The impact of the COVID-19 pandemic on the employment situation and financial well-being of families with children in Austria: Evidence from the first ten months of the crisis

Objective: This study investigates the impact of the COVID-19 pandemic on the employment situation of parents and in turn on the subjective financial well-being of families with children in Austria. Background: The pandemic had strong repercussions on the Austrian labour market. The short-time work (STW) programme covered a third of employees in the first half of 2020 and helped to maintain employment levels. We provide evidence on how an unprecedented labour market crisis of this sort and in particular the exceptionally wide use of STW had affected the employment situation of parents and the financial well-being of different types of families. Method: The study draws on register data and representative panel survey data. The latter cover 905 families with minor children and include information on the employment situation of parents and the financial well-being of families before crisis onset, three months and ten months after its onset. Results: Register data show that mothers were not more strongly affected by the labour market crisis of 2020 than childless women or fathers. According to survey data, about a third of couples with minor children experienced income losses. Despite the wide use of STW and government support to families, the share of families in financial difficulties has substantially increased, especially among those with many children and single parents, many of whom were at risk of poverty already before the crisis. Conclusion: Substantial shares of dual-earner families that had low poverty risks before the crisis were in financial difficulties in 2020. Potential spill-over effects of financial shocks on children are discussed.


Introduction
The COVID-19 pandemic had strong repercussions on the Austrian labour market. Starting with March 2020, a variety of measures have been introduced to contain the spread of the virus and to mitigate the economic impacts of the containment measures (e.g., the lock-downs), that severely restricted economic activities and suddenly endangered substantial numbers of jobs. To counter the rapid rise in unemployment at the outset of the pandemic, a generous short-time work programme (Kurzarbeit) was implemented in March 2020 (Schnetzer et al. 2020) and used by a large share of employers. At its peak in April 2020, the short-time work programme covered just below 30% of Austrian employees ) -a much higher share compared to Germany, where about 18% of employees were short-time workers in the same month (Bundesagentur für Arbeit 2021). The extensive use of short-time work in Austria was effective in keeping the number of job losses at bay, with about 5% of those employed in February 2020 losing their jobs in the first half of 2020 .
Given its wide use, the short-time work programme had a sizeable impact on household incomes in Austria. Although employees in short-time work received between 80% and 90% of their previous net salary, the reduction of income associated with shorttime work can put the livelihood of low-income families at risk. Among families with a tight household budget, a reduction of the household income by 20% may severely reduce financial well-being and heighten risks of financial deprivation. Moreover, some employees such as those working in the tourism and catering industries lost larger shares of disposable income when sent into short-time work, due to the omission of income components such as tips and overtime pay in the calculation of short-time work allowances (Theurl 2020). On top of short-time work, the rise in unemployment and the loss of income among many of the self-employed 1 had a direct impact on the financial situation of private households (Albacete et al. 2021). In Austria, unemployment benefits generally amount to 55% of previous net income but can be higher for those entitled to family supplements (MISSOC 2020). 2 To reduce crisis-related income losses for those who lost their jobs due to the pandemic, unemployment benefits were topped up by oneoff payments to the unemployed. 3 To mitigate the loss of income among the selfemployed and families with children, hardship funds were created (Heitzmann 2020). Unemployed parents received a bonus of EUR 100 per child. About 90,000 families with minor children and at least one parent unemployed or in short-time work (eligibility criteria) had received monies from the hardship fund in 2020, with an average amount of EUR 1300 as a lumpsum (APA-OTS 2021).
To date, there is a dearth of knowledge as to how families have been economically affected by the COVID-19 crisis in Austria. The available evidence on the effects of the crisis on household incomes is based on simulation studies that project into the year 2020 1 According to the AKCOVID study (Steiber 2021a), more than a third of the self-employed recorded a severe drop in revenue in the first half of 2020 and reported that the COVID-19 crisis put their business in financial trouble (own calculations). 2 Entitlement to and the duration of benefit receipt depend on the period of insurance (for details, see MISSOC 2020). 3 The unemployed received EUR 450 twice: in September and December 2020. (Albacete et al. 2021;Fink et al. 2020) or on survey data (Kalleitner & Schiestl 2020;Resch 2020). These studies are concerned with households more generally, not families with children in particular. The present study targets this population group. It draws on panel survey data from the AKCOVID project that allow for detailed analyses of families with children of different ages. We study crisis-related changes in the parental employment situation and the consequences for the subjective financial well-being of families during the first ten months of the COVID-19 crisis in Austria. We examine the proportion of families in which at least one parent either lost the job or was registered for short-time work due to the COVID-19 crisis, and what impact these changes had on the families' financial well-being. To this end, we combine measures of current financial well-being at the household level with measures of future expectations. The core focus of our study is on the implications of the COVID-19 crisis on the financial well-being of different types of families, depending on the age and number of children, the presence of a co-residing partner and the number of earners. We first look at the share of families affected by short-time work and/or job loss. And given that family types differ in terms of their financial vulnerability to income losses, the analyses of parental labour market outcomes and income losses are complemented with analyses of families' financial well-being using subjective indicators (such as whether the household was able to make ends meet). Our study has an explicit focus on the main drivers (i.e., labour market shocks such as job loss or short-time work) and extent of financial hardship that was experienced by families in Austria during the first ten months of the COVID-19 crisis. Gendered experiences within households are outside the scope of this study (see Berghammer 2021 andNaujoks et al. 2021 in this Special Issue for studies on the impact of the pandemic on the division of care work within families).
We proceed as follows. In section 2, we provide a brief overview of prior research on the impact of the COVID-19 crisis on the household income distribution more generally and on the incomes and financial well-being of families more specifically. This is followed by an outline of our theoretical approach and theoretical expectations in section 3. After a description of the data used and our analytical approach in section 4, we present our main results in section 5. We close with a discussion of the implications of our findings for child well-being.

