Modelling charitable donations to an unexpected natural disaster: Evidence from the U.S. Panel Study of Income Dynamics

Journal of Economic Behavior & Organization 84 (2012) 97–110

Contents lists available at SciVerse ScienceDirect

Journal of Economic Behavior & Organization journal homepage: www.elsevier.com/locate/jebo

Modelling charitable donations to an unexpected natural disaster: Evidence from the U.S. Panel Study of Income Dynamics Sarah Brown a , Mark N. Harris b , Karl Taylor a,∗ a b

Department of Economics, University of Sheffield, United Kingdom Department of Econometrics and Quantitative Modelling, Curtin University, Australia

a r t i c l e

i n f o

Article history: Received 30 November 2010 Received in revised form 26 July 2012 Accepted 14 August 2012 Available online 29 August 2012 JEL classification: D19 H24 H41 H31 Keywords: Charity Donations System tobit Tobit

a b s t r a c t Using household-level data, we explore the relationship between donations to the victims of the 2004 Indian Ocean tsunami disaster and other charitable donations. The empirical evidence suggests that donations specifically for the victims of the tsunami are positively associated with the amount previously donated to other charitable causes. This relationship exists when we decompose overall charitable donations into different types of philanthropy, with charitable contributions to caring and needy organisations having the largest positive association with donations to the victims of the tsunami. Furthermore, when we explore the impact of donations to the victims of the tsunami on future donations to charity, there is evidence of a positive relationship with the largest association with donations to caring and needy organisations. Hence, there is no evidence to suggest that unplanned spending on donations to an unforeseen natural disaster diverts future expenditure away from donations to other charitable causes. © 2012 Elsevier B.V. All rights reserved.

1. Introduction and background A plethora of empirical and theoretical studies exist in the economics literature exploring why individuals make contributions to charity, with much of the existing research focusing on charitable donations in the U.S. (see, for example, Andreoni, 2006). Given the economic significance of such donations and government intervention via tax regulation, such interest is not surprising. Recent figures from Giving U.S.A. 2011, for example, estimate total charitable contributions in the U.S. in 2010 at $290.89 billion, indicating a 2.1 percent growth compared to 2009 adjusting for inflation.1 Over the last four decades, the literature on the economics of charity has focused on the decision to donate at the individual or household level, with much attention paid to the impact of tax deductibility on such decisions related to charitable giving and the associated price and income effects. The empirical analysis of charitable donations has been influenced by methodological advances with respect to econometric techniques as well as increased availability and quality of data. Andreoni (2006) presents a comprehensive survey of the influences on charitable donations established in the existing literature. For example, Auten et al. (2002) find that income is an important determinant of donor responsiveness, whilst, according to Glenday et al. (1986), donations are expected to vary over the lifecycle increasing with age. In a similar vein, Schokkaert (2006) finds that older and more educated individuals tend to give more. In general, the findings from existing

∗ Corresponding author. Tel.: +44 (0) 114 222 3420. E-mail address: k.b.taylor@sheffield.ac.uk (K. Taylor). 1 The figure relates to total charitable contributions from U.S. individuals, corporations and foundations and includes both cash and in-kind donations. 0167-2681/$ – see front matter © 2012 Elsevier B.V. All rights reserved. http://dx.doi.org/10.1016/j.jebo.2012.08.005

98

S. Brown et al. / Journal of Economic Behavior & Organization 84 (2012) 97–110

studies suggest that married households, households with dependent children, households with a female head and religious households are all expected to give more in absolute terms. Our focus lies in exploring the influences on charitable donations at the household, i.e. donor, level. As stated by Schokkaert (2006), who presents a comprehensive survey of the empirical literature on charitable giving, much of the existing research at the donor level focuses on total contributions made to charity without distinguishing between different recipient causes. In our empirical analysis, we aim to explore the relationship between donations specifically related to an unexpected adverse shock in the form of a natural disaster and donations to other charitable causes. Specifically, we focus on donations to the victims of the 2004 Indian Ocean tsunami. As stated by Athukorala and Resosudarmo (2005, p. 1), who analyse the immediate economic impact of the tsunami and disaster management in its immediate aftermath, ‘with a death toll of about 350 thousand, the Indian Ocean tsunami . . . is by far the worst natural disaster of that kind in the recorded human history.’ Brown and Minty (2008), who find that media coverage of disasters has a large impact on donations to relief agencies, cite five reasons for the high level of donations to U.S. charities for the tsunami disaster relief (estimated at $1.6 billion in private donations). Firstly, the time of year coincided with a holiday period which may have increased the ‘warm glow’ associated with charitable giving; secondly, South East Asia has been an increasingly popular destination for U.S. tourists; thirdly, tax incentives in the U.S. motivate charitable giving and the tsunami occurred just before the deadline (31st December) for 2004 tax deductions and, furthermore, the Tsunami Disaster Aid Tax Relief Act extended the deadline to 31st January 2005; fourthly, the provision for online giving was extensive; and, finally, there was extensive media coverage. The importance of one-off appeals for disaster relief as a means to raise significant funds from relatively small contributions made by many individuals was noted in an early contribution by Sugden (1982), who cites the Cambodia famine appeal in 1980 as an example. One concern surrounding such disaster appeals relates to the possibility of donations being diverted away from existing charitable causes towards such relief funds. More recently, Eckel et al. (2007) explore the impact of Hurricane Katrina upon charitable donations within the context of an experiment conducted in a laboratory environment. Such studies are particularly interesting in the context of claims put forward by Wright (2002) that the majority of donations in the U.S. are regarded as a planned activity whereas in the UK donations tend to be more spontaneous. Hence, the novelty of our contribution to the literature on charitable giving lies in exploring the relationship between donations to a specific unexpected natural disaster, namely the 2004 Indian Ocean tsunami, and donations to other charitable causes. In Section 2 we investigate what factors influence the level of donations to the victims of the tsunami, i.e. unplanned donations, including the role of donations to other charitable causes, i.e. planned donations. Section 3 expands the analysis to explore the relationship between different types of charitable donations and donations to the victims of the tsunami. In Section 4, we explore the effect of tsunami donations on future donations to other charitable causes. 2. Donations to the victims of the tsunami and other charitable donations 2.1. Data and methodology We use data from the U.S. Panel Study of Income Dynamics (PSID), which is a panel of households ongoing since 1968 conducted at the Institute for Social Research, University of Michigan.2 In the PSID waves 2001, 2003, 2005 and 2007, there are a series of detailed questions relating to giving to charity.3 Due to our focus on tsunami donations, we restrict ourselves to the 2005 PSID, which after conditioning for missing observations, yields a sample for analysis of 6590 households. Households are asked about total donations to charity over the calendar year 2004.4 Excluding donations specifically related to the 2004 Indian Ocean tsunami, the average total value of the amount of donations in the calendar year 2004 is $1557 (Table 1A), with 40 percent of households not making any donations.5 The average amount of donations amongst households who do donate to charity is $2577. As a separate category, heads of household were asked to indicate the total dollar value of donations made between the end of December 2004 and the month of interview, to help the victims of the 2004 Indian Ocean tsunami (which occurred on the 26th December 2004). The average amount of these donations was $28 (Table 1A), with 22 percent of households making such donations. The average amount of donations for the tsunami victims amongst those households who made

