The Impact of a Financial Speculation Tax on GDP: Problems with the European Commission Model

December 14, 2011

While interest in a financial speculation tax, also known as a financial transaction tax (FTT), is picking up in the United States, the European Union is getting close to actually approving a tax. As part of this process, the European Commission (EC) had its staff put together a model that would project its impact on investment and growth.

This model, which the staff admits is very much a work in progress, projected that in the long-run the tax would lead to a 1.76 percent decline in output. Opponents of the tax were quick to seize on this projection as a basis for opposing the tax.

As I noted earlier, there are good reasons for questioning this projection from the model. It implies that the cost of financial transactions (a tax would have no different effect than brokerage fees and other trading costs) have an extraordinarily large impact on growth and productivity growth. This is completely at odds with nearly all of the economic literature on growth, which does not mention financial transactions costs at all.

Extrapolating from the model, the decline in the average cost of financial transactions in the United States over the last three decades would explain close to 15 percent of the productivity growth in over this period. If this is true, then the U.S. should anticipate slower productivity growth in the years ahead, since there is little room for transactions costs to decline further, now that they getting close to zero. (The Congressional Budget Office and the Office of Management and Budget have not incorporated any such slowdown into their growth projections.)

Another implication of the EC model’s projection is that the U.K. could quickly see a jump in its GDP of close to 9 percent if it got rid of its 0.5 percent tax on stock trades. It is unlikely that anyone really believes this. Of course, if the EC model’s projections are accurate then the UK should have seen its GDP increase by close to 9 percentage points above its baseline in the years following 1986. That was the year it lowered its tax from 1.0 percent to 0.5 percent. (There was no increase in growth that was close to this size.)

A closer examination of the EC model shows why its projections are so unrealistic. Essentially, the model includes (and exaggerates) all the ways in which a tax can harm the economy, while excluding the ways in which it could benefit the economy.

First the model’s assumptions imply that a tax will increase the cost of capital to firms through its effect on increasing the cost of buying and selling shares of stock. The model’s assumptions exaggerate the size of this effect by failing to note that the tax would reduce the volume of trading.

In fact, most research suggests that the trading volume will be reduced roughly in proportion to the increase in the cost per transaction. This means that if the tax has the effect of doubling transaction costs then trading volume will be reduced by close to 50 percent. This would mean that total transactions costs would be little changed for most investors, which would imply that there would be very little effect on the cost of raising capital to firms. Also, at least some of the tax would be borne by banks in the form of lower fees. However, by assumption, the EC model excluded the feedback effect that the tax would have on reducing transactions costs by reducing the volume of transactions and lowering fees.

The second major issue with the EC model is that it excluded the possibility that the tax could reduce market volatility by reducing the number of noise traders in the market. Noise traders are investors who buy and sell based on market gossip or the short-term momentum in the market. They do not trade based on the fundamentals of the market.

There is literature (e.g. this and this) that argues that noise traders can lead to greater volatility in the market, since they may exaggerate movements away from trend values. If this is the case, and a tax has the effect of driving noise traders out of the market, then an FTT may have the effect of reducing volatility in financial markets. If this is the case, it would reduce the cost of capital to firms, since it would make buying stock and other financial assets less risky.

The research on this point is ambiguous; however, there is at least anecdotal evidence (e.g. the flash crash and the 1987 crash that was caused by programmed trading) that suggests that high-volume, low-cost trading can lead to large divergences from fundamental values. The EC model rules out by assumption the possibility that an FTT will be a stabilizing force by reducing the share of noise traders in the market.

A third way in which an FTT could benefit the economy that is excluded from the EC model is that it could help to drive out super-informed traders who profit by getting news a short time ahead of ordinary investors. The logic of this issue can be seen by considering the case of insider trading.

In the case of insider trading, a portion of the returns that would otherwise go to ordinary investors are instead siphoned off by investors trading on inside knowledge. These inside trades reduce returns to ordinary investors, even though they are in fact giving information to markets.

The same situation would arise with highly informed investors, who do not engage in insider trading but simply move quickly based on publicly available information to take advantage of small gaps in prices before other traders. For example, a trader who has super-fast access to weather reports may be able to beat the market in movements of crop futures. This would mean that they would capture some of the gain that otherwise would have gone to farmers or consumers of agricultural products. (The assumption here is that the greater speed of adjustment is too small to affect production or consumption decisions.)

If an FTT makes this sort of trading less profitable, and thereby drives some highly informed traders out of the market, it will be increasing the returns to other actors. This would mean that ordinary investors could expect the same return even if businesses had a somewhat lower cost of capital, since money would not be getting siphoned away by these super-informed investors.  

Finally, the EC model does not in any way capture the drain of resources of the financial sector on the productive economy. There is no feedback that allows the reduced volume of trading in the financial sector to lead to increased output elsewhere in the economy. In the EC model, if Steve Jobs had spent his career running a hedge fund that arbitraged soy bean prices, there would be no negative impact on the rest of the economy.

In the real world, the financial sector does drain capital and labor (often highly skilled labor) from productive uses elsewhere in the economy. An accurate model would include some accounting for this drain and pick up the benefits to the real economy from reducing the drain. The EC model is constructed in a way that does not allow for an FTT to benefit the economy in this manner.

As noted earlier, the EC model is quite explicitly a work in progress. There has not been a full macro model designed to project the effect of an FTT, so the EC staff are breaking new ground. Hopefully they will improve this model in the months ahead and find ways to incorporate the potentially beneficial effects of an FTT. In the meantime, we are better off basing arguments on evidence in the world rather than the projections of an incomplete model.

Support Cepr


If you value CEPR's work, support us by making a financial contribution.

Si valora el trabajo de CEPR, apóyenos haciendo una contribución financiera.

Donate Apóyanos

Keep up with our latest news