We test a standard DSGE (Dynamic Stochastic General Equilibrium) model on impulse responses of hours worked and real GDP after technology and non-technology shocks in emerging market economies (EMEs). Most dynamic macroeconomic models assume that hours worked are stationary. However, in the data, we observe apparent changes in hours worked from 1970 to 2013 in these economies. Motivated by this fact, we first estimate a structural vector autoregression (SVAR) model with a specification of hours in difference (DSVAR) and then set up a DSGE model by incorporating permanent labour supply (LS) shocks that can generate a unit root in hours worked, while preserving the property of a balanced growth path. These LS shocks could be associated with very dramatic changes in LS which look permanent in these economies. Hence, the identification restriction in our models comes from the fact that both technology and LS shocks have a permanent effect on GDP yet only the latter shocks have a long-run impact on hours worked. For inference purposes, we compare empirical impulse responses based on the EMEs data to impulse responses from DSVARs run on the simulated data from the model. The results show that a DSGE model with permanent LS shocks that can generate a unit root in hours worked is required to properly evaluate the DSVAR in EMEs as this model is able to replicate indirectly impulse responses obtained from a DSVAR on the actual data. JEL Classification: C32, E32

中文翻译：

---

新兴国家和DSVAR中的技术冲击和非固定工时

我们对新兴市场经济体（EME）中技术和非技术冲击后的工时和实际GDP的冲激响应测试了标准的DSGE（动态随机一般均衡）模型。大多数动态宏观经济模型都假设工作时间是固定的。但是，在数据中，我们观察到这些经济体从1970年到2013年的工作小时数有明显变化。基于这一事实，我们首先估算具有差异小时数（DSVAR）的结构矢量自回归（SVAR）模型，然后通过合并可在几小时内产生单位根的永久劳动力供应（LS）冲击来建立DSGE模型。在保持平衡的增长路径属性的同时发挥了作用。这些LS冲击可能与LS的巨大变化有关，这些变化在这些经济体中看起来是永久的。因此，我们模型中的识别限制来自这样一个事实，即技术和LS冲击都对GDP产生永久影响，而只有后者会对工作时间产生长期影响。出于推理目的，我们将基于EMEs数据的经验冲激响应与基于模型的模拟数据运行的DSVAR的冲激响应进行比较。结果表明，需要使用具有永久LS冲击的DSGE模型，该模型可以在工作小时内生成单位根，以正确评估EME中的DSVAR，因为该模型能够在实际数据上复制从DSVAR获得的间接冲激响应。JEL分类：C32，E32 我们将基于EMEs数据的经验冲激响应与基于模型的模拟数据运行的DSVAR的冲激响应进行比较。结果表明，需要具有永久LS冲击的DSGE模型，该模型可以在工作小时内生成单位根，以正确评估EME中的DSVAR，因为该模型能够在实际数据上复制从DSVAR获得的间接冲激响应。JEL分类：C32，E32 我们将基于EMEs数据的经验冲激响应与基于模型的模拟数据运行的DSVAR的冲激响应进行比较。结果表明，需要使用具有永久LS冲击的DSGE模型，该模型可以在工作小时内生成单位根，以正确评估EME中的DSVAR，因为该模型能够在实际数据上复制从DSVAR获得的间接冲激响应。JEL分类：C32，E32