Saturday, February 18, 2012

JW Mason (2): What Context for Debt?

My continuing saga of reading JW Mason's Fisher Dynamics in Household Debt (PDF, 31 pages).

Part 2 of Mason's paper opens thus:

Traditionally, economists have attributed only a minor role to leverage (the ratio of economic units' gross liabilities to some some measure of their income or net worth) in understanding macroeconomic outcomes

A few days back I had some trouble understanding the interest coverage ratio, which looks at some portion of income as a multiple of total interest owed. Leverage is the leviathan relative of that, looking at total debt owed as a multiple of some measure of income.

Graph #1: Debt per GDP
Let me again make the point that looking at debt relative to income -- widely used by lenders in micro-economic encounters, even if it plays only a "minor role" in "macro" analysis -- looking at debt relative to income is like inbreeding: Every dollar of debt becomes somebody's income. Income is a false support for debt. A sudden urge to deleverage will undermine income (as we well know). And income can collapse faster than debt can be repaid. This is the truth that Irving Fisher spoke, and Steve Keen speaks today.

EDIT: Changed "Debt is a false support for income" to "Income is a false support for debt." Oops! 1 August 2014

Graph #2: Debt per Dollar
I prefer to look at debt relative to the quantity of unencumbered money: relative to M1 or Base. Relative to the money that must be used to repay debt. The money for which "the quantity of money" is reduced when debt is repaid. Spending money.

But Mason prefers to look at the same ratio that everybody else looks at, debt relative to income. Mason prefers the same guide used by lenders and policymakers in the run-up to the recent financial crisis. Debt relative to income:

What constitutes a sustainable ratio of government debt to GDP has long been a central concern for public finance; more recently the behavior of, and limits to, the debt-income (or debt-net worth) ratios of other economic units have become salient questions as well.

And again:

If units' assets are not reliable sources for either funding or market liquidity, then the capacity to service debt out of current income becomes paramount.

Sure. But total income is equal to the sum of its micro-economic parts. Aggregate money measures, on the other hand, are macro measures. Mason and the others prefer a micro-economic analysis of the macro-economic problem.

// See also: Context


JW Mason said...

Hey Arthur,

Can you spell out your criticism here? Arjun and I had quite a few discussions about the best way of measuring leverage, and I'm certainly open to the idea that we ended up with the wrong answer. But it's not at all clear to me why debt/M1 would be a better measure. I think most people -- including me -- would say that today there are so many "money-like" assets no included in M1, and such easy substitution between M1 and non-M1 assets, that the measure has lost more or less all economic meaning.

More broadly, I guess we're coming from a kind of Minskyan perspective where the macroeconomic effects of debt stem from its magnitude relative to households' ability to make their contracted payments, i.e. the scale of adjustment of real activity potentially required to meet financial obligations. So it seems pretty clear that the denominator in the leverage ratio should be some measure of "capacity to pay." ow I am very open to the idea in principle this should include the stock of sufficiently liquid assets as well as the flow of money income -- I think we even acknowledge this might be a better measure in a footnote -- but, leaving aside all the conceptual issues with assessing the degree of liquidity of various assets, you'd have to be talking about assets *of the household sector*. The extent to which financial institutions are selling assets to the Fed for additional reserves -- the main factor in recent movements in M1 -- can't have any bearing on households capacity to service their dot, I don't think.

The Arthurian said...

JW, hello.

"Can you spell out your criticism here?"

I'm not an economist and my terminology is not the best, so please bear with me. I still think (as I was taught in the 1970s) that we have two objectives for the economy: growth, and price stability.

If growth is a goal, and GDP is how we measure growth, then GDP is a result of policy -- hopefully the result we want, but in any case a result.

Monetary policy is a tool we use to achieve our goals. This means that the quantity of money -- be it Base or M1 or MZM or TCMDO debt -- is by design of policy. (As Keynes said, we can draw the line between money and debt at any convenient point.) (I include in monetary policy not only the actions of the Federal Reserve, but also the things Congress does to encourage both credit use and credit supply.)

When I look at a graph of "Debt divided by GDP" what I see is a policy divided by a result. Now, the Fed may have control over the quantity of money, but it certainly does not have control over GDP. So this graph is half useless from a policy-maker's perspective. More than half useless.

My "Debt per Dollar" graph, on the other hand, compares two elements of policy, and shows an imbalance between them. It shows where policymakers went wrong: They thought they could allow credit-in-use to increase faster than money-in-circulation on a continuing basis. They were wrong.

My first look at DPD, long ago, considered what is shown as the blue part on Graph #2 above, the Historical Statistics (Bicentennial Edition) part. If you look only at that part of the graph, it's pretty easy to see that by 1970 we were already perilously close to going beyond the monetary imbalance that brought on the Great Depression (or, that accompanied the Great Depression, if you like).

Looking at the full picture of Graph #2 today, the significance of that once-mighty peak is lost, largely as a result of all the financial innovation that has happened since 1970. But we can see now the result of all this innovation.

JW, the peaks on the "Debt per GDP" graph show problems, and little more. The peaks (and also the relative levels) on the "Debt per Dollar" graph are related to the performance of GDP -- depressions at peaks, recoveries on downslopes, and "golden" and "miracle" economic growth immediately following the recoveries. That little wiggle in the DPD from 1990 to 1993 was followed immediately by several years of improved productivity and also, as you will remember, by a briefly balanced Federal budget.

One man's use of credit is another man's income. If I borrow a dollar and buy your widget, and you use the money to pay down your debt, the economy as a whole has not seen a reduction of debt. If you *don't* use it to pay down your debt, the economy has *not* seen a reduction of debt-to-income.

If a helicopter drops money on us and we use it to pay down our debt, the debt and the money both cease to exist. If Bernanke had done QE a little differently and PAID OFF bad debts for people, rather than BUYING them from creditors, we would have eliminated some of the liability that still crushes the economy today.

I have held the number of links to a minimum in these remarks so as not to be overbearing. I hope it worked :)

The Arthurian said...

JW: "I think most people -- including me -- would say that today there are so many 'money-like' assets not included in M1, and such easy substitution between M1 and non-M1 assets, that the measure has lost more or less all economic meaning."

The problem has gone on for a very long time. So it is easy to think of parts of it as "normal". What you have described here is a result of the financial innovation that created the problem in the first place.

Today, we use credit for money. As a result, financial costs are a significant cost, along with wages and profits. Wages and profits are necessary major costs, buy finance is not -- Finance is necessary in a minor role only.

The world as it was before the crisis seemed "natural" to most people, and most people seem to think that a major role for finance is not a problem. That is why they didn't see the crisis coming.

If you like, replace M1 in my Debt-per-Dollar calculations with Base money. You'll see the same pattern, even more exaggerated. And every time the line goes up, it means financial costs are increasing in our economy. These costs have been partially offset for 30 years by a secular reduction of interest rates. Now that this policy is no longer available because we are at "the zero bound", the chickens come home to roost and the problem of excessive debt and its costs have become more evident.


I prefer to point the finger not at financial innovation, but at the policies that allowed and encouraged it -- and truly, at the mindset that allowed and encouraged those policies.