Friday, November 2, 2012

Slopes


Yesterday I put up the "hover craft" version of my "blueberries" graph. In the original Blueberries, there was a bunch of lines sloping down from left to right. In the hover craft version we got to see many little pieces of that, individually. But you were left... no, I was left with no coherent overview. Yeah yeah, in any business cycle or fragment thereof, it was obvious that as the debt/GDP ratio increased, the economy's performance (as measured by Real GDP growth) declined.

That's Total Debt slash GDP, by the way. Not just government debt.

So we were left with a general indication that a growing debt burden is not so great for growth. And, indeed, this seemed to be true repeatedly, in business cycle after business cycle and in fragment after fragment. But it was a fragmented indication.

Each separate graph used for the hover craft display included a slope value which described the general trend of the fragment displayed. But I was clever and planned ahead. I had the Excel file (or, really, the VBA code contained in the Excel file) write out the start- and end-date of each fragment and the slope value for that fragment, write it out to a CSV file, a Comma Separated Values file.

For today's post what I did was open up that CSV file in Excel and make a graph from it. I made an extra column in there for the dates. And in the new column I used the average value of the start- and end-dates for the slope value. (That worked easily in Excel by the way, taking the average of two date values, but it didn't work at all in the Open Office spreadsheet.)

Oh, also the first graph (4th Quarter 1950 thru 2nd Quarter 1953) didn't have a slope value -- I think because some of the quarterly data was missing for the first year or two. So I skipped that fragment and started with the 2nd Quarter 1953 thru 3rd Quarter 1955 fragment. That one gave me an "average" date some time in 1954. So that's where the graph starts.

I love this shit.

So anyway, when you take all the slope values from all the fragments and plot those values as a time series, this is what you get:


First glance, the trend is going up. Oh but wait. Those are all minuses on the vertical axis, all negative values. The slope is getting less negative over time.

What does that mean? It means at the start the fragments showed a big difference in Real GDP  growth for a given change in the burden of debt. But by the end there was NOT a big difference in Real GDP growth, for a given change in the burden of debt. And the general trend from start to finish was that there was less and less of a difference.

So we see that the fragments show for the short term that as the debt/GDP ratio increased, the economy's performance declined. And we see that for the big picture, the whole half-century and more, again the economy's performance declined as the debt/GDP ratio increased.

Over all that time by the way the burden of debt was growing. And it grew like crazy for a while before the financial crisis, though the economy did not.

And then -- surprise! -- there was the crisis.


I would say that probably this graph does not "prove" that the burden of debt is the thing that hinders economic growth. But I would also say it is good solid evidence, and it is time economists STOP IGNORING THE PROBLEMS CREATED BY EXCESSIVE DEBT.

That's what I would say, you know, if I had a blog.

No comments: