Credit impulse vs automatic x-leveraging

In this post, I want to argue briefly that Jason Benderly’s automatic deleveraging / releveraging framework is superior to Deutsche Banks’s notion of the credit impulse, even though the arithmetic involved in each concept is similar.

I won’t get into great detail here, because that would be tedious and because I don’t want to steal Benderly’s intellectual property.  My points here might be interesting to those of you who already have access to these approaches and wonder why they are sometimes deployed to support different macro outlooks.

For the rest of you, if this interests, I suggest you get in touch with Jason at Applied Global Macro Research.  His deleveraging thesis was very helpful back when the bears were worried about deleveraging, before — for example — Bridgewater woke up, smelled the coffee and retroactively called it “beautiful.”  (Oops would have been more honest, especially from a firm favoring radical transparency.)

Incidentally, the timing of this post reflects JPM’s calendar-Q2 results and the contemporaneous upturn recently of bank lending growth.  It is not that there has been some meaningful inflection in the economy itself.

Ok, on to my simple point.

The main contribution of DB’s credit impulse perspective is to look at the right derivative, the one that correctly relates credit to aggregate demand growth.   The deleveraging bears wanted to look at the ratio of total credit outstanding to GDP, note that the line was going down, and draw dire implications.  DB (and perhaps others) corrected that wrong intuition with its credit impulse.

They pointed out (and still do) that the flow of aggregate demand supported by credit is determined by the flow of credit itself.  If that flow quickens, then the contribution to growth should quicken and — if you think credit determines, which I don’t — nominal demand growth should speed up.

Closely related, if the flow starts off extremely depressed, then its quickening might leave the ratio of credit to nominal demand on a declining trend, and yet the contribution of the credit impulse might still be positive.

This ain’t rocket science, but doing the arithmetic here correctly can steer you away from some pretty big mistakes. The most costly mistake was the notion that deleveraging meant the rate of growth of credit-financed spending would be low.

One problem with the credit impulse perspective, though, is that it is too often presented as causal and leading.  For example, if the credit pulse is positive, then we should be bullish the economy and vice versa. That gets the causation wrong, IMV, and leaves some useful information on the table, which Benderly’s automatic deleveraging / releveraging perspective picks up.

For Benderly’s approach, let’s just focus on the deleveraging idea, which was probably the more helpful, although both were helpful. And then let’s just say that releveraging is the converse.

As mentioned, automatic deleveraging starts with a bit of arithmetic very close to the credit impulse approach. But it is fussier about relating credit growth to the components of aggregate demand that are actually credit-financed (basically all forms of investment). And more to the point, it gets causation right, by my lights anyways.

If the flow of spending in credit-intensive components of aggregate demand is currently at a depressed level, then we can make two forecasts. First, we can project, that the ratio of credit outstanding to nominal demand will continue to fall “automatically”, pretty much irrespective of the outlook for nominal demand growth.

Second and very closely related, we can then insist that there is no logical contradiction between a high-conviction view that deleveraging (as conventionally defined) will continue and an upbeat view on the economy. *  There is no need to wait until the credit flow data — measured even as impulse  — have turned before getting upbeat.   This was the key advantage that Benderly had over DB.

That was then, and this is now.  In the current environment, credit-financed spending ** has recovered towards normal levels relative to the broader economy and so automatic deleveraging has predictably been replaced by releveraging.  I say “predictably” because I predicted it and actually mentioned to Jason that he might want to update his perspective to emphasize releveraging.  The student had learned from that master and had seized the pebble!

So US credit growth seems likely now to outpace overall GDP growth, for at least as long as investment (of all sorts) spending continues to track with GDP or slightly stronger, which probably means basically til the eve of the next recession.  I guess that is “good” if you are cheering for credit growth.

But it is no more bullish the economy than deleveraging was bearish.  In fact, to the extent that maintaining the flow of credit-intensive spending requires a bit more credit growth, if makes the economy slightly more exposed to any disruption in the flow of credit, from the supply or demand side.

I have suppressed here any discussion of the distinction between credit supply and demand — in order to simplify.  I have assumed that credit demand is largely endogenous to real-side developments, while ignoring supply, which — in fairness — has not been a major determinant of wiggles in the economy post crisis.  That could change obviously.  The discussion around it is a separate one left for another time, but it is not necessarily unimportant.

This post just goes over the difference between x-leveraging and impulse.  The latter seems to do the arithmetic properly, but the former has the additional advantage of getting the direction of causation right and is more helpful for forecasting.

* After the crisis, Benderly had an upbeat view of the economy, partly because of his insight into automatic deleveraging. And for the part of the economy where credit was relevant (and which the stock market disproportionately capitalizes) he was spectacularly right.  GDP growth was been held back largely by weakness in credit non-intensive spending, particularly during the early post-crisis period during which Benderly was pushing automatic deleveraging.

** This is more true of business investment than household outlays on housing and durables, which have recovered from their lows by briefly achieving quite strong growth, but remain fairly low in level terms by historical standards. I will update the relevant charts when I have access to more tech.