GST import threshold – a few principles

By Eric Crampton 07/08/2015


Since the government is looking set to lower the threshold for GST application at the border, or at least will be releasing a discussion paper on it next week, a few principles to keep in mind:

  • Were it possible to collect GST at the border seamlessly and at low cost, it should be applied at the border.
  • There are two classes of losses that obtain where GST is not collected at the border on lower valued imports. 
    • First, we have allocative inefficiency where slightly too many goods are imported relative to a first best. The 15% GST-based price difference may induce some extra imports at the margin, and this is inefficient.
    • Second, we have effects on the overall tax base. This is separate from the allocative efficiency issue. Suppose that each and every import currently undertaken is one that would also have happened if GST were applied at the border and that there is zero distortion. There remains erosion of the tax base in that world – the government is less able to rely on consumption taxes and consequently has to rely more heavily on other, potentially less efficient, taxes. The deadweight costs of having to rely on less efficient taxes will also here matter. 
      • That said, one thing that absolutely and emphatically does not matter is effects on the domestic company tax base or domestic income taxes. It potentially matters for thinking about fiscal incidence, but not for efficiency. Think of import leakages through online shopping here no differently than we think about import leakages when we buy cars that are built abroad. We do not, and should not, complain about lost company tax revenue where there isn’t a domestic car manufacturing industry in New Zealand. Resources that would otherwise have been used making cars instead are put to other and better uses. Similarly, resources that would be put to use in domestic retail can be put to other better uses when consumers can more efficiently engage in direct-to-consumer imports from abroad. To count effects on domestic company tax base is to adopt mercantalism. 
  • In assessing where the threshold should be, we need, in the first instance, to think about the effects on efficiency. And that will depend entirely on how you set up the collection mechanism. If you can collect GST at the border in a way that imposes no greater cost than collecting GST on domestic purchases, then there’s no allocative inefficiency tradeoff to be made and the optimal threshold is $0. As soon as the collection mechanism is costly, you start having tradeoffs. And recall that these costs are not just the costs on Customs or IRD for running the system: it’s also the hassles that some systems would impose on those wishing to buy goods from abroad. When I say costs, it’s inclusive of all of that. What’s the tradeoff?
    • If the import threshold is set at some arbitrarily high level, you will induce way too much importation of high value, low shipping cost goods relative to a first best. That’s an allocative inefficiency. 
    • If the import threshold is set at $0 and there are fixed costs in collecting GST at the border – say a fixed processing charge, or hassles for consumers, or hassles for foreign shippers, you will deter a lot of efficient imports. 
    • As you increase the threshold from $0, the allocative inefficiency from the fixed collection cost starts being trumped by the allocative inefficiency from too much higher valued import. At that point, you stop and set the threshold. What’s the right threshold? I do not know, but it is endogenous to your collection mechanism.
And so we come to the absolute importance of picking a non-stupid, non-wishful-thinking GST collection mechanism.
Here are some things I believe count as wishful thinking. 
  • That foreign suppliers other than the very large ones would have any interest in registering with IRD to pay tax on goods shipped to New Zealand.
  • That foreign suppliers other than the very large ones, and perhaps not even the very large ones, would have any interest in putting a tax stamp on boxes being shipped to New Zealand indicating to Customs that tax had been paid on those shipments. 
  • That you can run this through the credit card companies, when the credit card companies cannot tell whether a good purchased abroad on a NZ credit card was for consumption in NZ or for consumption abroad.
  • That New Zealand shoppers would not bear any cost worth mentioning if their goods were held up at customs for days while waiting for them to get a note from Customs advising that the goods were there waiting for tax payment. 
Now I understand that there is talk about some big international agreement where shippers would register with some central agency and that outfit would make it relatively easy for those suppliers to ship goods to any foreign country. To the extent that that is successful, that counts against my first point. But consider too how many American retailers are already unwilling to ship goods abroad because of the hassles – that’s why YouShop exists. That problem will not get better if we start thinking that foreign retailers are going to be happy to incur hassles in order to ship on to tiny tiny New Zealand.
Recall that there are a few classes of benefit from being able to access direct-to-consumer imports. Consumers doing the purchasing benefit from lower prices. Consumers continuing to purchase from domestic retailers very likely benefit from lower prices that come when domestic retailers have to compete with foreign websites. And, consumers get access to a broader range of niche goods that are not otherwise available in small markets. That last bit would be killed by proposals requiring foreign niche sellers to figure out how to get tax stamps to put on their packages. They’ll just stop shipping the couple hundred items a year they might each ship to New Zealand. 
The least bad way I’ve thus far seen for imposing GST on goods shipped into New Zealand is the one Bronwyn Howell proposed. The domestic leg of shipment in New Zealand has to collect the GST and can compete to find the best way of doing it. It still needs careful assessment and somebody has to talk with the shippers about feasibility, but I can see its working where other mechanisms wouldn’t. It doesn’t require wishful thinking about credit card companies’ willingness or ability to come to the table, or about foreign suppliers caring enough about the NZ tiny market relative to the hassles the NZ government thinks they’re willing to bear for the privilege of selling to us. 
But we still have to sort out better ways of handling the customs and biosecurity extra levies that would absolutely kill low value imports if they were levied at the same time as GST on low-value goods. And we have to figure out what the domestic shippers would charge customers for having to act as tax collectors. I do not know whether the Howell proposal passes cost-benefit, but it seems the one most likely to do so if any of them will. To the extent that domestic retailers lobby for mechanisms that are more costly to implement, you might wonder whether they are really trying to impose a trade barrier.