The effect of a Tobin tax on the exchange rate volatility

Sandra Askegaard & Flemming Søgaard

Studenteropgave: Kandidatafhandlinger


This thesis work analyses the impact of a Tobin tax on exchange rate volatility and evaluates the proposed sizes for such a tax. We furthermore suggest a possible solution to the inherent endogeneity problem between transaction costs and exchange rate volatility. Finally, we link our empirical results to the conclusions of a game theoretical model to comment on how the market composition corresponds to the contrasting theories underlying a Tobin tax. Using a multiple variable time-series model we estimate the effect of an increase in transaction costs, resembling the introduction of a Tobin tax, on the exchange rate volatility. We use a measure of cost based on the percentage bid-ask spread of the exchange rate analogous to a transaction tax, which furthermore does not require us to make any assumptions of the market efficiency. The measure of exchange rate volatility that we use captures the variation within a period by utilizing the minimum and maximum values of the exchange rate opposed to opening and closing values. As a possible solution to the endogeneity problem, we propose that the exchange rate can be used as an instrument for our transaction costs measure. We believe so because the exchange rate is unaffected by the volatility and because the exchange rate affects our measure of transaction costs. To test this, we set up a two-stage least squares model. From our multiple variable time-series model we conclude that the introduction of a Tobin tax into the currency market will have either a positive significant or negative insignificant effect on the exchange rate volatility. Furthermore, these results are emphasized by our analysis of our instrumental variable. Based on these results, we conclude that a Tobin tax is more likely to amplify than dampen volatility thereby contradicting the claim made by Tobin (1978) but agreeing with the view of Friedman (1953). Although comparing our empirical results to the game theoretical model suggest that the market is composed of a larger share of ‘noise traders’ than ‘informed traders’. This indicates that the view of Friedman can exist even if the majority of traders are not informed. Additionally, we use our empirical results to measure the impact of two proposed taxes of 0.5% and 0.005%. The first is motivated by the desire to change market behaviour and the second is motivated only by the desire to generate revenue to cover the costs of financial crises without affecting the market. We find that the first tax is up to seventeen times larger than the average transaction costs and even if it worked as intentioned we find a tax of this size to be unrealistically large. In contrast, the revenue generated by the second tax only covers a fraction of the price of the latest crises and increasing it will only influence the market behaviour and thereby most likely increase volatility.

UddannelserCand.merc.aef Applied Economics and Finance, (Kandidatuddannelse) Afsluttende afhandling
Antal sider152