The Tobin tax was a transaction tax first mooted in the 70’s, it was meant to be a tax that would reduce blatant speculation in currency markets.

The revived idea of a ‘financial transaction tax‘ is as flawed now as it was then, because it will likely hurt investment, reduce liquidity and institutions will pass on the cost to private individuals as was seen in the example of Sweden who implemented one for several years in the 80’s. .

However, there are issues in the current market which are distortionary and where a tax might aid the market and reduce volatility, I’m talking about algorithmic trading and high frequency trading. Much of the volume (on US exchanges estimated at c.60%) is not due to people making calls on the market, rather it is on computers that execute trades in short amounts of time, taking a minute profit each time – in many cases they do it just to earn the exchange commission (a fee the actual stock exchanges pays to active traders).

The financial arms race is at the point where people are putting machines in the actual stock exchanges just so that they can get an extra nano-second upper hand on people outside of the exchange, is this the ideal of price discovery or is it a rent seeking activity? I would go with the latter.

So the tax would instead be a Pigovian tax rather than a Tobin tax. A Pigovian tax is there to reduce undesirable outcomes (negative externalities in economist speak), if you decided that algorithmic trades are part of the issue then you could tax based on the time a share was held even if it sells at a loss. This would reduce liquidity provided by the algorithmic trades but is liquidity a universal good and would it exist without algorithmic trades?

This approach would be better than outright bans on short-selling, which do not aid in market transparency. It would also be a more easily understood (by the public as opposed to practitioners only) solution than the up-tick rule which is really only there to punish a short to some extent.

A tax based on the time an investment is held could go to zero after 24 hours, the point is that it would eradicate much of the black box activity which is dominated by investment banks at present. You could argue that this is a fair hedge, but was the idea of electronic trading to use the speed of light as the ‘investors edge’?

Ockhams razor applies on this one in my opinion, and leads to a ‘no’. It only exists and an investors edge because it can be utilized and lead to certain outcomes, not because it serves a market function of itself. Which is why a sliding scale based on the time a security is held may be a good idea (or maybe not?!).

A regular household investor would likely never face this tax, it would specifically be something that allows speculation but taxes the brand of it that is operating in an area of the market where only machines trade, there are not humans capable of doing what the algorithmic traders do, they can programme the function but a human can’t do it as fast.

This may reduce the efficiency of an organization, one could argue that having a machine do a trade faster is efficient and therefore reduces the bottom line, while I empathize with that opinion, I cannot sympathize with it because the frequency and the machine actioned trades are the problem, not the platform.

Algorithms are not the devil, nor can we blame them, but they are not there to create an efficient market, they are there to give a certain group of investment banks who spent a lot of money on them a rent collecting ability. Thus the need to tax that activity, if it must exist and they cannot prove the advantage to society general (because the disadvantages are clear) then pay up.