Why Can’t I keep my Surname? The Fairness and Welfare of the Japanese Legal System

ABSTRACT

This study examines the welfare and fairness implications of Japan’s current policy on marriage surnames versus the proposed revised family law, which would enable husbands and wives to retain their premarital surnames. The study compares welfare in these two legal states, with a married couple’s welfare dependent on marriage-surname choice. It reviews the external preferences of anti-revisionists by the fairness criteria of impersonality or extended sympathy. Utilizing web-based survey data, the study conducts nonparametric rank analysis and parametric analysis of willingness to pay (WTP) for surname retention and legal support. Moreover, it conducts a structural equation analysis via a multiple indicators multiple causes (MIMIC) model, incorporating surname attachment and fairness as latent variables. The study shows that the revised law can increase welfare and that external disutility of the legal revision is invalid on fairness grounds.

KEYWORDS:

• Fairness
• family law
• gender
• welfare
• external preferences
• surname

JEL:

• Codes: D63
• J16
• K36

ACKNOWLEDGMENTS

I am grateful to Flavio Comim, Shailaja Fennell, Hiroyuki Kuribayashi, Noriko Mizuno, and Isamu Yamauchi, as well as participants in various conferences and seminars, for their helpful comments and discussions. I also wish to express my sincere gratitude to the three anonymous referees for valuable and detailed comments and suggestions. All views expressed are mine, and all errors are my own.

SUPPLEMENTAL DATA

Supplemental data for this article can be accessed at https://doi.org/10.1080/13545701.2019.1588467.

ORCID

Makiko Omura http://orcid.org/0000-0002-6970-7181

Notes

1 As of April 2018, 189 states have ratified or acceded to the treaty. Of the OECD countries, the United States is the only country that has not done so.

2 The ban on remarrying within three hundred days was ruled unconstitutional by the Japanese Supreme Court in December 2015, although the restriction remained in place for one hundred days, and the ruling became effective after cabinet approval in March 2016. Nonetheless, the CEDAW Committee condemns the prohibition imposed only on women for being discriminatory against women.

3 The word “welfare” is utilized here. However, the concept applied here can be considered in a broader sense, such as “capability” in Sen’s (2001) sense.

4 This article applies to all cases in which both partners of the couple are Japanese citizens.

5 The ratio of children born outside of wedlock is extremely low in Japan compared to other OECD countries.

6 Households in the agricultural sector are excluded from employed households.

7 The marriage rate and divorce rate are defined as the number of cases per 1,000 population in any given year. The latest figure is 0.346 in 2016.

8 This fairness concern would correspond to “moral preferences” defined by Harsanyi, which enforces “a special impersonal and impartial attitude, that is, a moral attitude, upon himself” (1977: 635). The definitions and distinctions Harsanyi applied to “personal preferences” and “moral preferences” are not equivalent to those of Dworkin’s (1997) “personal preferences” and “external preferences.”

9 While Bentham (1823) regarded welfare straightforwardly calculable and comparable, Mill (1951) did not.

10 Elizabeth F. Emens (2007) discusses the costs of making an unconventional name choice in detail. In terms of administrative costs, based on our crude estimation considering various administrative procedures required for surname change and the associated opportunity costs, the cost of changing one’s surname ranges from JPY 7,137 to JPY 18,210, based on the minimum wage and average full-time wage, respectively. If I multiply these costs by the number of couples married in 2009 (approximately 360,000), the total cost to society becomes JPY 2.6 billion or JPY 6.6 billion.

11 Here, CS^p–unfair choices, y₀₁ and y₁₁, are considered unfair because the subject is not willing to change his surname despite his preference for a CS, that is, he expects his partner to change her premarital surname.

12 The WTP analysis methods for payment cards showing different ranges of amounts are discussed in Timothy C. Haab and Kenneth E. McConnell (2002). I employ categorical WTP ranges, since asking an open-ended question would likely confuse the respondents with too many options. In addition, since I am not interested in measuring their exact WTP, possible anchoring bias is considered less harmful. The notion of latent WTP* here implies that such WTP* for different choices, such as DS and CS, can be placed along a continuum of utility range, although WTP for each choice is considered separately here with non-negative values.

13 The logit model is considered here given the relatively small proportion of respondents who expressed any WTP above 0. Although there are criticisms about using contingent valuation to assess preferences for public goods, as reviewed in Peter A. Diamond and Jerry A. Hausman (1994), my examination of “comparative WTP” across different options that are mutually exclusive and collectively exhaustible is not considered vulnerable to most criticisms, unless certain options are systematically more prone to bias than other options. Some possible types of WTP bias are discussed where relevant.

14 The MIMIC model is an extension of the standard item response theory model, which extracts information through covariances (see Bengt Muthén [1988]).

15 The appropriateness of ordering is checked a posteriori via correlation coefficients and graphical examination. While all ordinal variables – q2, q5, and q6 – exhibit coherence with Fairness, and q4 and q6 exhibit coherence with Attachment, categorical variables - q1 and q12 exhibit no clear ordinal relationship with Attachment (see the Supplemental Online Appendix).

16 Any original disaggregated data of any survey conducted by the Japan Office of Public Relations are discarded after the production of its report, which provides only aggregated results, and thus, are unavailable even for academic purposes (personal communication with ministry staff).

17 The questionnaire used the POTORA site managed by NTT-Navispace. The registered users, called “monitors,” have to register their personal details, and their record undergoes a regular personal identification validity check. The monitor characteristics are highly similar to any other web-based survey monitor panels, with a larger representation of those living in major cities and those ages 30–40.

18 While there is a concern for measurement error and selection bias by using a stratified random sampling method, this method is considered to generate less bias compared to complete random sampling with posteriori weight application.

19 Respondents who took too short a time to complete the questionnaire were judged “non-serious” and were removed from the effective sample, together with obvious contradictory responses.

20 This survey was administered as a trial survey by Cross Marketing Inc., one of the largest survey companies in Japan, which has high quality-control measures. Because of the complete random sampling method, gender–age sampling weights are applied. Age categories are in five-year intervals and respondents’ ages vary between 18 and 69 years.

21 There seems to be no specific type of trend or bias across all questions, yet there are certain comparative characteristics. First, older generations tend to be more conservative, in the sense that they have more negative views regarding the DS system, explicitly or implicitly. This tendency is seen particularly for the government surveys, which have disproportionately higher ratios of senior respondents 60 plus years old. On the contrary, the younger respondents in the government surveys tend to be more “liberal” or “nontraditional” in accepting the DS system than their web-survey counterparts. It is worth noting that the summary results of government surveys are misleading in that neither the disproportionate age representation nor the respondents’ localities are adjusted a posteriori.

22 A study by John C. Whitehead (2006) on the WTP for a special fishing permit allowing holders to retain the present quota found the major reason for zero WTP in the contingent valuation method was that respondents fundamentally disagreed with the policy and perceived such a policy as unfair.

23 As aforementioned and specified in Figure 1, surname preference is considered to manifest one’s fairness and surname attachment, and the result is robust to the inclusion of either latent variable.

Additional information

Notes on contributors

Makiko Omura

Makiko Omura is Professor of Economics at Meiji Gakuin University in Tokyo. She earned her PhD from the University of Cambridge and has worked on various research fields, such as development economics, law and economics, and gender issues. Earlier in her career, she served as Programme Officer for the United Nations Development Programme.