of which requires a theory of money.

The real economy can create real assets without any money. You and I can decide to collaborate and create an asset (say a song) that attracts other people. We can even create a barter exchange. But we cannot monetize it without money. We CAN create net assets (real), but we cannot create net financial assets without money.

Without a coherent theory of money, economics is playing around equations which are only valid under laboratory conditions - but don't function well in the real world.

Accounting at least brings human behavior in the picture - that economic activity is based on a concept of equity - of equal exchange - and that recording these transactions is absolutely required for a stable economic system.

6:04 PM  

Post a Comment

Subscribe to Post Comments [Atom]

<< Home