Available online at www.sciencedirect.com
Economics Letters 99 (2008) 521 – 525 www.elsevier.com/locate/econbase
Social capital and windfalls: Empirical evidence Yannis Georgellis a , John G. Sessions b,⁎, Nikolaos Tsitsianis c a
Brunel University, Uxbridge, Middlesex UB8 3PH, England, United Kingdom University of Bath, Claverton Down, Bath BA2 7AY, England, United Kingdom University of Hertfordshire, Hatfield, Hertfordshire AL10 9AB, England, United Kingdom b
c
Received 10 March 2007; received in revised form 10 September 2007; accepted 27 September 2007 Available online 5 October 2007
Abstract We examine the correlates of windfall gains amongst a sample of British respondents. Our results suggest that such gains are not random, but are significantly related to a variety of individual characteristics and traits. In particular, proxies of social capital are significantly related to the probability of receiving a windfall. © 2007 Elsevier B.V. All rights reserved. Keywords: Windfalls; Social capital JEL classification: D1
1. Introduction Economists have for some time used windfall gains to analyse the effect of exogenous changes in wealth on individual behaviour. For example, there is mounting evidence that many individuals are frustrated in their desire to pursue entrepreneurship. Many researchers have attributed the shortfall to capital constraints that limit the access of nascent entrepreneurs to external sources of finance. Support for this thesis is derived from a number of studies that find the probability of transition into selfemployment to be significantly lower amongst less wealthy individuals, with the gap between actual and desired entrepreneurship greatest in those areas most afflicted by unemployment and poverty [see Georgellis et al. (2005)].1 Using wealth as an explanatory variable in this way is problematic. Pre-transition wealth may be correlated with an indi-
⁎ Corresponding author. Department of Economics and International Development, University of Bath, Bath, BA2 7AY, England, United Kingdom. Tel.: +44 1225 384517; fax: +44 1225 383423. E-mail address:
[email protected] (J.G. Sessions). 1 Other work has focussed on the effect of windfalls on labour supply, consumption and savings [see, e.g., Imbens et al (2001)]. 0165-1765/$ - see front matter © 2007 Elsevier B.V. All rights reserved. doi:10.1016/j.econlet.2007.09.036
vidual's ability or determination, and it may be these unobserved and omitted factors that are determining the success of the transition. With this in mind, attention has turned to the analysis of windfall gains. These are far less likely to be so correlated, being, in the words of Blanchflower and Oswald (1998), about as close to a natural experiment as economists are likely to get in this area. In this letter we address the question as to whether windfalls are indeed true exogenous shocks to income, or do attributes such as risk averseness, emotional intelligence and social awareness, cause individuals to behave in such a way as to affect their exposure to ostensibly random events. If they do, then a pertinent question, and one not addressed here, is how this social capital impacts upon individual economic behaviour. 2. Data and methodology Our data are derived from the British Household Panel Survey. This is a nationally representative survey of some 5500 private households, comprising some 10000 individuals, interviewed annually every fall since 1991 We restrict our attention to reposndents aged 18 to 60, thereby ensuring that the sampled members are always of working age, and are not apt to the receipt of lump sum payments that are associated with retirement at the usual age. The question concerning windfalls
