I have a problem with most commentary on China’s debt. Allow me to propose a solution inspired by Hyman Minsky’s framework of the credit cycle.

But first, the problem. Analysis of China’s debt typically begins by anchoring on either an estimate of the credit-to-GDP ratio or the so-called “credit gap.” From that starting point, commentators then roll directly into serious-sounding speculation on China’s dystopian future or hopes for a timely rebalancing.

But this starting point is unscientific.

One, the absolute level of debt is irrelevant. While the US continues to rack up foreign debt, Somaliland can’t even get a bank with a SWIFT code. To quote a proverb of unknown origin, “Running into debt isn’t so bad. It’s running into creditors that hurts.”

Two, the credit gap is also irrelevant. This measure, while cloaked in mild computational complexity, is essentially a statistical artifact, no different than the technical indicators employed by chartists. The Bank for International Settlements observes that countries with egregious credit gaps sometimes descend into financial crisis. I see correlation, but no causal mechanism explaining why aggregate credit growth drives individual borrowers into default.

While both pseudo-empirical measurements provide some descriptive power, neither offers a fundamental understanding of China’s borrowers.

My new set of evidence, inspired by Hyman Minsky’s stylized framework of the credit cycle, illuminates the incentives and constraints imposed on the individual economic agents central to China’s debt problem.

The main insight of the Minsky Cycle is that firms can be sorted into three categories based on their level of indebtedness: Hedge Units repay debt solely with cashflows from their business. Speculative Units make interest payments from their cashflows, but must borrow again to repay principal. And finally, Ponzi Units, which must borrow to repay both principal and interest. In the Minsky Cycle, the accumulation of firms in the ‘Ponzi Unit’ classification precipitates financial crisis.

I will not speculate on the probability of financial crisis in China (Michael Pettis does a good job laying out future scenarios). Instead, I will utilize Minsky’s insight to propose a new set of evidence for evaluating the incentives and constraints of China’s borrowers. There are three pieces to the puzzle:

The proportion of new debt issued to pay down existing debt.The debt-servicing capabilities of China’s borrowers.The proportion of bank loans issued under benchmark.

Let’s begin with the use of funds for China’s corporate bond issuers. As China’s overall credit efficiency has dropped, the proportion of bond issuance used completely to pay off old debt has risen dramatically. Among other causes, a growing proportion of China’s debt goes to pay off old debt rather than invest in new productive capacity.

Now, firms might repay old debt with new for one of two reasons: Healthy firms can refinance debt at a lower cost, or zombie firms must roll-over existing debt to avoid default.

Over the last three years many of China’s healthy borrowers have exercised the first option, refinancing existing higher-interest debt with new low-yield bonds. In fact, from 2014 through the end of 2016 corporate bond issuance surged as bond yields marched downward. And, as yields rebounded in 2017, corporate bond issuance fell in turn. Therefore, to the first layer of analysis, it appears that many Chinese borrowers are merely optimizing their balance sheets when they repay old debt with new.

However, peeling back to the next layer of analysis reveals evidence that a subset of zombie issuers borrowed to avoid default. Even as Chinese corporate bond yields have rebounded and issuance stalled, the proportion of bond volume issued to pay off old debt reached an all-time high – not the behavior of healthy firms taking advantage of a low-yield environment.

Finally, we can also track the proportion of bank loans made below benchmark to place borrowers along the Minsky Cycle. Currently, the behavior of China’s banking sector implies a subsidization of zombie firms to avoid default. That is, banks restructure debt at lower rates so the borrower can continue to make interest payments. This happened in Japan in the 90’s, and it seems to be happening now in China.

In the absence of bank subsidies to zombie firms, we would expect the proportion of loans issued under benchmark to reflect the true market price for loans. That is, when the benchmark is artificially high, we would expect banks to extend loans at below benchmark. However, the current difference between benchmark and corporate bond yields is its narrowest since 2014, while the proportion of bank loans made below benchmark is at its highest since the financial crisis, implying that banks are restructuring loans at cheaper rates to avoid losses.

In their most recent annual filling, ICBC claimed that net interest margins were hurt by successive reductions in the benchmark lending rate in 2015. However, the growing proportion of loans issued under that benchmark shows banks racing to offer lower interest payments, even as corporate bond yields have rebounded, implying a Japan-style subsidization of zombie firms.

In sum, we as China-watchers need a scientific framework for analyzing the implications of China’s corporate debt. The proportion of corporate debt issued to pay off old debt, the financial health of borrowers, and the loan pricing by banks is a reasonable place to start.