The following model was provided by Wing Suen in 1994 and assumes that a woman will manage her time between three primary activities: housework, market work and leisure. In order to simplify the model and allow the demand for domestic services and the supply of market work to be separate decisions regarding wage rates, two main assumptions were made. The first assumption is that one can perfectly substitute the chores done by native woman with the chores done by domestic workers. The other assumption is the perfect divisibility of the services provided by domestic workers. The two assumptions imply that when the native woman’s own wage exceeds the domestic workers’ wage, she will purchase the housework service.
Let D be the time spent on housework by the native woman, which is assumed to be a fixed length of time. T is the time endowment for the woman who hires a domestic worker, and will change to T-D if she does not hire one. The cost of hiring a domestic worker is P.
Due to the perfect substitution of housework done by the native woman and by the domestic worker we can say that it will take the two the same length of time to do the chores. Therefore, the domestic worker’s payment per unit of time would be P/D. The woman’s utility function is only directly affected by consumption and leisure, which are assumed to be normal goods.
Figure 3.1 shows the effect of the domestic worker program on women who originally did not join the labor force market. Panel (a) assumes that a woman’s potential wage, w, equals the domestic worker’s payment per unit of time. CAF is the budget constraint of the woman before the domestic worker program. F equals T-D. I0 is her unearned income.
Because the slope of U0 at A is steeper than the budget constraint, she would spend 𝐹𝑇̅̅̅̅ to
‧
do the housework while using 0𝐹̅̅̅̅ for leisure. After the implementation of the domestic worker program, the woman’s budget constraint will change to CBT, and the cost of purchasing a domestic worker makes I0 drop to I1. The slope of U1 at B is flatter than the slope of U0 at A, which means that the lower reservation wage allows the woman join the labor force. When the policy is put into practice, she will purchase a domestic worker and reallocate her time to leisure 0𝑁̅̅̅̅, and market work 𝑁𝑇̅̅̅̅. In this case, the domestic worker program will stimulate the woman to participate in the labor market.
Panel (b) assumes that the woman’s potential wage is greater than the domestic worker’s payment per unit of time. The situation would be the same as Panel (a) before the implementation of the domestic worker program. When the policy starts, the new budget constraint C1BT will dominate the original one. As the illustration shows, the slope of U1 at B is flatter than the slope of U0 at A. Given this, a woman will purchase a domestic worker and reallocate her time to leisure and market work after the policy was implemented. As a result, the policy also encourages women to join the labor market.
The case where the woman’s potential wage is smaller than the domestic worker’s payment per unit of time is illustrated in Panel (c). When P is large enough, the woman will refuse to hire a domestic worker and will stay at home to do the housework herself.
But it would be another story if the marginal rate of substitution between leisure and consumption was greater than the wage rate. In this situation the woman will choose to hire a domestic worker, changing her utility from U0 to U1. The three cases in Figure 1 all point to the positive relationship between the domestic worker policy and women’s labor force participation.
Figure 3.2 illustrates the effect of the domestic worker program on women who had joined the labor force market before the implementation of the policy. Unlike the consistent results of the cases in Figure 1, the policy in Panel (b) benefits the woman while not
‧
國立 政 治 大 學
‧
N a tio na
l C h engchi U ni ve rs it y
12
changing the woman’s utility level in Panel (a) and even makes the woman worse off in Panel (c). The three illustrations explicitly show that the decision of whether to hire a domestic worker or not is different among women and depends on their wage rates.
In sum, the model predicts that the emergence of the domestic worker programs is more likely to benefit women who did not join the labor force before the policy than it is to benefit those who had previously participated in the labor market. That is, non-employed women would be more willing to hire a domestic worker after the implementation of the policy.
The model provides us with a means to predict the effect of foreign domestic worker programs on native women in Taiwan. The following sections of empirical research will only discuss the overall impact of foreign domestic worker programs and will not discuss the six cases in the model separately because of the limitation of micro data. In fact, the foreign domestic payment per unit of time is far lower than the minimum wage instituted by the government in Taiwan. So the situation which faces native women may be similar in Figure 1(b) and Figure 2(b). According to the model, we may predict that foreign domestic worker programs will increase the labor force participation rate and the working hours of native women in Taiwan.