Comments and Questions

An interesting exercise in collecting data and conducting a preliminary investigation, but I was left looking for more statistical analysis. In particular, there is no consideration of tax base factors that influence potential tax/GDP ratios. Most of the variables considered as potential explanations (in essence of deviations from an average, often a regional average) seem to relate to choice of tax rates (in an aggregate or even specific sense). One could argue that certain groups of countries have similar economic structures and income levels hence similar tax bases (e.g. East Asia, Europe, SSA). In this context, political economy factors (proxied by governance or democracy variables) may explain rates and hence performance relative to the group average. The weakness of the paper is that there is no control for the tax base. For example, as SSA countries have a low tax base one may expect relatively low tax ratios even if rates are as as high as feasible (thus authors like Keen argue that SSA countries may be collecting as much tax as feasible).Some of the effects or relationships posited are more ambiguous than implied. As a specific example, SSA countries may recieve aid because they are low income with low tax ratios, raher than having low tax performance because they receive aid. As another example, governance may not be associated with preferences for collecting taxes. High governance countries can have high or low tax ratios, and improving governance could increase collection efficiency (hence higher ratios) without altering preferences (or rates).A minor comment (p9 and table 3): as noted, SACU accounts for 50% of Lesotho's tax revenue. Lesotho has no influence over this as the SACU revenue sharing formula is based on shares of intra-regional trade plus a redistribution to poorer members (and South Africa chooses the external tariff rates that determine the size of the pool to share). Countries like Lesotho that source most imports from South Africa and are relatively poor therefore derive a disproportionately large share of recvenue (Swaziland is also an above average performer; even the relatively richer Botswana and Namibia, who benefit less from the sharing formula, are above the trend line). Given that South Africa makes the net transfers in SACU, perhaps the most surprising finding is that it also is above the trend line.