Guidelines for cost-effective use of SURE: Rule-based short-time work with workers’ consent and aligned replacement rates
As reaction to the COVID-19 crisis, the EU has implemented ‘temporary Support to mitigate Unemployment Risks in an Emergency’ (SURE), which consists of financial assistance up to € 100 billion in the form of loans to member states (see e.g. Garicano 2020, Grund et al. 2020). SURE loans have to be used for the preservation of employment, either in the form of short-time work (STW) or similar schemes. STW provides a subsidy to employers to compensate their workers for temporary working-time reductions. These working-time reductions may then be an alternative to layoffs for firms. Our article analyses how SURE loans can be used by national governments in a cost-effective way (i.e. how to save the maximum amount of jobs for a given amount of SURE loans). We derive guidelines based on macroeconomic research and institutional examples from the German STW system. These guidelines also provide useful perspectives if SURE is further developed into a European (un-)employment insurance system. However, we argue that the current exceptional situation due to COVID-19 justifies temporary additions in some dimensions.
Cost-effective short-time work
A cost-effective STW system defines eligibility criteria according to which firms can use this labour market instrument. Firms that are negatively affected by a recession will automatically fulfil these conditions. Thereby, STW acts as an automatic stabiliser as it changes firms’ expectations, which smooths hiring and firing over the business cycle. Balleer et al. (2016) find that the rule-based component of STW stabilises the economy without any additional discretionary government interventions. In a counterfactual model simulation for Germany, they show that unemployment fluctuations over the business cycle are around 20% smaller due to STW (compared to an economy without STW).
To reduce firing by a maximum amount for a given amount of spending, an effective rule would only allow for STW when the firm would be firing workers in the absence of STW. In addition, STW would only be used for cyclical reasons and not for structural ones in order to prevent any negative side effects. As it is non-trivial to implement these rules in practice, we provide several guidelines for how national governments can design rules to improve the cost-effectiveness of their STW systems.
National governments should define both a rule as well as simple eligibility criteria for STW, because one main advantage of this policy is its prompt availability in crises. Firms may for example have to prove that they face a substantial cyclical demand drop that would force them to initiate firings in the absence of STW. Obviously, there is asymmetric information, as firms know their financial situation much better than the government agency in charge. To reduce the asymmetric information problem, STW subsidies should only be paid if workers provide their consent, both on the decision of whether to implement STW as well as its extent (i.e. the actual hours reduction). This consent will only be given if a firm is faced with the option of firing these workers or going bankrupt. Thus, workers’ consent is a good protection against deadweight effects (i.e. STW for workers that would not have been fired). In the German case, the works council (elected worker representation) typically has to approve STW. In other countries, workers’ consent may have to be organised differently depending on industrial relations regimes. Another way to disincentivise excessive use of STW is partial cost sharing for firms, which implies that a working time reduction would not lead to a proportional cost reduction.
Furthermore, the publicly financed replacement rate (STW transfer relative to the wage) for the lost working time should be aligned with the replacement rate of the short-term unemployment benefit system in the respective country. The alignment with the replacement rate prevents excessive use of STW. Imagine the extreme case with a close to 100% replacement rate under STW and a much lower replacement rate for short-term unemployed. Under this scenario, workers would have no objections to a 100% working-time reduction, even though a much smaller working-time reduction may be sufficient to prevent firing. In addition, high STW replacement rates may give an incentive for workers to remain in declining industries during a recession and to wait until outside options have improved. Such a scenario would slow down structural change and thereby increase the (indirect) costs of STW. As a further protection against using STW for structural change, the use of the instrument should be limited to a reasonably short time horizon (e.g. one year).
Short-time work in severe crises
In severe recessions, such as the current COVID-19 crisis, a time limit on STW may be too restrictive and force firms to fire at the end of this period. Therefore, governments typically enact additional measures in these cases. These may include looser or quicker access to the instruments, increased subsidies (e.g. by increasing the replacement rate for workers or by reducing employers’ costs) or an extension of the maximum duration of STW. These discretionary changes certainly make the usage of STW more attractive for firms. However, with a well-designed rule-based system, where STW would only be activated if firing was the outside option, the extra effects of discretionary measures on unemployment can be expected to be limited in normal times. Based on a time-series analysis, Gehrke and Hochmuth (forthcoming) show that the state of the business cycle matters for the ability of discretionary policy to save jobs. In expansionary phases, these interventions can have negative effects. Discretionary STW may hence be very costly due to large deadweight effects. Put differently, in these cases STW pays subsidies to workers that would not have been fired otherwise or prevents an efficient reallocation of workers in times of structural change. However, in severe recessions, discretionary policy interventions can have positive extra effects, e.g. by protecting firm-specific human capital and preventing credit-constrained firms from bankruptcy.
To prevent discretionary measures from being used at the wrong point in time (with zero effectiveness), we recommend defining the case of ‘emergency’ or the condition for a severe recession ex ante. The ex-ante definition ensures that the policy changes affect firms’ expectations. Practically, the maximum duration of STW or the effective costs of STW could be automatically changed in case this critical aggregate state is reached. Balleer et al. (2019) show for Germany that such a rule could be defined based on a business confidence indicator such as the ifo business cycle index that is available in timely manner.
Extensions in the COVID-19 crisis
Are the described guidelines suitable for the current COVID-19 crisis? Countries that have several of our proposed guidelines in place (or implement them) certainly have a timely automatic employment stabiliser. Naturally, this will lead to increased fiscal expenses in recessions. SURE loans would support member states to finance these extra expenses. If the accessibility to STW is automatically increased in case of a severe recession, this further stabilises the labour market. This is also true in case of a pure discretionary intervention in a severe crisis like the current one, but the stabilisation effect should be expected to be smaller.
Importantly, there is one dimension wherein the COVID-19 crisis leads to a very different STW usage as compared to other crises. The German experience from previous recessions teaches us that STW is typically strongly concentrated in cyclical export sectors (such as the automobile industry or mechanical engineering, see e.g. Bellmann et al. 2010). STW is strongly used by medium-skilled blue-collar workers with relatively high incomes. Due to STW, these workers are already much better off than unemployed workers, also because the typical working time reduction in recessions due to short-time work is on average only around one third (i.e. these workers keep around two thirds of their previous income plus the STW benefit), see e.g. Balleer et al (2016). By contrast, during the COVID-19 crisis, many workers in low-paid sectors (such as certain retail sectors) are forced to a 100% working time reduction due to the lockdown.
For workers in locked-down sectors, excessive use of STW is less of an issue. In addition, inequality considerations come to the forefront. Although a more generous STW replacement rate makes the system more costly (without saving extra jobs), the COVID-19 crisis provides a justification to use STW systems as a temporary social transfers vehicle to quickly channel funds to groups that are hit particularly hard by the crisis (see also Giupponi and Landais 2020). In this case, potential top ups (relative to the short-term benefit replacement system) should be targeted towards low-income groups (see Schnetzer et al. 2020). These extra transfers may for example be oriented towards the generosity of national social assistance systems. This is particularly relevant for workers whose income falls below the subsistence level. In this deep and sharp crisis, channeling the funds via the STW system has the advantages of prompt availability and prevention of potentially complicated and time-consuming means tests. However, once the lockdown-induced crisis due to COVID-19 is over, national STW systems should return to the above guidelines if funds are meant to be used cost effectively.
Summary and outlook
Overall, we argue that SURE loans can be a useful instrument to stabilise the labour market. We have defined guidelines for national governments for effective usage of SURE loans. Most importantly, STW should be implemented with clear rules, as Balleer et al. (2016) find substantial automatic stabilisation of unemployment at very low costs due to a rule based STW. The German STW system follows most (but not all) of our guidelines. As SURE is potentially a cost-effective automatic stabiliser at the European level (if used properly), it may be the starting point for a more ambitious European (un)employment insurance system (see Vandenbroucke et al. 2020).