Friday, April 28, 2017

Issues with Figure 1 in Bezemer and Hudson's Finance Is Not the Economy

In Finance Is Not the Economy, Dirk Bezemer and Michael Hudson write:

These correlations suggest a one-on-one ratio between bank credit and the non-financial sector’s economic activity (Figure 1). Growth in credit to the real sector paralleled growth in nominal U.S. GDP from the 1950s to the mid-1980s — that is, until financialization became pervasive. Allowing for technical problems of definitions and measurement, growth of bank credit to the real sector and nominal GDP growth moved almost one on one, until financial liberalization gathered steam in the early 1980s.

Figure 1 shows how, after the mid-1980s, the real sector was borrowing structurally more than its income — a remarkable trend noted by few.

Here is their Figure 1:

Figure 1 from the Article by Bezemer and Hudson
When I presented this graph before, in If the growth of debt was one-for-one with the growth of GDP before the 1980s, the line would be flat, I was looking at the dashed line and ignoring the jiggy lines. This time, just the opposite: Look at the jiggy lines and ignore the dashed line.

According to Bezemer and Hudson, the graph "shows how, after the mid-1980s, the real sector was borrowing structurally more than its income". I don't know what "structurally more" means; I think it just means "more": The real sector was borrowing more than its income, after the mid-1980s. That's what they are saying, I think. But I find it hard to see that on the graph.

If what they are saying is true, then the bold jiggy line ("nominal credit") should run higher than the faint jiggy line ("nominal GDP") after the mid-1980s. And if the real sector was borrowing "one for one" with income before the mid-1980s, the two lines should run together. Are these conditions met?

I dunno. The lines are too jiggy to tell. The bold and faint lines seem to run more or less together until about 1985, and after that the bold is clearly higher, sometimes. But I am looking at the graph, looking for what they told me I should see in it, and I'm trying to see it. That's not objective evaluation. It's hokum.

So I did what I always do: I tried to duplicate their graph. It's a little awkward because I don't know exactly what data they are using, nor the value of their smoothing constant. I took non-financial corporate business debt and non-financial non-corporate business debt, and added them together to get a number for what they call "nominal credit to nonfinancal business". Then I found "nominal GDP" and put that on the graph too. And I showed both data sets as "YoY growth". In other words, I did my best to make my graph look like theirs (without the smoothing). Here's what I got:

Graph #2: Trying to Duplicate Bezemer and Hudson's Figure 1
Perhaps not the best of matches. The two graphs are somewhat similar. It does real harm, not knowing exactly what data they used. I can only discuss what my graph shows, not what their graph shows.

The blue line is credit growth; the red line is GDP growth. In the latter half of the graph the blue line is definitely higher than the red... sometimes. Sometimes definitely lower. And sometimes, quite the same. But these are the years for which Bezemer and Hudson say "the real sector was borrowing structurally more" than GDP. Their words put the blue line reliably above the red. My graph doesn't show it.

In the early years, where Bezemer and Hudson say we should see a "one-on-one" correspondence, the blue line runs generally higher than the red. Specifically, blue is higher for essentially all of the 1960s and the first half of the 1970s -- and for most of the 1950s as well. Credit growth was mostly higher than GDP growth in the early years, not one-for-one as Bezemer and Hudson say.

But maybe it will be easier to see if we look at the data a different way. The next graph shows the same source data. But instead of showing the "YoY growth" rates, it just shows dollars, billions of dollars. And I let the graph do the comparison for us, showing the credit number as a percent of the GDP number.

So for example, when credit is growing one-on-one with GDP, the blue line will run flat. (If the graph starts out with credit at 30% of GDP, the line will stay close to 30% as long as the "one-on-one" holds good.) And when credit is growing faster than GDP, the blue line will go up. Here is the graph:

Graph #3: Non-Financial Business Debt as a Percent of GDP
Non-financial business debt starts out at around 30% of GDP, and climbs 20 percentage points to around 50% by the mid-1970s. It is pretty much a straight-line increase all the while. After the mid-1970s it runs in fits and starts -- still with a trend of increase, but slower than before, and with a lot more variation in the path of the data.

The increase in debt, relative to GDP, is more rapid and more persistent before the mid-1970s than after. There is a 20 percentage point increase in about 25 years early on (1951-1975), and a bit less than a 20 percentage point increase in the 40 years thereafter.

The trends indicate that non-financial business debt was growing faster than GDP for the whole period shown on the graph, but more rapidly before 1975 than after. And you know, maybe that's what we should expect to see if the economy slowed in the mid-1970s and never fully recovered.

Bezemer and Hudson claim that non-financial business debt grew no more rapidly than GDP during the 1950-1985 period, and much more rapidly thereafter. I don't see it.

No comments: