| kwwa |
2002-10-28 15:07:15, 조회 : 490 |
An Analysis of Determinants in Female Labor Market Participation / by Taehong Kim
/ KWDI Research Reports/Women's Studies Forum, Vol.17 / December 2001
* This article modifies and summarizes Chapter 2 of Change of Female Employment Structure and Policy Directions(2000).
Taehong Kim, Senior Fellow
INTRODUCTION
Female labour market participation has been gradually increasing since 1963. However, the changes in female labor market participation show differences according to age, marital status, and household type. In Korea, various studies have been conducted on the determinants of female labor market participation in the 1990s. However, these studies concentrated on specific years and the results of the studies were different due to the use of different estimation models or explanatory variables. Accordingly, it has been difficult to judge what are the major determinants in female labor market participation and how these determinants influence the participation rate. In this article, I have analyzed the structure and trends of female labor market participation rates, determinants and the degrees of influence of the determinants during the mid 1980s to 1990s considering the limitations mentioned above.
TRENDS AND CHARACTERISTICS OF LABOR MARKET PARTICIPATION RATE
A. Long-term Trends in Labor Market Participation Rate by Sex
An examination of look into the labor market participation rate beginning with 1963 reveals it was 56.6% in 1963, 57.6% in 1970, 59.0% in 1980, 60.6% in 1990 and 62.2% in 1997, right before receiving IMF relief aid. Up through 1997, the rate has gradually increased. However in 1998 and in 1999, the rate decreased to 60.7% and 60.5% respectively. Like many other countries, the participation rate in Korea shows a big difference according to sex. The male participation rate decreased constantly from 78.9% in 1963 to 72.1% in the mid 80s. After 1986, the male participation rate recovered a little to 74.0% in 1990 and 75.6% in 1997. But in 1998, the year of large scale layoffs, the male participation rate dropped again to 75.2% in 1998 and 74.4% in 1999. This phenomuna of continuous decline in the male participation rate is seen in advanced countries, too. This is probably because people engage in full-time employment later than in the past as the years of education have been lengthened and they are being retired earlier than in the past. Additionally, the number of working days and hours has decreased, the number of holidays and vacations have increased, and social welfare programs have expanded (Researchers hold different opinion concerning some factors. O. Ashenfelter, and R. Layard, eds. Handbook of Labor Economics, Vol. 1 (North-Holland, 1986), 3~10). In Korea, as you can see in the following analysis, the rate change has resulted from the lengthened number of years of education. More people go to college or university and the average age of male workers who enter the labor market is older than before. Therefore the participation rate of the men between 15 and 29 has decreased. However, more middle aged and elderly men participate in the labor market.
[Figure 1] Trend in Labour Market Participation Rate by Sex
The female labor force participation rate has increased unlike that of the male rate. It was 37.0% in 1963, 39.3% in 1970, 42.8% in 1980, 47.0% in 1990 and 49.5% in 1997. However, the female labor force participation rate has also decreased drastically in 1998 and 1999, being 47.0% and 47.4% respectively. In the late 1980s the female labor force participation rate increased the most. (It increased from 40.7% in 1984 to 47.0% in 1990, a 6.3% increase). There are two main reasons for the increase in the female participation rate. One is the changes in the structure of demands for labor, such as changes in the industrial structure which make women’s participation easier. The other is the changes in the labor supply, such as higher educational levels, decreased birth rates, consciousness change among working women, and a better child care system(Ibid, J. Humpheries (1995), Taehong Kim, (1996), Jiyeon Jang (1998), Haesun Bae (1995).
There are some who think the small electric appliances for household are the important factors to increase the number of working women. But some scholars think it is the result of the increased female labor force participation). An examinination of the labor force participation rate by region reveals that the labor force participation rate of rural and urban males is almost same, although there are trivial differences by year. The change trend is similar, too. In other words, the labor force participation rate of men from both rural and urban sectors decreased between 1963 and the mid 1980s (Since 1963, it was the lowest in 1984. In 1984, the participation rate of men from the rural sector was 72.0% and that of men from the urban sector was 72.0%.) After 1984, it has gradually increased. Especially, after 1985, the men from the urban sector show a little higher participation rate than those of the men from rural sector.
If we look into the female labor force participation rate by region, the female labor force participation rate has gradually increased since 1963. In 1963, it was 41.6%, but in 1998 it became 68.2%, which was a 26.6% increase. The periods that showed the biggest increases in the female labor force participation rate from the rural sector were 1963~76 (15.4% point increase) and 1984~98 (15.9% point increase). The labor force participation rate of women from the urban sector showed little difference from 1963 until 1972. However, it began to increase from 1972, being 28.9% in 1972, 44.4% in 1990 and 47.5% in 1997. The participation rate change by region after receiving the IMF relief aid in 1997 reveals the participation rate of urban sector decreased regardless of sex. In 1997, the male and female participation rate decreased from 75.2% and 47.5% to 74.6% and 44.6% respectively because of layoffs.
Comparatively, the participation rates of the rural sector increased. Rates were 79.1% and 67.3% in 1997 but those increased to 80.2 %, and 68.2% in 1998. In other words, the large-scale layoffs caused a decrease in the labor force participation rate in the urban sector and an increase in the rural sector; the decrease of participation in the urban sector because of IMF was bigger in females that in males. In 1999, the participation rate of males from the urban sector decreased further and that of males from the rural sector increased a little. On the other hand, the participation rate of women was the opposite of men. While more women from the urban sector participated in the labor force, fewer women from the rural sector participated in the labor force(In 1999, the total labor force participation rate of the urban sector was 59.1% and that the rural sector was 73.4%. If analyze it by sex, the female labor force participation rates are 45.4% and 66.3% in the urban sector and the rural sector respectively. However, the rates of men are 73.7% and 80.3% respectively).
If we analyze the cause of fluctuations by de-composing the female labor force participation rate as rural and urban sector(As the labor force participation rates of women from the urban sector and the rural sector are significantly different from each other, the increase and decrease of female labor force participation rates can be divided into female population structure changes between the sectors and female labor force participation rate in the sectors. In other words, the division of the labor force participation rate by effect can be identified in the following formula.
If we divide the population (P) into the rural sector (P₁) and the urban sector (P₂), the total labor force participation rate (1fpr) can be symbolized in a formula (1). If we integrate formula (1) against time, if will be formula (2) that classifies the effect as mentioned above. From here, Pit=(Pit/Pt) after 1980, the change in female population distribution between the rural and urban sectors has played a role of decreasing the female labor force participation rate. (see Figure 2). As the ratio of the female population is decreasing in the rural sector where the female labor force participation is relatively high, the female population in the urban sector is increasing where the female labor force participation is relatively low.
This resulted in the female labor force participation rate constantly decreasing. However, in the late 1980s, the female labor force participation rate regardless of region has increased remarkably. Although the difference was bigger in the rural sector than urban sectors, the latter influences the total labor force participation rate more, as it occupies a high erpercentage of the population. This trend was maintained till 1998, the year of IMF relief aid. After 1998, both the component ratio effect and the participation rate effect have changed negatively.
B. Labor Force Participation Rate by Age
The traditional statistics that reveal the sex differences are found in the "labor force participation rate curve by age" which shows that men enter the labor market in their 20s. By the time they are 29, most men engage in economic activities which will be maintained until they are 50, the retirement age. After 50, the rate decreases drastically, so the appearance of the male participation rate becomes is a plateau. The male participation rate curve by age like a plateau does not show a big difference from 1980. However, as more and more men go to university or college and graduate school, the participation of men between 15 to 29 has decreased constantly (In Korea and Japan, to enter university or college, one should dedicate oneself to the examination preparation. Therefore, the increase of the rate of those going to college leads to decreased of labor force participation rate. Taehong Kim (1996)). Consequently, the male labor force participation rate between 15 and 19 decreased from 27.3% in 1980 to 8.6% in 1997. Comparatively, more males over 55 have participated in the labor force from 1985.
[Figure 2] Factors Influencing the Female Labor Force Participation Rate The Korean female labor force participation rate curve is the M shape or twin peak type indicating a exit from the labor market because of marriage and child rearing and a return to the market after rearing the children. In the period from 1985 to 1990, when the female labor force participation rate increased dramatically, all age segments except 15 to 19 shifted upward. The participation rate of females from 20 to 24 and 25 to 29 and over 60 increased remarkably. The increase in those age segments influences the female labor force participation rate increase more compared to other age segments. However during the same period, the population structure change effect (The effect that female population structure change by age segment has on the female labor force participation rate) acts negatively on the participation rate(The influence of the labor force participation rate because of population structure change by age segment under the same circumstances was -.0.15% point, -0.18% point, -0.02 point, -0.03% point and 0.19% point in 1985~86, 1986~87, 1987~88, 1988~89 and 1989~90 respectively). Between 1990 and 1997, the participation rate of the women between 15 and 19 has decreased constantly and that of 25~29 and 40~44 have increased a lot.
[Table 1] Changes in Female Labor Force Participation Rate by Year and Age
Unit: %
+-----+-----------------------------------------------------------------------------+
| | 15~19 20~24 25~29 30~34 35~39 40~44 45~49 50~54 55~59 60+ |
+-----+-----------------------------------------------------------------------------+
|1963 | 37.4 43.6 36.4 39.4 41.8 48.6 45.1 38.7 32.7 10.8 |
|1970 | 44.1 47.1 34.6 38.3 42.7 47.0 46.5 41.2 37.1 14.7 |
|1980 | 34.4 53.5 32.0 40.7 53.0 57.0 57.3 54.0 46.2 17.0 |
|1985 | 21.1 55.1 35.9 43.6 52.9 58.2 59.2 52.4 47.2 19.2 |
|1990 | 18.7 64.6 42.5 49.5 57.9 60.7 63.9 60.0 54.4 26.4 |
|1995 | 14.5 66.1 47.8 47.5 59.2 66.0 61.1 58.3 54.3 28.9 |
|1997 | 13.0 66.4 54.1 50.9 60.4 67.1 62.3 58.1 54.1 30.3 |
|1998 | 12.0 61.0 51.8 47.3 58.5 63.5 61.5 55.2 51.0 28.1 |
|1999 | 11.8 60.8 52.3 48.2 58.6 63.1 62.8 55.5 51.3 29.4 |
+-----+-----------------------------------------------------------------------------+
Source: National Statistical Office (of the year), Annual Report on The Economically
Population Survey
In the twin peak curve (M-shaped curve), the labor force participation rate dropped from the first peak (about 20~24) until 1990. But after 1990, the average age of first marriage became higher, and the first exit of the working women occurred in 30~34 since 1995. Additionally, the decrease after the first peak has become gradually smaller. If we compare the first peak and the second peak in the M-shaped curve, the difference between the peaks is getting bigger as the participation rate of 40~45 has been dropping after 1990. In 1998, the participation rate of 20~24 dropped drastically because of IMF and layoffs, so the two peaks became similar. The participation rate of women over 60 is increasing gradually.
In females, there are different characteristics between the labor force participation curve of the urban sector and that of the rural sector. Accordingly, I have reviewed the female participation rate curve by age in the urban sector and the rural sector. First, women’s participation rate curve in the urban sector is the twin peak curve (M-shaped curve) showing the first peak in 20~24 and the second peak in the 40~44 age brackets. The first peak is always higher than the second peak. If in the urban sector, we compare women’s participation rate curve by age in 1980 with that in 1997, the participation rate of the 15~19 age bracket decreased a lot while the older age segmen t’s participation rate increased dramatically. The decrease of the 15~19 age bracket resulted from university or college education. The age bracket that shows the biggest increase in participation rate is 25~29, which increased from 27.0% in 1980 to 53.9% in 1997.
This is probably because fewer women chose to exit owing to marriage or delivery and the average age of the marriage and the first delivery shifted from 25~29 to 30~34(In major advanced countries, the decrease of the labor force participation rate of women because of marriage and delievery is being mitigated). Because of delayed marriage and delivery, the lowest peak of the curve in 1997 shifted from 25~29 to 30~34.
Additionally, the participation rate of 40~44 increased from 44.8% in 1980 to 65.5%. The second peak was nearly as high as the first peak. Regardless of the increase in the female labor force participation rate, the increase in 20~24 and over 60 was relatively low. (They were 52.8% and 8.8% in 1980 and 1997 66.3% and 20.0% in 1997 respectively.).
[Figure 3] Women’s Participation Rate Curve by Age Group in Urban Sector
The women’s participation rate curve in the rural sector is different from that of women in the urban sector, although it is an M-shaped curve, too. The participation rate of women in the 20~24 bracket is similar to that of the urban sector. However, 85~90% of the women in the second peak (40s and 50s) in the curve were working as of 1997. This means that most women in their 40s and 50s in the rural sector are participating in labor. Especially the participation rate of the women over 60 in the rural sector drastically much from 25.2% in 1980 to 61.7% in 1997, 36.5% increase, which means that the labor force in the rural sector is getting older. Besides, the participation rate of the women in their 50s increased rapidly, about 20%.
The other characteristics of the women’s labor force participation rate curve in the rural sector is the lowest peak of the M-shaped curve. Between 1980~85, the participation rate in the first peak (20~24) is same as that in the lowest peak (25~29).
Strictly speaking, we cannot say that it is a M-shaped curve. The M-shaped curve appeared from 1990. Even here, fewer women left the labor market because of marriage and delivery than in the urban sector. Additionally, in the rural sector, the participation rate of women in their 20s has increased remarkably but those in their 30s do not show a big difference.
C. Labor Force Participation Rate by Marital Status
The labor force participation rate according to marital status, shows that married men have maintained a rate around 88.0% from 1980 until 1997. But the participation rate of single men has decreased from 52.4% in 1980. It was 43.5% in 1985 and 43.7% in 1990. Of course, the participation rate of single men has increased from 1990, becoming 49.6% in 1995 and around 49% from then on.
The participation rate of single women has shown a continuous decrease from 1980 (50.8%), like single men, dropping to 46.5% in 1990. But in the beginning of the 1990s, it recovered again and in the late 1990s it has maintained the level of 1995. Married women showed a big increase from 1985 to 1990. In the 1990s it was increasing a little before the IMF financial crisis. Conclusively, the participation rate of singles, regardless of sex, decreased in the 1980s but recovered in the beginning of the 1990s. In contrast, other hand the participation rate of married men did not show much difference from that of 1980. But that of married women increased dramatically in the late 1980s(In major advanced countries, such as the USA. Canada, England, and Germany, the continuous increase of female labor force participation rate is due to that of the married women O. Ashenfelter and R. Layard, eds.(1986), 104~107).
