Friday, October 05, 2012

Follow-up on the financial transactions tax

Yesterday, as mentioned in the previous post, I participated in a panel discussion at the International Fiscal Association’s annual meeting (held this year in Boston), entitled: “Towards an EU Financial Transactions Tax?” The discussion was chaired by Malcolm Gammie of the UK, and the panelists, besides me, were (1) Manfred Bergmann of the European Commission, who is the principal author of the Commission’s financial transactions tax (FTT) proposal, (2) Francesco Guelfi, an Italian tax lawyer, (3) Axel Haelterman, a Belgian tax professor and tax lawyer, and (4) Urs Kapalle, a Swiss tax lawyer. Thanks are also due to Eric Roeder of the Max Planck Institute, who made sure the whole thing worked out.

A word on the state of the play for the European FTT: as an EU-wide proposal, it is dead as a doornail, having been conclusively rejected this past summer due to some member states’ unyielding opposition (the UK being a key player here). But a number of EU countries are adopting FTTs anyway. For example, Germany and France have already acted, and Spain may joint them soon. FTT proponents are hoping for coordinated adoption by a significant subset of EU countries. So the European FTT is alive, kicking, and potentially still relevant, even if the European Community as a whole is not going to act.

Wholly without regard to the merits, this is a tricky scenario to sustain. Everyone agrees that, given the mobility of financial transactions, a pretty wide scope is needed for them to work. On the other hand, insofar as countries follow the EC proposal’s structure of imposing the taxes on a residence basis, not just where the deal is executed, single-country FTTs (at least for relatively large countries) may actually succeed in generating non-trivial tax revenues, on an ongoing basis, at a relatively low administrative cost.

This, of course, doesn’t tell us that the proposals are a good idea – and not just because there will still be plenty of exit (the English may well be delighted to see FTTs on the books in Germany and France) but due to the broader efficiency costs of a cascading tax that is levied on a gross, per-transaction basis. Case in point: the European Commission estimated that its tax on derivatives would lower trading volume by 75%. They believe that this would mostly have reflected fewer deals, as opposed to tax avoidance via exit. This is not a result that one should like, to the extent that the parties to the eliminated derivative trades are consenting adults who are optimizing for themselves without hurting anyone else. So liking this result has to rest on believing that many of the eliminated transactions are bad ones, e.g., due to externalities (such as their increasing market volatility – except, it is hard to generalize, either theoretically or empirically, about how they affect volatility overall).

As noted in prior posts and in my article on the subject, I am not a huge FTT fan. The two points I see in its favor are as follows:

(1) I believe that there is too much trading, from a social standpoint, in that people invest too much effort in seeking trading gains at the expense of other traders.

(2) That issue aside, I view the FTT as an inefficiently designed tax on capital, borne mainly by investors in general. This hardly sounds like an endorsement so far, but suppose it is more politically feasible than better choices, that a given country has desperate revenue needs that would otherwise be politically hard to meet, and that at least the thing is fairly progressive (albeit more poorly designed than other progressive levies). Then I would hate to be a purist and say no, based on superior alternatives that are unavailable, unless the efficiency costs (relative to the revenue) are on the high end, rather than the low end, of the range that I think possible.

All this makes me in the end a bit agnostic about whether or not to adopt the FTT if the available alternatives are very highly constrained. Sorry to dodge a more definite answer here, but I am not running for office, and spouting a bottom line when one is unsure can turn into mere spitballing.

OK, what I probably would do if less constrained is (a) address income tax (and, outside the U.S., VAT) under-taxation of financial services, (b) adopt the “FAT-2” version of the financial activities tax, in response to apparently durable rents in the financial sector that make me very suspicious of what is going on there (see here for further details), and (c) maybe do something – through the income tax?? somewhere else?? – about the excessive trading issue.

One of the core problems with the FTT not only affects the merits of adopting it, but impedes figuring out how it should be designed. I thought I made this point more articulately at the IFA panel yesterday than I have managed previously, so here’s an effort to repeat what I said (at characteristically high words-per-minute volume) during the session.

Consider having an income tax, a consumption tax, or a carbon tax. In each case we know what we are trying to do and why, and this really helps to inform design choices. The income tax and the consumption tax are attempted (albeit imperfect) measures of material wellbeing, or ability, or ability to pay, or whatever you want to call it. It’s not that we want to discourage income or consumption, but we believe they have informational content about, or correlation with, something else. To a large extent though not 100%, we know what we mean by “income” or “consumption,” and this helps a lot – although not all the time – in making design choices for the system, albeit that we still need to consider a host of other issues (e.g., tax planning, compliance, and administrative costs, not to mention other types of information that are relevant to distribution).

