Essays on household and public finance
- Autores: Isabel Mico Millan
- Directores de la Tesis: Juan José Dolado (dir. tes.), Evi Pappa (codir. tes.)
- Lectura: En la Universidad Carlos III de Madrid ( España ) en 2023
- Idioma: inglés
- Tribunal Calificador de la Tesis: Miguel Almunia Candela (presid.), Nuarpear Lekfuangfu (secret.), Effrosyni Adamopoulou (voc.)
- Programa de doctorado: Programa de Doctorado en Economía por la Universidad Carlos III de Madrid
- Materias:
- Enlaces
- Tesis en acceso abierto en: e-Archivo
- Resumen
Households are central actors in the economy. First, studying the complex and heterogeneous nature of household financial decisions is key to better understanding the impact of different public policies or regulations in a world of imperfect markets. Second, partly because of recent developments in behavioral economics, economic research has developed a more profound appreciation of the role of economic expectations in shaping household financial behavior and their implications for the propagation of macroeconomic shocks. The first chapter of this thesis exploits rich household survey data from the Spanish Survey of Household Finances to examine how inheritance and gift taxes affect wealth mobility in Spain. The second chapter uses the same survey data to understand how marital property regimes influence household financial investment. The last chapter analyses the effects of regional income shocks on household economic sentiment and durable purchases and highlights optimism as a propagation mechanism of local demand shocks.
In the first chapter, "The Effects of Inheritance and Gift Taxation on Upward Wealth Mobility at the Bottom: Lessons from Spain" I study the impact of inheritance and gift (IG) taxation on intragenerational wealth mobility by exploiting rich regional variation in effective tax rates in Spain. At the heart of the ongoing debate on the sharp rise in wealth inequality is the use of inheritance and inter-vivos gift taxation as one of the main available policy tools to redistribute wealth and guarantee equal opportunities (OECD, 2021; Piketty et al., 2013). Yet, empirical research on this topic is very limited since isolating the causal impact of IG taxation on wealth distributional outcomes is challenging due to identification and measurement issues. First, inheritance and gift tax reforms that could be used in a quasi-experimental setting are rare. Second, even if they have occurred, rich administrative or survey data containing detailed information on heirs¿ and donees¿ wealth has often been unavailable to researchers. I overcome these two challenges by using Spain as a laboratory, a country that offers quasi-experimental regional variation in effective tax rates and rich household panel data on wealth.
The Spanish IG tax is designed at the national level with a progressive tax schedule. In 1996 the ad- ministration and regulation of this tax were decentralized to regional governments, which were awarded regulatory power to introduce tax credits and deductions as well as to modify the marginal tax schedule at their will. Regions started to exercise this right in the mid-2000s resulting in substantial regional variation in the effective tax rates for any bracket due to differences in (i) the timing of the tax reforms, (ii) the number of tax brackets affected, and (iii) the magnitude of the tax discounts introduced. I collect information on all regional IG tax reforms between 2002-2018 relying on different official data sources. With this novel information, I construct a tax simulator for inheritance and gift taxes for all Spanish regions which was previously unavailable. Next, I apply this tax calculator to household panel data from the Spanish Survey of Household Finances (or EFF for its acronym in Spanish) which contains detailed information on the wealth and debt of Spanish households, including information on pre-tax inheritances and inter-vivos gifts amounts and their asset composition. I estimate the average treatment effect of inheritance and gift tax changes, as well as their dynamics, using an event-study specification. By comparing heirs and donees who pay taxes in different regions, I find that higher inheritance taxes have a negative impact on net wealth mobility, but only at the bottom of the wealth distribution. Specifically, a one percentage point increase in inheritance tax rate makes households below the 50th net wealth percentiles between 0.01 to 0.33 percent less likely to improve their position in the net wealth distribution. For heirs at the very bottom of the wealth distribution, these point estimates represent a wealth mobility decrease of 36 to 77 percent in the years after the tax payment relative to their average pre-inheritance wealth mobility. Interestingly, this negative effect is persistent, remaining statistically significant during 3 to 6 years after the inheritance receipt for heirs at the first two percentiles. Instead, gift taxes on cash transfers do not seem to affect differently wealth mobility at any part of the wealth distribution. Next, I investigate more deeply the empirical drivers behind these wealth mobility dynamics by studying debt and gross wealth responses to inheritance taxation for different groups of households depending on their position within the wealth distribution before the tax payment. I provide evidence that a one percentage point increase in the inheritance tax rate decreases heirs¿ gross wealth by 9 to 12 percent in the years after the tax payment for households at the bottom of the wealth distribution. This negative effect of taxes on gross wealth is mostly driven by a reduction in their financial wealth, particularly in liquid assets, that goes in parallel with a rise in personal credit debt-to-wealth ratio by 3.2 and 4.7 percentage points in the years after the tax payment.
