‘Zombie lending’ occurs when a lender supports an otherwise insolvent borrower through forbearance measures such as repayment holidays and temporary interest-only loans. The phrase was first coined for Japan in the late 1990s, but more recently several authors have documented that zombie lending to European firms has been widespread following the sovereign debt crisis (see Acharya et al (2019), Adalet McGowan et al (2018), Banerjee and Hofmann (2020), Blattner et al (2018) and Schivardi et al (2017)). In a recent paper, I examine whether these lending practices contributed to the subsequent low output experienced by the euro area. My findings suggest that zombie lending had negative consequences for output, investment and productivity in the euro area over the period 2011 to 2014.
How does zombie lending affect economic performance?
Zombie lending can lower output by misallocating funds across firms. By tying up credit to zombie firms, the flow of lending to more productive firms in the economy may be restricted, which lowers overall productivity. Several authors have argued that zombie lending to inefficient firms is not just a by-product of economic conditions but has in fact hampered the economic recovery in the euro area in recent years (see Acharya et al (2019), Blattner et al (2018) and Caballero et al (2008)). And these lending practices may ultimately have deleterious effects on banks’ balance sheets if they merely forestall firm liquidations.
But there are also potential benefits associated with zombie lending. It may not always involve lending to genuinely insolvent firms but rather firms with good long-term prospects experiencing a temporary period of weak demand. The Covid-19 crisis is a good example where these benefits are more likely to exist than in the economic setting covered by my paper. While it is too early to ascertain whether this applies to zombie lending activities in the euro area, there is some evidence that it may have been relevant for the Japanese case. Fukuda and Nakamura (2011) show that the eventual bankruptcy of Japanese firms who benefited from zombie lending was rare and that most of these firms recovered in the 2000s. They argue that if these supposed zombie firms were truly inefficient, the zombie lending they received would not have prevented their ultimate insolvency.
A model of zombie lending
To assess the impact of zombie lending practices on firm behaviour and economic performance, I develop a dynamic model of heterogeneous firms that incorporates a role for forbearance lending.
The model includes the following four key ingredients: (i) firms can borrow from a lender to finance their operations; (ii) firms can default on their loan obligations, leading to either their liquidation or a form of ‘zombie lending’; (iii) zombie lending takes the form of renegotiation over a reduction to the outstanding loan repayment; and finally (iv) lenders have incomplete information when making a loan because they do not know precisely whether a firm is ‘low quality’ or ‘high quality’, where the former firm type have lower average productivity than the latter. Rather, lenders rely on the partial information they can observe to determine the probability that a firm is low quality (ie the ‘firm-type risk’).
What I find
The model is fitted to euro-area firm-level statistics over the period 2011 to 2014, a period marked by higher levels of zombie lending (Acharya et al (2019)). The fitted model represents a ‘benchmark scenario’ with zombie lending. To evaluate the impact of zombie lending, I conduct a counterfactual exercise in which firms still have the option to default and liquidate but no longer have access to any zombie lending option.
The results can be summarized as follows. I first show that the rate of firm liquidation is only a little higher in the counterfactual scenario with no zombie lending as compared to the benchmark scenario. This is the net result of two effects. First, in the absence of zombie lending, the probability of liquidation for a given level of capital increases because liquidation is now the only default option (Chart 1). Second, and in turn, loan schedules become more expensive, as lenders take account of this increased risk of liquidation in their loan pricing (Chart 2), which results in firms taking on significantly less leverage, ie they choose smaller loans. The overall effect is that the counterfactual liquidation rate (ie proportion of firms that are actually liquidated) is little changed from the benchmark scenario.
Chart 1: Firm liquidation probabilities
Note: The figure plots how the liquidation probabilities vary with the size of a firm’s capital choice. The liquidation probabilities are on average higher in the counterfactual with no forbearance. But firms respond by making different choices about their leverage and so the liquidation rate is little changed in the counterfactual. A firm’s choice of capital depicted by the x-axis is scaled by the average level across firms.
Chart 2: Loan interest rate schedules
Note: The figure plots how the loan interest rates vary with the size of the loan repayment. The interest rate loan schedules are on average higher in the counterfactual with no forbearance. A firm’s loan repayment depicted by the x-axis is scaled by the average level of capital across firms.
The average of firms’ growth, investment rates and total factor productivity are higher in the counterfactual scenario with no zombie lending. A key driver of these results is that there is a larger proportion of high-quality firms in our counterfactual scenario; the zombie lending that is present in the benchmark scenario prevents low-quality firms from liquidating, which is a drag on aggregate output, investment and total factor productivity. Specifically, I find that aggregate output, investment and total factor productivity are around 5%, 8% and 1% higher in steady state in our counterfactual scenario with no zombie lending. These findings suggest that zombie lending practices may have been one of the contributing factors in relation to the recent low output across the euro area.
Why do zombie firms get credit?
An important question is: why do primarily low-quality firms receive zombie lending in our benchmark scenario, given a profit maximising bank would not ordinarily want to throw good money after bad by lending to a firm it judges to have poor prospects?
Several authors argue that the answer lies in the fact that lenders face zombie lending incentives in order to avoid the declaration of non-performing loans on their own balance sheets (see Andrews and Petroulakis (2019), Acharya et al (2019), Blattner et al (2018), Caballero et al (2008), Jordà et al (2020), Schivardi et al (2017) and Storz et al (2017)). My model highlights that information asymmetry faced by lenders is another key driver. Lenders do take account of firm-type risk. But because they cannot perfectly predict firm quality, the interest rate schedule is more favourable than otherwise for low-quality firms, and vice versa for high-quality firms. The end result is that the vast majority of firms in receipt of zombie lending are low quality because credit is (mis-)allocated to these firms with higher default risk.
While the model provides new findings about the impact of forbearance lending on firms’ financial decisions and their dynamics, some related questions remain unanswered. For example, the model does not incorporate a role for ‘zombie lending incentives’ but rather focuses on how zombie lending affects the firm behaviour and economic performance. Similarly, the model does not include any role for the potential increased unemployment associated with higher firm liquidation rates. Zombie lending could mitigate aggregate demand externalities if these lending practices can prevent an increase in unemployment (see Haldane (2017)). But given that I do not find that liquidation rates increase significantly in the absence of zombie lending, this suggests a smaller role for such a potential offsetting channel. Nonetheless, a full examination of whether the costs of zombie lending, due to lenders misallocating credit, outweigh the benefits, due to preventing an increase in unemployment, presents an exciting avenue for future research.
Belinda Tracey works in the Bank’s Research Hub.
If you want to get in touch, please email us at firstname.lastname@example.org or leave a comment below.
Comments will only appear once approved by a moderator, and are only published where a full name is supplied. Bank Underground is a blog for Bank of England staff to share views that challenge – or support – prevailing policy orthodoxies. The views expressed here are those of the authors, and are not necessarily those of the Bank of England, or its policy committees.