Types of data and indicators of financial well-being used
Disposable household income is a core objective indicator of financial well-being. Until now, the impact of the pandemic on household incomes, both nationally and internationally, has mostly been discussed based on simulations using pre-crisis structural data (e.g., Fink et al. 2020;Martin et al. 2020). The few studies available that have drawn on current data on how household incomes have changed in the course of the pandemic have tended to use survey data (e.g., Resch 2020 for Austria; Blom & Möhring 2021 for Germany; Clark et al. 2021 for a comparison of European countries). Given the susceptibility of survey-based income data to social desirability bias (Angel et al. 2019), income data from administrative registers would be preferable. Yet, register household income data will only be available with considerable delay. Moreover, an important advantage of survey data is the possibility to include subjective indicators of financial wellbeing, allowing for an analysis that accounts for how individuals and families experience their current financial situation (Brüggen et al. 2017).

Previous findings
Simulations of the impact of the pandemic on disposable household incomes for Austria suggest that the lowest income households were least affected by crisis-related income losses, because they have tended to source their income from state transfers already before the crisis (Fink et al. 2020). The level of social security benefits and pension payments has so far not been affected by the crisis. 4 Conversely, absolute and relative income losses were highest in the top income quintiles, partly because of the lower wage replacement in short-time work for high earners. Similar conclusions have been drawn by Resch (2020) based on survey data from the Austrian Corona Panel Project. Overall, these studies show that the lowest income households were most likely to maintain their pre-crisis disposable income and that the early crisis has thus caused a slight decrease in inequality between households. Such findings are in line with survey-based international evidence (Clark et al. 2021), estimating that absolute income inequality between households has declined in 2020 in Germany, Italy, Spain and Sweden, because policy responses have been more to the benefit of lower income groups (for similar results, see Almeida et al. 2020;Bruckmeier et al. 2020 for Germany). In summary, evidence on the impact of the pandemic on household incomes tends to suggest that the lowest income households, especially those strongly reliant on transfer income, have been shielded from financial crisis effects in 2020.
Families with minor children (below 18 years of age) only represent about a fifth of private households in Austria (Statistik Austria 2021a) and are underrepresented among the low-income households that were found most likely to have maintained their pre-crisis income. 5 Therefore, the extent to which general studies on household incomes provide relevant background information for our study on families with children is limited. Available studies do however provide important background information on the financial situation of different family types before crisis onset, from which we can infer their degree of financial vulnerability (i.e., risk of financial hardship if affected by income losses due to the pandemic, cf. Albacete et al. 2021). Single parents and families with more than three children for instance are since long overrepresented among those at risk of poverty. If hit by sudden income losses, most of these families are unable to maintain their standard of living even in the short-term (Albacete et al. 2021). 4 Some government measures have even slightly improved the incomes of the lowest income households (e.g., an increase in the benefit level of unemployment assistance). 5 Among low-income households (based on equivalized disposable income) that were found most likely to have maintained their pre-crisis income, single households are overrepresented (Albacete et al. 2021: 119).
While to date we lack knowledge about the extent to which different types of households have been affected by crisis-related income losses in Austria, research from the United Kingdom pertaining to the beginning of the crisis suggest that negative labour market and income shocks were more common among working parents than among workers without children (Cheng et al. 2021; see also Lofton et al. 2021 for the United States) and most common among low educated parents . In line with such observations there is evidence for the UK that in April/May 2020 parents tended to be more distressed financially and more pessimistic about their financial future than the childless (Cheng et al. 2021), whereas no such differences by parental status were discernible before the crisis. In a similar vein, a simulation study for Germany suggests that couples with minor children were more strongly affected by crisis-related income losses compared to childless couples (Beznoska et al. 2021).
Studies for Austria that focus on subjective indicators of financial well-being in times of COVID-19 so far analysed all private households and not families with children in particular. They show that in April to June 2020 the largest shares of those expecting their financial situation to deteriorate looking three to twelve months ahead were located in the lowest disposable household income quintile in Austria (Albacete et al. 2021). These findings point to widespread economic insecurity that mostly affected the lowest income households, even though they were least likely to record losses in disposable income (Albacete et al. 2021). 6 To date, neither survey data nor studies using register data provide insights into the financial well-being in times of COVID-19 among families in Austria. How did the pandemic affect the employment situation of parents? How do families with children evaluate their financial well-being in the COVID-19 crisis? This study aims to answer these questions, using measures of how the financial situation at the household level was perceived by the individuals sharing their financial resources within the household (Hamplova & Bourdais 2009;Ponthieux 2013).

Background and theoretical considerations
To capture the impact of the COVID-19 crisis not only on family incomes but on financial well-being, we need to combine families' differential risk of income losses with their financial vulnerability. Some families are more likely to have lost income than others (e.g., those with above-average household incomes, cf. Albacete et al. 2021: 122). Yet such differential risk is not informative about families' financial well-being. Families differ greatly in their capacity to cope with financial shocks (Lusardi et al. 2011: 5) and thus in terms of the risk that income losses during the year 2020 threatened their financial well-being (Demertzis et al. 2020). Both theoretically and empirically it is important to distinguish between the risk and the vulnerability perspective, as outlined in more detail in what follows. 6 This can be explained with reference to wide-spread worries about job losses in the further course of the pandemic among those still employed (Steiber 2021b: 28) and the fact that once government measures such as short-time work and temporary tax exemptions are retracted and rent referrals lifted, low income families are likely to face economic shocks.

Financial risk: job losses
In Austria, about a third of those employed in February 2020 have been in short-time work or became unemployed in the first half of 2020 (Steiber et al. 2021: 4). A simplistic risk appraisal predicts families with two earners to have more commonly been affected by crisis-related job and income losses than single-earner families (risk accumulation: two instead of only one job at risk, whereas transfer income was shielded from crisis effects so far). Yet, labour market risks are stratified. As in other countries, in Austria the young and lower educated population was the one most strongly affected by the crisis-related labour market risks of job loss and short-time work (Resch 2020;Steiber et al. 2021: 6). Given preferences to mate with partners who are similar in terms of age and education and thus often also in terms of labour market risks (assortative mating, Greenwood et al. 2014), we may thus find an accumulation of the risk of income losses during 2020 in particular among lower educated and younger couples (e.g., couples with two low-educated partners, younger couples with small children).

