In-Work Benefits for Married Couples: an Ex-Ante Evaluation of EITC and WTC Policies in Italy

Transcription

1 Ministry of Economy and Finance Department of the Treasury Working Papers N 12 - November 2013 ISSN X In-Work Benefits for Married Couples: an Ex-Ante Evaluation of EITC and WTC Policies in Italy Giuseppe De Luca, Claudio Rossetti, Daniela Vuri

2 Working Papers The working paper series promotes the dissemination of economic research produced in the Department of the Treasury (DT) of the Italian Ministry of Economy and Finance (MEF) or presented by external economists on the occasion of seminars organized by MEF on topics of institutional interest to the DT, with the aim of stimulating comments and suggestions. The views expressed in the working papers are those of the authors and do not necessarily reflect those of the MEF and the DT. Copyright: 2013, Giuseppe De Luca, Claudio Rossetti, Daniela Vuri. The document can be downloaded from the Website and freely used, providing that its source and author(s) are quoted. Editorial Board: Lorenzo Codogno, Mauro Marè, Libero Monteforte, Francesco Nucci, Franco Peracchi Organizational coordination: Marina Sabatini

3 In-Work Benefits for Married Couples: an Ex-Ante Evaluation of EITC and WTC Policies in Italy 1 Giuseppe De Luca, Claudio Rossetti**, Daniela Vuri*** Abstract This paper investigates labor supply and redistributive effects of in-work benefits for Italian married couples using a tax-benefit microsimulation model and a multi-sectoral discrete choice model of labor supply. We consider in-work benefits based on the Earned Income Tax Credit (EITC) and the Working Tax Credit (WTC) existing in the US and the UK, respectively. The standard design of these income support mechanisms is however augmented with a premium for two-earner households to avoid potential disincentive effects on secondary earners. Revenue neutral policy simulations show that our reforms may greatly improve the current Italian tax-benefit system in terms of both incentive and redistributive effects. Furthermore, neglecting sector-specific attributes of the various job opportunities may lead to an oversimplified representation of the choice set that does not allow to capture some labor market transitions and thus results in attenuated labor supply responses. JEL Classification: I38; H31; H53. Keywords: In-work benefits, sectoral labor supply, poverty, microsimulation, married couples. 1 This paper uses EconLav, a tax-benefit microsimulation model developed by ISFOL with the support of the Ministry of Economic and Finance and the Ministry of Labor. The authors are responsible for all errors and interpretations. Corresponding author. giuseppe.deluca@unipa.it, Tel: , Department SEAS, University of Palermo, Italy Department of Economics and Finance, LUISS Guido Carli, Italy. Department of Economics and Finance, University of Rome Tor Vergata, Italy, IZA, CESIfo.

4 CONTENTS 1 INTRODUCTION IN-WORK BENEFITS IN THE US AND THE UK Earned Income Tax Credit Working Tax Credit POLICY REFORMS The Italian FA program The EITC policy reform The WTC policy reform STRUCTURAL MODEL OF LABOR SUPPLY A multi-sectoral model of labor supply for married couples Empirical specification and estimation Labor supply predictions DATA AND MICROSIMULATION MODEL ESTIMATION RESULTS SIMULATION OF EITC AND WTC REFORMS CONCLUSIONS REFERENCES APPENDIX A APPENDIX B... 33

5 1 Introduction In-work benefits are usually promoted as income support mechanisms that encourage employment in the low-skilled population, while maintaining high levels of social protection. This twofold objective is achieved by providing mean-tested transfers to low-income households with eligibility conditional on some employment requirement to avoid the harmful disincentive effects of the welfare trap. Pioneering in-work benefit schemes are the Earned Income Tax Credit (EITC) in the US and the Working Tax Credit (WTC) in the UK, but similar policies have been recently implemented in a number of OECD countries. Despite the general consensus on effectiveness of these welfare instruments for lone parents, the design of in-work benefits for married couples is still muddied by important concerns. Economic theory and previous empirical evidence suggest that family-based schemes, where the benefit is mean-tested against household income, generally promote employment among primary household earners (Eissa and Liebman 1996; Blundell 2000; Blundell and Hoynes 2004; Eissa and Hoynes 2004; Bargain and Orsini 2006). Such schemes are likely to create, however, negative labor supply effects on secondary earners as their earnings may move households in regions of the budget set with high marginal tax rates (Eissa and Hoynes 2004). To contrast these unintended disincentive effects, some countries have experienced individual-based schemes where the benefit is mean-tested against individual income (Immervoll and Pearson 2009). In this case, however, the major concern is on a less efficient targeting of redistributive effects since transfers may also be provided to lowincome workers in well-off households. Whether labor supply incentives and redistributive effects can be reconciled into a single welfare instrument is still an open issue. This paper contributes to the literature on in-work benefits by proposing an innovative design of EITC and WTC schemes for married couples aimed to overcome potential disincentive effects on secondary earners. Our strategy consists of introducing a new benefit premium for two-earner household within otherwise standard family-based schemes that preserve an efficient targeting of redistributive effects towards poor working households. Effectiveness of the resulting in-work benefit schemes is assessed by means of policy simulations on the 2008 Italian tax-benefit system using a national tax-benefit microsimulation model and a structural model of labor supply. Unlike most of the previous studies using a similar policy evaluation approach, our structural model of labor supply draws upon the multi-sectoral discrete choice framework by Dagsvik and Strøm (2006) that jointly accounts for nonlinear and nonconvex budget sets, hourly wage differentials among jobs of different sectors, prediction errors in the hourly wages, observed and unobserved heterogeneity, and sector- 2

6 specific quantity constraints. The original setup of the multi-sectoral model for a single decision maker is extended, within a unitary framework, to the case of two decision makers to capture labor supply responses of both spouses. Further, we assess the implications of using alternative model specifications by comparing our multi-sectoral model with a simpler benchmark that ignores sector-specific attributes of the various job opportunities. Differently from earlier ex-ante evaluations of in-work benefits for Italy (Figari 2011 and Colonna and Marcassa 2011), our policy simulations focus on EITC and WTC in-work benefits for Italian married couples with a special emphasis on the new benefit premium for two-earner households. Policy reforms are simulated under tax revenue neutrality by considering the abolition of Italian family allowances (FA) for dependent employees and contingent workers. We find that in-work benefits with a suitable set of incentives for secondary earners may have strong positive effects on labor supply of wives, weak negative effects on labor supply of husbands, and strong positive effects on equity. The EITC is more effective than the WTC in boosting employment of wives, while the WTC is more effective than the EITC in fighting poverty. Most labor supply responses take place in the private sector where jobs are characterized by lower hourly wages than the public sector. Furthermore, we show that neglecting sector-specific attributes of the various job opportunities may lead to an oversimplified representation of the choice set which does not allow to capture some labor market transitions and thus results in attenuated labor supply responses. The remainder of the paper is organised as follows. Section 2 provides a brief review of the EITC in the US and the WTC in the UK. Section 3 describes the Italian FA and formalizes the EITC and WTC schemes simulated in our study. Section 4 presents our multi-sectoral discrete choice model of labor supply for married couples. Section 5 describes the data and the tax-benefit microsimulation model. Estimates of the two labor supply models and their labor supply elasticities are presented in Section 6, while policy simulations are presented in Section 7. Finally, Section 8 concludes. 2 In-work benefits in the US and the UK This section reviews main features, incentive effects, and previous empirical findings of the in-work benefits existing in the US and the UK, respectively. A common feature of both welfare programs is that low-income households are entitled to a refundable tax credit (or benefit) provided that at least one adult member works and other eligibility criteria are satisfied. 3

