A friend in need is a friend indeed? Nonbanks and lending relationships during crises

Understanding the implications of the rising footprint of non-banks on financial stability and the real economy is a key concern for policy makers. Using data from the global syndicated loan market, we show that non-banks cut their credit by significantly more than banks during financial crises. Differences in the value of lending relationships explain most of the lending gap. While lending relationships with banks benefit borrowers, relationships with non-banks do not improve borrowers’ access to credit during crises. Our results imply that the rise of non-banks leads to a shift from relationship towards transaction lending and could exacerbate the repercussions of financial crises.

The importance of non-bank financial institutions, or non-banks, has been growing tremendously since the Great Financial Crisis of 2007/08. They now hold about half of the assets in the global financial system. This shift from banks to non-banks has raised concerns about the impact on credit supply, financial stability, and the real economy, especially during times of market turmoil. However, so far little is known about non-bank lenders’ behaviour in crises.

In a new study (Aldasoro et al (2023)), we use data on global syndicated loans to examine non-bank lending during financial crises. Non-banks are important players in the syndicated loan market: they make up about one-third of global syndicated lenders and have been increasing in number since the early 1990s. By now they originate about 20% of all new syndicated credit to non-financial firms. Non-banks, which include investment banks, finance companies and mutual funds, serve firms in all regions and sectors (Figure 1), but their loan originations are more concentrated in certain countries and industries than those of banks (see Aldasoro et al (2022)).

We identify financial crises occurring in borrowers’ countries using Laeven and Valencia’s (2018) classification of systemic banking crises. Lenders’ exposure to financial crises is measured as the stock of outstanding loans extended by a lender to firms in a given crisis country over the lender’s total stock of outstanding syndicated loans. This measure reflects that some lenders are more exposed than others to the same financial crisis.

Figure 1: Country-level share of non-bank lending

Our study begins by establishing that non-banks reduce their credit to non-financial firms in crisis countries by considerably more than banks (see Figure 2). Concretely, non-banks cut lending by about 50% more than banks. This “lending gap” is observed even after controlling for differences in lender characteristics such as their funding models.

In principle, the stronger contraction in non-bank lending could be due to non-banks serving firms that fare worse during crises themselves. Indeed, in our large cross-country sample, borrowers connected to non-banks are significantly riskier than those of banks, even in the same country and industry. To account for such differences in borrower characteristics, including profitability, risk, or size, our analysis uses fixed effects. Effectively, this helps us separate loan supply from loan demand effects. We find that while the lending gap narrows when we take borrower quality into account, it remains economically large and statistically significant.

To understand the drivers of the lending gap, we then investigate the value of lending relationships for banks and non-banks during crises. Previous research has shown that lending relationships with banks can lead to better loan terms and improve borrowers’ access to credit, particularly during crises (Bolton et al (2016)). In other words, relationships with banks mitigate the negative effects of financial crises on bank credit to firms.

The analysis shows that relationships with non-banks do not improve borrowers’ access to credit to the same extent as those with banks. Unlike for banks, lending relationships with banks do not benefit borrowers through lower loan spreads during crises. Moreover, after accounting for lending relationships in our regressions, the lending gap between bank and non-bank lending significantly decreases.1 These results suggest that non-banks tend to act more like transaction lenders, even if they have a history with the borrower. Indeed, despite their specialization and potential informational advantage in the riskier segment of the market, non-banks significantly reduce lending especially to riskier borrowers during crises. These findings support the argument that relationships with non-banks provide limited value.

Figure 2: Non-bank lending during crises

The contraction in non-bank lending during crises has real effects: firms connected to non-banks see a significantly stronger decline in overall syndicated lending during financial crises. Consequently, their investment rates decline by relatively more. Taken together, our findings suggest that the increasing presence of non-bank lenders may cause greater financial instability and harm the broader economy during times of economic stress. The rise of non-bank lenders may also result in a shift from relationship-based lending to transaction-based lending, which could negatively impact borrowers’ access to credit during crises. Moreover, the rising footprint of non-bank lenders and the strong contraction in their lending to highly-leveraged borrowers during crises are particularly concerning, given the historically high levels of corporate debt.

Our findings carry two important implications for policy. First, policymakers should keep a close eye on non-bank lending to non-financial companies, in addition to non-banks’ impact on money markets. Second, while regulations enacted after the GFC have made banks more resilient, the greater presence of non-bank lenders and their stronger contraction in lending during crises may offset some of these gains.