Financial vulnerability: impact on financial well-being
For Austria, Albacete et al. (2021) show that households with a particularly low financial margin (i.e., net income after deducting basic living costs) are single parents living with dependent children and two-parent families with three or more dependent children. On average (median value), these families can save or spend only EUR 100 a month on consumption above basic needs. Households with two parents and one or two dependent children, by contrast, have a median financial margin of about EUR 1200 (Albacete et al. 2021). From these estimates, it becomes clear that single parents and families with many children were, on average, not able to sustain their standard of living during a period of short-time work, even if that lasted only for several months. Single parents are furthermore twice as likely to have a negative financial margin (i.e., debts accruing by the month) compared to two-parent families and most of them have barely any liquid assets (Albacete et al. 2021). As a consequence, single parents would on average not be able to maintain their standard of living in the event of a complete income loss for even one month, whereas couples with up to two dependent children would on average be able to compensate by drawing on liquid assets for 12 months or more and couples with three or more children for about four months. 7 Such greater financial vulnerability of families with single earners and/or many children is reflected in at-risk-of-poverty rates 8 before crisis onset: Based on yearly incomes in 2019, it has been estimated that 14% of private households in Austria were at risk of poverty; with particularly high shares of at-risk families among single parents (31%), families with three or more children (30%, Statistik 7 For this estimate, Albacete et al. (2021) computed a combined indicator of households' financial margin and their liquid assets to measure the number of months during which a median household would be able to compensate for financial losses (in the event of a complete income loss). 8 Definition: someone is at risk of poverty if their equivalized household income is less than 60% of the national median household income (Statistik Austria 2021b: 16).
Austria 2021b: 78) and couples with children that involve a single male breadwinner and a non-employed mother (more than 20%, Statistik Austria 2020a). Given such differences in the degree of financial vulnerability across families (Lusardi et al. 2011), some families that have lost less income in this crisis than other families (in absolute terms) may still have been harder hit in terms of their financial well-being. For instance, single earners may have been more strongly affected by the crisis in terms of their financial well-being compared to dual-earner families, because a second earner in the household can potentially compensate for financial losses (Thomas & Sawhill 2005) and because single-earner families, in particular single parents, were at a higher risk of poverty already before the crisis. Overall, we would expect the share of families reporting low financial well-being to have increased compared to before crisis onset, with financially vulnerable family types like single earners and families with many children likely to have seen the steepest increase in rates of poor financial well-being. Moreover, given their high level of financial vulnerability we would expect family types such as single-earner families and those with many children most likely to worry about their financial future situation.

Parental employment
The extent to which parental job loss or short-time work affects families' income depends on whether the father, the mother or both are affected. The loss of the mother's job tends to incur smaller income losses compared to the father's, given the pronounced gender pay gap (Bergmann et al. 2019) and the fact that in the majority of dual-breadwinner parents of minors the father worked full-time and the mother part-time before crisis onset. 9 Female part-time work tends to be more precarious than full-time work (Hinterseer 2013). However, it is not clear if (female) part-timers had a higher risk of income loss in this crisis compared to full-timers. 10 In fact, prior research for Austria shows that throughout 2020, the crisis had affected women's and men's employment to a rather similar degree: Compared to 2019, employment declined by about 2% with negligible gender differences ) and only slightly more men than women were on shorttime work during the year 2020 (Steiber et al. 2021: 5; for similar results in the German context, see Naujoks et al. in this Special Issue). Some have argued that the COVID-19 crisis led to a re-traditionalisation of gendered employment and that mothers in particular faced a greater risk of job loss or hours reductions compared to men and childless women (e.g., Hipp & Bünning 2021;Hanzl & Rehm 2021, but Knize et al. 2021. Evidence for the United States (Lofton et al. 2021;Landivar et al. 2020) shows that in this crisis mothers' employment has declined more steeply than fathers' while the recovery of jobs has been faster for men and childless women than for mothers. Similar 9 In about 17% of families with minor children both parents worked full-time in 2019; 46% involved a fulltime working father and a part-time working mother (Statistik Austria 2020b). In the remaining 37% of these families either only the father was employed (about 27%, incl. families in which the mother was on parental leave), only the mother worked (4%), both were non-employed (4%), both were part-time workers (2%), or the woman worked full-time and the father part-time (1%). 10 The exception are those who held marginal jobs (i.e., earning less than €461 per month), who were not eligible to receive short-time work benefits and thus faced a greater risk of job loss, mostly affecting women (Bock-Schappelwein et al. 2020).
results have been reported for the United Kingdom (Andrew et al. 2020) and Canada (Fuller & Qian 2021). Against the backdrop of a conservative gender regime and the predominant role of mothers as secondary earners (Riederer & Berghammer 2020)combined with a sudden increase in family-based care work during the pandemic (Berghammer 2021 in this Special Issue) -we may expect to see a similar pattern in Austria. Yet, to date, we lack data on how the pandemic has affected the employment of fathers and mothers in Austria.