7 2.1 Earned Income Tax Credit The EITC was introduced in the 1970s as a negative income tax program with eligibility conditional on employment. Although it started as a modest program for low-income families with children, the EITC went through several expansions in 1986, 1990, and 1993 and is now considered a key pillar of the welfare system in the US. To be eligible for the EITC, the taxpayer needs meet two conditions: (i) positive earned income from employment or self-employment, and (ii) earned income, adjusted gross income and investment income below certain thresholds. Conditional on eligibility, the credit entitlement depends on family earned income according to separate schedules by filing status and number of eligible children. The left panel of Figure 1 illustrates the 2011 EITC schedule of married filers by number of eligible children. Each schedule includes a phase-in region where the credit is initially increased at a certain subsidy rate, a flat region where the credit is kept constant at the maximum amount, and a phase-out region where the credit is reduced at a certain tapper rate. Subsidy rate, maximum credit amount, and tapper rate vary with the number of eligible children leading to a more generous welfare program in favor of large households. 1 As discussed in Eissa and Liebman (1996) and Eissa and Hoynes (2004, 2011), economic theory predicts an unambiguously positive effect of the EITC on labor force participation of primary household earners. The incentive effects on hours worked are expected to be ambiguous for people in the phase-in region and unambiguously negative for people beyond the phase-in region. Secondary earners of married couples, many of whom are women, have instead an incentive to reduce both labor force participation and hours worked since their earning may move the household in the phase-out region of the EITC schedule. These theoretical predictions are supported by empirical findings from quasi-experimental studies based on the various EITC expansions. For single parents, there is evidence of strong positive effects on labor force participation and small negative effects on hours worked (Dickert et al. 1995; Eissa and Liebman 1996; Liebman 1998; Meyer and Rosenbaum 2001). For married couples, Eissa and Hoynes (2004) find a reduction of labor force participation due to a small rise in participation of husbands and a sizable reduction in participation of wives, while Neumark and Wascher (2001) find a negative effect on hours of work among poor households with a working adult. 1 For additional details on the EITC see 4

8 2.2 Working Tax Credit The UK has a long history of in-work benefits. The Family Income Supplement was introduced in 1971 and then replaced by the Family Credit in After a number of reforms during the early 1990s, the Working Families Tax Credit was introduced in 1999 and then replaced again by the Child Tax Credit and the Working Tax Credit (WTC) in Below, we focus on the more recent WTC which extended eligibility to childless households. To be eligible for the WTC, the claimant must work a minimum number of hours per week (either 16 or 30) and fulfill other conditions related to either age, disability, responsibility for children or previous periods of unemployment benefits. The maximum credit entitlement is made up of a basic element for eligibility plus a set of additional elements for the presence of a second adult, being a lone parent, working at least 30 hours per week, having a basic or severe disability, and sustaining formal childcare costs. The effective amount of the credit is mean-tested against gross annual income before tax and national insurance contributions (jointly assessed in the case of couples). Specifically, it is equal to the maximum credit if income does not exceed a fixed disregard and is reduced at a certain taper rate above this disregard. As illustrated in the right panel of Figure 1, an important difference with the EITC is the lack of a phase-in region. Moreover, as the claimant works 30 hours or more per week, one can observe an upward shift of the maximum credit entitlement due to the 30 hour element of the WTC schedule. 2 Theoretical predictions and empirical findings on the incentive effects of the WTC line up with those discussed above for the EITC. There is evidence of strong incentives on employment of single mothers, weak disincentive effects on employment of married women whose husbands work (see Blundell et al. 2000, Brewer et al. 2006, and Francesconi and van der Klaauw 2007, among others), and positive effects on hours worked by those in employment (Leigh 2007) and lone mothers working part-time (Gregg et al. 2009). As pointed out by Blundell et al. (2002), the British WTC has two distinguishing features. First, the minimum hours limit placed on eligibility may weaken the incentive effects on labor force participation. This conjecture is supported by a differencein-difference analysis of the 1992 reform to the Family Credit which reduced the minimum hours limit from 24 to 16 hours per week. Second, the additional element for full-time work (the 30 hour element) may help offset the disincentive effects on hours of work which are typically observed in the EITC. 2 For additional details on the WTC see 5

9 3 Policy reforms Family-based in-work benefits which are mean-tested against household income are usually well targeted to poor working households. For the same reason, however, these welfare programs are likely to create disincentive effects on secondary earners and hence are not advisable in cases where married women are the population group of interest. In this section, we present an innovative design of the EITC and the WTC to deal with this issue. Before formalizing the schedules of these in-work benefit schemes, we first describe the Italian FA program to be abolished for financing the new welfare programs. 3.1 The Italian FA program The Italian FA program provides family-based benefits that are exempt from taxation and meantested against number of eligible household members, household composition and gross household income. Its schedule distinguishes 15 household typologies depending on the number of adult members, their civil status, and the presence of children and disabled individuals. Our analysis focuses on four household typologies consisting of a married couple, no disabled member and (i) no children, (ii) one child, (iii) two children, and (iv) three children. People entitled to claim for FA are dependent employees, contingent workers, unemployed covered by the unemployment benefit system, and former-employees pensioners. Self-employed workers are excluded from the program since at least 70% of gross household income is required to be from wages, salaries, former-employee pensions and other benefits granted to dependent employees. Gross household income is defined as the sum of earnings from employment and self-employment, taxable non-labor incomes, and other non-taxable incomes if above 1, 033 Euro. Figure 2 shows that FA are strictly increasing with household size and non-increasing with gross household income. With the exception of childless couples, each schedule is characterized by a flat region where poor households receive the maximum benefit and three subsequent phase-out regions where the benefit is reduced at certain tapper rates. Although this is considered a welfare program to alleviate poverty, one can see that the benefit expires at rather generous income cut-off points. As discussed below, our in-work benefit reforms impose moderately lower income limits to collect resources from the right-hand-side of the household income distribution. These resources are then used to finance a suitable scheme of incentives for secondary earners of low-income households. 6

