The economics of altruism: The old, the rich, the female

ABSTRACT

This study examines whether people of different observed characteristics exhibit different degree of altruism. The article utilizes a National Mental Health Survey that gathered questions about respondents’ self-reported altruism along with their demographic, labor force, and income information. The empirical results revealed the following: (1) Older people are more altruistic. This could potentially be explained by the Socioemotional Selectivity Theory (SST) that the older people shift towards emotion-related goals that can be internally rewarding in the present. (2) Higher-income people are more altruistic. This could potentially be explained by the fact that higher-income people could better afford to be more altruistic compared to lower-income people. (3) Women are more altruistic. This could potentially be explained by either biological evolution or social imposition. The main results are robust once the potential endogeneity problem of the income variable is mitigated by the use of the instrumental variable estimation method.

KEYWORDS:

• Behavioral economics
• altruism
• old age
• gender

JEL CODES:

• D64
• J16
• J40

Acknowledgments

The author would like to thank Jessica Vechbanyongratana and the two anonymous referees for their helpful comments that allowed the author to improve the article significantly. The author also would like to thank the National Statistical Office of Thailand for allowing and facilitating the use of Thailand’s Mental Health Survey.

Funding

This research is supported by the Ratchadaphiseksomphot Endowment Fund Part of the “Research Grant for New Scholar CU Researcher’s Project.”

Notes

1. Source: Merriam-Webster Dictionary.

2. Source: Becker (1976) pp. 818.

3. The, earlier waves of the MHS were conducted as attachments to other surveys (such as the Survey on Conditions of Society and Culture, the Health and Welfare Survey, etc.). Depending on which surveys they were attached to, some of MHS waves did not have large number of observations or did not contain all the labor force information.

4. For 2014, the MHS was conducted for months 2, 5, 7, 8, and 11 (but month 11’s information was not released to the public). For 2015, the MHS was conducted for months 2, 5, 7, 8, 9, and 11. However, the LFS was conducted for every month of the year.

5. For the monthly data, only the summary statistics at the region level (5 regions including Bangkok) are available (due to insufficient observations from each of the 77 provinces). Summary statistics at the province level are available on a quarterly basis.

6. 300 EAs from Bangkok, 1,902 EAs from the Central region (excluding Bangkok), 1,278 EAs from the Northern region, 1,476 EAs from the Northeast region, and 1,014 EAs from the Southern region.

7. 4,800 households from Bangkok, 26,424 households from the Central region (excluding Bangkok), 17,856 households from the Northern region, 20,640 households from the Northeast region, and 14,160 households from the Southern region.

8. Observations with missing information, observations that belonged to extremely small industry groups, and the observations that were outliers were dropped.

9. An ideal instrumental variable could be the employment rate in each area. However, since the province information was removed from the data by the NSO (not sufficient observations for each province for the monthly data), the only location variations available were regions and urban/rural areas which were not sufficient.

Additional information

Funding

This research is supported by the Ratchadaphiseksomphot Endowment Fund Part of the “Research Grant for New Scholar CU Researcher’s Project.”