WTaxO Lesson #3: Trade Without Stats

A third takeaway from this week’s World Trade Organization’s ruling against Argentina’s anti-tax haven policies is:

Proving protectionism doesn’t require statistics or demonstrated injury.

One would think that a body tasked with enforcing open trade would have a very clear and consistent methodology for examining trade impact. Not so.

When development economists talk about trade disputes and protectionism, they think of situations where:

1. there is an actual good or service
2. from the complaining country
3. being harmed in the challenged host country
4. in a significant
5. and demonstrable way
6. to the benefit of host country providers
7. that sell the same thing.

As it happens, this is also consonant with part of the broader justification for trade agreements: create legal penalties that can deter countries from injuring one another. Less conflict, more peace – everybody wins.

As I write in this book chapter, the WTO has developed an odd approach to evaluating trade distortions. Instead of a more traditional economic analysis of protectionism, the panel report released this week used a “conditions of competition” (COC) framework. This was developed by trade lawyers over many WTO cases.

Here, there needn’t be suppliers or goods present in the host country – the logic being that the protectionism might be so bad that they the foreign products don’t get into the local market in the first place. That has a certain appeal. (Although mostly in the the case where a country is blocking market access (which was not the case here).

Where I get lost is the rest of the formulation. Under COC, no amount of discrimination is too little to be a violation, and it needn’t be proven that the discrimination generates a tangible benefit reaped by any identified domestic interest. What products or services get compared in the likeness tests is a freewheeling enterprise with substantial discretion for panelists.

The recent WTO decision revealed some of the quirks of COC.

Minimal evidence

Panama made allegations that eight different anti-tax haven measures were discriminatory, but the panel spent almost no time identifying what types of Panamanian entities (and where located, what modes of supply*, and in what financial sectors) would actually be hurt. While finding that Argentina’s policies discriminated, the panel presented no evidence of cancelled contracts, diminished business, or named any firm actually affected.

Measure imposed on nationals, not foreigners

Another strange factor was that several of the anti-tax haven measures were not imposed on foreign firms, but on Argentine consumers. For example, one measure increased the default amount of tax an Argentine would have to withhold when making payments to tax haven country entities (one presumes banks, but the transaction details are not discussed in any depth).

This is something that hurts Argentines in Argentina. From here, the panel posits an indirect impact on foreign service providers because presumably Argentines would choose firms in countries without this requirement. But no consumers were polled, no costs tallied up. It’s mere conjecture about incentive structures.

Distorting incentives only marginal

Trade with tax havens was not prohibited. Argentina’s policies created presumptions that the individual taxpayer could rebut by providing paperwork. This seems like a fairly non-trade restrictive way to enforce a tax haven measure. But the WTO panel found that even having to keep additional paperwork altered the “conditions of competition” (para 7.300).

Again, the economist in me would really like to see a number attached to this, so we could understand how important a distortion this is and whether it is really worth the WTO spending time and energy on.

Positive discrimination

Argentina lost the case in part because it – by November 2013 – was extending treatment to Panamanians that was better than that extended to operators from other tax haven countries, and perhaps even than Argentine operators.

This is historic: a country brings a WTO case and wins it because it was benefiting too much from a government’s policy relative to its competitors. The panelists determined that there is no necessary reason why Panama couldn’t challenge policies on behalf of the other tax havens (para. 7.93). (Still, no naming or evidence of harm on other foreign businesses from third countries was provided.)

I’ll say it a third time: Panama won a case about discrimination of which it was the beneficiary!

One could argue that Panama was being magnanimous. My read is a bit different: Panama throughout has treated this as a test case. It wanted to send a signal to countries that there would be a litigation price to following Argentina’s lead. This might make sense for Panama’s tax haven industrial policies, but not so clear that it’s an efficient or optimal use of WTO resources.

Existence of discrimination examines effectiveness of regulatory scheme

The panel showed a concern throughout their decision in emphasizing the deep interconnections between services and regulation. This reads partially as an appreciative nod to the role of government.

