Last week, the International Monetary Fund published the results of its 2017 Article IV consultation with China concluding that the credit gap, the difference between the country’s current credit-to-GDP ratio compared to its historical trend, has begun to moderate. However, I believe this is a flawed indicator of debt-related macroeconomic risk. As I pointed out in my framework for China’s debt problem, the credit-gap ignores the firm-level debt structures that precipitate a liquidity crunch.

As I mentioned in that note, the late economist and Washington University professor Hyman Minsky stylized the credit cycle by sorting firms into three categories based on debt structure. These included hedge units, firms that can repay debt with cashflow from their business; speculative units, firms that can make interest payments using cashflow from their business, but rely on more debt to repay principal; and ponzi units, which rely on more debt for both interest and principal repayments. In Minksy’s framework, the accumulation of ponzi units precedes a corporate liquidity crisis and macroeconomic bust.

The question we should ask ourselves is, how many of China’s corporate borrowers are paying off existing debt with new debt?

Since complete firm-level bank loan data is hard to come by, we can use corporate bond data to approximate the buildup in China’s liquidity risk. Analyzing the prospectuses for 29,600 corporate bonds issued between 2010 and the present, I found that, even as yields rebounded and issuance tanked, the proportion of newly issued debt used entirely to pay off existing debt remained near its all-time high.

While the IMF maintains that the top-line numbers on China’s debt accumulation are moderating, I believe that underlying liquidity issues remain.