Analysis shows that the female labor force participation rate increase in the period of 1985~1990 resulted from female population structural changes due to marital status or due to the actual increase of participation of women regardless of their marital status. According to the analysis, although the component ratio of married women has increased since 1985 because of the decrease of young women, the population structure effect according to marital status seems to be trivial because there is not much difference in women’s participation by marital status.
[Table 2] Trend of Labor Force Participation Rate b Sex and Marital Status
Unit: %
+-------+-------------------------------------+-------------------------------------+
| | Male | Female |
| +-----------+------------+------------+------------+------------+-----------+
| | Unmarried | Married | Total | Unmarried | Married | Total |
+-------+-----------+------------+------------+------------+------------+-----------+
| 1980 | 52.4 | 88.3 | 76.4 | 50.8 | 40.0 | 42.8 |
| 1985 | 43.5 | 86.8 | 72.3 | 44.7 | 41.0 | 41.9 |
| 1990 | 43.7 | 88.4 | 74.0 | 46.5 | 47.2 | 47.0 |
| 1995 | 49.6 | 88.8 | 76.5 | 50.4 | 47.6 | 48.3 |
| 1996 | 49.5 | 88.4 | 76.1 | 49.5 | 48.5 | 48.7 |
| 1997 | 49.5 | 88.0 | 75.6 | 49.3 | 49.5 | 49.5 |
| 1998 | 48.5 | 86.8 | 75.2 | 46.0 | 47.3 | 47.0 |
| 1999 | 49.1 | 85.5 | 74.4 | 45.9 | 47.9 | 47.4 |
+-------+-----------+------------+------------+------------+------------+-----------+
Source: National Statistical Office (1998), Annual Report on The Economically Active Population
Survey: 1963~1997; National Statistical Office (1999), Labor Force Population Survey, raw data set
However, as the labor force participation rate of married women occupies a bigger part of the population that increased between 1985 and 1990, they contributed much to the increased participation rate of women in the late 1980s. But between 1990 and 1993, single women contributed to the actual increase and between 1994 and 1997, married women contributed more to the actual increase again. An examination of the factors that influence the female labor force participation rate in 1998 and 1999 shows that the female labor force participation rate increased as a component ratio of married women due to the high participation rate increase in 1998. But if we examine the female labor force participation rate by marital status, that of the singles and the married decreased respectively.
Accordingly, the decrease in 1998 resulted from the decrease in married women working. This is because although the decrease rate itself is smaller than that of singles, the component ratio of thepopulation is bigger. In 1999, the labor force participation rate is converted into an increasing trend focusing on singles.
D. Male and Female Labor Force Participation Rate by Education
An examination of the female labor force participation rate by education, reveals that college graduates have the highest rate and university graduates, high school graduates, and under middle school graduates follow. The change trend of the participation rate indicates that the female labor force participation rate has a large increase between 1985 of 1990 regardless of a woman? education level. Between 1990 and 1997, the participation rate all women groups except under middle school graduates increases.
Especially, university graduates show the highest growth. But in 1998, when Korea experienced economic depression and received IMF funds, the participation rate of women university graduates decreased greatly (-4.3% point) but the under middle school graduate segment showed a relatively low decrease. The reason why it shows a smaller decrease is that many of the these persons were working as unpaid workers in agricultural sector and were less influenced by the economic depression.
The participation rate of males, shows a big difference in participation rate according to education. While 90% of college graduates and university graduates participate in the labor force, only 80% of high school graduates and 60% of middle school graduates or under participate in the labor force. Since 1985, there was a small increase from 1985 to 1990 except for university graduates. In 1998, all the segments show a decrease. The reason why they show a smaller decrease regardless of the economic depression in 1998 seems that the laid off male employees remain as an unemployed labor force rather than turning into a non-labor force (As the result of unemployment analysis also shows that there is a difference of the probability of transition from employment to unemployment and from employment to non-labor force according to sex in 1998).
The male and female labor force participation rate change factors can be divided into structure effect (effect of the increases when a highly participating group is increasing) and participation rate effect (effect of the participation rate increases in each educational levels). First, the structure effect in the changes of female labor force participation rate, shows that since 1985 more females have finished college and university. Accordingly, even though the female labor force participation rate has not changed, the female population structure change according to education has yielded a 0.1% point annual increase of the labor force participation rate continuously.
However, the major factor to increase the female labor force participation rate was the increases of female labor force participation rate in each education level. That is, the reason why the female labor force participation rate from 1985 to 1990 dramatically increased is due to the increased female labor force participation rate in the under middle school graduates segment. The increase in the 1990s is mainly owing to the increase of the female labor force participation rate in the high school graduate segment. In 1998, the female labor force participation rate decreased a lot mainly because the female labor force participation rate decreased in the high school graduate segment and under middle school graduate segment, both of which occupy a high portion of population. However, the portion of college and university graduates in the population has increased, which mitigates this participation rate change (According to the education level of the female population, 51.3% are under middle school graduates, 37.0% are high school graduates, 4.8% are college graduates and 6.9% are university graduates as of 1997).
[Table 3] Labor Force Participation Rate Change by Sex and Education Level
Unit: %
+-----+-------------------------------------+---------------------------------------+
| | Men | Women |
| +--------+--------+--------+----------+---------+---------+--------+----------+
| | Under | High | | Over | Under | High | | Over |
| | Middle | School |College |University| Middle | School | College|University|
| | School |Graduate|Graduate| Graduate | School | Graduate|Graduate| Graduate |
| |Graduate| | | |Graduate | | | |
+-----+--------+--------+--------+----------+---------+---------+--------+----------+
|1985 | 61.1 | 77.1 | 89.3 | 93.5 | 39.5 | 42.1 | 60.8 | 46.3 |
|1990 | 63.2 | 80.0 | 93.4 | 93.2 | 45.6 | 47.5 | 66.1 | 53.1 |
|1995 | 62.3 | 81.2 | 94.2 | 93.9 | 44.6 | 50.2 | 63.5 | 57.9 |
|1996 | 61.2 | 80.9 | 93.4 | 93.4 | 44.2 | 50.9 | 65.6 | 59.6 |
|1997 | 60.5 | 81.0 | 94.6 | 91.8 | 44.4 | 52.0 | 68.1 | 61.0 |
|1998 | 58.4 | 80.0 | 93.3 | 90.6 | 42.4 | 48.4 | 65.0 | 56.7 |
|1999 | 57.8 | 79.2 | 93.0 | 88.3 | 42.9 | 48.7 | 63.9 | 56.6 |
+-----+--------+--------+--------+----------+---------+---------+--------+----------+
Source: National Statistical Office (each year), Annual Report on The Economically Active Population Survey, Raw material
In 1998, the labor force participation rate itself decreased a lot. Therefore the female participation rate decreased drastically, regardless of population structure effect. In 1999, as the participation rate improved, especially in the low-education segment, it improved the labor force participation rate with a concomitant population structure effect. The labor force participation rate of low education females will increase due to the increase employment opportunities as temporay and casual workers in 2000. But in the mid to long term point of view, it will decrease from late 2000 because the middle school graduate population will be older (The female population in the under middle school graduate segment was 5,109,000 as of 1995 according to the census. According to age group as of 2000, the number over 65 years of age is 920,000 (18%), 60~64 is 634,000, and 55~59 is 686,000 of middle school graguates or under are over 50 as of 2000. If the present educational trend continues and the death rate of middle school graduates does not increase rapidly, it is expected that the portion of the middle school graduates will increase from currently 18% to 30~40% at the end of 2000).
On the contrary, the structure effect and participation rate effect are similar in males. That is, the male labor force participation rate change is greatly influenced by the change in the portion of university and college graduates: 50% of male labor force participation rate increase in from 1985 to 1990 was due to the actual increase of the male labor force participation rate and remaining 50% was due to the increase of the portion of university and college graduates. In 1990s the actual rate of male labor force participation decreased. But the participation rate increased if we add the structure effect and participation rate effect together, because the participation rate increase was significantly high due to the change of the portion of university and college graduates.
Comparatively, in the late 1990s the population effect was relatively small. Therefore male labor force participation rate showed a decreasing trend.
DETERMINANTS OF FEMALE LABO FORCE PARTICIPATION RATE
A. Estimation Model
We need to analyze the labor force participation model to identify the determinants and their level of influence. The major factors that influence female labor force participation were analyzed using the model that explains the participation as the relationship between market wage and reservation wage, that is equation (1-1). In equation (1-1), T is the variable showing whether a person participates in the labor market.
If he/she participates in the labor market, it is 1. Otherwise it is 0. wm and wr represent market wage and reservation wage and x and y are explanatory variables that decide market wages and reservation wages respectively. In ε0i=(εli-ε2i), if ε0i shows normal distribution where average is 0 and variance is σ0i, Equation (1-1) that shows the probability that an individual will participate in the labor force can be represented as a probit model like Equation (1-2).
Let us assume that we have “n” samples and “m” out of “n” participated in the labor force and the rest of “n-m” did not participate. The likelihood that we can observe the samples is like Equation (2-1) and index estimation of the variable that explains labor force participation is done by applying maximum likelihood method to Equation (2-2). The first-order condition of the Equation (2-2) is like Equation (2-3).
