The LPC did not find a substitution effect in favour of the younger group, but rather, estimated that employment share of 16-17 year olds fell while that of 18-21 year olds has risen. This result may be somewhat inconclusive, as the New Deal for Young people may have had a role to play. .
Dickens et al. used regional variations in wages across the UK to analyse aggregate employment changes due to the NMW. They expected to find that regions most affected by the introduction of the NMW would exhibit the greatest changes in employment. Using country level data, they examined differences in employment in the six months after the introduction with six months one year earlier. There results show that regions that were the most effected have had a small increase in employment, however, none of the coefficients were significant. They also found that the NMW had a small negative impact on total hours worked; but again their results were statistically insignificant. Although their regressions were insignificant, they found no evidence that the NMW had adversely affected employment. .
Stewart compares the probability of remaining in employment from one period to the next between workers directly affected by the NMW and those not directly affected. Choosing appropriate comparison groups may be problematic due to substantial measurement error in earnings data. Nonetheless, Stewart uses longitudinal data from three contrasting data sets in conjunction with a difference in differences estimator to compare observations surrounding the introduction of the NMW with observations from periods entirely before the introduction. The difference in differences approach relies on there being a meaningful difference between the treatment group and the untreated comparison group. This distinction may be threatened by spillover effects or by measurement error in the wage variable. Fortunately, recent literature has accused the NMW of having non-existent spillover effects.
