Comment on Lof, Mekasha, and Tarp (2014)

Rejoinder to Herzer, Nowak-Lehmann, Dreher, Klasen, and Martinez-Zarzoso (2014)

“…These PVAR results suggested, unlike the results by Nowak-Lehmann et al (2012), a positive and significant impact of aid on income. Herzer, Nowak-Lehmann, Dreher, Klasen, and Martinez-Zarzoso (2014) have now commented on our paper. Regarding the first point (missing observations due to a logarithmic transformation), Herzer et al (2014) acknowledge the mistake but argue that this mistake did not distort the results as the results are qualitatively similar when no logs are taken or the missing observations are controlled for in another way.…”

“…Herzer, Nowak-Lehmann, Dreher, Klasen, and Martinez-Zarzoso (2014) have now commented on our paper. Regarding the first point (missing observations due to a logarithmic transformation), Herzer et al (2014) acknowledge the mistake but argue that this mistake did not distort the results as the results are qualitatively similar when no logs are taken or the missing observations are controlled for in another way. This is indeed true, but Herzer et al (2014) leave out some fundamental points.…”

“…Unfortunately, Herzer et al (2014) do not tackle the causality issue at all while this was one of the two main points in our article Lof et al (2014). Instead, they raise a number of concerns regarding our PVAR analysis.…”

“…Regarding the first point (missing observations due to a logarithmic transformation), Herzer et al (2014) acknowledge the mistake but argue that this mistake did not distort the results as the results are qualitatively similar when no logs are taken or the missing observations are controlled for in another way. This is indeed true, but Herzer et al (2014) leave out some fundamental points. Nowak-Lehmann et al (2012) arrived at the log-linear specification by considering a Solow-type framework with a Cobb-Douglas production function.…”

“…Most of these concerns, however, seem to be based on misunderstandings. First, Herzer et al (2014) claim repeatedly (first time in footnote 8) that our panel is unbalanced, because we only include observations for which full time-series are available. This is incorrect.…”