The Threshold Effect of Finance on Growth: Reassessing the Burden of Evidence

Ronald Coase famously stated, “If you torture the data long enough, it will confess,” underscoring the need for the replication of well-accepted empirical results. In Economics, replication is more honoured in the breach than the observance. As a departure, this paper assesses whether the ‘burden of evidence’ is met for a recent, widely-cited finding, with potentially deep policy implications — that the finance-growth relationship is non-monotonic and has a credit threshold above 100% of GDP that reduces economic growth. If this empirical fact is established beyond a reasonable doubt, then it could be pathbreaking in further developing our understanding of the link between finance and growth. We assemble the ‘burden of evidence’ through the comprehensive scrutiny of several vital aspects, viz., (i) an exhaustive list of 14 absolute and relative measures of financial development, (ii) replications and extensions across two global datasets, (iii) near exhaustive analytical trajectories, (iv) different functional forms, (v) unifying analytical approach, and (vi) analytical rigor. The ‘burden of evidence’ from almost 3,000 well-structured cross-sectional and panel estimates do not support the threshold effect, and where evidence is uncovered, the parameters imply the questionable policy implication that advanced economies need to scale back their relative levels of financial development to those of Eastern Europe to avoid the growth costs associated with over-developed financial systems. The findings reject the assertion that finance is excessive and reduces economic growth.