Writing about the UK government’s consultation on Local Government Finance reform a couple of months ago, SERC affiliate Teemu Lyytikäinen talked about the fact that government had no intention of allowing councils to set their own business tax rates. He argued that the problem was partly political, but that the bigger issue is that it’s not clear what the effects of full localisation would be. There are several main fears. One is a race to the bottom – local authorities undercutting their competitors and undermining tax basis. The other fears relate to the opposite scenario, a `race to the top'. Would some councils set very high tax rates and waste the revenues on useless programmes and bureaucracy? Would these very high tax rates bring more pain to firms already struggling in the face of the recession? Some of my recent research, published in the Economic Journal this week, provides evidence that this fear is to some extent justified.
This issue has been the focus of an extensive theoretical literature and our study is not the first one to consider these issues empirically. Evidence from the 1960s and 1970s suggested that there was no effect of taxes on firm location decisions. Work focusing on the 1980s suggested a negative relationship and a number of subsequent papers have confirmed that finding.
The existing literature, however, has failed to resolve a central problem when assessing this impact. Specifically, there are many things about firms and local authorities that we do not observe, so any correlation between taxes and firm growth needs to be interpreted cautiously because some third factor (e.g. the remoteness of the location) might explain both. In addition, tax setting may be driven by firm choices, rather than vice-versa. That is taxes may be high because employment is low.
Our methodology solves these problems by using firm level data and comparing changes over time for firms located on either side of local authority boundaries. Comparing sites close to local authority borders eliminates major differences (because we assume closely located sites are similar). Comparing firms over time allows us to identify ‘good’ and ‘bad’ firms and so eliminate the problems due to sorting of firms. That is, these techniques help eliminate things about local authorities and firms that we do not observe that may be responsible for the correlation between taxes and firm growth. Finally, we use the electoral make up of the local authority to predict local taxes (some parties consistently set higher taxes) so we can adjust our results to allow for the possibility that taxes might be high because employment is low.
We use our methodology to study the impact of the UK business rates between 1984 and 1989. We find a negative (statistically significant) relationship between employment and taxes. Higher local authority taxes lower employment in existing firms. In contrast, we find that local taxation has no effect on the entry of new establishments, probably because landlords have to lower rents in high tax local authorities to continue to attract tenants. In contrast, because existing firms are less likely to move away, landlords don’t necessarily change their rents, hence the negative effect on employment in existing firms.
Our methodology is applicable more widely, but results for the UK suggest that the government may be right to worry that local taxation can negatively affect local employment. One important caveat, however – our results can’t tell us whether local authorities would actually set such high taxes. They only provide a warning about the negative employment effects of doing so.
(If you are not in an academic institution and would like to see the full paper, please get in touch with my colleague Max Nathan: m dot a dot nathan [at] lse dot ac dot uk.)