On the measurement of tasks: does expert data get it right?

Journal for Labour Market Research

Table 13 Task measures as wage predictors: narrow task definitions

Dependent variable: log hourly real wage	(1)	(2)
NRA (Ind.)	0.40***	0.42***
NRA (Ind.)	(0.04)	(0.04)
NRI (Ind.)	0.28***	0.31***
NRI (Ind.)	(0.04)	(0.04)
RC (Ind.)	0.28***	0.29***
RC (Ind.)	(0.04)	(0.04)
RM (Ind.)	− 0.03	0.01
RM (Ind.)	(0.04)	(0.04)
NRA (Exp.)	0.31***
NRA (Exp.)	(0.03)
NRI (Exp.)	0.12***
NRI (Exp.)	(0.04)
RC (Exp.)	0.08**
RC (Exp.)	(0.03)
RM (Exp.)	− 0.02
RM (Exp.)	(0.04)
Controls	\(\checkmark\)	\(\checkmark\)
Occupation dummies		\(\checkmark\)
Adj. \(R^{2}\)	0.20	0.22
Observations	27,777	27,777

Robust standard errors in parentheses
* \(p<0.10\), ** \(p<0.05\), *** \(p<0.01\)
This output is based on Eq. (10), but with a narrow definition of tasks instead. “NRA” represents Non-Routine Analytic Tasks, “NRI” reflects Non-Routine Interactive, “RC” is short for Routine Cognitive. Lastly, “RM” represents Routine Manual tasks. For brevity, this robustness exercise is restricted to a model comprising individual survey as well as expert data (column 1) and a model based on individual-level tasks and occupational FE (column 2). I use BIBB/BAuA data that has been collected in 2011–12 and 2017–18, and data from BERUFENET, covering the years 2011–13. For the BIBB/BAuA data, see Hall et al. (2020b) and (Hall et al. (2020a), respectively. For the BERUFENET data, see Dengler et al. (2014)