شوک های فناوری و پویایی های بازار کار: برخی شواهد و نظریه

A positive technology shock may lead to a rise or a fall in per capita hours, depending on how hours enter the empirical VAR model. We provide evidence that, independent of how hours enter the VAR, a positive technology shock leads to a weak response in nominal wage inflation, a modest decline in price inflation, and a modest rise in the real wage in the short-run and a permanent rise in the long-run. We then examine the ability of several competing theories to account for this VAR evidence. Our preferred model features sticky prices, sticky nominal wages, and habit formation. The same model also does well in accounting for the labor market evidence in the post-Volcker period.

Understanding labor market dynamics has been an important goal for business cycle studies, at least since Dunlop (1938) and Tarshis (1939). While the historical debates have focused on the cyclical behavior of wages, a more recent strand of literature has focused on the effects of technology shocks on per capita hours, with starkly different empirical conclusions. One side of the empirical literature suggests that a positive technology shock leads to a short-run fall in per capita hours. This result has been obtained from structural vector autoregression (SVAR) models, where a technology shock is identified as the only shock that affects long-run labor productivity (e.g., Gali, 1999; Francis and Ramey, 2005). 1 A similar result has also been obtained with technology shocks measured by a “purified Solow residual” that controls for non-technological factors that may affect measured total factor productivity (e.g., Basu, Fernald, and Kimball (BFK), Basu et al., 2006). Yet another line of research argues that a positive technology shock triggers a rise, not a fall, in per capita hours, even when the technology shock is identified using the same long-run restrictions as in Gali (1999) (e.g., Christiano, Einchebaum and Vigfusson (CEV), Christiano et al., 2004). The difference in conclusions arises from different treatments of hours in the SVAR: whether hours rise or fall following a positive technology shock depends on whether hours enter the SVAR in log-levels or log-differences. 2 The lack of consensus in empirical findings renders it difficult to assess competing macroeconomic theories based upon the effects of technology shocks on hours. If a technology improvement indeed leads to a fall in hours, then such evidence would cast doubt on the empirical relevance of the standard real business cycle (RBC) theory, which predicts that hours and output should be positively correlated following a technology shock. This evidence, however, seems to be consistent with the predictions of a sticky-price (SP) model with weak monetary-policy accommodation to technology shocks (e.g., Gali, 1999; BFK, 2006), or an RBC model augmented with habit formation and investment-adjustment costs (e.g., Francis and Ramey, 2005). If, on the other hand, a positive technology shock indeed leads to a rise in hours, then the standard RBC model (without habit formation) would do just fine and, as we show below, so would a pure sticky-wage (SW) model. The present paper proposes a way out of this dilemma. We go a step further in exploring the empirical evidence—from a broader perspective of the labor market. In particular, we examine the effects of technology shocks not only on hours but also on wages and prices. We present in Section 2 a four-variable SVAR model that includes the growth rate of average labor productivity, per capita hours, nominal wage inflation, and price inflation. We allow hours to enter in either log-differences or log-levels. This approach allows us to avoid taking a stand on the debate between the difference specification vs. the level specification. It also allows us to examine the robustness of the effects of technology shocks on wages and prices across the two alternative empirical specifications. Some empirical regularities emerge from the SVAR model. Following a positive technology shock, hours may rise or fall, depending on how hours enter the SVAR model. This finding confirms what others have found in the literature. More interestingly, independent of how hours are modeled, a positive technology shock leads to a weak response in wage inflation, a modest decline in price inflation, and a modest short-run rise in the real wage, which continues rising until reaching a permanently higher level. Aside from our own SVAR evidence, we find corroborating evidence from the work of BFK (2006), who obtain similar empirical regularities in wages and prices based on an independent measure of technology shocks. We argue that a reasonable business cycle model should be able to account for the dynamic adjustments of wages and prices documented here. Upon establishing these stylized facts, we develop a baseline model that captures the spirit of a broad class of DSGE models studied in the recent literature in the context of the technology-hours debate. We then use our baseline model to assess the abilities of several popular business-cycle models in accounting for the observed labor-market dynamics driven by technology shocks. Our general framework (presented in Section 3) features monopolistic competition in both the goods market and the labor market (e.g., Blanchard and Kiyotaki, 1987). We assume, in the spirit of Calvo (1983), that pricing decisions are staggered across firms and wage-setting decisions are staggered across households.3 We further incorporate in our model a source of real rigidity in the form of habit formation in preferences. We assume that monetary policy follows a Taylor rule under which the nominal interest rate responds to variations in inflation and output.4 Under calibrated parameters, we find that the baseline model does well in capturing the observed labor-market dynamics (Section 4). The model predicts that, following a positive technology shock, hours fall modestly in the short-run; the real wage rises modestly on impact and keeps rising until reaching a permanently higher steady state; nominal wage inflation responds weakly; and price inflation declines modestly in the short-run. These patterns of labor-market adjustments are broadly in line with the SVAR estimates. Quantitatively, the model's predicted impulse responses of the labor-market variables lie mostly within the 95% confidence bands estimated from the SVAR. Further, when we take into account of changes in monetary policy rules, the model does well in accounting for the SVAR evidence not only for the full sample period in the postwar U.S. economy, but also for the sub-sample period under the Volcker–Greenspan regime. To understand the driving mechanism in our baseline model, we consider three special cases of the model to show how each friction (sticky prices, sticky wages, and habit formation) matters for the results (Section 5). We first examine the SP channel emphasized by Gali (1999). In the SP model, price level inertia, along with weak monetary policy accommodation, implies small short-run adjustments in aggregate output following a positive technology shock, so that hours decline. The decline in hours lowers the marginal disutility of working, which, coupled with small adjustments in consumption, leads to a fall in the marginal rate of substitution between leisure and consumption and thus a fall in the real wage as well. Since price inflation also falls, nominal wage inflation declines by even more than does the real wage. These wage dynamics arise in the SP model with or without habit formation, and they are at odds with our SVAR evidence.5 The inability of the SP model to produce a weak response of nominal wage inflation to technology shocks suggests that nominal wage rigidities can be important in capturing the observed labor market dynamics, especially the wage dynamics. To isolate the role of nominal wage rigidities, we next examine a version of the baseline model with flexible prices and sticky nominal wages (SW). The SW model predicts that a positive technology shock leads to a weak (and more realistic) adjustment in nominal wage inflation. However, since prices are flexible, price inflation declines sharply and the real wage rises sharply on impact of the shock, implications that are at odds with our evidence. We also find some interesting interactions between nominal wage rigidities and habit formation. In the absence of habit formation, as the price level falls, aggregate output rises and, with monetary policy accommodation, the rise in output can exceed that in productivity, so that hours can rise following a positive technology shock. The presence of habit formation dampens the rise in output, so that hours can fall in the short-run. Quantitatively, with the help of habit formation, the SW model is able to generate responses of hours and nominal wage inflation that lie within the empirical confidence bands, although the model has difficulties explaining the dynamic responses of the real wage and price inflation. To further examine the role of habit formation, we consider two polar cases: one with no nominal rigidities, but with habit formation; and the other with sticky prices and sticky nominal wages, but without habit formation (SPW). The first case corresponds to a version of the RBC model studied by Francis and Ramey (2005) and, not surprisingly, we confirm their finding that hours decline modestly in the short-run. Yet, the RBC model augmented with habit formation fails to deliver plausible responses of wages and prices. The SPW model does well in capturing the observed patterns of the labor market dynamics but, quantitatively, it does not perform as well as does our baseline model. These results suggest that, in the absence of nominal rigidities, habit formation alone is not enough to generate realistic labor market dynamics; however, incorporating habit formation in a model with sticky prices and sticky wages improves the model's quantitative fit. We conclude that all three types of frictions, including sticky prices, sticky wages, and habit formation are important for capturing the observed labor market dynamics following technology shocks.

Competing business cycle theories have often been evaluated based on their implications on labor-market dynamics. Some recent empirical studies find that a technology improvement typically leads to a fall in per capita hours. Yet, other studies suggest that a technology improvement may also be consistent with a rise in hours. These findings have stimulated a lively debate, both on the empirical validity of such findings and on the theoretical implications. The lack of a consensus about how hours respond to technology shocks presents difficulties in evaluating competing business cycle theories based solely upon their implications on hours dynamics. We have taken up this issue from a broader perspective. We have examined the effects of technology shocks on a broader set of labor-market variables, including wages and prices, not just hours. We find that, although a positive technology shock can lead to a rise or a fall in hours depending on the details of empirical specifications, the shock consistently leads to a weak response of wage inflation, a modest decline in price inflation, and a modest rise in real wages on impact and a permanent rise in the long-run. These patterns of adjustments in wages and prices are remarkably consistent across our alternative empirical models. We argue on this ground that, for the purpose of evaluating competing business cycle theories, it is more informative to examine the models’ predicted dynamic responses of wages and prices following technology shocks than focusing solely on hours dynamics. Under this criterion, we have shown that a model without nominal rigidities, such as a standard RBC model (or its variants), does not do well. We have also shown that a pure SP model fare no better, as it predicts the wrong sign of the response of real wages and the wrong magnitude of the response of nominal wages. We further propose that, an alternative model with both price and nominal wage rigidities, coupled with habit formation in preferences, has a better chance of succeeding in explaining the observed labor-market dynamics following technology shocks. In a broad sense, the new generation of DSGE models with micro-foundations and nominal wage and price rigidities has been fairly successful in explaining the dynamic effects of monetary shocks (e.g., Huang et al., 2004 and Christiano et al., 2005). Our results suggest that this class of models, which marks a significant departure from the traditional RBC paradigm, can also be useful in explaining the dynamic effects of technology shocks.