Data and methods
The study draws on register data (AMDB) and survey data (AKCOVID, cf. Steiber 2021a).
The register data are used to provide information on gendered employment risks in the COVID-19 crisis. For the analyses of income losses and financial well-being, we use a representative longitudinal survey of 2,000 respondents aged 20-64 living in Austria (AKCOVID Panel Survey with two waves of data collection and a retention rate between waves of 71%). In both waves, data were collected using a mixed-mode-design (80% of respondents taking part in an online survey; 20% were interviewed by telephone). This design provided for a good representation also of those unable or unwilling to participate in web surveys. To survey a representative sample of the population aged 20-64, quota sampling has been applied (population strata defined by gender, age, education, household size, number and age of children, and NUT2-region). Post-stratification weights are applied in all descriptive analyses to reduce potential sampling errors and non-response bias. To control for potential selective attrition between survey waves, longitudinal weights have been constructed using information collected in the first wave to predict participation in the second wave. 11 The first wave of data collection took place in June 2020, i.e., three months after crisis onset and at a time when the pandemic was temporarily under control (low infection rates). The second wave was carried out in January 2021, at a time when the crisis had already lasted ten months, and in the midst of the third (extended) lockdown that was in place since mid-November with only a short break at the end of December. At the outset of the crisis unemployment reached a historic high (increase by about 66% compared to March 2019, cf. Huemer et al. 2021). Yet, by April 2020 about 30% of employees were in short-time work and further job losses could be held at bay. The number of short-time workers decreased swiftly before the first survey wave in June 2020, when about 14% of employees were in short-time work. This share further decreased to about 8% in January 2021 ).

Defining the sample of analysis: families
To allow for a detailed analysis of different family constellations, the AKCOVID study aimed at an over-representation of families with minor children in the survey. We define as families only those living in the same household and our attention focuses on adults living with at least one minor child (below age 18). The sample of analysis includes 905 families with minor children (i.e., 106 single parents and 799 parental couples living in the same household irrespective of the legal status of their union, cf. Table 1 for sample description). These can be compared to 565 couples who do not live in the same household with a minor ('childless' couples, who may have children not living in the household). One person per household was interviewed and asked about the household's financial well-being (4.3 for discussion); the employment situation of the partner was measured by partner-proxy reports.

Employment situation and income losses
Respondents reported on their own and their partner's employment situation before and after crisis onset. Combining direct and proxy information on mothers' and fathers' employment status in each interviewed household, we construct a variable for parental work arrangements, differentiating between dual earners (both partners employed before crisis onset) and male breadwinner couples (the man was employed full-time or selfemployed, while the woman was not employed or on parental leave). Case numbers did not allow for a further differentiation of dual earners depending on the full-or part-time status of the woman (predominance of maternal part-time work in this group). We distinguish couples according to their level of education (both high/low educated based on whether or not both partners had a higher education entrance qualification). 12 Case numbers allow for this differentiation among dual-earner couples only. Whether or not households have experienced income losses until June 2020 13 was measured by agreement to the statement "We have to manage on a lower household income since the start of the Corona crisis" (1-does not apply at all, 2-does not apply, 3-in part, 4-applies to some degree, 5-applies completely).

Measures of financial well-being
In line with Brüggen et al. (2017: 2), we define financial well-being as the "perception of being able to sustain current and anticipated desired living standards and financial freedom". We use several items to measure different dimensions of subjective financial well-being. Based on a classical item, included in the European Social Survey, we tap respondents' feelings about their household income -at the time of the interview compared to before the crisis in February 2020 (1-living comfortably, 2-coping, 3-difficult to manage, 4-very difficult to manage). Answers are dichotomized with values 3 to 4 denoting financial fragility (Demertzis et al. 2020). We measure financial distress asking if families had to draw on savings or incur debt to meet living expenses and if families started to miss payments (e.g., failure to pay bills, rent or mortgages). As financial well-being is also influenced by future expectations and job insecurity (Choi et al. 2020;Vera-Toscano et al. 2006), we furthermore examine financial anxiety, i.e., worries about future financial problems as a result of the crisis using an 11-point scale (0-not at all worried, 10-extremely worried, dichotomized with values of 7 or higher denoting a strong worry).
We focus on measuring subjective financial well-being at the household level; the investigation of potential gender differences in individual financial well-being within households (Knittler & Heuberger 2018) is outside the scope of this study. To test our assumption of income pooling at the household level, we carried out regression analyses, which confirm that the household's financial well-being is assessed similarly, irrespective of whether it is reported by a female or male interviewee.

Analytical strategy
In a first step, we use register and survey data to investigate the impact of the crisis on the employment situation of parents and to ascertain if working women and in particular mothers were more likely to lose their jobs or to register for short-time work than men or the childless. In a second step, we move the analysis to the household level and estimate the risk of income losses in connection with job loss or short-time work (STW) in different family constellations. The Austrian family hardship fund granted financial support to families with minor children in which at least one co-residing parent either lost the job due to the crisis, was sent into STW, or had substantial losses of income from selfemployment. Since most of those eligible for hardship funds were parents experiencing one of the first two, we computed a combined risk of STW or job loss. We investigate if this risk was affected by the number of earners in the family, the age and number of children and the level of parents' education. Since parents of younger children tend to be younger themselves, the age of the youngest child serves as a proxy for parental age to investigate our expectation that young parents have been particularly hard hit. Third, we estimate the shares of (different types of) families who reported having lost part of their household income due to the crisis. Finally, we turn to analyse subjective financial wellbeing at the household level, based on the assumption that in most families with minor children, financial resources are pooled and shared among family members (Hamplova & Bourdais 2009;Ponthieux 2013). We investigate which family types were most at risk of low financial well-being due to crisis-related income losses. Section 5 provides descriptive results based on register and survey data. Tests for the statistical significance of differences between family types are provided in the supplementary material.

Employment situation of mothers and fathers
As shown by the register data (Table 2), in Austria crisis-related labour market risks in 2020 were rather similar for women and men. Of all those aged 20-64 and employed in February 2020, about 27% of men and 24% of women were in short-time work (STW) in April 2020 and about 6%-7% of each gender lost their job and became unemployed (UNE). By June 2020, about 16% of men and 14% of women were on STW and 4%-5% were unemployed (and about 4% of each gender left the labour force, mostly into retirement). Overall, about 39% of men and 34% of women experienced a spell of STW or UNE in the first half of 2020. Moreover, mothers have not been more strongly affected by job loss or STW than the childless or men. Between crisis onset and the end of 2020, about 43% of pre-crisis employed men and 37% of women had a spell of STW or UNE, irrespective of motherhood status (Table 2). Our survey data from June 2020 are close to the register data in terms of employment changes across 2020, thus proving highly representative. Limiting our sample to those aged 20-64 and employed in February 2020, we find that about 17% of the pre-crisis employed were in STW in June 2020 and about 4% were newly unemployed, with nonsignificant differences between women and men (Steiber et al. 2021: 5). Robustness analyses based on our survey data confirm that the presence and age of children had no impact on the risks of STW or UNE in June 2020, either for men or for women (cf. regression analysis in the supplementary material, Table S1). That is, in contrast to what has been found for countries such as the UK or the US, in Austria parents have not been more strongly affected by crisis-related job losses or STW compared to the childless. Moreover, we find no difference in the odds of STW or UNE between full-timers and parttimers (see supplementary material, Table S1).
The survey data further show how the crisis has affected parental working hours. As illustrated in Fig. 1, the share of mothers with minor children working either full-time, part-time or short-time remained stable at its pre-crisis level. That is, the sum of the shares of mothers in one of these three statuses was constantly at about 60% from February 2020 through to January 2021. In June 2020, about 12% of mothers were on STW and this share declined to about 7% by January 2021. The share of part-time work first declined -concomitant to the rise in STW -but then recovered to the original level of about 35%-36%. Yet, the share of full-time work declined from about 25% pre-crisis to 19%. It thus appears that the return-flows from STW to regular employment among mothers worked better in case of part-time work. Unless most of the mothers who were in STW in January 2021 will eventually return to full-time work, this may result in a further spread of part-time work among mothers as a result of the COVID-19 crisis. The first half of 2020 also saw a rise in the share of unemployed mothers from about 4% to 9% in the first three pandemic months, which then declined to about 6% by January 2021.

Figure 1:
Parental employment situation pre-crisis, three and ten months after crisis onset Sample: Mothers and fathers with children <18 years in February 2020 (N=785 for 02/2020 and 06/2020 and N=529 for 01/2021). As the share of fathers on leave is so small (<1%), they were added to those outside the labour force (OLF).
Among fathers, the summative share of those working either full-, part-or short-time also remained stable (sum of the shares in these statuses at about 77% throughout, cf. Fig.  1). In June 2020, about 16% of fathers were in STW and this share declined to about 7% by January 2021. In the supplementary material (Fig. S1), we provide ancillary analyses that differentiate by the age of children, showing that those most likely to move from fullto part-time work in the second half of 2020 were mothers of children below schooling age. The mid-year increase in the share of unemployed mothers can be traced to a combination of a dampened level of job creation that may have delayed labour market reentry after parental leave, some declines in self-employment and the loss of marginal jobs (Bock-Schappelwein et al. 2020;Bock-Schappelwein, Fink, et al. 2021).

Employment and income situation of families
In a next step, we investigate the degree to which parental income loss during the crisis accumulated at the household level. In June 2020, in about 19% of single parents and 36% of couple households with minor children at least one parent was affected by either UNE or STW (Table 3). About a third of families with two employed parents saw at least one parent lose the job or move to STW by June 2020. However, we find heterogeneity within dual earners: those involving two high-educated parents were less at risk (21% in 06/2020; 8% in 01/2021) compared to those that involved two low-educated parents (38% and 40%, respectively). This difference is statistically significant, suggesting that labour market risks accumulated in low-educated couples (in about 10% of these couples both parents were affected). We furthermore find higher shares of affected parents among (younger) parents with small children (about 40% in 06/2020) than among parents of older children (33%). Male breadwinner families (one job to lose) were not less likely to be affected by parental STW or UNE than dual earners in June 2020, showing that male breadwinners are a select group with heightened labour market risks. Yet, by January 2021 many of the affected male breadwinners appear to have moved back to regular employment, while the share of those in STW or UNE remained very high among low-educated dual earners (about 40%).

Income losses among families
Parental labour market dislocations, and in particular the exceptionally wide-spread experience of STW, led to income losses at the household level. Among couples with minor children, about a third (33%) reported that they had to manage on a lower household income in June 2020 compared to just before the crisis (Fig. 2). We find no difference between dual-earner and male breadwinner families, but within dual earners, the share of families with income losses was substantially higher among couples with two low-educated parents compared to those with two high-educated parents (25% versus 37%, significant gap at p<0.05). The share of families with income losses was particularly high among couples with 3+ children (about 40%) and single parents (42%). The latter is in part the result of reduced or postponed alimony payments by non-resident fathers, who may have been affected by income losses (Fuchs & Premrov 2020). In summary, we find the largest shares of families with income losses among those who were already more likely to have low incomes and higher poverty risks before the crisis. Whereas in February 2020 about 8% of couples with minor children had difficulties managing with their household income, this share of financially fragile families had increased to about 21% by June 2020 and remained at this elevated level until January 2021 (Fig. 3). In line with findings for the United Kingdom (Cheng et al. 2021), shares of financially fragile couples were higher among parents than the childless (Table S2 in the supplementary material for significance test). Among couples with minor children, the largest shares of financially fragile families in June 2020 were those with 3+ children (32%) and single parents (36%). We generally find higher shares of male breadwinner families than dual-earner parents with difficulties making ends meet in times of COVID-19. Yet, among the latter, a great deal of heterogeneity is observed: Among dual-earner households with two high-educated parents the share of those with financial problems remained low during the crisis with at most one in ten families having difficulties making ends meet (10% in 06/2020, 4% in 01/2021), whereas among low-educated parents this share amounted to about a fifth at both points in time (Table S2 for significance tests).
Overall, we find those family types to experience the greatest levels of financial fragility in times of crisis who were already at a higher risk of poverty before the crisis: single parents and families with many children. We do however also observe a sizable share of financially fragile families among those who were relatively well protected from financial hardship before the crisis: dual-earner parents.

Running into debts and failure to meet financial obligations: financial distress
About a quarter of single parents and about a fifth of parental couples had to draw on savings or to accrue debts due to financial losses since the start of the crisis (Fig. 4, graph [*] on top). Since not all families have savings to fall back on, the share of those who had to go into debt to avoid falling into arrears will be higher among families with a small financial margin such as single parent families. About 14% of single parents and 8% of couple parents were not able to meet their financial obligations (e.g., paying bills, rent or mortgages) in June 2020 (Fig. 4, graph in the middle). In terms of subgroup analysis, we find parents of minors, especially single parents, at significantly higher risk of financial distress compared to childless couples (Table S2 for significance tests). Moreover, the share of parents in financial distress (drawing on savings or accruing debt to maintain living standard) was significantly higher among low-educated than among high-educated dual earners (about 21% and 12% respectively in June 2020) and it appeared that male breadwinner families saw less of an improvement in their financial well-being until January 2021 compared to other family types.

Worries about the financial future: financial anxiety
Financial anxiety was wide-spread in the first ten months of the pandemic: about every third family with minor children strongly worried about their financial situation in the further course or aftermath of the pandemic (Fig. 4, graph at the bottom). Such financial anxiety was significantly more pronounced among families with children than among childless couples (Table S2 for significance tests). No significant differences emerged depending on the number or age of children. In couple households with minor children, significantly higher shares of low-educated dual earners (about 30%) and male breadwinners (29%) than higher educated dual earners (17%) expressed financial future worries in June 2002. While the concerns of the latter further decreased until January 2021 (to about 13%), those of low-educated two-earner and male breadwinner families remained at the high level of June 2020.