10 3.2 The EITC policy reform Our EITC policy reform grants a benefit exempt from taxation which is mean-tested against gross household income and household size. Eligibility to the EITC is restricted to households with positive earnings and the underlying policy coefficients are determined to ensure tax revenue neutrality for each household type after taking into account both the abolition of the FA program and potential labor supply responses. The EITC schedule for one-earner households has the standard form E c = Ec max{0, min{ec, t 1c (G 1c G)}} max{0, min{ec, t 2c (G G 2c )}}, c = 1, 2, 3, 4, where Ec is the maximum benefit provided in the flat region, t 1c = Ec /G 1c is the subsidy rate of the phase-in region, t 2c = Ec /(G 3c G 2c ) is the tapper rate of the phase-out region, G tc, t = 1, 2, 3, are the income cut-off points, and G is gross household income. The EITC schedule for two-earner households presents two important differences. First, it provides a higher maximum benefit Ē c = (1 + p c ) Ec, where p c 0 is the new benefit premium for two-earner households. Second, it extends the phase-in and phase-out regions to avoid too binding constraints on gross household income. By imposing that the length of the flat region, the subsidy rate and the tapper rate do not change, the income cut-off points of the schedule for two-earner households are set to Ḡ 1c = (1 + p c ) G 1c, Ḡ 2c = G 2c + p c G 1c and Ḡ3c = (1 + p c) G 3c p c (G 2c G 1c ). Figure 3 illustrates the EITC schedule used in the first scenario our policy simulations. The benefit premium for two-earner households is fixed at p c = 1 for all household types. The income cut-off points G 1c, G 2c, and G 3c are fixed to 50%, 100% and 150% of the poverty line in the baseline tax-benefit system and are allowed to vary across household types according to the coefficients of the Carbonaro equivalence scale. Compared to FA, our EITC policy imposes more binding income cut-off points. Poor households may receive a lower benefit, but as gross household income increases the EITC schedule provides a higher benefit. 3.3 The WTC policy reform Our WTC policy reform grants a benefit exempt from taxation which is mean-tested against hours of work, gross household income and household size. Eligibility to the WTC is restricted to households with at least one adult person working 16 hours per week or more. The maximum benefit entitlement is made up by three components: the basic element W1c, the second adult element W 2c and the 30 hour element W3c. We assume that W 1c is provided to all eligible households, W 2c is 7

11 provided to households where both spouses work at least 16 hours per week, and W3c is provided to households where both spouses work at least 30 hours per week. Similarly to the EITC, the basic element W1c is determined to guaranty tax revenue neutrality for each household type. The other two elements are instead fixed to W 2c = p c W 1c and W 3c = q c W 1c, where p c 0 is the benefit premium for two-earner households and q c 0 is the benefit premium for working full-time. form The WTC schedule for households who are only entitled to the basic element has the standard W c = W c max{0, min{w c, t c (G G 2c )}}, c = 1, 2, 3, 4, where W c = W 1c is the maximum benefit provided in the flat region, t c = W c /(G 3c G 2c ) is the tapper rate of the phase-out region, and G 2c and G 3c are the income cut-off points delimiting the phase-out region. Notice that, in this schedule, the first income cut-off point of the flat region is defined implicitly by the eligibility condition placed on hours of work and it varies across households as a function of hourly wages of both spouses and non-labor household income. Households who are entitled to the second adult element face a similar WTC schedule with maximum benefit W c = W 1c + W 2c and income cut-off points Ḡ2c = G 2c and Ḡ3c = G 3c + p c (G 3c G 2c ). For households who are entitled to the 30 hours element, the maximum benefit is W c = W1c +W 2c +W 3c and the income cut-off points are G 2c = G 2c and G 3c = G 3c +(p c +q c ) (G 3c G 2c ). Thus, two-earner households are entitled to a higher maximum benefit and are subject to a larger phase-out region. The flat region and the tapper rate of the phase-out region are instead kept constant. Figure 4 illustrates the WTC schedule used in the first scenario of our policy simulations. For comparability with the EITC, we set the benefit premium for two-earner households to p c = 1, the benefit premium for working full-time to q c = 0.5, and the income cut-off points G 2c and G 3c to 100% and 150% of the poverty line in the baseline tax-benefit system multiplied by coefficients of the Carbonaro equivalence scale for household type c. Compared to the EITC, the WTC schedule is particularly targeted to households with low hourly wages and includes additional incentives to encourage full-time work. 4 Structural model of labor supply Our approach to estimate labor supply responses of married couples to changes of the tax-benefit system relies on a generalization of the multi-sectoral discrete choice model of labor supply for a single decision maker developed by Dagsvik and Strøm (2006) and Dagsvik et al. (2011). 8

12 4.1 A multi-sectoral model of labor supply for married couples Like the bulk of the labor supply literature, we consider a unitary framework where husband and wife maximize a common utility function that depends on their own leisure, their spouse leisure, and disposable household income. 3 The utility of the couple is U = V (l m, l f, y) ϵ(s, k), where V (l m, l f, y) is a systematic component, l m is the husband s leisure, l f is the wife s leisure, y is disposable household income, ϵ(s, k) is a positive random taste shifter, and s = (s m, s f ) and k = (k m, k f ) index, respectively, the sector combination and the combination of jobs for the husband and the wife. The random error ϵ(s, k) is assumed to be identically and independently distributed across households, sectors and jobs with distribution function P r{ϵ < x} = exp( 1/x), x > 0. The budget constraint of the couple, which incorporates earnings of the husband and the wife, non-labor household income, taxes and benefits, is represented by y = ψ (h m ω m (s m ), h f ω f (s f ); I), where h m and h f are the hours of work of the husband and the wife, ω m (s m ) and ω f (s f ) are their sector-specific hourly wages, I is non-labor household income, and ψ is a tax-benefit function mapping gross incomes into disposable household income. Provided that the utility is increasing in y, one can substitute the budget constraint into the utility function to obtain U = ν(z m, z f ; I) ϵ(s, k), where z i = (h i, s i ), i = m, f, and ν(z m, z f ; I) = V (l m, l f, ψ(h m ω m (s m ), h f ω f (s f ); I)). The choice sets of jobs offered to the husband and the wife are assumed to be independent from each other. Each of them consists of a finite number of alternatives including non-participation and several job opportunities in different sectors. The hours of work of each job are fixed, but there can be several job opportunities with the same working requirement that differ because of other nonpecuniary characteristics. Each job is then characterized by a number of working hours, a wage 3 A more general approach would be a collective model where the two spouses have their own preferences and they are involved in an intra-household bargaining process. This approach implies weaker behavioral restrictions than the unitary approach and allows investigating intra-household welfare allocation (Vermeulen 2006; Francesconi et al. 2009). Model identification requires however additional assumptions or specific data. 9

13 rate, and other nonpecuniary (qualitative) attributes. The size of the choice sets is unknown to the researcher. Let Q(z m, z f ) denote the set of job pairs with hours of work (h m, h f ) in sectors (s m, s f ) and q i (z i ) denote the number of jobs available to spouse i = m, f with hours of work h i in sector s i. From standard results in discrete choice models, the probability π(z m, z f ) of selecting a specific job pair k = (k m, k f ) with characteristics (z m, z f ) is equal to the sum of the selection probabilities of all jobs in Q(z m, z f ) with the same observable characteristics (z m, z f ). Thus, for job pairs with h m > 0 and h f > 0, we obtain π(z m, z f ) = k Q(z m,z f ) ν(z m, z f ; I) P where P is a normalization constant of the form P =ν(0, 0, 0, 0; I) + + x m >0 x m>0 x f >0 s m s m ν(x m, s m, 0, 0; I) q m (x m, s m ) + = ν(z m, z f ; I) q m (z m ) q f (z f ), (1) P x f >0 s f ν(x m, s m, x f, s f ; I) q m (x m, s m ) q f (x f, s f ). s f ν(0, 0, x f, s f ; I) q f (x f, s f ) The choice probabilities of the other job pairs have a similar form. Equation (1) suggests that the reduced form of the model is analogous to a multinomial logit model where the systematic part of the utility function is weighted by the number of jobs available to the two spouses. As in Dagsvik and Strøm (2006), we can employ an alternative parametrization of the number of available jobs to foster interpretation. Let q i (z i ) = θ i (s i ) g i (z i ) with θ i (s i ) = x i >0 q i(x i, s i ) for alternatives with h i > 0 and θ i (s i ) = g i (z i ) = 1 for alternatives with h i = 0. It follows that θ i (s i ) measures the number of job opportunities relative to nonworking opportunities that are available to spouse i in sector s i, whereas g i (z i ) measures the share of jobs with h i hours of work that are available to spouse i in sector s i. 4.2 Empirical specification and estimation The empirical specification of our model distinguishes between jobs in the public and private sectors to capture the sizeable wage differentials existing in these two sectors of the Italian labor market (Giordano et al. 2011). The choice sets of both spouses consist of 13 alternatives (non-participation plus 6 hours of work intervals for each sector with middle points at 20, 30, 35, 40, 50 and 60 hours per week), yielding 169 possible job combinations for each couple. The logarithm of the systematic component of the utility function has a quadratic form, ln V = v A v + b v, 10

14 where v = (l m, l f, y), and A and b are a symmetric 3 3 matrix and a three-dimensional vector of unknown parameters. This functional form is a good compromise between flexibility and ease of estimation because it is locally second order flexible and linear in its parameters (Callan et al. 2009). Preferences variation across couples is introduced through the linear coefficient of each spouse s leisure (i.e. the coefficients on l m and l f ) b i = Xi β i + ζ i, i = m, f, where X i are vectors of exogenous individual and household characteristics, β i are vectors of unknown parameters, and ζ i are random errors accounting for unobserved heterogeneity in the preferences. The covariates in X i include a second-order polynomial in age of spouse i, number of children and an indicator for children aged less than 3 years. The errors ζ m and ζ f are assumed to be mutually independent, independent of other errors in the model, and normally distributed with mean zero and constant variance. The husbands hours of work densities g m (z m ) are taken to be uniform except for two distinct peaks corresponding to full-time jobs with 35 and 40 hours of work per week, while the wives hours of work densities g f (z f ) include an additional peak at part-time jobs with weekly hours of work. The baseline levels of these uniform densities are normalized to ensure that g m (z m ) and g f (z f ) are proper density functions. The job availability measures ln θ m and ln θ f are linearly related to gender-specific regional unemployment rates, two indicators for education attainment, plus interactions with the corresponding education scores. Wage rates are allowed to vary across jobs in different sectors but not across jobs within the same sector. To deal with the issue of unobserved wages, we estimate a hourly wage equation for each sector in an early stage. Estimation is done separately for husbands and wives using a variant of the Heckman two-step procedure to account for the selection bias (Dagsvik and Strøm 2006). In the first step, we estimate a multinomial choice model with three alternatives: non-participation, working in the public sector and working in the private sector. In the second step, we estimate a sector-specific hourly wage equation including the logarithm of the probability of working in that sector (with negative sign) as bias correction term. Identification is attained through a set of exclusion restrictions: a second-order polynomial in non-labor income, number of children, and indicators for children aged less than 3 years and working status of the spouse. The covariates in the second step include second-order polynomials in age and experience, two indicators for education attainment plus interactions with the corresponding final grades at school, gender-specific regional 11

15 unemployment rates, and two indicators for leaving in the Center and the South of Italy. Hourly wage predictions incorporate random errors for unobserved heterogeneity in wages. Thus, our model includes four additional random components η = (η m1, η m2, η f1, η f2 ) for prediction errors in the sector-specific hourly wages of the two spouses. Conditional on the random components ξ = (ζ m, ζ f, η m1, η m2, η f1, η f2 ) for unobserved heterogeneity in preferences and prediction errors in the hourly wages, our model is equivalent to a multinomial logit model with 169 alternatives. The errors in ξ are integrated out by taking the expectation of the conditional likelihood contributions π(z m, z f ξ) with respect to ξ. {ξ r : r = 1,..., R} denote R draws from the distribution of ξ. Model parameters are estimated via simulated maximum likelihood (SML) by approximating the unconditional likelihood contributions with simulated means of the conditional likelihood contributions π(z m, z f ξ r ) across the R draws. Provided that R goes to infinity faster than the square root of the number of observations, the resulting SML estimator is asymptotically equivalent to the exact maximum likelihood estimator (Hajivassiliou and Ruud 1994). 4.3 Labor supply predictions For any given tax-benefit system, labor supply choices are predicted through the four-step procedure suggested by Creedy and Duncan (2002). First, we draw 50 realizations from the Extreme value distribution and the estimated distribution of the parameter estimates such that, under the baseline tax-benefit system, predicted probabilities are optimal at the observed choices. 4 Let Second, we use the maximum probability rule to allocate each couple to the alternative with the highest predicted probability. Third, we repeat the first two steps after a policy reform and create transition frequencies using the same random draws selected in the first step. Forth, we compute point estimates and standard errors of labor supply responses taking the mean and the standard deviation of the transition frequencies across the 50 replications. 5 Data and microsimulation model The data are from the 2008 wave of the Survey on Household Income and Wealth (SHIW), a biannual household survey administrated by the Bank of Italy. The accounting rules defining the 4 For each couple, we continue drawing until the optimal choice under the baseline tax-benefit system corresponds to the observed choice. If after 750 trials a successful realization cannot be found, then labor supply is considered inelastic and hence kept fixed at the observed choices. 12

16 2008 Italian tax-benefit system are simulated through EconLav, a national microsimulation model which covers social security contributions, taxes and public transfers. 5 The working sample employed in our analysis consists of 1, 982 married couples in which the husband and the wife are both aged between 20 and 65 years and any of them is neither disabled, retired, in education, nor engaged in self-employment activities. The descriptive statistics presented in Appendix A show that labor supply is higher for husbands than for wives due to both the higher participation rate (93% and 53%, respectively) and the larger number of worked hours (40 and 32 hours per week, respectively). Hourly wages are also higher for husbands than for wives (13.34 and Euro, respectively) and for both spouses there is a public-private wage differential of about 4 Euro. The hours of work distributions in Figure 5 show that the peaks associated with full-time jobs in the private sector occur at 40 hours per week for both spouses. In the public sector, the peaks at 40 hours per week are sizeable, but the mode of the distribution is 38 hours for wives and 36 hours for husbands. In both sectors, there is also a conspicuous fraction of wives employed in part-time jobs with weekly hours of work. 6 Estimation results Table 1 compares the estimates of our multi-sectoral model of labor supply (labeled as Model 2) with those of a benchmark model (labeled as Model 1) which ignores sector-specific attributes of the jobs offered to the two spouses. 6 Both models are estimated via SML using R = 50 draws from the distribution of the random errors in ξ. The upper panel of Table 1 refers to estimated coefficients of the utility function, including the interactions between the linear coefficient of each spouses s leisure and socio-demographic characteristics. The estimates of these coefficients do not differ much between Models 1 and 2, their standard errors are remarkably small, and the random effects for unobservable heterogeneity in the preferences do not seem to play an important role. The first and the second order conditions in van Soest (1995) implies that the deterministic part of the utility function is increasing in disposable household income and quasi-concave for all sample observations. A positive coefficient on the interactions between leisure terms and socio-demographic characteristics can be interpreted as a positive effect on the marginal utility of leisure or, equivalently, as a negative effect on labor supply. Accordingly, we find that labor supply has an inverted U-shaped age profile with a maximum at 5 Additional information on the Italian microsimulation model EconLav can be found in Cipollone et al. (2013). 6 These estimates are conditional on preliminary two-step estimates of the sample selection models for hourly wages which are presented in Appendix B. 13

17 38 years for husbands and 41 years for wives. The number of children has a negative impact on wives labor supply, while the presence of young children has negligible effects. These findings line up with earlier results in Aaberge et al. (1999) and Figari (2011) and they partly reflect a widespread cultural tradition of Italian married couples with breadwinner men and women taking care of children. Estimated peaks on the hours of work densities suggest that jobs are concentrated around full-time working opportunities with 35 and 40 hours of work. The full-time peak at 35 hours is statistically significant only for husbands working in the public sector, while the full-time peak at 40 hours is statistically significant for both spouses and sectors. One can also notice that estimated peaks from Model 1 are always in the range of the sector-specific estimates from Model 2. Estimated coefficients on the job availability measures confirm that couples living in regions with high unemployment rates face a less favorable labor market environment. Except for husbands working in the private sector, the relative number of jobs offered to the two spouses is positively related to the indicators for secondary and tertiary education attainment. The quality of education, as measured by the interactions between the indicators for educational attainment and the corresponding final grades at school, has a positive effect on the job availabilities of wives in the public sector. Table 2 presents own and cross elasticities of the participation probability (PP) and the unconditional hours of work (HW) computed by increasing the hourly wages of all job opportunities by 10%. The elasticities obtain from Model 2 are higher than those obtained from Model 1 mainly because of differences at the extensive margin of labor supply and at the bottom of the disposable household income distribution. A plausible justification for this result is that our multi-sectoral framework exploits a finer representation of the choice sets available to the two spouses. As a consequence, this model allows to capture labor market transitions from non-participation to job opportunities that are indeed excluded from the choice sets of Model 1. Coherently with previous empirical studies, we find that the elasticities from both models are considerably higher for wives than for husbands, are decreasing across deciles of the disposable household income distribution, and are mainly driven by changes at the extensive margin of labor supply. Own elasticities of the participation probability are equal to 0.48 for husbands and 2.32 for wives, whereas own elasticities of the unconditional hours of work are equal to 0.50 and 2.71, respectively. Cross elasticities are negative and smaller in absolute value than own elasticities ( 0.26 for husbands and 0.55 for wives). 14

18 7 Simulation of EITC and WTC reforms As explained in Section 3, our policy reforms involve substituting the Italian FA program with a new in-work benefit program based on either the EITC or the WTC. Overall, we carry out four policy simulations for each in-work benefit scheme. In the first scenario, the benefit premium for twoearner households is fixed to 1. Other things being equal, this means that two-earner households are entitled to a double maximum benefit with respect to one-earner households. In the other three scenarios, we analyze sensitivity of our results to the benefit premium for two-earner households by setting this policy coefficient equal to 0, 0.5 and 1.5. The remaining coefficients of the EITC and WTC schedules are always determined to ensure tax revenue neutrality for each household type after accounting for labor supply responses of the two spouses. Labor supply effects of the simulated reforms are measured through the relative variations of the participation probability (PP) and the unconditional hours of work (HW), while redistributive effects are measured through the relative variations of disposable household income (DHI), poverty head count ratio (HCR), and poverty gap ratio (PGR). 7 Tables 3 and 4 present the effects of the first EITC and WTC reforms on the participation probability and the unconditional hours of work of husbands and wives by model specification, household type, decile of disposable household income and sector. Labor supply effects for wives are positive and mainly concentrated at the extensive margin, while labor supply effects for husbands are negative (although not always significant) and mainly concentrated at the intensive margin. According to the estimates from Model 2, the participation probability of wives increases by 2.19% under the EITC reform and by 1.28% under the WTC reform. The effects on the participation probability of husbands are instead negligible. After accounting for labor supply responses at the intensive margin, the incentive effects generated by the two reforms are closer because of the WTC benefit premium for working full-time. The unconditional hours of work of wives vary by 1.83% under the EITC reform, and by 1.10% under the WTC reform. Heterogeneity of labor supply responses across population groups seems to be substantial. The incentive effects of both reforms are strongly increasing with household size and decreasing with disposable household income. Under the EITC (WTC) reform, the relative variation of the unconditional hours worked by wives ranges from a minimum of 0.28% (0.19%) for childless couples to a maximum of 8.03% (7.37%) for couples with three children. Since these reforms are calibrated 7 The HCR is defined by the proportion of households with disposable income below the poverty line, while the PGR is defined by the mean gap between poverty line and disposable household income as a proportion of the poverty line by counting the non-poor as having zero poverty gap. 15

19 to ensure no financial transfer across household types, this positive gradient mostly reflects the generosity of FA in favor of large households. The negative gradient with respect to disposable household income reflects instead two effects: the decreasing labor supply elasticities across income deciles and the redistributive nature of our reforms that impose more binding income cut-off points than the FA program. As a consequence, most labor supply responses are found at the bottom of the disposable household income distribution where wives increase their unconditional hours of work by 14.78% under the EITC reform and by 10.66% under the WTC reform. Models 1 and 2 lead to qualitative similar policy considerations, but our multi-sectoral model (i.e. Model 2) leads to a substantially higher labor supply effects. These results are consistent with our previous findings on the labor supply elasticities. The most striking discrepancies between the two labor supply models occur again for the participation probabilities of wives, especially those living in large households and at the bottom of the disposable household income distribution. The redistributive effects in Table 5 suggest that our in-work benefit reforms may also improve upon the baseline tax-benefit system in terms of equity. The mean of disposable household income increases by 0.41% under the EITC reform and by 0.45% under the WTC reform. Moreover, such effects are increasing with household size. The effects of the EITC on poverty are less clear-cut: the HCR decreases by 0.67%, but the PGR increases by 6.44%. Hence, we find a reduction in the number of poor households but also a deterioration in the living standards of those who remain poor. This undesirable effect on the PGR may be a consequence of the phase-in region of the EITC schedule which implies that households with disposable income far below the poverty line are usually entitled to a lower benefit than that received with the original FA program. Uncontroversial results on poverty reduction are achieved by the WTC reform which leads to a reduction of 8.99% of the HCR and 2.86% of the PGR. Table 6 presents the results of our sensitivity analysis on the benefit premium for two-earner households. Here, we consider four alternative scenarios of the EITC and WTC schemes by varying p c between 0 and 1.5 with step 0.5. The income cut-off points G tc, t = 1, 2, 3, and the WTC benefit premium for working full-time q c are left unchanged, while the maximum benefit entitlements Ec and Wc are always determined to ensure tax revenue neutrality for each household type. As expected, higher values of the benefit premium for two-earner households lead to stronger labor supply effects and weaker redistributive effects. Hence, under tax-revenue neutrality, the choice of this policy coefficient generate a trade-off between these two competitive objectives. The EITC and WTC scenarios with p c = 0 are particularly interesting as they correspond to standard in- 16

20 work benefit schedules with no benefit premium for two-earner households. In this setting, the participation probability of wives decreases by 1.03% under the EITC reform and by 0.92% under the WTC reform. For the other simulation scenarios, we find that labor supply effects for wives are increasing with the size of the benefit premium for two-earner households. However, too large values of this policy coefficient may also induce harmful effects on poverty. 8 Conclusions In this paper, we have provided an ex-ante evaluation on labor supply and redistributive effects of EITC and WTC in-work benefits for Italian married couples using a tax-benefit microsimulation model and a structural model of labor supply. The novel contribution to the literature on in-work benefits is twofold. First, we have augmented the standard design of EITC and WTC in-work benefits with a new benefit premium for two-earner households to avoid the potential disincentive effects that these family-based schemes typically generate on secondary earners. Unlike the usual dichotomization between family and individual-based in-work benefits, this policy design provides a smooth mechanism for handing the well-known trade-off between labor supply and redistributive effects within a single welfare instrument. Second, the ex-ante evaluation of our in-work benefit reforms is based on a multi-sectoral discrete choice model of labor supply that jointly accounts for nonlinear and nonconvex budget sets, observed and unobserved heterogeneity, hourly wage differentials among jobs of the public and the private sectors, prediction errors in the sectorspecific hourly wages, and sector-specific quantity constraints from the demand side on the number and the type of jobs offered to the two spouses. Comparisons with a simpler model of labor supply that ignores sector-specific attributes of the various job opportunities suggest that the finer representation of the choice set provided by a multi-sectoral framework has important consequences on the estimates of labor supply responses. Our revenue neutral policy simulations reveal that in-work benefits with a suitable scheme of incentive for secondary earners may greatly improve the current Italian tax-benefit system in terms of both incentive and redistributive effects. The EITC is more effective than the WTC in boosting employment, while the WTC is more effective than the EITC in fighting poverty. In both schemes, the trade-off between incentive and redistributive effects is crucially related to the new benefit premium for two-earner households. Whether this trade-off can be formalized as a problem of optimal mechanism design is a interesting research topic which is left to future work. 17

21 References Aaberge, R., Colombino, U., Strøm, S., Labour supply in Italy: An empirical analysis of joint household decisions with taxes and quantity contraints. Journal of Applied Econometrics 14, Bargain, O., Orsini, K., In-work policies in Europe: Killing two birds with one stone? Labour Economics 13, Blundell, R., Work incentives and in-work benefit reforms: A review. Oxford Review of Economic Policy 16, Blundell, R., Duncan, A., Costas, M., Evaluating the working families tax credit, background Paper for structural versus non-structural approaches to evaluation, Social Policy Monitoring Network, IFS. Blundell, R., Duncan, A., McCrae, J., Costas, M., The labor market impact of the working families tax credit. Fiscal Studies 21, Blundell, R., Hoynes, H., Has in-work benefit reform helped the labour market? In: Blundell, R., Card, D., Freeman, R. (Eds.), Seeking a Premier League Economy. University of Chicago Press, Chicago, pp Brewer, M., Duncan, A., Shepard, A., Suarez, M., Did working families tax credit work? The impact of in-work support on labour supply in Great Britain. Labour Economics 13, Callan, T., van Soest, A., Walsh, J., Tax structure and female labour supply: Evidence from Ireland. Labour 23, Cipollone, A., Coromaldi, M., Curci, N., De Luca, G., Depalo, D., Rossetti, C., The non-behavioral module of the italian microsimulation model econlav. Working Paper , Ministry of Economy and Finance. Colonna, F., Marcassa, S., Taxation and labor force participation: The case of Italy, mimeo. Creedy, J., Duncan, A., Behavioural microsimulation with labour supply responses. Journal of Economic Survey 16, Dagsvik, J., Jia, Z., Orsini, K., Camp, G. V., Subsidies on low-skilled workers social security contributions: the case of Belgium. Empirical Economics 40, Dagsvik, J., Strøm, S., Sectoral labour supply, choice restrictions and functional form. Journal of Applied Econometrics 21, Dickert, S., Houser, S., Scholz, J., The earned income tax credit and transfer programs: A study of labor market and program participation. In: Poterba, J. M. (Ed.), Tax policy and the economy. National Bureau of Economic Research and the MIT Press, Cambridge, MA, pp Eissa, N., Hoynes, H., Taxes and the labor market participation of married couples: The earned income tax credit. Journal of Public Economics 88, Eissa, N., Hoynes, H., Redistribution and tax expenditures: The earned income tax credit. National Tax Jornal 64, Eissa, N., Liebman, J., Labor supply responses to the earned income tax credit. Quarterly Journal of Economics 111,