We analyzed the three years (1985, 1992, and 1997) to discover the factors that influence female labor force participation rate and their level of influence, using the same explanatory variables in estimating Labor Force Participation Model of all three years to make comparison possible (see Table 4 for the explanatory variables used in the analysis) (The local female labor demand variable is divided as high and low according to the third industry component ratio. If it is higher than the nation average, the area is sorted as high, otherwise it is sorted as low. There was another explanatory variable besides those used in the analysis. This was the attitude of husbands' toward the working wives. But they asked this question only of employees, with the unemployed and non-labor force hoping to work.
Therefore, I did not use this variable as an explanatory variable in this article). The date for the analysis came from the first, second, and third “Female Employment Status Survey” by KWDI (I used to scope and target of survey from first, second, and third survey (1986, 1992, 1997). The survey items were referred to the third Survey of KWDI). According to the National Statistical Office, the labor force participation rate of the married women has increased a lat from 41.0% in 1985 to 47.0% in 1992, and in 1997, it increased to 49.5%. That is, the labor force participation rate of married women has increased dramatically. Generally, most married women in the rural sector are participating in the labor force as non-wage earning workers. Accordingly, if they want to, they can be employed. It is relatively easy to manage labor activity and housework or child nurturing. The distance between work and house is relatively short. Therefore, the factors influencing the labor force participants in the rural sector can be different from those in the urban sector.
Accordingly, some researchers analyze the labor force participation of women not as binary choice (whether to work or not) but as multiple choice (working as paid workers, working as unpaid family workers, and non-working). In this article, I included the married women in the urban sector only because of the characteristics of the female labor force in the rural sector and the decrease of the married women in rural sector with the decrease of the rural sector in 1997. According to KWDI's Employment Status Survey, the labor force participation rate of the married women was 41.7% in 1985, 42.7% in 1992, and 44.7% in 1997).
[Table 4] Explanatory Variables used in Analysis
+-------------------------+---------------------------------------------------------+
| Type of Variable | Description |
+----------------+--------+---------------------------------------------------------+
| | |- It shows that they are high school graduates |
| |Dedu1 |- If he/she is high school graduates, it will be 1 |
| | | otherwise, it will be 0 |
|Education +--------+---------------------------------------------------------+
| | |- It shows that they graduated from college or over |
| |Dedu2 |- If he/she is college school graduates, it will be 1 |
| | | otherwise; it will be 0 |
+----------------+--------+---------------------------------------------------------+
|Age |Age |- Sequential variable representing age (15 to 64 years) |
| |Agesq |- Square of the age variable (Age X Age) |
+----------------+--------+---------------------------------------------------------+
|Experience |Exp |- It shows total experience after 15 years old |
| | | ※ It is achieved through retrospective method |
+----------------+--------+---------------------------------------------------------+
|Children under 6|Ch6 |- It shows whether there is a child under 6 |
| | |- If yes, it will be 1; otherwise, it will be 0 |
+----------------+--------+---------------------------------------------------------+
|Incoem of other | |- Income of the other household members except the |
|members in the |Oincome | respondent (Unit: 10,000 won) |
|household | | ※(Total income of the household - Income of the |
| | | respondent) |
+----------------+--------+---------------------------------------------------------+
|Whether has |Dmar |- It shows whether one has spouse or not |
|spouse | |- If yes, it will be 1; otherwise, it will be 0 |
+----------------+--------+---------------------------------------------------------+
|Local demand on |Fdem |- It shows the local demand for female labor |
|female labor | |- If high, it will be 1; otherwise, it will be 0 |
+----------------+--------+---------------------------------------------------------+
B. Results of Analysis on Determinants of Labor Force participation Rate
The function to estimate the labor force participation of the married women can be seen in [Table 5]. The estimation results of the labor force participation model of married women show that, most of the explanatory variables seem to be significant.
When the explanatory variables are examined one by one, the dummy variable showing education level was found to be statistically significant in a negative sign in 1985 and 1992. However, in 1997 the estimation coefficient education variables were all positive numbers. Especially, the estimation index of university graduate dummy variable showed statistically significant results in positive direction. Generally speaking, if one’s education level is high, one can earn higher market wages. But a high education level also raises the value per hour of housework and the reservation wages, which lowers the participation rate. But in most of the preceding studies, the education level raises the labor force participation rate (According to Taehong Kim (1998), the education variable makes a statistically significant influence the female labor force participation rate in a negative direction. In this analysis, I estimated firstly the wage equation after correcting the problem of selectivity bias. Then, I estimate female participation equation with independent variable that are determainants of the reservation wage (for examples, the presence of children with 6 years and under, the number of family, education level, etc.) and market wage derived from estimated wage equation. In Korea the education variable influences the female labor force participation rate in both directions. Therefore there are many studies that say the education variable is statistically meaningless or the sign of the estimation coefficient is negative).
According to the estimation of the education variable, high school graduate married women show 13.5% less participation rate than middle school graduate married women assuming other conditions are the same (Marginal effect of the explanatory variable can be calculated through Formula (2-3). University graduate married women show an in significant 4.5% less participation rate than middle school graduate married women. In 1992, the gap between high school graduates and middle school graduates was lessening.
The participation rate of high school graduate married women was 5.4% lower than that of the middle school graduate married women. However the gap between the middle school graduate married women and university graduate married women became wider. The participation rate of university graduate married women was 8.9% lower than that of middle school graduate married women. But in 1997, the participation rate of university graduate married women was higher than that of high school or middle school graduate married women, being 10.5% higher than that of middle school graduate married women. But the estimation result of the high school graduate dummy variable was not significant statistically. The estimation result shows the influence of the education level of married women on the female labor force participation rate has changed a lot since 1985. That is, in 1985 and 1992, the higher the education of married women, the more their reservation increased over the market wage.
Therefore the labor force participation rate of middle school graduates is higher than that of high school graduates or university graduates. But in 1997, the market wage increase exceeded the reservation wage increase. Therefore, the more they were educated, the higher was their participation in the labor force participation. Additionally, an examination of the component ratio of the female population over 15, shows that the component ratio of high school and college graduates was growing continuously because more and more women went on to higher studies. (Refer to Appendix [Table 1]) (According to the Female Employment Status Survey, 16.1% of the women in the urban sector in 1992 was college graduates, and it increased into 19.0% in 1997. But the average of the dummy variable of college and university graduate married women included in the explanatory variable dropped to 13.6% in 1997. This was because the survey missed a portion of females who were over college graduates). Accordingly, it seems that the high education level of the female population influenced the labor force participation rate negatively in 1985~1992 but positively in 1992~1997.
[Table 5] Estimation of Yearly Labor Force Participation Factors
+--------------+----------------------+----------------------+----------------------+
| | 1985 | 1992 | 1997 |
| +------------+---------+------------+---------+------------+---------+
| | Estimates | t-value | Estimates | t-value | Estimates | t-value |
+--------------+------------+---------+------------+---------+------------+---------+
|Intercept | -2.5376*** | -6.764 | -3.0036*** | -5.977 | -2.2740*** | -4.842 |
|Dedu1 | -0.3567*** | -5.906 | -0.1388** | -1.924 | 0.0549 | 0.802 |
|Dedu2 | -0.1159 | -1.093 | -0.2307*** | -2.418 | 0.2856*** | 3.006 |
|Age | 0.1744*** | 9.472 | 0.1866*** | 7.667 | 0.1386*** | 6.425 |
|Agesq | -0.0025*** |-11.457 | -0.0026*** | -9.153 | -0.0020*** | -8.241 |
|Exp | 0.0662*** | 16.156 | 0.0597*** | 13.909 | 0.0440*** | 12.628 |
|Ch6 | -0.3382*** | -5.252 | -0.2827*** | -3.956 | -0.5456*** | -7.206 |
|Oincome | -0.0062*** | -7.552 | -0.0014*** | -3.915 | -0.0016*** | -6.054 |
|Dmar | -0.3682*** | -4.014 | -0.3912*** | -3.145 | -0.1848** | -1.916 |
|Fdem | 0.1548*** | 3.095 | 0.1925*** | 3.169 | 0.2338*** | 4.383 |
+--------------+------------+---------+------------+---------+------------+---------+
|Chi Squared | 3854.30 | 2139.58 | 2424.56 |
|Number of | 3,070 | 2,082 | 2,370 |
|Observations | | | |
+--------------+----------------------+----------------------+----------------------+
Note: ***, **, * are statistically significant at 0.01%, 0.05% and 0.10% respectively
The experience variable’s estimated coefficient shaws that it yielded statistically significant positive values in 1985, 1992 and 1997. That is, experience influences positively the labor force participation of married women because more experienced women can get higher market wages. However, there were small differences by year. In 1985, the participation of those with one year of experience increased about 2.5%, in 1992 it increased 2.3% and 1.7% in 1997(The marginal effect of the experience variable was calculated based on the average years of experience).
This shows that the influence of the experience variable on the participation rate is decreasing up to present. The same result appears in the estimation of the function to calculate the wages of working wives if we use same data (According to the estimation of the wage function, the wage increase per experience year was 0.0234 in 1992. But in has droppd to 0.0187). With this, the years of experience of married women is increasing. The life-long employment of married women has influenced much the increase in the female labor force participation rate.
One of the major factors that influence the labor force participation rate of smarried women is whether they have children under 6. If there is a child under 6, the reservation wage of a married woman increases. Therefore, the labor participation of a married woman in such a household is discouraged. That is, having a child under 6 influences negatively the labor force participation rate of married women. In the model estimating labor force participation of married women in Korea, the existence of a child under 6 has a significant negative influence on the participation rate. According to the estimated coefficient, married women with a child under 6 showed a 13.0% lower participation rate than those who did not have a child under 6 in 1985. The difference in the labor force participation rate was narrowed to 10.9% in 1992, but it was expanded to 20.6% in 1997. That is, the existence of a child under 6 negatively influences the labor force participation rate of married women.
