Money and credit : a long-term view

Schularick and Taylor (2012) documented a sizeable increase in the ratio between credit and broad money since the end of WWII, which they interpreted in terms of a progressive disconnect between the two aggregates. I show that this interpretation is incorrect, since, as I demonstrate mathematically, this evidence is uninformative for the issue at hand. In fact, Jordà, Schularick and Taylor's (JST) data show that, since the XIX century, fluctuations in broad money and credit have exhibited an extraordinarily strong correlation within each single country in the dataset, to the point that (e.g.) either Shin's (1994) or Wright's (2000) test consistently detects cointegration between the multipliers of the two aggregates. I also show that, after WWII, there has been no change in the relative prediction power of credit and broad money for financial crises compared to the pre-WWII period, and that the change in the multiplier of either aggregate has been more powerful than credit growth, the variable considered by Schularick and Taylor. My results imply that (1) for the "traditional" banking sector there has been no change, since WWI, in the relationship between its monetary liabilities, and the amount of credit it extends to the private non-financial sector; and (2) only the comparatively recent ascent of the "shadow" banking sector - which is not covered by either JST's, or the Bank for International Settlements' data - introduced a "wedge" between broad money and credit. Contrary to Schularick and Taylor's interpretation, the ascent of "shadow banking" is the only reason why, today, we live in the "Age of Credit".