Tax Evasion: Responses to Tax Information Exchange Agreements

In late 2008 and early 2009, the G8 and the G20 pressured remaining non-cooperative jurisdictions with sanctions if they failed to commit to the OECD standards of transparency and exchange of information for tax purposes and sign 12 agreements that met these standards. Leading up to this event was the continued work of the OECD to construct a treaty network to combat tax evasion by means of tax information exchange agreements (TIEAs). In this thesis, I investigate if the pressure to implement the network of tax information exchange agreements reduces tax evasion activities in jurisdictions providing banking secrecy. To measure the effect of these tax information exchange agreements I use the value of service charges banks and other financial institutions in a country apply to their foreign clients that own financial assets there. These charges are for each country included in the category financial services export in the Balance of Payments (BOP) data set. And this is the data used to analyse potential effects. In other words, I use financial services export as a proxy for foreign financial assets. The statistical software package used in this thesis is Stata 14.1. There are two methods that I use to estimate an effect of the pressure to commit to the OECD standards and sign 12 agreements that meet the standards. The first approach is the fixed-effects (FE) model, and it is used to measure the change in financial services export as a percent of goods and services export for all countries that goes from being uncooperative to cooperative in exchanging tax information. The overall effect on the data set is then measured by this method. The second approach is called the synthetic control method and is used on the data to estimate a potential effect on an individual country's financial services export. I apply it to Switzerland as it historically has been the place to hide financial assets, with a banking secrecy law enacted in 1935, and an almost complete hegemony over the international wealth-management market until the beginning of the 1980s. By using the fixed-effect model and, in this context, the previously unused Balance of Payments data set, I find evidence that committing to the OECD standards and signing at least 12 tax information exchange agreements to the OECD standards has no significant effect on the cumulative amount of foreign financial assets. I theorise that this is a result of heterogeneous effects, caused by spillover effects of foreign financial assets from highly cooperative jurisdictions to less cooperative jurisdictions. By using the synthetic control method on an individual country known to provide banking secrecy, Switzerland, I find a large decrease in foreign financial assets from committing to the OECD standards and signing tax information exchange agreements to the OECD standards. This strengthens the theory of heterogeneous effects from the fixed-effects model, as a reduction in Switzerland necessitates a similar opposite effect in other countries that have committed to the OECD standards and signed at least 12 agreements that meet the standards. The concerns that the bilateral agreement network does not cover enough country pairs to effectively prevent tax evasion seems to be valid. The network for each country is far from complete, and a large part of the world is completely left out because of administrative costs of implementing and enforcing agreements. This continues to provide the opportunity for a tax evader to transfer financial assets, with little risk of detection, to countries not covered by an agreement with the home country of the tax evader