522
Y. Georgellis et al. / Economics Letters 99 (2008) 521–525
in year t is: Since September 1st ‘t-1’ have you received any payments, or payments in kind, from anything listed on this card? If answering Yes, respondents, were subsequently asked: (i) Which ones? And (ii) About how much did in total did you
receive / was this worth?’ The list of possible payments in waves five to twelve, includes a life insurance policy, a lump sum pension payout, a personal accidental claim, a redundancy payment, an inheritance or bequest (including inherited property),
Table 1 Dependent variable: receipt of windfall (yes/no) probit (pooled) estimates Model (1) Mean Constant Demography Male White Age Age-squared/100 Married Couple Widowed Divorced Separated Number of Children Labour force (ref: employment) Self-employment Unemployment Long-term sick Maternity care Education (ref: no education) Higher A Levels O Levels Other higher Other qualification Income (ref: o0 V dividends and interest b o100) o100 b dividends and interest V o1000 Dividends and interest z o1000 Private personal pension Savings (dummy variable) Social capital Provides care for non-resident person Active in religious group Active in voluntary group Active in other community group Active in social group Is there someone who will listen? Is there someone to help you in a crisis? Is there someone who really appreciates you? Subjective well-being Happinessb Regions (ref: south east) South west East Anglia Midlands North west North east Wales Scotland Northern Ireland
Model (2) Coef
T-ratio
Coef
T-ratio
0.254
− 1.551
−14.90
Mean 0.263
−1.571
− 15.14
0.512 0.961 38.623 16.403 0.554 0.132 0.013 0.060 0.020 0.632
0.143 0.372 0.016 − 0.014 0.063 0.097 0.067 − 0.037 − 0.083 − 0.049
10.34 9.60 3.68 −2.52 2.74 3.95 1.07 −1.05 −1.55 −5.92
0.453 0.960 40.127 17.768 0.590 0.120 0.021 0.063 0.019 0.686
0.152 0.383 0.019 −0.018 0.043 0.073 −0.171 −0.018 −0.094 −0.057
10.59 9.61 4.51 − 3.55 1.79 2.74 − 3.04 − 0.51 − 1.71 − 6.98
0.100 0.053 0.048 0.067
− 0.051 − 0.045 − 0.171 − 0.154
−2.22 −1.41 −4.52 −5.10
0.084 0.045 0.037 0.053
−0.068 −0.106 −0.201 −0.156
− 2.70 − 3.04 − 4.67 − 4.72
0.145 0.147 0.196 0.265 0.081
0.091 0.175 0.147 0.180 0.140
3.41 6.65 6.21 8.04 4.80
0.124 0.133 0.203 0.249 0.090
0.078 0.167 0.170 0.166 0.145
2.85 6.31 7.49 7.55 5.21
0.116 0.030 0.182 0.443
0.130 0.227 0.086 0.179
6.30 6.12 4.90 12.79
0.133 0.039 0.166 0.413
0.079 0.082 0.089 0.164
3.87 2.36 4.73 11.42
0.105 0.099 0.038 0.015 0.097 0.958 0.957 0.952
0.075 − 0.108 0.097 0.121 0.204 − 0.001 0.056 0.056
3.49 −4.60 2.84 2.32 9.45 −0.04 1.44 1.62
0.113 0.102 0.041 0.016 0.097 0.959 0.959 0.959
0.081 −0.132 0.077 0.113 0.184 −0.014 0.118 0.030
3.81 − 5.57 2.25 2.10 8.24 − 0.34 2.85 0.78
2.023
0.030
2.61
2.021
0.037
3.18
0.067 0.029 0.140 0.083 0.118 0.122 0.168 0.050
− 0.015 − 0.040 − 0.025 0.045 − 0.088 − 0.199 − 0.031 − 0.448
−0.54 −0.99 −1.15 1.74 −3.73 −7.96 −1.41 −10.89
0.076 0.033 0.157 0.090 0.136 0.111 0.150 0.007
−0.030 −0.052 −0.040 0.036 −0.101 −0.210 0.012 −0.732
− 1.07 − 1.31 − 1.85 1.40 − 4.37 − 7.96 0.53 − 6.60
Year Dummies
Yes
Yes
Likelihood ratio chi-squared Log likelihood Pseudo R-squared Number of observations
2444.28 45 − 23951.497 0.0485 44442
1969.56 45 −22919.142 0.0412 41468
Notes: (a). The right-hand side variables are set at time t (t − 1) in specifications (1) and (2) respectively; (b) happiness’ is measured on a five-point scale with 1 (5) indicating the highest (lowest) level of happiness.
Y. Georgellis et al. / Economics Letters 99 (2008) 521–525
a win on the football pools, national lottery or other sort of gambling, or anything else. In wave five the list also includes an annual/seasonal bonus from employment, whilst in wave seven it
523
includes a lump sum from a building society conversion. Given the specificity of these questions we incorporated them into the ‘anything else’ category.
Table 2 Dependent variable: receipt of windfall (yes/no) probit (pooled) estimates: model 1 Type of windfall
Constant Demography Male White Age Age-squared/100 Married Couple Widowed Divorced Separated Number of children Labour force (ref: employment) Self-employment Unemployment Long-term sick Maternity leave Education (ref: no education) Higher A levels O levels Other higher Other qualification Income (Ref: o0 V Dividends and Interest b o100) o100 b Dividends and Interest V o1000 Dividends and Interest z o1000 Paid into private personal pension Savings (dummy variable) Social capital Provides care for non-resident person Active in religious group Active in voluntary group Active in other community group Active in social group Is there someone who will listen Is there someone to help in a crisis Anyone who really appreciates you Subjective well-being Happinessb Regions (ref: south east) South west East Anglia Midlands North west North east Wales Scotland North Ireland Likelihood ratio chi−squared Log likelihood Pseudo R-squared Number of observations
(1) Insurance
(2) Pension
Coef
Coef
T-stat
(3) Accident
T-stat Coef
(4) Layoff
T-stat Coef
T-stat
(5) Bequest
(6) Lottery
(7) Other
Coef
Coef
Coef
T-stat
T-stat
T-stat
− 3.257 − 14.76 −2.926 − 7.58 −2.425 − 8.48 − 3.767 − 13.82 − 2.861 −13.42 − 1.550 − 13.29 − 3.013 − 14.69 − 0.010 0.136 0.019 − 0.004 0.054 − 0.018 0.402 − 0.009 0.016 0.001
− 0.37 1.72 2.19 − 0.43 1.13 − 0.32 4.48 − 0.13 0.16 0.08
0.060 0.103 −0.034 0.069 0.147 0.207 0.698 −0.103 0.068 −0.051
1.25 0.72 − 2.30 4.15 1.58 2.04 5.36 − 0.76 0.34 − 1.57
0.169 −0.056 0.007 −0.024 −0.054 0.026 0.105 0.040 0.014 −0.005
4.28 − 0.59 0.56 − 1.49 − 0.87 0.42 0.55 0.42 0.09 − 0.22
0.080 0.108 0.049 − 0.058 0.140 0.104 − 0.030 0.039 − 0.045 − 0.061
2.32 1.10 4.28 − 4.13 2.43 1.68 − 0.17 0.46 − 0.33 − 3.04
− 0.090 0.436 0.010 − 0.008 0.032 − 0.003 0.144 − 0.001 − 0.038 − 0.060
−3.38 4.98 1.08 −0.75 0.72 −0.07 1.30 −0.01 −0.36 −3.56
0.202 0.313 0.017 − 0.019 0.049 0.099 − 0.094 − 0.060 − 0.121 − 0.043
13.02 6.97 3.44 − 3.12 1.87 3.63 − 1.26 − 1.50 − 1.96 − 4.59
0.042 0.258 0.022 − 0.027 − 0.005 0.077 − 0.238 0.118 0.034 − 0.001
1.63 3.42 2.54 − 2.53 − 0.11 1.70 − 1.62 1.87 0.35 − 0.09
− 0.054 − 0.241 − 0.100 − 0.234
− 1.29 − 3.04 − 1.03 − 3.93
0.195 0.179 −0.149 0.096
2.91 1.81 − 0.66 1.24
0.169 −0.056 0.007 −0.024
4.28 − 0.59 0.56 − 1.49
− 0.116 0.753 − 0.525 − 0.185
− 1.96 14.69 − 2.83 − 2.07
0.064 − 0.064 0.208 − 0.163
1.52 −0.90 3.30 −2.29
− 0.069 − 0.198 − 0.222 − 0.128
− 2.69 − 5.30 − 5.12 − 3.76
− 0.038 − 0.091 − 0.074 − 0.033
− 0.92 − 1.33 − 0.98 − 0.51
0.002 0.066 0.071 0.065 0.069
0.05 1.30 1.62 1.56 1.25
0.081 −0.059 0.119 0.171 0.116
0.91 − 0.61 1.63 2.56 1.32
0.040 0.027 0.019 0.094 0.045
0.52 0.36 0.28 1.42 0.53
− 0.002 0.226 0.179 0.127 0.216
− 0.02 3.53 3.09 2.26 3.12
0.500 0.327 0.211 0.303 0.107
9.24 5.86 3.99 6.11 1.58
− 0.133 0.094 0.086 0.079 0.086
− 4.34 3.23 3.30 3.18 2.66
0.343 0.220 0.189 0.293 0.174
6.56 4.02 3.74 6.18 2.81
0.117
3.21
0.165
2.62
0.080
1.36
0.180
3.72
0.243
6.91
0.039
1.70
0.141
3.91
0.327 0.110 0.186
5.79 3.47 6.90
0.269 2.86 0.142 −0.233 − 3.49 0.116 −0.001 − 0.03 0.039
1.38 2.45 0.98
0.071 0.76 0.166 4.10 − 0.055 − 1.54
0.238 0.019 0.089
3.82 0.58 3.27
0.016 0.023 0.165
0.37 1.18 10.53
0.466 0.156 0.126
8.59 5.09 4.79
− 0.023 − 0.58
−0.120 − 1.66 0.045
0.72
− 0.065 − 1.18
− 0.026 −0.62
0.114
4.77
0.060
1.52
0.119 0.079 − 0.148 0.061 − 0.126 0.057 0.260
3.02 1.31 − 1.38 1.51 − 1.68 0.74 3.27
−0.084 0.069 −0.105 0.095 0.159 0.083 −0.069
− 2.28 0.71 1.18 2.28 − 0.17 − 0.45 0.92
− 0.082 0.165 − 0.027 0.126 0.166 − 0.014 0.069
0.100 0.163 − 0.009 − 0.029 − 0.001 0.039 − 0.089
2.50 2.89 −0.09 −0.63 −0.01 0.49 −1.36
− 0.237 0.023 0.123 0.250 0.007 0.047 0.043
− 8.38 0.58 2.08 10.78 0.15 1.05 1.09
0.037 0.076 0.194 − 0.077 − 0.016 0.027 − 0.002
0.90 1.27 2.34 − 1.72 − 0.21 0.36 − 0.03
0.051
2.31
−0.039 − 0.97 −0.011 − 0.35 − 0.037 − 1.32
0.013
0.62
0.031
2.42
− 0.004 − 0.21
−0.060 − 0.56 0.218 1.85 0.135 1.82 −0.009 − 0.10 0.151 1.96 0.067 0.83 −0.099 − 1.21 0.139 1.17 425.67 45 −1595.5738 0.1177 44442
0.013 0.24 0.096 1.35 − 0.040 −0.94 − 0.091 −1.76 − 0.057 −1.26 − 0.055 −1.18 − 0.168 −3.89 − 0.365 −4.32 447.48 45 − 5207.8694 0.0412 44442
0.054 0.97 0.174 2.39 0.085 1.94 0.123 2.42 0.094 2.05 0.053 1.11 0.104 2.44 0.012 0.16 592.19 45 − 5296.8775 0.0529 44442
− 1.11 0.67 − 0.58 1.43 1.07 0.60 − 0.61
−0.180 0.069 0.159 0.132 −.0182 −0.046 0.092
−.0197 − 0.21 0.124 1.12 0.082 1.27 0.310 4.63 0.207 3.24 0.059 0.83 −0.198 − 2.61 0.269 2.99 177.73 45 −2291.7046 0.0373 44442
− 1.32 2.07 − 0.20 2.50 1.56 − 0.15 0.77
0.037 0.54 − 0.109 − 1.03 − 0.013 − 0.24 0.025 0.39 − 0.014 − 0.24 − 0.047 − 0.79 − 0.112 − 1.98 − 0.125 − 1.38 390.61 45 − 3043.548 0.0603 44442
− 0.058 − 1.85 − 0.139 − 3.04 − 0.014 − 0.59 0.024 0.82 − 0.119 − 4.51 − 0.254 − 8.77 0.025 1.05 − 0.631 − 11.22 2514.22 45 − 18258.273 0.0644 44442
0.076 1.57 − 0.042 − 0.58 − 0.101 − 2.44 − 0.045 − 0.93 − 0.158 − 3.45 − 0.229 − 4.77 − 0.111 − 2.76 − 0.415 − 4.92 485.25 45 − 5569.8902 0.0417 44442
524
Y. Georgellis et al. / Economics Letters 99 (2008) 521–525
Our analysis is based on windfall data recorded in waves five (1995), seven (1997), eight (1998), nine (1999), ten (2000), and eleven (2001) — no windfall data was collected in wave six
(1996). We estimate two pooled probit models with the dependent variable taking the value of 1 (0) to denote the receipt (nonreceipt) of a windfall and our various explanatory variables set
Table 3 Dependent variable: receipt of windfall (yes/no) probit (pooled) estimates: model 2 Type of windfall
Constant Demography Male White Age Age-squared/100 Married Couple Widowed Divorced Separated Number of Children Labour force (ref: employment) Self-employment Unemployment Long-term sick Maternity leave Education (ref: no education) Higher A levels O levels Other higher Other qualification Income (ref: o0 V dividends and interest b o100) o100 b dividends and interest V o1000 Dividends and Interest z o1000 Paid into private personal pension Savings [dummy variable) Social capital Provides care for non-resident person Active in religious group Active in voluntary group Active in other community group Active in social group Is there someone who will listen Is there someone to help in a crisis Anyone who really appreciates you Subjective well-being Happinessb Regions (ref: south east) South west East anglia Midlands North west North east Wales Scotland North ireland Likelihood ratio chi-squared Log likelihood Pseudo r-squared Number of observations
(1) Insurance
(2) Pension
(3) Accident
Coef
Coef
Coef
T-stat
T-stat
(4) Layoff
T-stat Coef
T-stat
(5) Bequest
(6) Lottery
(7) Other
Coef
Coef
Coef
T-stat
T-stat
T-stat
0.212 − 13.74 − 4.289 − 10.65 −2.231 − 7.80 −3.761 − 13.47 − 3.348 − 15.43 − 1.618 −13.89 − 2.708 − 13.51 0.027 0.076 0.008 0.009 0.049 0.059 0.094 0.070 0.114 0.016
0.73 1.14 1.93 − 0.23 1.45 − 0.29 − 0.39 − 0.03 − 0.70 0.46
0.157 0.066 0.018 0.017 0.176 0.160 0.113 0.108 0.135 − 0.089
3.68 0.53 1.17 1.09 1.94 1.46 0.83 0.94 0.73 − 2.74
0.187 −0.086 0.002 −0.019 −0.004 0.059 0.182 0.075 0.031 0.008
4.49 − 0.89 0.16 − 1.18 − 0.06 0.85 1.03 0.73 0.20 0.35
0.164 0.131 0.052 −0.059 0.004 0.003 −0.152 −0.022 0.056 −0.070
4.66 1.25 4.70 − 4.49 0.07 0.04 − 0.97 − 0.25 0.43 − 3.37
− 0.103 0.489 0.028 − 0.028 − 0.014 − 0.014 − 0.295 − 0.110 − 0.156 − 0.054
− 3.66 5.12 3.30 − 2.83 − 0.30 − 0.27 − 2.38 − 1.53 − 1.37 − 3.25
0.191 0.320 0.023 − 0.028 0.041 0.087 − 0.138 − 0.017 − 0.109 − 0.048
11.90 6.93 4.77 −4.99 1.52 2.96 −2.09 −0.42 −1.75 −5.34
0.044 0.330 0.005 − 0.005 − 0.012 0.037 − 0.176 0.076 0.068 − 0.030
1.65 4.05 0.64 − 0.50 − 0.27 0.76 − 1.50 1.18 0.71 − 1.93
0.048 0.076 0.145 0.062
− 2.26 − 1.61 − 2.90 − 2.04
− 0.150 0.054 0.071 − 0.014
− 2.01 0.49 0.30 − 0.17
−0.051 −0.006 −0.158 0.123
− 0.68 − 0.06 − 1.32 1.33
−0.756 −0.076 −0.451 −0.728
− 6.86 − 0.88 − 2.68 − 4.87
0.097 − 0.152 0.296 − 0.147
2.12 − 1.84 4.22 − 1.93
− 0.023 − 0.096 − 0.231 − 0.076
−0.82 −2.49 −4.72 −2.07
0.024 − 0.022 − 0.186 − 0.138
0.54 − 0.32 − 2.16 − 1.91
0.035 0.054 0.034 0.027 0.035
2.94 3.92 3.23 7.18 2.94
0.072 0.050 0.159 0.209 0.112
0.89 0.59 2.51 3.65 1.43
0.056 0.047 0.092 0.063 0.103
0.68 0.59 1.32 0.93 1.26
−0.027 0.122 0.119 0.018 0.189
− 0.38 1.85 2.09 0.33 2.83
0.508 0.389 0.243 0.333 0.156
9.55 7.16 4.87 7.05 2.49
− 0.160 0.076 0.098 0.065 0.082
−5.05 2.60 3.88 2.64 2.65
0.358 0.214 0.191 0.255 0.093
7.09 4.06 4.09 5.75 1.58
0.035 2.94
0.120
2.33
0.043
0.70
0.105
2.23
0.080
2.16
0.038
1.68
0.092
2.52
0.054 3.92 0.034 3.23 0.027 7.18
0.168 0.098 0.101
2.30 1.83 2.43
0.206 2.11 −0.150 − 1.52 0.153 2.97 0.108 2.51 −0.007 − 0.17 0.085 2.39
0.040 0.047 0.072
0.65 1.33 2.61
0.001 0.027 0.141
0.04 1.31 8.75
.205 .096 .088
3.63 2.89 3.29
0.037 1.38
− .080
− 1.37
0.055
0.85
−0.139 − 2.38
0.226
6.12
0.069
2.85
− 0.048 − 1.16
0.039 0.059 0.103 0.041 0.076 0.077 0.080
− 0.086 0.037 − 0.230 0.017 0.084 0.187 − 0.000
− 1.34 0.42 − 1.29 0.29 0.62 1.37 − 0.00
−0.192 0.098 −0.109 0.141 −0.088 −0.079 0.061
− 2.21 0.97 − 0.59 2.31 − 0.81 − 0.72 0.56
−0.075 0.001 −0.011 0.115 0.172 0.055 0.003
− 1.21 0.01 − 0.07 2.26 1.46 0.49 0.03
0.107 0.131 0.093 − 0.143 − 0.050 0.102 − 0.119
2.67 2.29 1.03 − 0.31 − 0.62 1.21 − 1.69
− 0.275 0.027 0.057 0.231 − 0.014 0.122 0.060
−9.57 0.68 0.93 9.63 −0.31 2.60 1.37
0.040 0.021 0.267 − 0.033 0.010 0.037 − 0.069
0.98 0.35 3.25 − 0.75 0.13 0.47 − 1.00
0.023 0.69
0.038
1.06
−0.016 − 0.47 0.021
0.70
0.073
3.33
0.014
1.07
0.034
1.60
0.054 − 0.09 0.069 2.51 0.042 0.73 0.050 1.11 0.043 2.11 0.050 − 0.44 0.043 2.60 0.156 0.64 512.97 45 − 5313.4756 0.0460 41468
0.039 0.50 0.026 0.24 0.015 0.23 − 0.013 − 0.17 0.075 1.17 − 0.138 − 1.81 − 0.122 − 1.68 − 0.315 − 0.86 638.55 45 − 2100.4186 0.1319 41468
2.76 1.14 − 0.77 0.41 − 1.17 0.16 2.26
−0.048 − 0.54 0.100 0.91 0.052 0.82 0.290 4.29 0.205 3.23 0.036 0.47 −0.203 − 2.49 0.144 0.64 175.94 45 −2045.2899 0.0412 41468
0.018 0.26 −0.105 − 1.00 0.059 1.11 0.031 0.48 −0.006 − 0.10 −0.030 − 0.48 −0.110 − 1.82 −0.450 − 1.29 282.71 45 −2885.4024 0.0467 41468
0.026 0.50 0.010 0.13 − 0.049 − 1.15 − 0.076 − 1.47 − 0.073 − 1.62 − 0.050 − 1.03 − 0.163 − 3.57 − 0.586 − 2.25 416.96 45 − 4916.129 0.0407 41468
− 0.079 −2.53 − 0.113 −2.54 − 0.028 −1.16 0.007 0.25 − 0.138 −5.34 − 0.230 −7.50 0.087 3.44 − 1.081 −5.89 2072.58 45 − 17520.975 0.0558 41468
0.040 0.98 0.066 1.37 − 0.054 − 0.75 − 0.119 − 2.87 − 0.030 − 0.63 − 0.126 − 2.86 − 0.254 − 4.96 − 0.115 − 2.71 334.81 45 − 5311.7842 0.0306 41468
Y. Georgellis et al. / Economics Letters 99 (2008) 521–525
to either their current (Model 1) or previous year (Model 2) values. 3. Empirical results It is apparent from Table 1 that the two models are very consistent, with many of the estimated coefficients on both the current and lagged values of the explanatory variables being similar in both size and significance. A number of variables are highly significant — respondents who are older, white, male, in paid employment, more educated, and with some level of personal wealth, are all more likely to receive a windfall in either the current or following year ceteris paribus. There are also significant correlations between the receipt of a windfall and respondents' social characteristics. For example, happier respondents, those who provide care for a non-residential person, those who are active in a social, voluntary, or other community group, and those who report that they have someone who would potentially help them in a crisis, are all significantly more likely to enjoy a concurrent or future windfall ceteris paribus. In contrast, respondents who are active in a religious group are less likely to so benefit. Tables 2 and 3 apply our two probit models to the seven specific windfalls mentioned in the BHPS. Again, there is a remarkable degree of consistency between the two models, as well as a number of interesting, and perhaps surprising, correlations. 4. Final comments In this letter we have found evidence to suggest that windfall gains are significantly related to a variety of individual char-
525
acteristics and traits. In particular, proxies of social capital appear to play an important role, even after controlling for a plethora of demographic and work variables. Our results cast doubt on the presupposition that windfalls are indeed exogenous shocks to income or wealth. As such, their validity as instruments for modelling the impact of wealth on individual labour supply or consumption behaviour is questionable. More generally, they allude to the importance of ethereal characteristics in individual behaviour. Acknowledgements We are grateful to the ESRC Data Archive for supplying the British Household Panel Study Survey. References Blanchflower, D., Oswald, A.J., 1998. What makes an entrepreneur? Journal of Labor Economics 16 (1), 26–60. Georgellis, Y., Sessions, J.G., Tsitsianis, N., 2005. Windfalls, wealth, and the transition to self-employment. Small Business Economics (5), 407–428. Imbens, G.W., Rubins, D.B., Sacerdote, B., 2001. Estimating the effect of unearned income on labour supply, earnings, savings and consumption: evidence from a survey of lottery players. American Economic Review 91 (4), 778–794.
