Does high unemployment rate result in a high divorce rate?: a test for Japan.

AuthorKawata, Yukichika
PositionReport
Pages149(16)
  1. Introduction

    When and why we bring an end to our marriage? It is a tough question to answer. Nevertheless, there is an indication in Japan that as the unemployment rate increases, the divorce rate registers a hike as well. In fact, this tendency seems statistically evident when we apply a simple correlation analysis. Some public institutions and other sources have already pointed out this fact on the Internet (1). Such evidence can be classified into two types: one, using time series data, and the other, using cross-section data. Both indicate a close relationship between the unemployment rate and divorce rate.

    However, these calculations are very amateur with regard to the methods employed in their analysis. Therefore, certain issues remain unresolved. First, according to our knowledge, there are no academic papers that study this phenomenon in the context of Japan. Besides, existing studies have not applied the appropriate procedures such as cointegration analysis (unit root test and cointegration test); therefore, we cannot confirm whether this correlation is genuine or spurious. This paper, therefore, uses a statistically rigorous procedure to validate the authenticity of the relationship.

    Fortunately, there have been some existing studies in other countries on this matter. Our analysis can be classified as a branch of family economics whose roots can be traced to the prominent works of Becker (1974a, b) as they comprise the fundamental literature that deals with marital instability. Though other methods exist, Becker's model is the one that is most used in studies dealing with marital issues that use similar statistical methods (Eliason, 2004). The basic explanation provided for divorce is that the people in concern discontinued their marriage when they felt that the net benefit from divorce is more than the net benefit from continuing the marriage.

    Some researchers argue that the impact of job loss goes as follows. When couples decide to marry, they may not know each other completely. As time goes by, couples' understanding of each other grows and sometimes as well, unexpected occurrences take place or unobserved characteristics of the spouses are revealed. Therefore, if the newly-gained information and/or unexpected incidents decrease the net benefit from the marriage, a divorce takes place. A job loss can be one of such unexpected incidents (Jensen and Smith, 1990; Boheim and Ermisch, 2001; Kraft, 2001).

    However, a job loss does not necessarily increase the possibility of a divorce. This is because the dismissed spouses may find it difficult to remarry due to their low income, as pointed out by Eliason (2004). Therefore, an empirical study is crucial; in this paper, we use time series data along with cross section data for Japanese couples.

  2. Brief Review of Existing Studies

    In a related move, we briefly take a look at the former studies on this issue that have been conducted in other countries. Studies on this issue are relatively scarce, and almost all come from Europe: Denmark (Jensen and Smith, 1990), Germany (Kraft, 2001), United Kingdom (Boheim and Ermisch, 2001), Sweden (Eliason, 2004), and Norway (Hansen, 2005; Rege, Telle, and Votruba, 2007). Naturally, if we include papers dealing with issues that are less related, the numbers increase. To name a few, Charles and Stephens (2004) and Stevenson and Wolfers (2007) overview divorce from an economic point of view including the relationship between job loss and divorce in their studies. Strom (2003) reviews this topic from sociological and psychological perspectives.

    Jensen and Smith (1990) used panel data of about 3000 Danish married couples from 1979 to 1985 and revealed that husband's unemployment immediately influences the probability of marital dissolution. They also suggested that the wife's unemployment does not significantly affect marital dissolution. It was estimated that around 6% of the divorces were due to reasons of unemployment.

    Kraft (2001) used panel data for a sample of 7,300 from the 39,959 available German household records from 1987 to 1996. Compared with Jensen and Smith (1990), Kraft's work featured a better analysis in that (1) it applied panel method for probit and logit, (2) it used separation instead of divorce as a variable, and (3) the number of observations was higher (Jensen and Smith, 1990, p. 68). He concluded that an unemployment spell of several months significantly increases the possibility of divorce.

    Boheim and Ermisch (2001) used panel data of 5,500 households for the time period 1991-1998. Their results suggested that a negative economic change increases the probability of marital dissolution. Boheim and Ermisch also stated that unexpected financial gains have a negative affect on marital dissolution. They compared cohabiting and married couples and found that the possibility of marital dissolution for the former is twice as large as it is for the latter.

    Eliason (2004) analyzed administrative data of Sweden and examined both immediate and long term implications of job loss over marital dissolution. Eliason showed that a job loss not only has an immediate impact on the possibility of marital dissolution but also that this effect is long lasting.

    Hansen (2005) used panel data of 8,933 Norwegian couples from 1989 to 1996 and suggested that job loss increases the possibility of divorce. He found that in Norway, the impact of the husband's job loss is lower whereas the job loss of the wife is higher if compared to the other countries. However, in Norway, as in other countries, the impact of the husband's job loss is more than that of the wife's job loss.

    Rege, Telle, and Votruba (2007) used panel data of 80,932 married Norwegian couples and showed an 11% rise in the number of divorce cases until 2003 in households affected by the husband's job loss due to plant closures, between 1995 and 2000, as compared to households wherein the husband did not have suffered job loss. They also revealed that households affected by plant closures do not show a drastic reduction in earnings that "suggests that the effect of husband's plant closure on divorce is not due to an unexpected reduction in earning" (Rege, Telle, and Votruba, 2007, p. 20). The authors then argue the applicability of role theory in this case.

    Compared to the above existing studies, our paper has the following distinctive features. First, we use both time series and cross section data. This can be advantageous as results can be empirically obtained using a lesser amount of data and disadvantageous in the sense that the reliability of the method hinges on limited aggregated data. Second, we apply cointegration analysis to avoid spurious regression. Finally, we searched for lag periods between the occurrence of high unemployment and increase in divorce rate.

  3. Materials

    3.1. Divorce Situation and Variable...

To continue reading

Request your trial

VLEX uses login cookies to provide you with a better browsing experience. If you click on 'Accept' or continue browsing this site we consider that you accept our cookie policy. ACCEPT