Sedigh Golnaz, Devlin Rose Anne, Grenier Gilles, Deri Armstrong Catherine
Department of Economics, University of Ottawa, 120 University Private, Ottawa, Ontario, K1N 6N5, Canada.
Econ Hum Biol. 2017 Feb;24:125-139. doi: 10.1016/j.ehb.2016.11.010. Epub 2016 Dec 1.
This paper uses the 2005 and 2010 Canadian General Social Surveys (Time Use) to investigate the effect of wages on the sleep duration of individuals in the labour force. The endogeneity of wages is taken into account with an instrumental variables approach; we find that the wage rate affects sleeping time in general, corroborating Biddle and Hamermesh's (1990) main conclusion. A ten percent increase in the wage rate leads to an 11-12min decrease in sleep per week. But this number masks several effects. The responsiveness of sleep time to wage rate changes depends upon the sex of the individual, whether or not sleep problems are present and general economic conditions. By far the largest adjustment is found for insomniacs in 2010, a year of general economic downturn in Canada. We also investigate the non-randomness of insomnia in the population by using a Heckman procedure, and find that the sleep time of female non-insomniacs is even more responsive to wage rate changes once account is taken of this selection bias, but otherwise selection was not a problem in our samples.
本文利用2005年和2010年加拿大综合社会调查（时间利用）来研究工资对劳动力中个体睡眠时间的影响。采用工具变量法考虑工资的内生性；我们发现，一般而言工资率会影响睡眠时间，这证实了比德尔和哈默梅什（1990年）的主要结论。工资率每提高10%，每周睡眠时间会减少11 - 12分钟。但这个数字掩盖了几种影响。睡眠时间对工资率变化的反应取决于个体的性别、是否存在睡眠问题以及总体经济状况。到目前为止，在2010年加拿大总体经济衰退的这一年里，失眠者的调整幅度最大。我们还使用赫克曼程序研究了人群中失眠的非随机性，发现一旦考虑到这种选择偏差，女性非失眠者的睡眠时间对工资率变化的反应甚至更大，但在我们的样本中，否则选择不是一个问题。
