A fundamental question in predicting the economic effects of the Covid-19 pandemic is understanding whether or not the effects are persistent. That is, we must determine if the economy will return quickly to its pre-crisis level, or if it will take a long time to re-absorb the fall in output. This depends to a large extent on how many companies will go bankrupt from the liquidity crisis (due to the drop in sales). It is therefore important to provide firms with the liquidity needed in order to avoid bankruptcies on a large scale. This is a shared goal, and the response by policymakers has generally been to provide ‘whatever it takes’. In fact, most governments have set up some form of credit guarantee, particularly for small and medium enterprises (SME) (OECD 2020). However, these policies need to be credible to be effective (Beck 2020). As a result, it is important to determine ‘how much it takes’ – are the schemes that governments provide sufficient to avoid massive liquidity-induced bankruptcies or not? It is also important to determine ‘how long it takes’, as the impact of the lockdown on the liquidity shortage can be very fast. An estimation of firms’ month-by-month liquidity needs can be useful in evaluating the cost of slow implementation of effective measures.
Given the unprecedented character of this crisis, estimating its short-run effects on firms is difficult, as reliance on other historical episodes can only be of limited support. The few attempts that have been proposed rely on stock market reactions (De Vito and Gomez 2020), or on estimates of revenue shortfalls coupled with assumptions on the response of costs (Carletti et al. 2020). In Schivardi and Romano (2020), we develop a simple accounting framework to determine which firms will be liquidity constrained, when they will become constrained, and to what extent. The general logic is very straightforward and is based on three ingredients: the initial stock of liquidity, an estimate of the evolution of cash flow month-by-month, and the budget equation determining the evolution of liquidity.
We apply the method to the population of Italian incorporated businesses (around 650,000 companies producing three quarters of the Italian private sector output) and consider the period from March 2020 until the end of the year. Sales growth from more than 500 sectors is forecasted by Cerved,1 in a consensus macroeconomic scenario (baseline scenario). As the possibility of a second wave of the epidemic arriving after the summer is material, we also consider a pessimistic scenario based on the assumption that the pandemic reappears in September, before picking up in October. The forecasts are carried out by Cerved sectoral experts, who consider the legislation (the lockdown) in addition to other economic factors (drop in demand, effects of social distancing, disruption of supply chains etc).
The framework uses firms’ balance sheets to obtain pre-pandemic output and obtains costs and the initial stock of liquidity in the same way. As an input, it requires the month-by-month estimates of sales growth at the sectoral level which, given each firm's sales from the previous year, make it possible to forecast a ‘sales evolution’ at firm-level. Costs are predicted using inputs' elasticities, which allow us to use sales growth forecasts (mediated by the elasticity of each input) to determine monthly outflows. We assume that, following the Italian government decrees enacted during the crisis, all financial payments and taxes are suspended. Moreover, we also assume that firms freeze their investment expenditures.
Given an initial stock of liquidity, the budget equation determines the stock of liquidity on a month-by-month basis. When this value turns negative, we classify a firm as illiquid. The method is transparent and straightforward to implement and has been used by various institutions (Bank of Italy 2020, European Commission 2020, OECD 2020 – a code with a mock dataset is available here).
The expected impact of Covid-19 on Italian firms’ sales is large. Table 1 reports descriptive statistics of firms, dividing them according to the predicted sectoral drop in sales for 2020, with respect to 2019. We separate sectors into groups with a drop of 20% or larger, between 20% and 10%, between 10% and 0%, and with non-negative sales growth. The group with the largest drop is by far the most populated, with more than 300,000 firms. The two intermediate groups have approximately 130,000 firms, and less than 60,000 firms record sales increases. Firm characteristics are very similar across groups. Average liquidity is around €400,000, but the median is much lower, at around €30,000 in all groups. Further, the 25th percentile is always below €10,000. This indicates that many firms have small liquidity buffers. Next, we consider two indicators of financial fragility: leverage (which is defined as debt over equity) and Cerved Group Credit Score (a riskiness indicator computed by Cerved that takes values from 1 (very safe) to 10 (very risky)). As it turns out, firms are also very similar across groups in terms of financial conditions. Therefore, the statistics in Table 1 indicate that, perhaps not surprisingly, the crisis hit sectors with an intensity uncorrelated to sectoral characteristics, at least in terms of size and financial health.
Table 1 Firm Statistics by Change in Sales
Note: The table reports descriptive statistics for firms’ characteristics, split according to the drop in sales. Leverage is debt over equity and is trimmed at the 1st and 99th percentiles. Risk class is from the Cerved Score, which takes discrete values between 1 (very safe) to 10 (very risky), with unit intervals.
Once we apply our scheme, we find that the effects of the pandemic are very quick, with more than 180,000 firms (employing 3.1 million workers) already becoming illiquid in April 2020 (Figure 1). The number of illiquid firms peaks at 200,000 (employing 3.3 million workers) in September 2020, and then it decreases very slightly for the rest of the year. The amount of liquidity shortage (that is, the value of negative liquidity of illiquid firms) is €40 billion in April 2020. This level then keeps increasing until the end of the year, when it reaches €72 billion. Of this, more than €50 billion are held by firms with less than 500 employees (Figure 2). In the case of a second wave of the epidemic after the summer, the number of illiquid firms surges to 236,000 in October 2020. In this scenario, the number of workers in illiquid firms also jumps to 4 million, and the negative liquidity jumps to €106 billion by December 2020.
Figure 1 Illiquid firms and workers
Panel a Number of illiquid firms
Panel b Workers in illiquid firms
Note: The figure reports the total number of illiquid firms (Panel a) and the total number of workers in such firms (Panel b) when using the budget equation to detect the firms for which liquidity has hit the zero constraint.
Figure 2 Total liquidity shortage for all firms and by firm size
Note: The figure reports the value of the total liquidity shortage for all firms and distinguishes between firms above and below the 500 employees threshold.
Next, we use the scheme to evaluate the coverage provided by the Italian Liquidity Decree, which supplies public guarantees to bank loans issued as a response to the pandemic. The decree designs different facilities to provide loan guarantees, with the amount offered in the guarantee decreasing with the size of the loan and with different conditions for small (less than 500 employees) and large firms.2
The government claims that this scheme mobilises €400 billion, more than enough to cover liquidity shortages. To check for this, we let firms borrow the maximum amount according to the different measures and check which firms cannot cover their liquidity shortage with such borrowing. The coverage is indeed complete. At peak, just 153 firms (employing less than 13,000 workers) cannot cover their liquidity shortages. Even in the case of a second wave after the summer (which would increase the liquidity shortfall substantially), firms' liquidity needs are manageable under the current schemes of liquidity provision.
Of course, this is the maximum theoretical coverage assuming that firms have access to the maximum loan supply the decree allows for. However, the procedural complexity of the measure increases with the amount it supplies. For example, the simplest measure offers SMEs up to €30,000 fully guaranteed. This measure is being implemented rather quickly, as it entails no risk for banks. As the loan amount increases, the government guarantee stops being complete, and loans need to be approved by a government agency. This means that banks might need some time to process applications in addition to obtaining the approval from the government agency. However, many firms become illiquid very quickly, so it is essential that credit flows to firms with the necessary urgency.
To check the amount of coverage from the different measures, Figure 3 reports the liquidity shortages after borrowing the maximum amount on measure 1, and after adding measures 2, 3 and 4 sequentially. We perform this exercise only for SMEs, as large firms only have access to measure 4. At peak, more than 100,000 firms are not fully covered by measure 1, and around 65,000 are not covered by measure 2 (1.5 million workers). It is only with measure 3 that we obtain almost full coverage (Panel a), where the number of uncovered firms and workers drastically drops to 1,430 firms and 197,473 workers. However, the measure only entails a 90% guarantee, so banks will have to screen borrowers. Additionally, the measure requires approval from the government. According to our calculations, over 60,000 firms will need it, implying that the banking system and the government agency will have to process a large number of applications in a short period of time.
Figure 3 Firms with liquidity shortfalls according to the liquidity measure
Panel a Number of illiquid firms
Panel b Number of illiquid workers
Note: The figure reports the liquidity shortfalls with no borrowing, with borrowing from measure 1, then when adding measure 2, 3 and 4. We only consider SMEs.
One way to speed up the process is to provide a two-stage procedure, using algorithms that measure credit risk in a timely manner (based on scoring models in the first stage). If a company has a positive score, the credit should be given with a lean and fast screening procedure. Banks' specific skills in credit assessment should be committed to companies with weak scores to distinguish between companies that still have development prospects, despite their negative quantitative indicators, from those that do not. By their nature, the scores do not incorporate soft information (that is, information that banks develop through direct relationships with their customers or collect through direct investigation of the applicants). Soft information is important in cases in which hard information raises a red flag.
To assess how much this approach would reduce the preliminary investigation, we used the Cerved Credit Score. Of the 110,000 companies that will need liquidity in April 2020, around 90,000 fall into the first seven classes, considered solvent. These firms should get credit quickly with simplified procedures. The reduction in the number of detailed investigations would allow banks to devote more time and resources to carefully (but quickly) screen the 20,000 remaining companies in the risk area.
References
Bank of Italy (2020), “Rapporto sulla stabilità finanziaria”, Number 1/2020, April.
Beck, T (2020), “Finance in the times of coronavirus”, in Baldwin, R and B Weder di Mauro (eds) Economics in the Time of COVID-19, A VoxEU.org Book, CEPR.
Carletti, E, T Oliviero, M Pagano, L Pelizzon and M G Subrahmanyam (2020), “The equity shortfall of Italian firms in the COVID crisis: A first assessment”, VoxEU.org, 19 June.
De Vito, A and J-P Gomez (2020), “Estimating the COVID-19 cash crunch: Global evidence and policy”, Journal of Accounting and Public Policy 39(2).
European Commission (2020), “Identifying Europe’s recovery needs”, Commission Staff working document.
OECD (2020), “Corporate sector vulnerabilities during the Covid-19 outbreak: Assessment and policy responses”, OECD policy briefs on Tackling Coronavirus.
Schivardi, F and G Romano (2020), “A simple method to estimate firms liquidity needs during the COVID-19 crisis with an application to Italy”, CEPR Covid Economics: Vetted and Real-Time Papers.
Endnotes
1 Cerved is a data provider and credit rating agency that also supplies firm balance sheets.
2 The decree n. 23 of April 8, 2020 offers public guarantees that decrease with the amount of the loan. In particular, for firms with less than 500 employees (SMEs in what follows), it offers (i) full guarantee up to the minimum between 30,000 and 25% of 2019 (measure 1); (ii) for firms with less than 3.2 million turnover (in addition to meeting the SME criteria), it offers 25% of 2019 sales with 90% government guarantee and 10% Confidi (an association for mutual guarantees) guarantee (measure 2); (iii) up to 5 million with 90% government guarantee (measure 3); (iv) up to the maximum between 25% of sales and twice the labor costs of 2019, with a guarantee from 90% to 70% according to firm and loan size.
