What Happened to Integration and Financial Stability after the Global Financial Crisis? An Answer from Machine Learning

The impact of financial integration on financial stability is an intricate subject that has garnered extensive scholarly debate regarding its advantageous and deleterious dimensions. The crux of the discourse has predominantly centered around appraising the effects of financial openness on middle-income emerging markets that aspire to attain levels of prosperity and stability that characterize advanced industrial economies.

The literature unequivocally suggests a positive relationship between financial integration and economic growth, financial development, financial stability. Financial globalization offers two main advantages. First, it homogenizes the costs of capital across countries, which mostly benefits those countries where capital is scarce. Second, financial globalization tends to facilitate more streamlined international risk-sharing mechanisms. The reduction of financing constraints afforded by openness can enhance the resilience of financial systems, fostering smoother consumption and investment patterns that, in turn, mitigate volatility. Finally, financial integration has been associated with diminished macroeconomic volatility, which further fortifies the case for its positive influence on financial stability through enhanced risk-sharing capacities.

However, the extant empirical investigations present a somewhat discordant narrative, revealing conflicting evidence regarding the anticipated outcomes. Optimal risk-sharing outcomes during the recent globalization era have been largely confined to industrialized nations, leaving developing countries largely excluded from this advantage.

Instances such as the global financial crisis, the European debt crisis, and the Asian crisis at the end of the Twentieth century have highlighted some of the disadvantages that are associated with heightened financial globalization. The drawbacks of financial openness on financial stability include heightened volatility in financial markets, susceptibility to external shocks, and difficulties in effectively managing capital flows. The process of financial openness often results in a deeper integration with global financial markets, which exposes domestic financial systems to external shocks and contagion stemming from international financial crises. Furthermore, the surge in capital flows accompanying financial openness can amplify the level of volatility in exchange rates, asset prices, and interest rates, posing challenges for the management of domestic monetary and fiscal policies.

Recognizing these challenges, numerous countries have undertaken reforms, including the implementation of macroprudential policies, to fortify the resilience of their financial systems against shocks, particularly against those shocks that emanate from the volatility of international capital flows. This holds particular significance in economies that rely on commodities, where these capital flows are closely linked to fluctuations in international commodity prices.

Taking all this into account, in our most recent study, to be published in the coming weeks, we revisit the debate regarding the effects of financial openness on financial stability. The primary objective of our paper is to examine the effect of increased openness on financial stability while acknowledging the multifaceted dimensions inherent in these two variables. In contrast to most studies that focus on a single type of measure, we adopt a comprehensive approach by incorporating both de jure and de facto measures in our analysis. 

We use several different measures of economic and financial openness, considering that each of these proxies is used to measure different aspects of financial integration. We examine five proxies for openness, sourced from the IMF, UNCTAD, and Chinn and Ito ¹. This set includes de facto as well as de jure measures for openness. We consider different proxies for financial stability in acknowledgement of the various aspects entailed by this unobserved variable, such as capital adequacy, non-performing loans, provisions, and liquidity ratios.

Our estimates enable us to isolate the focal effects while controlling for a comprehensive set of macroeconomic, political, and institutional variables. We employ a double-debiased machine learning algorithm to estimate the effects of financial openness across its various dimensions on the financial soundness indicators. Double machine learning models are valuable for estimating direct effects in observational studies because they can effectively address high-dimensional sets of confounders.

Our primary findings indicate that, overall, financial openness tends to be advantageous for financial stability. Heightened levels of openness are associated with diminished ratios of nonperforming loans to total loans and/or increased capital adequacy ratios. Additionally, greater openness generally leads to a heightened level of bank liquidity, which is a favorable attribute for maintaining financial stability. Interestingly, the impact of openness on bank profitability is inconspicuous, suggesting that intensified competition in global markets does not notably affect the profitability of local banks. Notably, the outcomes underscore the fact that when openness involves mere receptiveness to increased capital inflows, it may pose a risk to financial stability due to heightened vulnerability to the abrupt stops that are linked with substantial capital surges.

Our results hold substantial policy implications and suggest pathways for further exploration. On the policy front, our results signify that a deeper integration with global financial markets has a positive influence on financial stability, while preserving bank profitability. This implies that policy-makers, especially those in nations with emerging financial markets, should contemplate implementing measures to facilitate integration with global markets, including easing the restrictions on capital inflows. While macroprudential policies may be advantageous in specific market conditions, our results advise against their permanent adoption as a strategy for insulating financial systems from global markets.

From an academic perspective, our findings stress the importance of prudence in selecting a proxy for openness. The specific choice of this variable can yield distinct outcomes that may not be reproducible under alternative proxies.