Wednesday, March 7, 2012

This is the graph I had in mind

UPDATE: In a comment, Jim points out that my change-in-debt numbers are quarterly, not annual. So all the changes I think of as annual changes in the post are actually quarterly changes. The annual changes are about 4 times bigger. Wow.

Two new graphs at bottom of post.

My Graph #1 from Monday shows "change in debt" compared to "change in the size of the economy". It shows that, as a rule, debt increases more than GDP does. Not really a surprise, that.

But if we do more than just the math, if we mix in a little economics -- if we admit we think credit use is good for growth -- then we have a problem. For debt is evidence of credit use, and Monday's graph says credit use is not good for growth. It says that credit use has become less and less effective in generating economic growth.

But that wasn't the graph I wanted. What I wanted to look at was "change in debt" compared to "size of the economy". This graph:

Graph #1: Quarterly Change-in-Debt compared to the Size of the Economy

Here's an interactive of it:

Source: FRED data, my Google Docs spreadsheet

Starting from an average of around 2 percent -- meaning that in the early 1950s, the annual quarterly increase in debt was about 2% of the size of the economy -- the trend sweeps up to a peak near 7.5% in the mid-1980s, drops to around 3% by 1993, and rises again to near 10% just before we broke the economy in 2008.
That's total US debt, not just government debt.
Three percent -- that's about how much GDP grows in a pretty good year.
3% GDP growth in a good year... 3% debt growth in a slow quarter!
The graph shows a general trend of increase, with two major interruptions. The more significant interruption follows the recent crisis, an extremely rapid and deep decline. Nothing else on the graph compares.

But the more interesting interruption, I think, is the 1986-1993 decline. It is a large drop, easily half the size of the recent one, but it took near a decade to work itself out. So there was no sudden and severe crisis to deal with. Only a relatively mild recession in 1990.

This unusual decline actually shifted the whole trend line downward. Thereafter the uptrend resumed, but from a lower starting point.

This unusual decline is something we have looked at before. The decline relieved pressure and created some slack in the financial system. This unusual decline in my view is the key change that created the opportunity for rapid growth of debt to resume in the latter 1990s.

That rapid debt growth -- unwise as it may have been, in hindsight -- opened the door for the rapid economic growth of the 1995-2000 years, growth that economist Robert J. Gordon has called a macroeconomic miracle.

Unfortunately, the miracle growth soon ended, but the rise in debt growth did not. Debt growth continued right up to the moment when everything fell apart, when jets of debt crashed the World Trade Center of our economy and turned it to dust and rubble.

Erectile Function

Today's graph shows several things. It shows that what goes up, must come down. (But I don't want to focus on the crisis just now.) It shows the unusual 1986-1993 decline that has such a prominent place in my argument (but that also is for another day). And it shows a general tendency of increase. This is my focus today.

It is the 'coming down' that gets everyone's attention, because it creates problems for people. But for the economy, for the system we rely on without recognition (except in the 'coming down' moments)... for our economy the coming down is a solution.

For the economy, the 'going up' is the problem. The increasing monetary imbalances are the problem. The increase of debt, relative to GDP and relative to income and relative to circulating money is the problem.

Today's graph shows that our new uses of debt tend to increase with time. Why? It's what we do. But I say policy is to blame. Passively, policy let it happen. Actively, policy encouraged it. We have plenty of policy devices intended to provide more credit and to encourage the use of credit. Why? Because policymakers think we need credit for growth.

But we have no policy that encourages the repayment of debt, which would free up credit so it could be borrowed again, put newly to use, and help to give us the growth we're looking for.

We have many policies that engorge the supply of credit and many policies that arouse the demand for credit, but no policies that provide protection by encouraging the paydown of accumulating debt.

This is the reason the graph shows uptrend. Policy makes it so.

Like a Black and White Cookie

Monday's graphs show a decline in "debt productivity". The graphs show that it took more and more extra debt, as the years went by, to get an extra dollar of output.

Today's graph shows that we've been trying really hard to get that extra dollar of output. We've been pushing debt up and pulling debt up and levering debt up every which way we can, just to get more output. Fools that we are, we think accumulated debt has the same effect as new uses of credit, while just the opposite is true.

Our grand plan to expand debt has been undermined by the declining effectiveness of credit-use as a tool for boosting economic growth. In the end, we have achieved only a massive expansion of debt and the worse economic growth in a lifetime.

Isn't it time to change the way we think about debt? For credit use is evidently not so good for growth. Not any more. And the reason is simple: There are two sides to debt.

There is the borrowing, the new use of credit, the spending, and the economic boost we get from it. The bright side.

And there is the dark side: the obligation, the accumulated debt, the payback, the reduction of boost that must come eventually. The residual effect.

Despite all the inducements and enhancements and encouragements of policy to get us borrowing more today, we do not borrow more. They pushed us beyond our limits, and we refuse. So the economy is sluggish.

It will remain sluggish until debt is no longer excessive.


Graph #N, Quarterly (blue) and Yearly (red) change in Debt, relative to GDP

Wow. And (just to see it) I compare yearly to quarterly, dividing the yearly by four:

Graph #N: Testing for Similarity, I divided the Yearly by Four


jim said...

Hi Art
Just a clarification on your graph of change in debt/GDP. I believe the
chart shows quarterly changes. This means that the annual change in debt in 2007 was somewhere round 30%-35% of GDP on an annual basis.


Jazzbumpa said...

We have plenty of policy devices intended to provide more credit and to encourage the use of credit. Why? Because policymakers think we need credit for growth.

I'm not sure it's that straight forward. What are the policy specifics? I look at reductions in top marginal tax rates, and deregulation. What do you have in mind?

Meanwhile, Look at finance sector debt as a function of total debt.

An exponential increase until this century, then flat into the crash.

That has to mean something. Is 32% as high as it can get? Did a decade at that level sap the economy? Wages stagnated and GDP growth was anemic.

Imagine if total debt were reduced by 20 or 25%, and all of that came from the finance sector. Wouldn't that ease downward pressure on the productive sectors?

If a bloated finance sector isn't the ultimate root cause, then what is?


The Arthurian said...

Wow. Good eyes, Jim!

Right you are. My first reaction was that you cannot be right, the numbers cannot be that big. But I looked at the annual numbers. And you are absolutely right.

$24 billion increase from 1950 to 1951: 7% of 1951 GDP... 8% of 1950 GDP!

$4421 billion increase from 2006 to 2007: 31.5% of 2007 GDP... 33% of 2006 GDP.

The 1985-86 annual change was more than 26% of the latter year's GDP... By 1990-91 this fell to just under 12%... By 1997-98 it was up over 21%...

Holy crap. Thanks. Jim.

I can't concentrate on what you are saying right now. Jim still has me flabbergasted!