22 Figari, F., From housewives to independent earners: Can the tax system help Italian women to work? Working Paper , ISER. Francesconi, M., Rainer, H., van der Klaauw, W., The effects of in-work benefit reform in Britain on couples: Theory and evidence. Economic Journal 119, Francesconi, M., van der Klaauw, W., The socioeconomic consequences of in-work benefit reform for British lone mothers. Journal of Human Resources 42, Giordano, R., Depalo, D., Pereira, M., Eugéne, B., Papapetrou, E., Perez, J., Reiss, L., Roter, M., The public sector pay gap in a selection of EURO area countries. Working Paper Series 1406, European Central Bank. Gregg, P., Harkness, S., Smith, S., Welfare reform and lone parents in the UK. Economic Journal 119, Hajivassiliou, V., Ruud, P., Classical estimation methods for LDV models using simulation. In: Engle, R., McFadden, D. (Eds.), Handbook of Econometrics IV. North-Holland, New York, pp Immervoll, H., Pearson, M., A good time for making work pay? Taking stock of in-work benefits and related measures across the OECD. Social, Employment and Migration Working Papers 81, OECD. Leigh, A., Earned income tax credits and supply: Evidence from a British natural experiment. National Tax Journal 60, Liebman, J., The impact of the earned income tax credit on incentives and the income distribution. In: Poterba, J. (Ed.), Tax Policy and the Economy. Vol. 12. MIT Press, Cambridge. Meyer, B., Rosenbaum, D., Welfare, the earned income tax credit, and the labor supply of single mothers. Quarterly Journal of Economics 116, Neumark, D., Wascher, W., Using the EITC to help poor families: New evidence and a comparison with the minimum wage. National Tax Journal 54, van Soest, A., Structural models of family labor supply: A discrete choice approach. Journal of Human Resources 30, Vermeulen, F., A collective model for female labour supply with non-participation and taxation. Journal of Population Economics 19,

23 Table 1: SML estimates of labor supply models for married couples Model component Variable Model 1 Model 2 Preferences (DHI/1000) ** (Husband s leisure) ** ** (Wife s leisure) * (DHI/1000) (Husband s leisure) ** ** (DHI/1000) (Wife s leisure) * (Husband s leisure) (Wife s leisure) * ** (DHI/1000) (Husband s leisure) ** ** (Age/10) * (Age/10) ** * No. children Children aged (0 3) Standard deviation ζ m (Wife s leisure) ** ** (Age/10) * * (Age/10) ** ** No. children ** ** Children aged (0 3) Standard deviation ζ f Hours of work Husband Full-time peak (35h) ** densities Full-time peak (40h) ** Wife Part-time peak Full-time peak (35h) Full-time peak (40h) * Husband - public sector Full-time peak (35h) ** Full-time peak (40h) ** Husband - private sector Full-time peak (35h) Full-time peak (40h) ** Wife - public sector Part-time peak Full-time peak (35h) Full-time peak (40h) * Wife - private sector Part-time peak * Full-time peak (35h) Full-time peak (40h) * 20

24 Table 1: SML estimates of labor supply models for married couples (continued). Model component Variable Model 1 Model 2 Job availability Husband Constant ** measures Secondary educ ** Secondary educ. score Tertiary educ * Tertiary educ. score Reg. unemployment ** Wife Constant ** Secondary educ ** Secondary educ. score Tertiary educ ** Tertiary educ. score * Reg. unemployment ** Husband - public sector Constant ** Secondary educ ** Secondary educ. score Tertiary educ ** Tertiary educ. score Reg. unemployment Husband - private sector Constant ** Secondary educ Secondary educ. score Tertiary educ Tertiary educ. score Reg. unemployment ** Wife - public sector Constant Secondary educ ** Secondary educ. score Tertiary educ ** Tertiary educ. score ** Reg. unemployment ** Wife - private sector Constant ** Secondary educ ** Secondary educ. score Tertiary educ ** Tertiary educ. score Reg. unemployment ** Notes: Model 1 ignores sector-specific attributes of the various jobs, while Model 2 distinguishes between jobs in the public and the private sectors. SML estimates are based on R = 50 draws from the distribution of the random effects in each model. DHI denotes disposable household income. Asterisks denote * a p-value between 5% and 1% and ** a p-value below 1%. 21

25 Table 2: Labor supply elasticities of husbands and wives by decile of disposable household income. Husbands Wives Model Outcome Decile Baseline Own Cross Baseline Own Cross 1 PP I-III ** ** ** ** IV-VII ** ** ** ** VIII-X ** ** ** Total ** ** ** ** HW I-III ** ** ** ** IV-VII ** ** ** VIII-X * ** ** ** Total ** ** ** ** 2 PP I-III ** ** 0.40 IV-VII ** ** ** ** VIII-X ** ** ** ** Total ** ** ** ** HW I-III ** ** 0.24 IV-VII ** ** ** ** VIII-X ** ** ** ** Total ** ** ** ** Notes: Elasticities are computed increasing the hourly wages of all jobs by 10%. PP and HW denote the participation probability (in percentage points) and the unconditional hours of work per week, respectively. Symbols denote a relative variation with respect to the baseline, * a p-value between 5% and 1% and ** a p-value below 1%. 22

26 Table 3: Effects of EITC and WTC reforms on participation probabilities of husbands and wives by model specification, household type, decile of disposable household income and sector. Husbands Wives Model Variable Baseline EITC WTC Baseline EITC WTC 1 0 children ** 0.15 ** 1 child ** ** 0.38 ** 2 children ** 0.91 ** 3 children ** 2.79 ** I-III * ** 4.25 ** IV-VII ** 0.58 ** VIII-X ** * Total * ** 0.51 ** 2 0 children ** 0.27 ** 1 child ** 0.58 ** 2 children * * ** 2.93 ** 3 children ** 7.57 ** I-III ** ** IV-VII ** 1.53 ** VIII-X ** ** ** Public ** 0.45 ** ** 0.84 ** Private ** ** ** 1.57 ** Total ** 1.28 ** Notes: Symbols denote a relative variation with respect to the baseline, * a p-value between 5% and 1% and ** a p-value below 1%. 23

27 Table 4: Effects of EITC and WTC reforms on unconditional hours of work of husbands and wives by model specification, household type, decile of disposable household income and sector. Husbands Wives Model Variable Baseline EITC WTC Baseline EITC WTC 1 0 children ** ** 0.09 ** 1 child ** ** 0.26 ** 2 children ** ** ** 0.85 ** 3 children ** 2.60 ** I-III ** ** 4.10 ** IV-VII ** ** ** 0.48 ** VIII-X ** ** ** ** Total ** ** 0.41 ** 2 0 children * ** ** 0.19 ** 1 child ** ** 0.36 ** 2 children ** ** ** 2.82 ** 3 children ** 7.37 ** I-III ** ** IV-VII ** ** ** 1.32 ** VIII-X ** ** ** ** Public ** 0.42 ** ** 0.72 ** Private ** ** ** 1.33 ** Total ** ** ** 1.10 ** Notes: Symbols denote a relative variation with respect to the baseline, * a p-value between 5% and 1% and ** a p-value below 1%. 24