2 One key advantage of the PSID is that it includes households which itemize charitable donations in their annual tax return as well as those who do not. In contrast, some existing studies, such as Auten et al. (2002), analyse individual tax returns collected by the U.S. Internal Revenue Service. One drawback of such data, however, relates to the fact that the sample is restricted to those tax payers who itemized deductions. Consequently, the sample potentially suffers from sample selection bias given that itemizing charitable contributions leads to a lower price of making a donation (this is discussed in detail below). Furthermore, Wilhelm (2006) explores the quality of the PSID data on charitable donations in terms of two dimensions: missing data and the amounts reported. He compares the PSID charitable donations data with data on charitable deductions from the Internal Revenue Service and finds that the reported amounts generally compare well across the data sources except above the 90th percentile. He thus confirms that the PSID data on charitable donations are of ‘high quality.’ 3 The definition of a charitable organisation in the PSID includes ‘religious or non-profit organisations that help those in need or that serve and support the public interest’. It is clearly stated that the definition used does not include political contributions. 4 Specifically, in the PSID heads of household are asked to indicate the total dollar value of all donations made by themselves and their family where donations ‘include any gifts of money, assets or property/goods made directly to the organisation, through payroll deduction or collected by any other means on behalf of the charity.’ 5 All monetary values are given in 2005 prices, using a CPI deflator available from the Bureau of Labor Statistics http://www.bls.gov/cpi/cpi dr.htm#2007.

S. Brown et al. / Journal of Economic Behavior & Organization 84 (2012) 97–110

99

Table 1A Summary statistics – monetary variables. Mean

Std. dev.

Min

Max

Levels – 2005 $ Tsunami victims donation 2005 (T) Total other charitable donations 2004 (T − 1) Religious charitable donation 2004 (T − 1) Needy charitable donation 2004 (T − 1) Multi-purpose charitable donation 2004 (T − 1) Caring charitable donation 2004 (T − 1) Other charitable donations 2004 (T − 1) Total charitable donations 2006 (T + 1) Religious charitable donation 2006 (T + 1) Needy charitable donation 2006 (T + 1) Multi-purpose charitable donation 2006 (T + 1) Caring charitable donation 2006 (T + 1) Other charitable donations 2006 (T + 1) Household labour income Household wealth Household non-labour income Household permanent income (HPI) Variance of HPI

28 1557 815 116 120 149 32 1368 878 139 128 160 62 34,093 56,934 49,408 54,433 25,175

142 6386 2572 543 576 947 581 3494 2627 597 662 1032 530 67,060 447,129 651,434 53,331 47,903

0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 1313 289

6000 113,667 100,000 15,000 20,000 30,202 43,000 74,703 64,032 13,873 26,680 53,360 24,545 2,710,000 13,000,000 1,100,000 1,238,063 1,867,183

Log transformations Log tsunami victims donation 2005 (T) Log total other charitable donations 2004 (T − 1) Log religious charitable donation 2004 (T − 1) Log needy charitable donation 2004 (T − 1) Log multi-purpose charitable donation 2004 (T − 1) Log caring charitable donation 2004 (T − 1) Log other charitable donations 2004 (T − 1) Log total charitable donations 2006 (T + 1) Log religious charitable donation 2006 (T + 1) Log needy charitable donation 2006 (T + 1) Log multi-purpose charitable donation 2006 (T + 1) Log caring charitable donation 2006 (T + 1) Log other charitable donations 2006 (T + 1) Log household labour income Log household wealth Log household non-labour income Log HPI# Log variance of HPI

0.90 4.04 2.77 1.28 1.30 1.55 0.87 3.83 2.59 1.35 1.21 1.51 0.69 7.98 2.38 1.79 10.62 9.69

1.77 3.48 3.40 2.35 2.35 2.45 1.99 3.58 3.43 2.44 2.35 2.49 1.80 4.35 3.81 3.50 0.77 0.87

0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 7.18 5.67

8.69 11.64 11.51 9.62 9.90 10.32 11.52 11.22 11.07 9.54 10.19 10.88 10.11 14.81 16.38 16.21 14.03 14.44

Observations

6590

Note: # household permanent income.

such donations was $125.3.6 Fig. 1 presents the distributions of the (natural logarithm of the) total amount donated for the victims of the tsunami and total donations to other causes, both in the censored data, i.e. including non-contributors (since there are no donations between zero and unity, for households reporting a zero donation, the value is recoded to zero), and for all those who make positive contributions. Initially, we investigate what factors are associated with making a donation to an unforeseen disaster, i.e. to the victims of ∗ denote the underlying latent propensity to donate to the victims of the tsunami disaster of household i, the tsunami. Let tsiT at time T (that is post 26th December 2004 through to the month of interview in 2005). In linear form, this latent propensity, which is defined over the whole real number line (− ∞ , ∞), can be written as: ∗ tsiT = X i   + yi(T −1) + i

(1)

where X is a vector of covariates, which are thought to influence the level of tsunami donations and  is a normally distributed random error term. If this latent propensity is negative or zero, we observe individuals at the corner solution point of zero, ∗ ). Accordingly, this model is estimated as a univariate otherwise observed donations equal the latent propensity (tsiT = tsiT tobit model with censoring (from below) at zero (Maddala, 1983). We include household donations to other charities, yi(T−1) , in order to explore the relationship between donations to the victims of the tsunami and donations to other causes. The subscript T − 1 denotes the timing difference as compared

6 Given the nature of the question related to the tsunami donations, it is not surprising that there is some variability in the level of donations across the month of interview, which is as follows for those who do make such donations (all individuals): March interview, $137 ($32); April interview $118 ($27); May interview $119 ($27); June interview $109 ($25); July interview $77 ($13); August interview $299 ($46); September interview $129 ($25). Hence, we control for the month of interview throughout the empirical analysis that follows.

100

S. Brown et al. / Journal of Economic Behavior & Organization 84 (2012) 97–110

Fig. 1. The distributions of tsunami donations and all other charitable donations.

to tsunami donations in that other donations are made over the calendar year 2004. Hence, a novelty of our contribution to the literature lies in exploring the relationship between different types of giving within the context of a large sample of households drawn from the PSID. The following demographic variables, which have previously been employed in the literature (see, for example, Andreoni, 2006; Auten and Joulfaian, 1996), are included in X: dummy variables for the head of household’s age (with over 60 as the base category); the number of adults in the household; the number of children in the household; the years of completed schooling of the head of household; the gender of the head of household; the marital status of the head of household (with all states other than married or cohabiting as the base); whether the head of household is currently employed, self employed or unemployed (not currently in the labour market is the reference category); the natural logarithm of household wealth; the natural logarithm of household labour income, the natural logarithm of household non-labour income (including benefit income)7 ; whether the house is owned outright or with a mortgage (rental and other types of housing tenure form the base category); and the ethnicity of the head of household (where groups other than white and black form the reference category). Auten et al. (2002) highlight the importance of distinguishing between permanent and transitory income effects. Their findings suggest that persistent price and income changes have much larger impacts on charitable donations than transitory changes. Hence, we include a measure of permanent income. To construct this, we follow Wilhelm et al. (2008), averaging family income over the recent past (using up to eight years depending on whether the household was in the panel over the period). We also include a control for the variance in permanent income over the period. In an early contribution, Schwartz (1970) analyses the price of donating to charity, which is determined by taxation as income donated to recognised charities in the U.S. is not subject to income tax. As a consequence, disposable income falls by less than the full amount donated: the price of the donation becomes the donation net of the saving in tax since each dollar donated to a recognised charity leads to less than one dollar sacrificed for consumption purposes. U.S. tax laws do specify an upper bound to deductibility with a maximum deductible percentage of the income tax base: 50 percent of gross income in 2006. The extent of the tax saving is determined by which marginal tax bracket the individual is in (Schwartz, 1970). In the context of the U.S., individuals who itemise deductions in their tax return reduce their taxable income in accordance with the level contributed to tax-exempt organisations. Hence, tax deductibility affects the price of donating to charity (Auten et al., 2002). Thus, we also control for the price of making a donation to charity. For households who itemise charitable donations in their tax return, the price of the donation is defined as one minus the household’s marginal tax rate on the contribution made, whereas for households who do not itemise charitable donations, the price of the donation is one: donating one dollar means that there is one dollar less for consumption. Hence, for these households the price of making a donation is less than one, which is the price of donating for those households who did not itemise such donations.

7 For household labour income, household non-labour income and wealth, the natural logarithms of zero values have been recoded to zero as there are no values that lie between zero and unity for these variables.

S. Brown et al. / Journal of Economic Behavior & Organization 84 (2012) 97–110

101

In the PSID, households are asked to indicate whether they made an itemised deduction for charitable contributions. One issue, however, which has arisen in the existing literature, is that the decision to itemise is arguably not fully exogenous, i.e. the decision to itemise may be influenced by the level of charitable donations. To account for this, as is common in the existing literature (see seminal contributions by Clotfelter, 1980; Auten et al., 2002), we exclude ‘endogenous itemizers’ who are defined as those who have itemised but would not have done so in the absence of their actual charitable donations. Households which itemise are assigned the relevant tax rate using the National Bureau of Economic Research TAXSIM programme (http://www.nber.org/∼taxsim/), which calculates federal state tax liabilities for survey data based on a range of factors such as earnings, marital status and children. Due to an additional source of possible endogeneity relating to the price of a charitable donation being a function of both the donation and income, which has been discussed extensively in the existing literature, following Auten et al. (2002), we calculate the price variable firstly by assuming that charitable donations equal zero (i.e. the first dollar price) and then after including a predicted amount of giving set at 1 percent of average income. As stated by Auten et al. (2002, p. 376), ‘this procedure yields a tax price consistent with the actual costs of giving, but not endogenous to individual donation decision.’ Following the existing literature, we then take an average of the two price variables. Additional controls included in X are: a set of dummy variables controlling for the health status of the head of household over the last 12 months, poor health (the omitted category), fair health, good health, very good health, and excellent health; and religious denomination of the head of household (with no religion as the base category). We also include binary controls for the month of interview and a binary indicator signifying whether the household donated to a disaster cause in 2002 in an attempt to capture past generosity to charitable causes.8 Full summary statistics for the monetary variables used in our empirical analysis are presented in Table 1A. These values are given in 2005 prices in levels (U.S. dollars) and logarithmic transformations for both donations and other monetary variables (labour income, non-labour income, permanent income and wealth, all defined at the household level). Summary statistics for the non-monetary variables are reported in Table 1B, where the majority of household heads in the PSID sample are: male (67 percent); employees (73 percent); white (67 percent); protestant (64 percent); and own their home either outright or with a mortgage (61 percent). We then investigate what factors are associated with total charitable donations to other causes (that is, excluding dona∗ tions to the victims of the tsunami disaster). As before, let yi(T denote the latent, partially observed, propensity to donate −1) to all other charitable sources of household i; and again yi(T−1) is the observed realisation of this (zero) corner solution model. This is determined by household characteristics X (as defined above), with unknown weights, ˛. This model is also estimated as a univariate tobit specification of the form: ∗ = X i ˛ + ωi yi(T −1)

(2)

where ω is a normally distributed random error term. We can consider Eqs. (1) and (2) as a simultaneous, bivariate recursive tobit model of the form: ∗ = X i ˛1 + ε1i yi(T −1) ∗ = X  ˛ + y tsiT i(T −1) + ε2i i 2

(3)

where the error terms jointly follow a bivariate normal distribution such that (ε1 , ε2 )∼N(0, 0, 12 , 22 , ), with covariance being given by  12 =  1  2 , so that  =  1  2 / 12 .9 The use of such a systems approach to overcoming potential endogeneity (especially with regard to two, or more, discrete and/or censored endogenous variables) is commonplace in the literature (see, for example, Maddala, 1983; Greene, 2008). With regard to a systems tobit model, this approach dates back to Amemiya (1974); and specifically with regard to the recursive tobit systems approach, see, for example, Greene (2007, pp. E27–58) and Wooldridge (2010, p. 685). In the current context the potential issue of reverse causality is, however, somewhat unlikely, due to the timing differential between standard charitable donations and donations to the victims of the tsunami. Here, arguably, conventional giving is a pre-determined covariate (having previously occurred) and hence exogenous. Indeed, this reduces the potential for reverse causality, as argued by Angrist and Pischke (2009), since conventional giving is measured ex ante, that is, it predates the outcome variable. Our prior is thus that the error terms are uncorrelated, i.e.  = 0 (and hence  12 = 0), although in estimation we will not enforce this. A hypothesis test of  = 0 is a test of endogeneity of the donations to all other charities variable in the tsunami equation (Greene, 2008). Standard Wald and likelihood ratio tests can be used here. We note that identification of the tsunami equation is econometrically based solely on the functional form of the bivariate (recursive) tobit system (Maddala, 1983). Clearly it would be preferable to identify this additionally on data: however, there are no obvious candidates for instruments here that are likely to affect tsunami donations independent of those for all

8 The information in the PSID 2003 relates to charitable donations over the calendar year 2002. In their 2002 October Press Release, the United National Environment Development Programme stated that natural disasters were set to cost over $70 billion in 2002. It is apparent that a number of disasters occurred in 2002 such as heat waves in India, floods in China and tornadoes in central and south east U.S., although there was an absence of disasters on a scale even approaching the 2004 Indian Ocean tsunami. 9 Clearly, this significantly complicates the resulting likelihood function, but such procedures are available in commercial software such as Limdep/Nlogit and Stata.

102

S. Brown et al. / Journal of Economic Behavior & Organization 84 (2012) 97–110

Table 1B Summary statistics – non-monetary variables.

HoH: aged <20 HoH: aged 20–30 HoH: aged 30–40 HoH: aged 40–50 HoH: aged 50–60 HoH: aged >60 (reference category) HoH: years of schooling HoH: male HoH: married/cohabiting HoH: unmarried (reference category) HoH: employee HoH: unemployed HoH: self employed HoH: not in labour market (reference category) HoH: white HoH: black HoH: other ethnic group (reference category) HoH: excellent health HoH: very good health HoH: good health HoH: fair health HoH: poor health (reference category) HoH: catholic HoH: Jewish HoH: protestant HoH: other religion HoH: non-religious (reference category) Number of adults in household Number of children in household Price = (1 − tax rate) Own home outright or via a mortgage Home rented or other (reference category) Whether donated to a disaster cause in 2002 Observations

Mean

Std. dev.

Min

Max

0.0058 0.1674 0.1912 0.2420 0.1953 0.1983 12.8618 0.6718 0.4921 0.5079 0.7288 0.0470 0.0970 0.2187 0.6728 0.3319 0.0486 0.1936 0.3131 0.3111 0.1303 0.0519 0.1903 0.0118 0.6436 0.0164 0.1310 1.8205 0.8220 0.8642 0.6149 0.3851 0.0364

0.0757 0.3733 0.3933 0.4283 0.3965 0.3988 2.7128 0.4696 0.4999 0.4999 0.4446 0.2117 0.2959 0.4134 0.4696 0.4709 0.2150 0.3952 0.4638 0.4630 0.3367 0.2218 0.3926 0.1359 0.4790 0.1270 0.3374 0.7850 1.1558 0.1143 0.4867 0.4867 0.1873

0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 1 0 0.62 0 0 0

1 1 1 1 1 1 17 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 7 8 1 1 1 1

6590

Note: HoH, head of household.

other charities. However, we also note that if, and only if,  = 0 then the donations to all other charities variable cannot be considered endogenous in the tsunami equation, so that the issue of identification on data is no longer an issue. Thus, due to the arguably weak identification with respect to the potential endogeneity of ‘donations to other charitable causes’, one interpretation of our results is that they are simply capturing correlations rather than causality.10 2.2. Results Table 2A presents the results of estimating a univariate tobit model of Eq. (1), tsunami donations, in the first column, and the results of estimating a univariate tobit model of Eq. (2), all other charitable donations, in the second column. Estimated coefficients and marginal effects are reported in Table 2A. Clearly, there is a positive association between other charitable donations, arguably planned philanthropy, and the amount donated to the victims of the tsunami disaster, which by definition is unplanned giving. The finding of a statistically significant positive relationship between donations to the natural disaster and the level of other charitable donations is consistent with the experimental evidence of Eckel et al. (2007) focusing on the influence of Hurricane Katrina on charitable donations. Furthermore, due to the timing of the tsunami (late December 2004), and the fact that donations to other causes are measured over the calendar year 2004, any potential reverse causality is arguably minimised.11 A one percent increase in donations to all other charities is associated with a 0.13 percentage point increase in tsunami donations, indicating a positive, yet inelastic, relationship. This is robust to joint estimation of both the tsunami and all other donations equations via a bivariate recursive tobit model, Eq. (3), where the results are summarised in Table 2B, showing the coefficient and the marginal effect for the potentially endogenous covariate. Moreover, in the bivariate framework the null hypothesis of independence of the error terms cannot be rejected by either the Wald statistic or the likelihood ratio statistic (Table 2B),

10

We are particularly grateful to an anonymous referee for bringing this to our attention. It should be acknowledged that the causality interpretation is less sanguine if there are unobserved factors that are correlated with both types of donation. We have attempted to control for this by including a binary control for past generosity which may capture unobserved fixed effects. 11

S. Brown et al. / Journal of Economic Behavior & Organization 84 (2012) 97–110

103

Table 2A The determinants of tsunami donations and all other charitable donations – univariate tobit model. Tsunami donations

Log all other charitable donations HoH: aged <20 HoH: aged 20–30 HoH: aged 30–40 HoH: aged 40–50 HoH: aged 50–60 HoH: years of schooling HoH: male HoH: married/cohabiting HoH: employee HoH: unemployed HoH: self employed HoH: white HoH: black HoH: excellent health HoH: very good health HoH: good health HoH: fair health HoH: catholic HoH: Jewish HoH: protestant HoH: other religion Number of adults in household Number of children in household Log household labour income Log household wealth Log household non-labour income Log HPI Log variance of HPI Price = (1 − tax rate) Own home outright or on a mortgage Whether donated to a disaster cause in 2002 Whether donated to charity in 2002 a

F(d, 6552) , p value Observations

All other donations

Coef.

TSTAT

M.E.

0.5777 0.0322 −0.2868 −0.5716 −0.0583 −0.3292 0.1518 −0.7998 0.7932 −0.0461 −0.7111 −0.1091 −0.6585 −0.2459 −0.3823 −0.1759 0.1777 0.4742 0.6024 2.1839 −0.2029 1.7898 0.0559 0.1774 0.0809 0.0257 0.0682 0.9327 −0.1098 −0.6950 −0.3993 2.1757

(17.55) (0.02) (0.71) (1.52) (0.18) (1.05) (3.69) (2.83) (2.61) (0.11) (1.15) (0.37) (1.63) (0.57) (0.74) (0.36) (0.36) (0.93) (1.87) (4.26) (0.71) (2.62) (0.40) (2.01) (2.13) (1.07) (2.76) (3.80) (0.67) (0.92) (1.72) (6.42) –

0.1271 0.0071 −0.0631 −0.1258 −0.0128 −0.0724 0.0334 −0.1760 0.1745 −0.0101 −0.1564 −0.0240 −0.1449 −0.0541 −0.0841 −0.0387 0.0391 0.1043 0.1325 0.4805 −0.0446 0.3938 0.0123 0.0390 0.0178 0.0057 0.0150 0.2052 −0.0242 −0.1529 −0.0878 0.4787

Coef. −2.3879 −2.3010 −1.1496 −0.8711 −0.5422 0.2884 −0.6034 1.4457 0.1473 −0.9879 0.0031 0.2872 0.1132 0.3927 1.1056 1.2937 1.2709 0.7606 0.5483 1.1775 0.0617 −0.2639 0.0486 0.0332 0.0820 0.0643 2.2004 −0.5940 −1.4634 1.4722 0.0001

36.19, p = [0.000]

TSTAT – (2.41) (8.71) (4.83) (4.13) (2.82) (10.86) (3.21) (7.42) (0.57) (2.44) (0.02) (1.03) (0.37) (1.12) (3.33) (3.86) (3.65) (3.57) (1.51) (6.09) (0.11) (2.77) (0.83) (1.37) (5.46) (4.03) (13.35) (5.58) (2.92) (9.27) – (5.45)

M.E. −1.4327 −1.3806 −0.6898 −0.5227 −0.3253 0.1730 −0.3620 0.8674 0.0884 −0.5927 0.0019 0.1723 0.0679 0.2356 0.6634 0.7762 0.7625 0.4564 0.3290 0.7065 0.0370 −0.1583 0.0292 0.0199 0.0492 0.0386 1.3202 −0.3564 −0.8780 0.8833 0.0001

90.34, p = [0.000] 6590

Notes: Month of interview controls are included. a d = 39 in tsunami equation and d = 38 in all other donations equation. Marginal effects are derived by multiplying the estimated coefficients through by the probability of being in the non-censored part of the distribution. Table 2B The determinants of tsunami donations and all other charitable donations – bivariate recursive tobit model. Tsunami donations Coef. Log all other charitable donations Chi2 (77), p value  Wald statistic  = 0, Chi2 (1), p value Likelihood ratio statistic  = 0, Chi2 (1), p value Controls Observations

0.7425

TSTAT

M.E.

(5.30)

0.1634

4824.63, p = [0.000] −0.1034 1.7770, p = [0.183] 1.7880, p = [0.181] As in Table 2A 6590

Notes: Month of interview controls are included. Marginal effects are derived by multiplying the estimated coefficients through by the probability of being in the non-censored part of the distribution.

hence  12 = 0 and thus the parameters of the tsunami model in the univariate framework are uniquely identified without explicit need for exclusion restrictions on explanatory variables across the two equations.12 Despite evidence of a positive relationship between planned and unplanned donations, there are noticeable differences in the determinants of each type of charitable donation. For example, focusing upon the univariate tobit results of Table 2A, in comparison to the amount donated to other charitable causes, age effects are apparent for all other charitable donations

12 To assess the adequacy of our approach, we test the residuals for normality and find that the null hypothesis that the residuals are jointly normally distributed cannot be rejected at the 1 percent level (chi squared statistic 0.11, p-value 0.948).

104

S. Brown et al. / Journal of Economic Behavior & Organization 84 (2012) 97–110

Table 3 The correlation between different types of charitable donations. Tsunami Tsunami Religion Needy Multi-purpose Caring All other

1 0.2797 p = [0.000] 0.2828 p = [0.000] 0.2555 p = [0.000] 0.3426 p = [0.000] 0.2519 p = [0.000]

Religion

Needy

Multipurpose

Caring

All other

1 0.2726 p = [0.000] 0.2833 p = [0.000] 0.3490 p = [0.000] 0.1848 p = [0.000]

1 0.2591 p = [0.000] 0.4214 p = [0.000] 0.2849 p = [0.000]

1 0.3461 p = [0.000] 0.1853 p = [0.000]

1 0.3416 p = [0.000]

1

Notes: Needy donations consist of organisations that help people in need of food shelter or other basic necessities; caring donations consist of donations to health care or medical research organisations, educational purposes, organisations that provide youth or family services, and organisations that support or promote the arts, culture or ethnic awareness; all other donations consist of organisations that provide international aid or promote world peace, organisations associated with preserving the environment, and organisations with all other purposes.

but are not evident in the case of donations to the victims of the tsunami. Specifically, compared to heads of household aged over 60, there is a monotonic decrease in the level of donations across younger households. For example, a household with a head aged less than 20 donates 143 percentage points less to charity than a corresponding household with a head aged over 60. Such findings are consistent with the evidence in the existing literature, such as Lankford and Wyckoff (1991), Auten and Joulfaian (1996), and Schokkaert (2006). This might be due to the fact that donations to planned causes are smoothed over the life cycle, which is consistent with the findings of Glenday et al. (1986), whereas this may not be the case for unplanned events such as the Indian Ocean tsunami, which may explain the absence of any significant age effects. Other differences between the determinants of planned and unplanned donations are that the level of giving to the tsunami victims is not influenced by either the price of donating or household wealth. Interestingly, housing tenure has a differential association with the types of giving. Owning the home outright or via a mortgage is inversely related to the amount donated to the victims of the tsunami, albeit at the 10 percent level of statistical significance, yet is positively related to donations to other causes. Arguably these differences might again reflect the unplanned spontaneous nature of donations to natural disasters. Both wealth and whether the home is owned outright are likely to reflect the stock of household wealth and are significant covariates for planned donations, whereas current available labour income matters for unplanned giving. The statistical insignificance of labour income for planned donations is also consistent with Auten et al. (2002), which might reflect transitory income effects. Factors associated with a positive and statistically significant effect across both planned and unplanned donating are: years of schooling of the head of household; a married or cohabiting head of household; a head of household in good health; a head of household who has a catholic faith; household non-labour income; and permanent income. Income effects are, however, found to be inelastic throughout. The only covariate which has a statistically significant inverse relationship with both types of charitable donations is the gender of the head of household where households with male heads donate less, on average, which is consistent with the existing literature. For example, male headed households donate around 18 percentage points less to the victims of the tsunami and 36 percentage points less to other charitable causes.13 3. Donations to the victims of the tsunami and types of charitable donation 3.1. Data and methodology In the PSID there is detailed information on the types of charitable donation made over the calendar year 2004 and so we are able to explore the relationship between donations specifically related to the victims of the tsunami and charitable donations to the following categories, where the percentages given in parentheses indicate the proportion each category represents within the total amount donated: religious purposes or spiritual development (47 percent); multi-purpose organisations (13 percent); organisations that help people in need of food shelter or other basic necessities (12 percent); donations for caring purposes – health care or medical research organisations, educational purposes, organisations that provide youth or family services, and organisations that support or promote the arts, culture or ethnic awareness (15 percent); and all other forms of donations, including donations to organisations that provide international aid or promote world peace and organisations associated with preserving the environment (13 percent). Donations for religious purposes are clearly the dominant category. Fig. 2 presents the distributions of each type of charitable donation for all households as well as for positive contributors only. In Table 3, we present simple bivariate correlations between the different types of charitable

13

See Yörük (2009) for a comprehensive analysis of the implications of gender differences and household bargaining for charitable donations.

S. Brown et al. / Journal of Economic Behavior & Organization 84 (2012) 97–110

105

Fig. 2. The distributions of the five different types of charitable donations in 2004.

donations. There are positive associations between all types of charitable donations and these are statistically significant at the 1 percent level. Total donations over the calendar year 2004 are then decomposed in order to examine how donations to different causes are related to donations specifically to the victims of the tsunami. Initially we re-estimate Eq. (1) as a univariate tobit model controlling for the five separate types of charitable donation (j): ∗ = Xi + tsiT

5 j=1

j yji(T −1) + i

(4)

where is a normally distributed random error term. We then examine each type of charitable donation and tsunami ∗ donations in a system framework. Let yji(T denote the latent propensity for charitable donation of type j (j = 1, . . ., J) of −1) household i, yji(T−1) is the observed amount of the type j charitable donation, X are variables which are thought to influence these propensities, defined in Section 2, which include a binary control for past generosity to charity type j. It is then possible to construct a system of tobit equations, which is estimated via a system approach following Huang (1999), where the key parameters of interest are j : ∗ y1i(T = X i ˇ1 + ε1i −1) ∗ = X i ˇ2 + ε2i y2i(T −1)

.. . ∗ yJi(T −1)

(5) =

X i ˇJ

∗ = X + tsiT i

+ εJi

5

j=1

j yji(T −1) + ε(J+1)i

106

S. Brown et al. / Journal of Economic Behavior & Organization 84 (2012) 97–110

Given that each of the J dependent variables has left hand censoring, there are 2J possible combinations at their censoring points. Assuming ε ∼ iid(0, ˝), where ˝ is a J × J symmetric positive matrix, Huang (1999) shows that the likelihood function, which encapsulates all censoring combinations, is given by: L(Y ; ˇ, , ˝) = 



n i=1 

LiSq (yi ; ˇ, , ˝)

(6)



where Y = (y1∗ , y2∗ , . . . , yJ∗ , ts∗ ) and LiSq gives the likelihood of the case that the ith observation falls into regime q.14 Following Barslund (2007), Eq. (6) is estimated via simulated maximum likelihood using the MVTOBIT command in STATA 11. Given that Eq. (5) is recursive and that donations other than those to the victims of the tsunami are arguably predetermined, one might expect that the error terms between the tsunami and all other donations are uncorrelated, i.e. j6 = 0, where j = 1, . . ., 5. Indeed, if this were to be the case, this uniquely identifies  and j of the tsunami equation, without reliance on identification on data. Again the null of j6 = 0 (and hence  j6 = 0) is formally tested via Wald and likelihood ratio statistics. 3.2. Results Table 4A reports the results of estimating the univariate tsunami donations model, Eq. (4), where the log amount of each type of planned donation is included simultaneously, whilst Table 4B reports a specification based upon the multivariate recursive system tobit model of Eq. (5). There is a positive, yet inelastic, statistically significant relationship between the level of the donations for the victims of the tsunami and the level of each type of charitable donation reported in Tables 4A and 4B. In line with a priori expectations, the correlations between the error terms of the tsunami donations equation and all other types of donation are not significantly different from zero in the multivariate setting (Table 4B), which is consistent with the decision making process of donating to victims of the tsunami being independent of that of donations to other charitable causes, conditional on the covariates. Thus endogeneity is (statistically) not an issue here, since both the Wald and the likelihood ratio statistics fail to reject the null hypothesis that j6 = 0 (and hence  j6 = 0) enabling the  and j parameters in Eq. (5) to be uniquely identified without reliance on identification through data. The results suggest that the strongest positive association exists between the tsunami donations and those in the caring and needy categories, which might reflect similar motivations for giving between planned and unplanned philanthropy. These positive effects exist when we control for past generosity to charity type j in 2002. Focusing upon the marginal effects for the univariate model (Table 4A), a one percent increase in donating to a caring or needy charity is associated with a 0.07 and 0.06 percentage point increase in the level of the tsunami donation, respectively.15 There are clearly a range of elasticities with the smallest being between donating to the tsunami disaster and religious organisations. Moreover, a joint test that the coefficients for planned giving are equal is rejected at the one percent level, hence heterogeneity exists between the association of the type of planned donation made and unplanned giving. The finding of positive associations between multiple types of charitable donations is consistent with the experimental evidence of de Oliveira et al. (2011) who find a positive correlation between giving to one organisation and another. 4. Donations to the victims of the tsunami and future donations to charity 4.1. Data and methodology The PSID 2007 includes information on donations to charity over the 2006 calendar year, i.e. 1st January 2006 to 31st December 2006, where the average donation was $1368 in 2005 prices (Table 1A) with 44 percent of households not making any donations. The average donation amongst those households donating to charity in 2006 is $2441. Hence, we are able to explore the implications of tsunami donations in 2005 (T) for donations to charity in the calendar year 2006 (T + 1). This is potentially important in that donating to an unplanned event might divert spending away from other charitable causes. ∗ Initially, we investigate what factors influence total charitable donations in 2006. Let yi(T denote the latent propensity for +1) total charitable donations of household i at time T + 1 (2006), then ∗ yi(T = X i + tsiT + ωi +1)

(7)

where X is a vector of covariates as defined in Section 2 and ω is a normally distributed random error term. The model is estimated as a univariate tobit model, with the level of tsunami donations made at time T (i.e. 2005), tsiT , included in the

14 Representing the 2J possible combinations by the 2J × 1 vector Si , i = 1, 2, . . ., 2J , we can define the following: S1 = (0, 0, . . . , 0), . . . , Sh = (0, . . . , 0, +, . . . , +), . . . , S2J = (+, +, . . . , +), where ‘+’ means the observed value is positive and equals the true value and ‘0’ implies the true value is non-positive. The first regime S1 is when all observations are censored and coded as zero; the second regime Sh is where the first r (r < J) dependent variables of the ith observation are zero, and the regime S2J is where the dependent variable is uncensored and equal to the true values, yi∗ , which are all positive (see Huang, 1999). 15 In Table A1 in Appendix A, we include each charitable donation one at a time and find similar results in that the coefficients all remain positive.

S. Brown et al. / Journal of Economic Behavior & Organization 84 (2012) 97–110

107

Table 4A The determinants of tsunami donations and different types of charitable donations – univariate tobit model.

Log religious charitable donation Log needy charitable donation Log multi-purpose charitable donation Log caring charitable donation Log other charitable donations HoH: aged <20 HoH: aged 20–30 HoH: aged 30–40 HoH: aged 40–50 HoH: aged 50–60 HoH: years of schooling HoH: male HoH: married/cohabiting HoH: employee HoH: unemployed HoH: self employed HoH: white HoH: black HoH: excellent health HoH: very good health HoH: good health HoH: fair health HoH: catholic HoH: Jewish HoH: protestant HoH: other religion Number of adults in household Number of children in household Log household labour income Log household wealth Log household non-labour income Log HPI Log variance of HPI Price = (1 − tax rate) Own home outright or on a mortgage Whether donated to a disaster cause in 2002

Coef.

TSTAT

M.E.

0.1918 0.2757 0.2316 0.3021 0.2147 −0.0566 −0.0493 −0.3430 0.0196 −0.3193 0.1155 −0.6527 0.6877 −0.0470 −0.8872 −0.0415 −0.8062 −0.5311 −0.3277 −0.1114 0.2453 0.5264 0.3563 1.7055 −0.1670 1.5546 0.0540 0.1224 0.0804 −0.0067 0.0694 0.7883 −0.1439 −0.5019 −0.3416 1.5592

(10.04) (6.48) (5.45) (7.89) (5.35) (0.04) (0.13) (0.92) (0.06) (1.03) (2.89) (2.35) (2.35) (0.12) (1.51) (0.15) (2.01) (1.23) (0.64) (0.23) (0.51) (1.05) (1.13) (3.04) (0.60) (2.37) (0.38) (1.40) (2.18) (0.28) (2.78) (3.27) (0.90) (0.67) (1.52) (4.14)

0.0422 0.0607 0.0510 0.0665 0.0472 −0.0125 −0.0108 −0.0755 0.0043 −0.0702 0.0254 −0.1436 0.1513 −0.0103 −0.1952 −0.0091 −0.1774 −0.1168 −0.0721 −0.0245 0.0540 0.1158 0.0784 0.3752 −0.0367 0.3420 0.0119 0.0269 0.0177 −0.0015 0.0153 0.1734 −0.0317 −0.1104 −0.0752 0.3430

F(43, 6550), p value Observations

42.61, p = [0.000] 6590

Notes: Month of interview controls are included. Marginal effects are derived by multiplying the estimated coefficients through by the probability of being in the non-censored part of the distribution.

Table 4B The determinants of tsunami donations and different types of charitable donations – multivariate recursive tobit model. Coef. Log religious charitable donation Log needy charitable donation Log multi-purpose charitable donation Log caring charitable donation Log other charitable donations Chi2 (215), p value Chi2 (10) 12 = 13 = · · · = 15 = 23 = · · · = 45 = 0, p value Wald statistic 16 = 26 = · · · = 56 = 0, Chi2 (5), p value Likelihood ratio statistic 16 = 26 = · · · = 56 = 0, Chi2 (5), p value Controls Observations

0.1646 0.5554 0.2691 0.6538 0.2599

TSTAT

M.E.

(3.49) (6.61) (3.56) (7.34) (3.16)

0.0362 0.1222 0.0592 0.1438 0.0572

6419.77, p = [0.000] 1003.60, p = [0.000] 1.66, p = [0.883] 2.86, p = [0.270] As in Table 4A 6590

Notes: Month of interview controls are included. Marginal effects are derived by multiplying the estimated coefficients through by the probability of being in the non-censored part of the distribution.

108

S. Brown et al. / Journal of Economic Behavior & Organization 84 (2012) 97–110

Table 5 Tsunami donations and future charitable donations by type.

Panel A: total charitable donations Log tsunami donation Donated to charity type j in 2004

Coef.

TSTAT

M.E.

0.1942 3.9615

(6.45) (24.84)

0.1088 2.2184

F(39, 6552) Panel B: religious charitable donations Log tsunami donation Donated to charity type j in 2004

116.06, p = [0.000]

0.1275 7.0767

F(39, 6552) Panel C: needy charitable donations Log tsunami donation Donated to charity type j in 2004

0.3510 3.7565

0.2638 5.1971

F(39, 6552) Controls, Panels A–F Observations

0.0878 1.3148

(5.96) (23.78)

0.0765 1.1434

70.18, p = [0.000]

0.3514 4.0278

F(39, 6552) Panel F: all other charitable donations Log tsunami donation Donated to charity type j in 2004

(6.24) (16.92) 48.59, p = [0.000]

F(39, 6552) Panel E: caring charitable donations Log tsunami donation Donated to charity type j in 2004

0.0485 2.6891

113.13, p = [0.000]

F(39, 6552) Panel D: multi-purpose charitable donations Log tsunami donation Donated to charity type j in 2004

(2.92) (14.04)

(6.22) (22.11)

0.0773 1.1681

86.90, p = [0.000]

0.4611 4.7045

(6.73) (16.56)

0.0645 0.6586

37.22, p = [0.000] As in Table 2A 6590

Note: Each panel represents a separate univariate tobit model. Marginal effects are derived by multiplying the estimated coefficients through by the probability of being in the non-censored part of the distribution.

set of explanatory variables. We also repeat this analysis (of estimating Eq. (7)) but focusing upon each type of charitable donation. 4.2. Results The first panel of Table 5 focuses upon the determinants of total donations and each type of donation in 2006 and presents the parameter estimate of  within the univariate context. The results in Table 5 Panel A imply a positive association between donating to the victims of the tsunami at time T and future charitable donations at time T + 1. This effect exists when we condition on whether the individual donated to charity of type j in 2004, where a one percent increase in donations to victims of the tsunami is associated with a 0.11 percentage point increase in giving to all other charities. Interestingly, this marginal effect is similar in magnitude to that found in Section 3 where we considered how prior planned giving influences unplanned donations. As such, the evidence implies that donating to the victims of the tsunami does not divert future household expenditure away from donating to other charitable causes. This implies that these unplanned contributions to charity do not have a crowding out effect on planned charitable donations.16 This is a potentially interesting addition to the literature on charitable giving as Ribar and Wilhelm (2002), who investigated crowding out across different types of non-profitable charitable organisations, noted the absence of studies which consider multiple forms of philanthropy (Fig. 3). The analysis of estimating the influence of tsunami donations on donations to different causes essentially decomposes this overall effect in order to ascertain whether the relationship is uniform across donations to different charitable causes. The results of estimating Eq. (7) for each constituent element of total donations are shown in Panels B–F of Table 5. There is evidence that donating to the victims of the tsunami at time T increases expenditure on each type of charitable donation

16 An important caveat to mention is that it is possible crowding out may occur but that the effects are not persistent, i.e. the effect of unplanned donations might influence donations to planned charities at the same point in time. However, we are unable to explore this in the PSID data since the information on donations is for the prior calendar year, i.e. we do not know the point in time at which a specific donation was made.

S. Brown et al. / Journal of Economic Behavior & Organization 84 (2012) 97–110

109

Fig. 3. The distributions of total donations and the five different types of charitable donations in 2006.

in the future (T + 1), although the effect is not uniform across the different charitable causes. The largest impacts in terms of magnitude are for needy and caring donations, where a one percent increase in donations to victims of the tsunami is associated with a 0.09 and 0.08 percentage point increase, respectively.17 5. Conclusions We have investigated the relationship between charitable donations related to an unexpected adverse event and donations to other types of charity. The importance of one-off appeals for disaster relief as a way to raise significant funds has been discussed in the existing literature. A concern surrounding such appeals relates to the possibility of donations being diverted from existing charitable causes towards such relief funds. Our empirical evidence allows us to investigate not only whether the level of charitable donations prior to the natural disaster is associated with the amount donated to the victims of the natural disaster, but also to consider whether donating to the victims of the tsunami influences the level of future charitable donations. Our empirical analysis makes three main contributions to the literature. Firstly, we find a positive association between planned and unplanned giving, i.e. donating to an unexpected natural disaster is associated with increased expenditure to other charitable causes in the future, an effect which exists even when past generosity is controlled for. Secondly, we find that certain covariates (such as age) have different influences on the level of giving to the victims of the tsunami and the level of giving to other charitable causes, which may reflect life cycle effects. Finally, there is no evidence in the analysis that giving to an unplanned natural disaster diverts future expenditure away from other types of giving: there is no evidence of crowding out. Hence, it would appear that our results are supportive of ‘joy-of-giving’ motivations and the existence of ‘giving types’, where the largest correlations with donating to the tsunami occur between planned donations to arguably similarly related causes, i.e. caring and needy charitable organisations. Our findings thus endorse the importance of distinguishing

17 A test of the null hypothesis that the size of the tsunami donations coefficient is equal across the j types of charitable donations is rejected at the one percent level.

110

S. Brown et al. / Journal of Economic Behavior & Organization 84 (2012) 97–110

between different charitable causes and hopefully will serve to stimulate further research into the motivations behind giving to different charitable causes. Acknowledgements We are extremely grateful to the Editor and three referees for excellent comments. We are also grateful to the Institute for Social Research, University of Michigan for supplying the Panel Study of Income Dynamics 1968–2009. We are also grateful to participants at the New Directions in Welfare Conference, Oxford, 2009, and the European Economics Association Annual Conference, Glasgow, 2010, and to seminar participants at the University of East Anglia for valuable comments. Finally, we are extremely grateful to the Australian Research Council for financial assistance. Appendix A.

Table A1 The determinants of tsunami donations and different types of charitable donations – univariate tobit model. Tsunami donation [1] Religious

Log religious charitable donation Log needy charitable donation Log multi-purpose charitable donation Log caring charitable donation Log other charitable donation F(39, 6551), p value

[2] Needy

Coef.

TSTAT

0.1721

(13.52) – – – –

34.66, p = [0.000]

Coef.

TSTAT

[3] Multi-purpose

[4] Caring

Coef.

Coef.

– 0.4212

(12.28) – – –

34.71, p = [0.000]

Controls Observations

TSTAT – –

0.3336

31.82, p = [0.000]

TSTAT

Coef.

– – –

(3.62) – –

[5] All other

0.5155

(8.04) –

37.69, p = [0.000]

0.3629

TSTAT – – – – (9.06)

31.46, p = [0.000]

As in Table 4A 6590

Note: Each column [1] to [5] represents a separate univariate tobit model.

References Amemiya, T., 1974. Multivariate regression and simultaneous equation models when the dependent variables are truncated. Econometrica 42, 999–1012. Andreoni, J., 2006. Philanthropy. In: Kolm, S.-C., Ythier, J.M. (Eds.), The Handbook of the Economics of Giving, Altruism and Reciprocity, vol. 2. Elsevier, North Holland (Chapter 18). Angrist, J.D., Pischke, J.-S., 2009. Mostly Harmless Econometrics: An Empiricists Companion. Princeton University Press, Princeton, New Jersey. Athukorala, P.C., Resosudarmo, B.P., 2005. The Indian Ocean tsunami: economic impact, disaster management and lessons. Asian Economic Papers 4, 1–39. Auten, G.E., Joulfaian, D., 1996. Charitable contributions and intergenerational transfers. Journal of Public Economics 59, 55–68. Auten, G.E., Sieg, H., Clotfelter, C.T., 2002. Charitable giving, income, and taxes: an analysis of panel data. American Economic Review 92, 371–382. Barslund, M., 2007. Estimation of Tobit Type Censored Demand Systems: A Comparison of Estimators. Discussion Papers. Department of Economics, University of Copenhagen, No. 07-16. Brown, P.H., Minty, J.H., 2008. Media coverage and charitable giving after the 2004 tsunami. Southern Economic Journal 75, 9–25. Clotfelter, C.T., 1980. Tax incentives and charitable giving: evidence from a panel of taxpayers. Journal of Public Economics 13, 319–340. de Oliveira, A., Croson, R., Eckel, C., 2011. The giving type: identifying donors. Journal of Public Economics 95, 428–435. Eckel, C., Grossman, P., Milano, A., 2007. Is more information always better? An experimental study of charitable giving and Hurricane Katrina. Southern Economic Journal 74, 388–411. Glenday, G., Gupta, A.K., Pawlak, H., 1986. Tax incentives for personal charitable donations. Review of Economics and Statistics 68, 688–693. Greene, W., 2007. Limdep Version 9, User Manuals, New York. Greene, W., 2008. Econometric Analysis, 6th International edition. Prentice Hall. Huang, H.-C., 1999. Estimation of the SUR tobit model via the MCECM algorithm. Economics Letters 64, 25–30. Lankford, R.H., Wyckoff, J.H., 1991. Modeling charitable giving using a Box–Cox standard tobit model. Review of Economics and Statistics 73, 460–470. Maddala, G.S., 1983. Limited Dependent and Qualitative Variables in Econometrics. Cambridge University Press, North Holland. Ribar, D., Wilhelm, M., 2002. Altruistic and joy-of-giving motivations in charitable behavior. Journal of Political Economy 110, 425–456. Schokkaert, E., 2006. Empirical analysis of transfer motives. In: Kolm, S.C., Ythier, J.M. (Eds.), The Handbook of the Economics of Giving, Altruism and Reciprocity, vol. 1. Elsevier, North Holland (Chapter 2). Schwartz, R.A., 1970. Personal philanthropic contributions. Journal of Political Economy 78, 1264–1291. Sugden, R., 1982. On the economics of philanthropy. The Economic Journal 92, 341–350. Wilhelm, M.O., 2006. New data on charitable giving in the PSID. Economics Letters 92, 26–31. Wilhelm, M.O., Brown, E., Rooney, P.M., Steinberg, R., 2008. The intergenerational transmission of generosity. Journal of Public Economics 92, 2146–2156. Wooldridge, J., 2010. Econometric Analysis of Cross Section and Panel Data. MIT Press, England. Wright, K., 2002. Generosity versus Altruism: Philanthropy versus Charity in the US and UK, Civil Society Working Paper Number 17. London School of Economics. Yörük, B., 2009. Charitable giving by married couples revisited. The Journal of Human Resources 45, 497–516.