In a profession that is controlled almost exclusively by people who completely overlooked the largest asset bubbles in the history of the world, Ken Rogoff earns at least a “B” for his early warnings of dangerous economic imbalances. However, his column criticizing financial speculation taxes (FST) is more on a par with the work of his hopelessly lost colleagues.
The column argues that an FST of the size being considered by the European Union (EU) would reduce the information content of prices, reduce liquidity, and have no appreciable impact on volatility. In the long-run they will raise the cost of capital and therefore slow growth, and not end up raising much revenue.
This is a serious list of charges against a tax of 0.1 percent on a stock trade (0.05 percent on each side) and 0.01 percent on derivative trades (0.005 percent on each side). The most obvious reason for skepticism about Rogoff’s attack is that the increase in transactions costs implied by the tax would just raise them back to the levels of early or even mid-90s.
Computerization and deregulation has led to a sharp decline in transactions costs over the last three decades. A tax of 0.1 percent on stock trades would just remove part of this decline. Trading costs would still be lower than they were in the 80s and much lower than they were in 50s or 60s when they were typically 1 percent of the share price or more.
Insofar as we think that lower transactions costs and higher trading volume are important for increasing the information content of prices, we must believe that the prices in financial markets in 2011 (and even in 2007) had far more information content than the markets of the high transactions costs 1950s and 1960s. That may be true, but it would be hard to find evidence for this view in growth data
Of course lower trading volumes mean that markets will be less liquid, but did we have serious problems with inadequate liquidity 15-20 years ago? While there are less liquid assets, where a reduction in liquidity would be a problem, a nice feature of the tax is that the impact on the liquidity of these assets would be very small.
The reason is that less liquid assets already have much higher trading costs. For the most heavily traded stocks, the tax might amount to a 50 percent increase in trading costs for large traders. However, with less liquid stocks, the transaction cost may already be in the 1 percent range. This means that the increase in costs implied by the tax would be around 10 percent, which could be expected to have limited impact on trading volume.
The impact on volatility is admittedly questionable. By making arbitrage more costly we may see somewhat larger daily gyrations. However, this is not likely to provide much cause concern among investors. The better question is whether an increase in trading costs may make episodes like the flash crash or the 90s tech bubble less likely. I don’t believe that we can answer this one definitively in either direction at this point.
This brings us to the cost of capital story. Do we really believe that a 0.1 percent increase in transactions costs will have a measurable impact on the cost of capital? If so, are there should be large gains to growth as a result of the sharp decline in transactions costs over the last three decades. If there are large economic gains from declines in transaction costs, most economists have forgotten to include this fact in their modeling of growth. Again, the data don’t provide much evidence for large gains in any obvious way – growth was much better almost everywhere in the high transactions costs 60s than the low transactions cost 00s, even before the crash. (There was also no evidence of any growth speedup as transactions costs fell.)
The other problem with the Rogoff story is that we actually do have examples of FSTs in action, most obviously in the UK which has been taxing stock trades since 1696. In the UK the tax rate is 0.5 percent on each trade, five times as high as the rate being considered by the EU. If the proposed EU tax would in fact have a terrible effect on the cost of capital and lead to a growth slowdown, then the UK must be suffering enormously because of its tax. This doesn’t in any obvious way seem to be the case. In fact it is striking that economists never even include this tax in regression models when they do cross country growth comparisons.
And, we also know that in spite of the dampening impact on growth, the UK still manages to collect between 0.2-0.3 percent of GDP ($30-40 billion in the U.S.) annually with this tax on stock trades. And this is not a short-term adjustment story. Presumably after three centuries it is reasonable to assume that markets have already adjusted to the imposition of the tax.
In addition, the revenue comes from the best possible source: the elimination of waste in the financial sector. We know that trading is relatively elastic. This means that if a FST doubles the cost of trading then the volume of trading will fall by roughly 50 percent. In this story, investors will pay twice as much per trade but only trade half as frequently, leading their total trading costs unchanged.
The loser is the financial industry, which will have considerably less revenue as a result of the tax. This should cause many in the industry to leave for greener pastures elsewhere. In other words, they will get jobs in productive segments of the economy. That will foster growth.
In short, it is hard to see how Rogoff has much of a case here. We know that FSTs can collect large amounts of revenue and are consistent with robust financial markets. What exactly is the problem?