An examination of the ratio of the women with a child under 6 based on the “Female Employment Status Survey” shows that the segment has decreased from 44% in 1985 to 39.0% in 1992 and 28.0% in 1997 because of the low birth rate. The decrease of the ratio of the women with a child under 6 will positively influence the female labor force participation rate. That is to say, because of the continuous decline in the birth rate, fewer women cannot participate in the labor force because of a child under 6. However, from the perspective of the individual woman, the married woman with a child under 6 has greater difficulty participating in the labor force because of the young child.
The estimation results of the age variable shows that the ages of married women comprises a reverse U-shape from 1985 to 1997. The labor force participation rate of married women in the urban sector increases until a certain age, and then it decreaseds. If other conditions are the same, the peak of the reverse U shape is about 35. If we analyze focusing on the samples of the average age of married women, a 1 year increase or decrease makes the participation rate increase or decrease about 1.0~1.5%.
The average age of the sampled married women, was 37.7 in 1985, 38.5 in 1992, and 40.8 in 1997 because of the lengthening of the life span. The average age increase after 1985 has lowered the labor force participation rate of the married women in the urban sector. The income of other household members except the respondents influences the labor force participation of women negatively. The same result has been achieved in the estimation of the labor force participation rate of married women for three years.
However, the negative influence of incomes made by other household members except the respondents has been weakened compared to that of 1985. That is, recently, the income of other household members does not influence as strongly as before. Additionally, whether one has a spouse or not is another variable influencing the financial status of the household. Divorced, widowed, and separated women tend to participate in the labor force because of financial needs. The spouse dummy variable negatively influenced the labor force participation of the married women in 1985, 1992, and 1997, but the influence of the spouse variable was weakened in 1997. This means that more and more married women participate in the labor force regardless of the presence of the spouse. The other factors that influence female labor force participation are economic and social conditions. The most representative factor is the demand for female labor.
Generally speaking, if there is high demand, the female labor force participation rate will be higher. Accordingly, I converted the service industry prosperity in the place where the married woman resides to an index and used it as proxy variable of female labor demand (Bowen and Finegan (1969) who used the Female Labor Demand Index for the first time, formulated the index through following method. Calculate the ratio of women against the total eomployees in the industry (F), multiple it by the total employees in the industry and get X. Finally, divide the sum of Xs (total sum of X of the individual industry) by the total employees in the place. In this article, I calcuated the ratio of the employees in the 3rd industry. If it is over average, the value is 1. Otherwise it is considered as 0). The estimation index of the labor demand variable, was statistically significant in a positive direction. The participation rate was high in the place of higher demand for female labor. By year, the influence of the female labor demand index is rising each year. In 1985, the ratio of the service industry increased 10% and it increased the female labor force participation rate by about 0.58%. In 1997, it increased by 0.87% (The average of the variables that show female labor demandn index was 0.613 in 1992 and 0.488 in 1997. This was because the service component ratio dropped under the average because Incheon was converted into a manufacturing zone in that period).
Generally, the attitude of the husband toward female labor force participation greatly influences the participation rate. In the “Female Employment Status Survey,” both the attitudes of the women themselves and their husbands were surveyed. But only the attitudes of women were surveyed in 1985 and some of the survey questions in 1997 were different from those of 1985. Therefore, the usefulness of the data in the analysis was limited. Additionally, only the husbands of the employed and those who want to be employed were surveyed. Therefore, the utilization of the data as explanatory variables for the labor force participation model was somewhat restricted. But if the attitude of the husbands of employed women is compared to those of unemployed who hope to work, the husbands of the employed show more positive attitude than those of the unemployed regardless of the year.
Additionally, the ratio of the positive attitude of husbands whose wives are performing economic activities is increasing up to the present. The husbands of the non-labor force women who hope to work show fewer objections to their participation in the labor force, but the ratio of those who “Agree and Welcome” remains the same. This shows indirectly that the change of the husbands’ attitude toward the labor force participation of the wife contributed to the participation rate increase of married women since 1985.
[Table 6] Husbands' Change of Attitude toward Working Wives
Unit: %
+-------+----------------------+---------+---------+---------+----------+-----------+
| | | Welcome | Agree | Depends |Objection |Impossible |
+-------+----------------------+---------+---------+---------+----------+-----------+
| | Working Women | 7.2 | 46.5 | 22.7 | 22.1 | 1.4 |
| 1985 +----------------------+---------+---------+---------+----------+-----------+
| | Non-working with the | 5.0 | 46.4 | 12.4 | 30.9 | 5.3 |
| | hope to work | | | | | &nb | | |