Now take the carbon tax. We know what we don’t like: carbon’s effect on global warming. Happily (or not), all carbon atoms are relevantly the same. So we know what we are trying to tax and why, and this really helps in designing a carbon tax – leaving aside the issue of global cooperation, since the scope of the public good (moderating climate change) is global.

Now let’s turn to the FTT. What are we trying to tax and why? Even if on balance adopting the tax would have good consequences relative to not adopting it, it’s not clear what the “thing” we’re after is. And this makes it really hard, for example, to say how derivatives should be treated under the FTT.

Suppose that what we are after is excessive effort to seek trading gains. Even with just old-fashioned stock trading, it’s not entirely clear how good a proxy gross trading volume is for the thing we have in mind. For example, do we want to tax repos that involve financial instruments? These are transactions in which I “sell” you a financial asset for $X, but we agree that I will buy it back at a given time for $X + $Y, where Y is essentially an interest return. This is really just a way of doing a secured loan, which principally wouldn’t be taxed if it weren’t structured as a repo. (The repo, by contrast, is potentially taxed twice: when I “sell” to you, and when you “sell” back to me.)

Sure, we could put an economic substance rule regarding sales into our FTT, just as the U.S. income tax has (requiring that at least sufficiently clearcut repos be treated as loans, not sales), but no one wants to run the FTT that way – it is supposed to be administratively much simpler and more automatic.

OK, I assumed that we don’t want to tax the repo or the secured lending under the FTT, but suppose it is partly speculation by the lender regarding the asset value. Thus, suppose the lender really is to some extent betting on the value of the secured asset in making the secured loan. Thus, the return of $Y might actually be a speculative bet on the financial asset’s upside – say, with cash settlement rather than physical settlement at the back end, so that there isn’t an asset transfer (in the secured loan version) even then.

Admittedly, this example is a bit picayune. It probably wouldn’t stop us from adopting the FTT, if otherwise we liked it enough, all by itself. But when you start considering the taxation of derivatives, the fact that the underlying goal is a bit murky becomes a real impediment to figuring out how the FTT should operate.

Case 1 is a total return equity swap. Rather than my buying 150 shares of Apple stock for $100,000, and financing it with a loan, the bank and I agree that in one year’s time we will net the following two obligations: (a) I owe them the interest that a $100K loan would have earned them, plus the amount, if any, by which 150 shares of Apple stock declined in value during the year; (b) they owe me an amount equal to the dividends paid plus the share appreciation, if any, for 150 shares of Apple stock during the same one-year period.

Counter-party risk aside, this is economically equivalent to my buying the 150 shares for the $100K loan. So, if we don’t tax it under our FTT – or even, like the European Commission’s proposal, if we tax it at a much lower rate (relative to the $100K “notional principal amount”) than a stock sale – we are simply telling people to replace stock sales with total return equity swaps.

But what about all sorts of other derivative instruments (which can of course be mixed and matched with the above swap if you like)? E.g., an interest rate swap: in one year, I’ll pay you the amount that I would have owed you on a $100K loan at a fixed interest rate, and you’ll pay me the amount that you would have owed me on a $100K loan at a floating rate.

Had we done the actual transaction (the offsetting loans) rather than the swap, the FTT wouldn’t have hit us. So here the derivative is being taxed worse than the underlying, for no obvious reason. What’s more, it isn’t entirely clear that we have a good reason for wanting to tax this swap. Is this socially excessive speculation on the likely direction of interest rate movement? What if I am simply trying to hedge my other economic positions from my business, and the bank is hedging or swapping out everything as well?

Anyway, this discussion has gone on long enough for a blog post (or perhaps too long), but the main point I want to make here is that it’s hard to bridge the gap between the underlying reasons we might have for imposing an FTT and the questions of how to tax various derivative transactions, which (if they were taxed differently) would be run through the blender and the Obscurathon anyway.


Anonymous said...

Why shouldn't we tax the repo?

If it is a disguise for a secured loan, why shouldn't such failures of transparency be taxable? Let them do a secured loan instead.

Daniel Shaviro said...

These are just different formal legal means of doing the same thing. The choice often turns on details of how the security interest is most cheaply exercised, etc. If these are private arrangements, there isn't a transparency issues. On the other hand, it is true there could be things going on such as firms attempting to keep their debt off the financial statements.

Unknown said...

One of the core problems with the FTT not only affects the merits of adopting it, but impedes figuring out how it should be designed. I thought I made this point more articulately at the IFA panel yesterday than. Matawan income tax preparation services