However, this linking popped up in some weird places. Here’s an excerpt from the decision:

the preamble to the GATS …refers to a balance between the objective of expanding trade in conditions of transparency and progressive liberalization, on the one hand, and, on the other, the right of Members to regulate the supply of services in their territories and to establish new regulations in this regard in order to meet national policy objectives…

we shall first determine whether Argentina accords different treatment to these two categories of services and service suppliers and then go on to examine whether this treatment is less favourable for like services and service suppliers of non-cooperative countries. In this connection, we consider that, in order to determine whether treatment is less favourable, it must be assessed whether the measure modifies the conditions of competition. We also consider that, in this particular case, such an assessment has to take into account regulatory aspects relating to services and service suppliers that may affect the conditions of competition; in particular, whether Argentina is able to have access to tax information on foreign suppliers. (para. 7.235) italics added

The panel went on to find that the conditions of competition were modified – because of the Panama-benefiting loophole that undercut the effectiveness of the regulation (i.e. that Argentina didn’t sanction tax havens in the process of signing information exchange agreements).

This is selective empiricism. As I noted above, a complainant is under no responsibility to show empirical evidence of injury. However, the panel gets very empirical when assessing the effectiveness of the regulatory regime.

I would be surprised if this aspect of the case – although wonky – doesn’t get appealed. My read of how regulatory effectiveness got stitched together with discrimination is from the very particular context of a string of cases against US consumer protection measures. In those cases (Tuna-Dolphin II, COOL, and Clove Cigarettes), the challenged measures were deemed to be technical standards falling under the WTO’s TBT agreement.

That pact – one of 17 at the WTO – does not have a general exception or defense from complying. I think WTO adjudicators saw that as a glaring omission in the treaty drafting, and so effectively cobbled one in. They found that technical standards that otherwise violate TBT rules are allowed so long as the “detrimental impacts on competitive opportunities stem[] exclusively from legitimate regulatory distinctions”. I’ve found this an impossible test to meet, so I gave it an impossibly pronounceable acronym: DIOCOSEFLRD.

The GATS, in contrast, does have a general exception. In my humble opinion, it is was unnecessary for the panel to consider – at the stage of establishing a violation of MFN – whether the anti tax haven regulations worked or not. Indeed, that’s something better to evaluate in the context of an explicit exception to the agreement.

In short, it’s a quirky case of the panel having a noble intention (recognizing importance of regulation to services trade and wanting to give a margin of deference), but this actually becoming redundant at best, counterproductive at worst.

Revenge of the Quantitative

The panel described COC as a “qualitative” approach to discrimination. Presumably, discrimination would be harder to prove under a quantitative approach – as it would have to be identifiable, substantial and systematic.

But the weird thing is that the panel does switch over to a quantitative approach in its assessment of whether Argentina could invoke defense provisions. As I wrote here, Argentina would have had to show that the impact on trade was quantitatively minimal to qualify for a number of the GATS defenses.

It would seem preferable to not import a trade restrictiveness test into the defense, but instead have a clear such test in establishing the initial violation. Or be qualitative in both, or quantitative in both. That ship has long since sailed in the jurisprudence, but it would be nice to see a change in the future.

As it turns out, the pane’s flip to the quantitative was not really that meaningful. The panel simply looked at the various rebuttable presumptions that Argentines wanting to do business in tax havens might have to face, and quickly noted that none of these was very economically important (paras. 7.719-7.724)

This would have been great information to present the reader on page 1!

Reading this WTO decision cost me a few hours of my life that I will never get back, but it cost the litigants, panelists and WTO system far more. Having a more consistent approach based on the most economically meaningful disputes would be a welcome improvement.

* The panel rushed through what services trade experts spend years learning: the different “modes of supply”. This refers to Mode 1 (ex: Consumer in Argentina buys services located in Panama), Mode 2 (Consumer in Argentina goes to Panama to consume goods), Mode 3 (Panamanian makes FDI in Argentina), and Mode 4 (Panamanian travels to Argentina to supply service).