Figure 3:
Families with difficulties managing on household income before and during the crisis

Conclusion and discussion
This study provided evidence on how the COVID-19 crisis in Austria had affected the employment situation of parents and the financial well-being of different family types in the first ten months of the crisis. Our analytical strategy involved the combination of objective data on parental labour market shocks with subjective data on the implications of income losses for the financial well-being of families that allow for an assessment of how families experienced their situation Based on register data, we found that parental employment rates remained relatively stable at pre-crisis levels, mainly due to the wide use of short-time work (STW). Mothers in Austria were not more strongly affected by the labour market crisis of 2020 than fathers or childless women, i.e., among working parents, equal shares of mothers lost their job or went on STW compared to men or childless women. In line with Germany, in fact, a slightly larger share of male than female workers were on STW in the year 2020. According to a systematic review by Gottardello and Mazrekaj (2021), these findings are in contrast to some studies carried out in countries such as the United States, the United Kingdom or Canada, that found substantive gender gaps in employment and income shocks to the disadvantage of mothers. Yet, in the other countries for which studies have been published in peer reviewed journal articles, the evidence of gendered effects of the pandemic on employment and earnings was highly mixed. None of the studies published and included in the systematic review had used register data. Adams-Prassl et al. (2020) used international survey data and found significant differences in job loss rates by gender in the United States and the United Kingdom, but not in Germany. Some studies found women -in particular mothers with young children -were more likely to reduce their working hours in the pandemic compared to men (e.g., Hupkau Yaish et al. 2021 for Israel). Our study found some evidence which points at a potential further spread of part-time work among mothers as a consequence of the pandemic also in Austria (i.e., in case mothers who were in STW in 2021 re-enter regular employment on a part-time basis). Time will tell if this will be among the longer-term consequences of STW for working conditions (Steiber 2021c).
Although the STW programme was able to keep job and income losses at bay, the labour market crisis had a stark impact on household incomes. About 36% of couples with minor children were affected by STW or job loss in the first half of 2020. Consequently, about a third of parental couples reported a drop in their household income, with particularly high shares among low-educated dual earners and male breadwinner families. About a fifth of single parents were affected by STW or unemployment in June 2020, yet due the shortfall of alimony payments, more than 40% reported income losses. A family type that was certainly also hard hit by the crisis were migrant families, given the greater risk of job and income loss among workers without Austrian citizenship (Steiber 2021b: 49). Due to limitations in terms of sample size, the present study was not able to provide separate analyses for these families.
Despite the wide use of STW and government support to families affected by crisisrelated income losses, the share of families in financial difficulties has substantially increased in the first ten months of the crisis, especially among those with a single earner (single parents and male breadwinners) and those with many children, many of whom were at risk of poverty already before the crisis. Although the use of STW decreased over the course of 2020 and despite government measures aimed to contain the financial fallout from the crisis (e.g., hardship funds) many families were still in financial distress in January 2021 and worried about their financial future. The persistence of financial distress highlights the longer-term impact of the financial shocks resulting from STW and job loss on household budgets.
Not only those families who were at risk of poverty already before the crisis were affected by financial distress due to the COVID19 crisis. Sizeable shares of dual-earner families with children, who appeared relatively well protected from financial hardship before the crisis had difficulties managing on reduced household incomes and had to draw on savings or accrue debt to compensate for the income losses, especially among lower educated dual earners.
Overall, we may see at-risk-of-poverty rates increase in Austria in the aftermath of the COVID-19 crisis, as predicted by Heitzmann (2020). This poses risks not only for the parents who are directly affected by labour market dislocations, but also for the welfare of their children (Misra et al. 2012). Based on evidence that parental job loss affects children's educational outcomes (Gregg et al. 2012;Kalil & Wightman 2011;Rege et al. 2011;Stevens & Schaller 2011), their subjective well-being and happiness (Nikolova & Nikolaev 2021), and their physical health (Mörk et al. 2014;Lindo 2011), it comes as no surprise to see first evidence on the negative implications of labour market shocks borne by parents during the COVID-19 pandemic on their children: children of those affected by financial shocks in 2020 have for instance been found to receive fewer learning resources such as tutoring  and to have lower mental well-being (Langmeyer et al. 2020) compared to children from families that were not in financial distress. Of course, families in financial distress during the pandemic experience negative effects beyond the purely economic: financial instability, especially at a low income, amplifies psychological distress in the household with negative implications for parental and child wellbeing (Gadermann et al. 2021). Mental distress and material hardship in early childhood, even if short-lived, can have lasting implications on children's development and lives.