These results altogether suggest that the negative effects of inheritance taxes on bottom-wealth mobility are mostly explained by lower financial wealth and higher debt of these households. I argue that liquidity constraints and restricted access to financial instruments are relevant factors in explaining the positive effect of inheritance taxes on personal credit debt at the time of the tax payment. Households below the 40th net wealth percentile in Spain receive on average inheritances as large as 6 times their gross wealth (or 86 times their liquid assets) at the time of their receipt. The liquidity constraints faced by bottom-wealth households at the time of the tax payment are reinforced by several Spanish IG tax law mandates, which limit heirs¿ access to different financial instruments. First, heirs are required to pay taxes in the next 6 months following the death event to gain ownership of the deceased person¿s estate, which becomes frozen by the bank system and public registry on the same day of the death (including bank accounts and deposits). Heirs can ask for a tax payment moratorium and/or installment but this comes with an additional cost and does not grant access to the deceased person¿s estate until the tax payment is completed. Second, the Spanish bank system does not allow heirs to put the yet-to-be-inherited real estate assets as collateral for loans, which reduces the number of debt instruments available for liquidity-constraint households.
Although the singularities of the Spanish IG tax system help rationalize the rise of personal credit debt of less-wealthy heirs, it is less obvious why the detrimental effects of inheritance taxes on bottom-wealth mobility persist over time. In combination with this channel, I provide evidence that the illiquidity of inheritances and delays in selling inherited real estate property help explain the persistence of the negative effect of taxes on bottom-wealth mobility. To do so, I leverage regional variation in tax-induced restrictions to sell the inherited dwelling. In Spain, the inheritance tax law allows heirs to benefit from generous tax credits applicable to the deceased¿s main dwelling under the condition that inherited property must not be sold for a certain amount of years. Although the default law establishes a 10-year period, regions have reduced this time restriction since the mid-2000s resulting in plausibly exogenous variation in the delay to sell real estate property due to differences in (i) the timing and (ii) the magnitude of these time limit reductions. I show that the effects of inheritance taxes on less wealthy heirs¿ personal credit debt and financial wealth are stronger in regions with longer restrictions to sell inherited real estate without cost. These results suggest that delays in selling illiquid inherited assets amplify the negative effects of taxes on financial wealth and personal credit debt and, therefore, on wealth mobility.
The second chapter, "When Wives Command: Household Portfolio Choices and Marital Property Regime", coauthored with Lidia Cruces and Susana Párraga, studies the link between married couples¿ portfolio choices and property division rules. In many countries, two main property regimes - community and separate property - coexist. Under separate property, each spouse maintains sole ownership of assets accumulated during the marriage and takes them upon dissolution. Contrary, in community property, most assets acquired during the marriage become jointly owned and split between spouses if the marriage ends. The type of marital property regime has relevant implications for savings decisions because (i) they determine the allocation of spouses¿ savings ex-post marriage (Voena, 2015) and (ii) affect the economic cost of terminating the marriage (Imre, 2022).
