NBER-技术银行_银行的技术采用及其影响_62页_1mb

Abstract Summary

The paper introduces a new model of endogenous growth where banks adopt technology embedded in capital goods produced by entrepreneurs. This adoption enhances bank efficiency, which affects occupational choices of agents (workers vs. entrepreneurs), firm productivity, and ultimately long-term economic growth. Empirical evidence from US bank and regional data shows that higher IT investment in banks reduces intermediation costs, increases lending to small businesses, and is associated with a more dispersed firm size distribution. A counterfactual simulation demonstrates that increasing bank technology adoption to the upper half of the data distribution accelerates GDP growth from 2% to 2.17%.

Model Overview

The model features banks and entrepreneurs competing in a general equilibrium setting. Banks choose to adopt capital goods, which affects their efficiency in transforming deposits into loans and monitoring borrowers. This efficiency spillover increases firm productivity by influencing agents' decisions to enter entrepreneurship. The occupation choice mechanism determines the growth of firm productivity based on aggregate ability, creating a feedback loop with bank efficiency.

Counterfactual Results

• Permanent increase in the capital goods share in loan production (e.g., to the top IT expense quartile) raises GDP growth to 2.17% by accelerating lending rates and credit supply.
• Bank efficiency grows faster than firm productivity initially but stabilizes, leading to persistent growth benefits.
• Firm size distribution becomes less skewed and more dispersed, consistent with creative destruction dynamics.

Empirical Evidence

• Bank-level data shows higher IT expenditure negatively correlates with the cost of financial intermediation (CFI), after accounting for labor substitution via IV.
• State/MSA-level data reveals that lower CFI banks grow small business loans faster and exhibit larger standard deviations/skewness in firm size, especially for sectors reliant on external finance, supporting the model's predictions.