CFO.com posted an interesting piece by Vincent Ryan that recaps the results of a new study by a researcher at the Bank for International Settlements (“The Central Banks’ Central Bank,” as it’s usually called). The BIS paper, by Mathias Drehmann, focus its attention on measuring not just the bank credit outstanding a country at any given moment but total credit outstanding, which would include non-bank and foreign lenders, for example.
Drehmann studied what he calls the “Bank Gap” and “Total Gap” in various countries ahead of crisis periods and found that the Total Gap did a better job of foretelling a crisis period (as well as a better job at not foretelling non-crisis periods). The original paper has a nice graphic, reproduced below, that illustrates the author’s point.
Graph 1 shows that the total and bank credit-to-GDP gaps (or, for convenience, the “total gap” and the “bank gap”) can give different signals about credit developments. For instance, in the United Kingdom (Graph 1, left-hand panel), the bank gap did not signal any large credit build-up ahead of the recent crisis. In contrast, the total gap clearly captured the run-up in credit from the early 2000s onwards. This reflects the part played by non-bank funding, eg via securitisation, as the boom’s main driver. And different signals also emerge from the total and bank gaps even for highly bank-based systems such as Germany’s, at least in certain periods (Graph 1, centre panel). If we look more specifically at the years ahead of the 33 crises in the sample, the right-hand panel shows that both gaps are generally elevated during this phase. But the total gap is on average higher and rises more strongly than the bank gap, suggesting that it may be the better indicator.
What’s fascinating to me is not just his finding but that it’s news that looking at Total Gap is a worthwhile metric. In today’s financial world of HFT, Shadow Banking, Dark Pools, SPV’s, etc, bank credit tells only part of the story. Total Gap, which is a relatively straight forward measure to calculate, would seem a natural and helpful risk metric to track.
That said, the metric is not perfect. Returning to the CFO piece, it notes that, “whether used in reference to China or elsewhere, the credit-to-GDP metric has its flaws…One possible flaw in the metric: it may signal crises too early.” A case in point is China, where the country’s credit-to-GDP ratio has recently increased to “176 percent, up from less than 120 percent in 2006. Not only is the magnitude of the ratio worrying but also the deviation above the long-term trend” both indicate systemic risk is increasing. This observation brings me back the a debate I have constantly with clients and colleagues: what is really happening in China. Assuming that a lot of credit in that country is not even captured in any statistic (a very safe bet), then the actual Total Gap is probably higher than 176%, which would suggest that the cost of risk of doing business there is even higher than the numbers may suggest. Caveat emptor…
Whatever the case in China, it’s all the pieces noted below are worth a read, and it’s good to see the continued development of more sophisticated risk metrics.
Read the BIS paper here: http://www.bis.org/publ/qtrpdf/r_qt1306f.pdf
Read an analysis of credit risk in China here: http://seekingalpha.com/article/1306551-credit-to-gdp-gap-shadow-banking-and-the-fragility-of-the-chinese-banking-system