Conclusion

Literally accepting the main results reported by Cournede and Denk (2015) would have the radical policy implication that the ideal amount of financial intermediation is no financial intermediation at all.

In contrast to the main results in linear specification of Cournede and Denk (2015), the quadratic specifications in Arcand, Berkes, and Panizza (2012) and Cecchetti and Kharroubi (2012) indicate a large initial range over which more finance brings higher rather than lower growth. Even those formulations yield what I suggest are unreliable turning points that advanced economies have supposedly already exceeded (Cline 2015b and appendix 6C).

That three empirical studies by teams at three international organiza­tions nonetheless come out with the same implication that there is already too much finance in such economies as that of the United States is poten­tially a powerful message for policy. In particular, in the Basel Committee reforms boosting required capital for banks, a key consideration has been whether too much required capital might discourage lending and curb growth because of a resulting decline in the rate of investment. But if the too much finance results were taken literally, that outcome could be a good thing rather than a bad thing, because reducing the amount of finance rela­tive to GDP would (at least over the highest range) boost growth rather than reduce it. A central purpose of this chapter is to show that these studies do not have sufficiently robust findings on negative effects of finance to warrant a general policy stance welcoming rather than seeking to avoid shrinkage of finance as a consequence of higher capital regulatory require­ments.

Some might argue nonetheless that the absence of a statistically sig­nificant positive effect of finance on growth in the higher ranges of fi­nancial depth means that the benefits of lesser risk of financial crisis as a consequence of even sharply higher capital requirements for financial in­stitutions could be obtained at no cost to the economy. But cross-country growth regressions, especially with insignificant coefficients, are not a reli­able basis for evaluating optimal capital requirements. Instead, a more reli­able basis for investigating this issue is to use a calibrated model comparing the costs associated with higher capital requirements against the benefits they provide through reduced risk of financial crises.^[240]