An aspect that has received less attention in the literature is how property division rules interact with couples¿ financial portfolio choices. This paper fills this gap by investigating the impact of property division rules on household financial investment choices. The Spanish institutional setting serves as an ideal testing ground to address this question as the marital property regime law is regulated at the regional level, resulting in variation in the default rules across Spanish regions. Separate property is the default regime in Catalonia and the Balearic Islands, while some form of community property is the default in the rest of the regions. Couples adopt the default marital property regime in their region of residence unless spouses agree on a different one by signing a prenuptial agreement. By means of an IV strategy, we exploit this regional variation in marital law in combination with rich survey data from the Spanish Survey of Household Finances to provide causal estimates of the effects of property division rules on couples¿ financial portfolio choices. The Spanish Survey of Household Finances (or EFF for its acronym in Spanish) gathers rich information on Spanish households¿ wealth, debt, and demographics. Particularly relevant for our study, it contains detailed information on household financial investment by asset class (i.e., bank deposits, shares, bonds, etc.) and on the marital property regime when households consist of married couples.
We find that separate-property couples take significantly more financial risk when wives are most knowledgeable about household finances. The definition of the household head in the EFF makes it very likely that this household member is the primary decision-maker regarding the household economy and finances. Specifically, the household head is the spouse most knowledgeable about the household economy and investments, being able to give detailed information about household wealth and debt holdings. We find that separate property couples are 9% more likely to participate in risky assets than their counterparts married under community property when wives are the household heads. We also find that these couples hold more diversified portfolios towards risky assets than those married in community property. On average, couples married under separate property hold a share in risky asset classes 5 percentage points higher than couples married under community property when wives take a primary role in household finance investments.
To rationalize these findings, we develop a two-period model of financial portfolio choice where couples differ in their marital property regime. Households consist of two spouses who are born married and face an exogenous probability of divorce. The household head decides on the level of consumption, which is public within the household, and her savings in safe and in risky financial assets given her spouse¿s savings decisions, forming expectations about both spouses¿ future labor income, asset returns, and marital status. In the model, property division rules directly influence the asset allocation rule upon divorce and the corresponding dissolution costs of marital assets. When separate property couples divorce, spouses take their individual assets according to the title of ownership and face no dissolution cost of marital assets. In contrast, community property couples must incur dissolution costs of marital assets as household total savings need to be equally split between spouses. We introduce the dissolution cost of marriage for community-property couples by assuming that an exogenous fraction of total household permanent income is destroyed in the event of divorce (Bacher, 2021; Cubeddu and Ríos-Rull, 2003). Divorce represents a source of financial risk in the model because it requires couples to split their assets and because it results in a state with lower income levels and higher income risk. However, the strength of the precautionary savings motive differs across marital property regimes and their associated dissolution costs of marital assets.
We calibrate the model to match key moments of Spanish married couples¿ financial behavior for which wives are the most knowledgeable about household finances. By means of counterfactual simulations, we show that divorce risk and gender heterogeneity in labor income profiles are the most important determinants through which marital property regime affects portfolio choices between safe and risky assets. Relative to separate property, community property¿s higher marriage dissolution costs induce spouses to increase precautionary savings and lower their demand for risky assets. Low labor income levels and higher income risk for wives further strengthen couples¿ precautionary savings motive under divorce risk. This explains why the property regime gap in risky financial savings arises between couples for which wives are the primary decision-makers.
The third chapter, The Sentimental Propagation of Lottery Winnings: Evidence from the Spanish Christmas Lottery, co-authored with Morteza Ghomi and Evi Pappa, uses the exceptional nature of the Spanish Christmas lottery to estimate jointly the individual and aggregate effects of lottery wins that are shared among many people living in the same province and accentuates the role of consumer confidence for the transmission of these shocks. We show that lottery shocks impact significantly consumers¿ sentiment which brings demand effects that improve macroeconomic conditions in the winning regions.
The Spanish Lottery has three characteristics that are different from other lotteries: (i) Large size 4 and quantity of prizes each year, (ii) Clustering of prizes to individuals living in the same Spanish province, and (iii) High level of participation. Each winner of the first prize, known as El Gordo (the fatty), receives around 20,000 euros per euro played, and the standard ticket costs 20 euros. Moreover, winners of the second and third top prizes receive 6,250 and 2,500 per euro played, respectively. Usually, one lottery outlay sells most (if not all) of the series of a single number in the lottery. The winning provinces receive an income shock equivalent, on average, to 0.2 percent of their GDP. For the provinces that receive the maximum lottery prize per capita, the income shock represents, on average, around 3.4 percent of the provincial GDP. Finally, because sharing Christmas lottery tickets is a social tradition, the lottery has an extremely high participation rate that is reflected in the high prizes for the winners.
We employ data from the monthly consumer sentiment survey conducted by the Center of Sociological Research (CIS). Each month, around 1,000-1,500 nationally representative households across Spain are asked questions related to their past and intended consumption behavior and their current views and expectations about their own personal finances, their employment status, and the overall economic outlook of Spain. Following the University of Michigan Survey, we construct regional indices of confidence for the current (ICC) and expected macroeconomic conditions (ICE) and show using local projections (See, e.g., Jordà (2005)) that confidence reacts positively and significantly on impact to lottery wins at the regional level. To explore in depth the sentimental propagation of lottery wins, we use ordinal regression models to study the effects of the lottery win on individual sentiment and consumption behavior using the same survey data. Households in the winning regions become temporarily more optimistic about their current and future income and employment and tend to update upwards their expectations about the evolution of the Spanish economy. In line with the results found in the existing literature (see, e.g., Kuhn et al. (2011) and Attanasio et al. (2020)), we also find that households in winning provinces increase significantly their consumption of durable goods, in particular, the consumption of furniture and vehicles - relative to household residing out of these provinces - the first six months after the lottery win.
The increase in sentiment and consumption can be attributed to an income effect. We provide evidence that this is not the case by (i) showing that households in the winning regions do not significantly improve their ability to pay bills and (ii) isolating the effect of consumer sentiment on durable consumption by using lottery wins as IV for confidence. The increase in optimism could be also driven by news about future economic fundamentals rather than animal spirits. To discard such an interpretation, we first argue that rational agents should not improve their expectations about the evolution of the Spanish economy as we find in the data as lottery wins have stimulative effects only at the local level. This is further confirmed when we look at regions with active secessionist movements, like Catalonia and the Basque Country, where people can clearly distinguish between national and regional conditions. We find that households¿ sentiment in these regions does not react differently from other regions after a lottery win. Hence, the positive reaction of expectations about the Spanish economy can only be attributed to increased optimism rather than news about regional fundamentals. Finally, we show that business sentiment does not react significantly to the lottery shock. Next, we show that lottery wins affect more significantly the sentiment and intended consumption of young, less educated, unemployed, and low-income households and that the effect of lottery wins on sentiment is stronger during recessions.
Finally, we examine the dynamic effects of the Spanish lottery shock on macroeconomic conditions using monthly Spanish province-level data. We find that lottery wins have significant and economically important stimulative effects at the provincial level. On average, after a province wins a lottery of 1000 euros per capita the unemployment rate falls sluggishly reaching its maximum fall (-0.3 percentage points) after a year and it remains significantly low 20 months after the initial impact. We provide evidence that this drop in unemployment can be attributed to an improvement in employment conditions rather than a fall in labor force participation. Furthermore, the price level in the winning province increases persistently reaching its maximum 17 months after the shock, returning to its pre-shock value after approximately two years. We do not find evidence suggesting that lottery shocks affect rental prices and mortgages at the provincial level.