28 Table 5: Redistributive effects of EITC and WTC reforms by model specification and household type. Household Model Outcome type Baseline EITC WTC 1 DHI 0 children ** 0.09 ** 1 child ** 0.36 ** 2 children ** 0.60 ** 3 children ** 1.50 ** Total ** 0.37 ** HCR 0 children ** ** 1 child ** ** 2 children ** ** 3 children ** ** Total ** ** PGR 0 children ** ** 1 child ** ** 2 children ** 1.99 ** 3 children ** 1.69 ** Total ** ** 2 DHI 0 children ** 0.09 ** 1 child ** 0.30 ** 2 children ** 0.86 ** 3 children ** 2.27 ** Total ** 0.45 ** HCR 0 children ** ** 1 child ** ** 2 children ** ** 3 children ** ** Total ** ** PGR 0 children ** 1 child ** ** 2 children ** 0.92 ** 3 children ** 0.51 Total ** ** Notes: DHI denotes (mean) disposable household income, HCR denotes head count ratio, and PGR denotes poverty gap ratio. Symbols denote a percentage variation with respect to the baseline, * a p-value between 5% and 1% and ** a p-value below 1%. 25

29 Table 6: Sensitivity analysis on labor supply and redistributive effects of the EITC and WTC reforms to the benefit premium for two-earner households (Model 2 only). Husbands Wives Redistributive effects Policy p c PP HW PP HW DHI HCR PGR Baseline EITC ** ** ** ** ** ** ** ** 0.68 ** 0.37 ** 0.04 ** ** ** ** 2.19 ** 1.83 ** 0.41 ** ** 6.44 ** ** 2.94 ** 2.68 ** 0.63 ** 7.94 ** ** WTC ** ** ** ** ** ** ** ** ** ** ** ** ** 1.28 ** 1.10 ** 0.45 ** ** ** ** 2.25 ** 2.00 ** 0.72 ** ** 1.16 ** Notes: p c denotes the size of the benefit premium for two-earner households. Symbols denote a percentage variation with respect to the baseline, * a p-value between 5% and 1% and ** a p-value below 1%. 26

30 Figure 1: The 2011 EITC and WTC schedules for couples. EITC WTC EITC children 1 child 2 children 3+ children WTC No childcare 2000 Euro childcare Household earned income Gross household income Notes: WTC is based on basic, second adult, and 30 hour elements assuming a hourly wage of 6 Euro and zero non-labor household income. EITC, WTC, household earned income and gross household income are expressed in 1000 Euro using an exchange rate of.7937 for dollars and for pounds. 27

31 Figure 2: The 2008 Italian family allowances by household type. Benefit children 1 child 2 children 3 children Gross household income Notes: Benefit amounts and gross household income are expressed in 1000 Euro. 28

32 Figure 3: Simulated EITC schedule by household type. 0 children 1 child Benefit One earner Two earners Benefit One earner Two earners Gross household income Gross household income 2 children 3 children Benefit One earner Two earners Benefit One earner Two earners Gross household income Gross household income Notes: The benefit premium for two-earner households is equal to p c = 1 for all household types. Benefit amounts and gross household income are expressed in 1000 Euro. 29

33 Figure 4: Simulated WTC schedule by household type. 0 children 1 child Benefit Basic el. Second adult el. 30 hours el. Benefit Basic el. Second adult el. 30 hours el Gross household income Gross household income 2 children 3 children Benefit Basic el. Second adult el. 30 hours el. Benefit Basic el. Second adult el. 30 hours el Gross household income Gross household income Notes: The benefit premiums for two-earner households and for working full-time are equal to p c = 1 and q c =.50, respectively, for all household types. The hourly wages of both spouses are fixed to 6 Euro and non-labor household income is fixed to zero. Benefit amounts and gross household income are expressed in 1000 Euro. 30

34 Appendix A: Summary statistics Table 7: Mean and standard deviation (SD) of the main variables by gender. Husband Wife Variable Mean SD Mean SD Weekly hours of work Participation rate - Public or private Public Private Weekly hours of work (among workers) - Public or private Public Private Gross hourly wage (among workers) - Public or private Public Private Age Education attainment - Primary Secondary Tertiary Years of experience Number of children - No children child children children Notes: Data are from the 2008 wave of SHIW. Our working sample consists of 1982 married couples. 31

35 Figure 5: Weekly hours of work distributions of husbands and wives by sector. Husbands Public Husbands Private Weekly hours of work Weekly hours of work Wives Public Wives Private Weekly hours of work Weekly hours of work 32

36 Appendix B: Hourly wage equations Table 8: Two-step estimates of sample selection models for log hourly wages of husbands. Model 1 Model 2 Variable Sel. Part. Wage Sel. Part. Sel. Pub. Wage Pub. Wage Pri. Age * ** ** ** ** * Age squared Experience ** ** Experience squared ** ** ** ** ** Secondary educ ** ** ** ** ** ** Secondary educ. score Tertiary educ ** ** ** ** ** Tertiary educ. score Center South * Unemployment ** * ** Non-labor income ** ** Non-labor income squared Wife work Wife public * ** Wife private Number of children Child < 3 yrs ln(p 2) ** ln(p 3 ) Constant ** ** ** ** ** ** Standard deviation η m ** Standard deviation η m ** Standard deviation η m ** Notes: Standard errors are computed using 500 nonparametric bootstrap replications. Asterisks * denote a p-value between 5% and 1%, and ** a p-value below 1%. 33

37 Table 9: Two-step estimates of sample selection models for log hourly wages of wives. Model 1 Model 2 Variable Sel. Part. Wage Sel. Part. Sel. Pub. Wage Pub. Wage Pri. Age ** ** ** * Age squared ** ** Experience ** ** ** Experience squared ** ** ** Secondary educ ** ** ** ** ** Secondary educ. score Tertiary educ ** ** ** ** ** ** Tertiary educ. score ** * Center * * * South Unemployment * ** * Non-labor income ** * Non-labor income squared Husband work * Husband public * Husband private * Number of children * Child < 3 yrs * * ln(p 2 ) ln(p 3) Constant ** ** ** ** ** ** Standard deviation η f ** Standard deviation η f ** Standard deviation η f ** Notes: Standard errors are computed using 500 nonparametric bootstrap replications. Asterisks denote a * p-value between 5% and 1%, and a ** p-value below 1%. 34

38 Ministry of Economy and Finance Department of the Treasury Directorate I: Economic and Financial Analysis Address: Via XX Settembre, Rome Websites: dt.segreteria.direzione1@tesoro.it Telephone: Fax: