on the cost of private debt

I'm a little sensitive about the appropriation by lenders of the term "private debt" to refer to private sources of lending. If you think of debt as an asset you miss the bigger picture. First of all, debt can't even exist as an asset unless you first convince somebody to take on debt as a liability. Then too, debt as a "toxic asset" doesn't arise except as a result of debt becoming a problematic liability.

So when I googled on the cost of private debt I expected to find much on private debt as an investment opportunity, and not much on private debt as a source of macroeconomic problems. I was pleasantly surprised to find Bailing out the people: The public cost of excess private debt (12 January 2019) by Samba Mbaye, Marialuz Moreno Badia, and Kyungla Chae, at VOX. The prefacing paragraph:

Since the financial crisis researchers have extensively explored the dangers of excessive public debt, but excessive private debt has received less attention. This column documents a common form of indirect private sector bailout that goes largely unnoticed. Whenever households and firms are caught in a debt overhang and need to deleverage, governments come to the rescue through a countercyclical rise in public debt. This indirect substitution takes place even in the absence of a crisis.

Reminds me of my work on the private-to-public debt ratio.

Again, in the article:

The basic mechanism works as follows – whenever the private sector is caught in a debt overhang, the ensuing deleveraging process weighs on activity, thus prompting the government’s response through higher borrowing to support the economy. We show that this is a recurrent pattern through history, where excess private debt systematically leads to higher public debt.

One thing I have trouble with is a definition they provide:

We define a private deleveraging spell as the distance from a peak to the ensuing trough in the private debt-to-GDP ratio...

Debt-to-GDP? Really?? GDP is highly unstable, and probably affects the ratio more than debt does. I'm reminded of something Scott Sumner said. In comments on Debt surges don’t cause recessions, Vivian Darkbloom "ran across another [debt-to-GDP] chart", one that goes "back to 1920". Vivian provided a link to an old Steve Keen site but the link no longer works, so I can't show the graph.

But based on Scott Sumner's response, I'm pretty sure the outstanding feature of that graph was the massive spike associated with the Great Depression:

Vivian, Thanks for that graph. Here’s my prediction: 99% of people will misread that graph. Most will think it shows a debt bubble before the Great Depression. In fact it shows there was no debt bubble before the Depression. Rather the debt ratio rose DURING the Depression, but only because the denominator (NGDP) fell.

Agreed. But if you were around in 1910 or 1920, the Roaring 20s probably looked very much like a debt bubble. It's not that "there was no debt bubble before the Depression", but that in hindsight the bubble is insignificant when compared to the debt spike of 1929-1933.

I also found this graph (which leaves no doubt about when the debt spike started) from Global Finance:

The graph is great! The vertical line just before 1930 shows the start of the Great Depression, and the graph shows what now looks like a slight increase in debt before the Depression, and the big spike during it as GDP fell.

People didn't suddenly start borrowing a lot more after the Depression hit. No. But GDP fell like crazy, and that pushed the debt-to-GDP ratio up. This graph shows it well. Unfortunately the Global Finance link no longer shows this graph.

Fortunately, the Financial Times still shows it.

I wouldn't want to use debt-to-GDP to show changes in debt. Some time back I compared growth rates for GDP and debt and wrote

They're all GDP, the lows. GDP falls faster than debt. Debt does not fall as quickly as GDP. That's one thing that makes the debt/GDP ratio go up...

and I gave an example: "like during the Great Depression".

If GDP is falling faster than debt, the debt-to-GDP ratio can only go up. I don't know how they worked it out in the PDF. (I didn't read it yet.) But they did come up with results I can accept.


After defining a private deleveraging spell, they move on and say

To get a firmer grasp of private debt dynamics, we break down changes in the private debt-to-GDP ratio into the impact of the interest rate-growth differential – which we refer to as ‘macro-related changes’ – and the remaining component – which we call ‘leverage flows’.

The "interest rate-growth differential" sounds to me like Piketty's r>g. Remember Vollrath's summary:

"Piketty says that if r>g, where r is the return to capital, and g is the growth rate of aggregate GDP, then wealth will become more and more concentrated."

Note that as the economy grows ever more financialized, the return to capital must be measured less in terms of profit and more in terms of interest.

So they look at debt both in terms of relative growth and in terms of leverage flows. They find that

while the debt ratio is rising at an increasingly faster pace before the deleveraging episode, leverage flows (blue bars) are growing at an increasingly slower pace...

If "leverage flows" means what I think it means (changes in debt levels) then it seems they are saying that a high debt level causes the economy to slow, increasing the debt-to-GDP ratio and creating the same sort of confusion among 99% of economists that Sumner predicted among 99% of people.

In a way, that's what the VOX article is all about.