型I糖尿病是幼年時即罹患主要是先天決定，前兆式糖尿病與型II糖尿病較會受生活型態影響，因此，有別於傳統文獻，本文特別關注前兆式糖尿病與型II糖尿病對美國就業的衝擊，進而對前兆式糖尿病與型II糖尿病合併的廣義型II糖尿病模型體重估計體重控制對於中老年就業的緩和效果。資料係取自2011~2012年美國國家健康與營養調查(National Health and Nutrition Examination Survey)，按全樣本、男性與女性分組估計。 研究方法係先估計各型糖尿病對美國就業的衝擊，使用Probit模型估計糖尿病在外生情況下之效果，進而考慮前兆式與型II糖尿病潛在內生性問題，使用帶有工具變數二元與三元Probit模型進行估計，其工具變數為近親是否罹患糖尿病；更進而以三元Probit模型估計體重控制對廣義型II糖尿病負向衝擊就業的緩和效果，使用近親是否罹患糖尿病與體重控制盛行率為其工具變數。 研究結果顯示:若忽視前兆式糖尿病，僅將糖尿病離析為型I與型II糖尿病，外生模型中型II糖尿病對於就業的衝擊效果為負向顯著且男性影響效果皆大於女性。而當採用預防疾病觀點加入前兆式糖尿病時，內生模型中前兆式糖尿病效果呈男性影響效果大於女性，前兆式糖尿病之負向就業影響已是不容忽視。而將前兆式糖尿病與型II糖尿病合併，進一步定義為廣義型II糖尿病進行估計，其男、女性外生效果皆為負向顯著，男性就業機率降低12.1%、女性則降低9.4%。當使用廣義型II糖尿病內生模型進行緩和效果估計，結果顯示男性中老群組中，以男性過重以上群組有7.6%的緩和效果，此結果驗證了體重控制能有效地緩和糖尿病對於就業負向影響。 The onset of type I diabetes is usually at the beginning of childhood or in adolescence. The prediabetes and type II diabetes are likely to be affected by lifestyles and are often detected for older people. This study is different from the traditional literature in the way that it is particularly concerned about the impact of prediabetes and type II diabetes on employment. This study defines prediabetes combined with with type II diabetes as the generalized type II diabetes. For elderly, the model of the generalized type II diabetes is estimated to examine whether employment moderating effect of weight control exists or not. Data are drawn from the 2011 ~ 2012 National Health and Nutrition Examination of the U.S.. All the full sample, male and female groups are estimated. This study estimates the impact of various types of diabetes on employment using the Probit model under exogenous assumption first. It then takes the potential endogenous problems of prediabetes and type II diabetes into account by using the bivariate and the trivariate probit models with instrumental variables. The instrumental variable is whether the close relatives suffering from diabetes. Furthermore, the trivariate probit model with instrumental variables is used to estimate the effect of weight control on the negative impact of generalized type II diabetes on employment. The instrumental variables inclue both whether the close relatives suffering diabetes and the prevalence of weight control . If this study ignores prediabetes, diabetes is only divided into type I and type II diabetes. The exogenous model of type II diabetes indicates that its effect on the employment is negative and significant for men. Its effect is larger for men than women. When the use of prophylactic view of the disease by taking the prediabetes into account, the endogenous model of precursor diabetes indicates that its effect is greater for men than women, the negative effects of prediabetes can’t be ignored. The exogenous model of the generalized type II diabetes indicates that both the effects for men and women are negative, the odds of employment is reduced by 12.1% for men and by 9.4% for women. When using the endogenous model of the generalized type II diabetes to estimate the moderating effect. The results show that the moderating effect is 5.3% for overweight (and obese) middle-aged and older men. The results confirm that weight control can effectively alleviate the adverse effects of the generalized type II diabetes on employment.
