Skip to content

On sovereignty and international agreements

“Save our sovereignty” is a rousing catch-cry, but it’s a useless guide to international agreements.

National sovereignty’s been a recurring theme in opposition to the Trans-Pacific Partnership (TPP). But the assertion rarely gets the examination it deserves, particularly for an argument that works against virtually every international agreement in existence.

It’s easy to see what makes the slogan attractive. There’s a strong, natural and relatively universal impulse to defend our right as a country to decide things for ourselves. This isn’t a bad impulse to have—indeed, I hold it strongly myself.

The problem is that any acceptance of an international agreement necessitates a loss of sovereignty. This includes every trade agreement that New Zealand is currently party to, every treaty and human rights convention. It would have included the International Covenant on Civil and Political Rights, as well as the Kyoto Protocol and its successor. Yet most TPP opponents, who happen to align politically left, would surely not believe that any country should have refused to sign those treaties. So what gives?

The difference between the “sovereignty” we gave up in signing Kyoto and what we will give up in signing the TPP is actually very obvious. The left likes what Kyoto binds us to do, but not what the TPP would. But in that case, national sovereignty is not a useful discriminant. The real question is whether the agreement is a worthwhile exercise in the first place.

Of course, that debate is far too complex to address in political rhetoric. As many have noticed, it depends on what’s in the agreement. What opponents probably mean when they talk about “sovereignty” is the tendency for TPP clauses to step on areas traditionally outside the domain of trade liberalisation. Indeed, almost everyone in New Zealand, including TPP proponents, opposes (for example) the intellectual property provisions sought by the United States. The question is whether such hurdles in negotiation should be cause to abandon the agreement altogether.

There’s probably a deeper question: many TPP opponents oppose free trade generally, but for other, better reasons of economic ideology. Still, it’s issues like intellectual property (which affects Pharmac, New Zealand’s drug-buying agency) that are the sticking points. Indications from Wellington and leaked documents are that New Zealand negotiators oppose these provisions. Free trade supporters, then, see it as just a challenge to work through; opponents see no hope.

But what’s sovereignty, anyway?
Would signing the TPP mean giving up our sovereignty? On closer reflection, it depends what we mean. Every time I sign a contract, say, accept a job offer, I’m giving up some self-determination: I accept an obligation to do something for someone else, in return for their commitment to doing something for me. In that sense, countries always lose some right to decide for themselves when they commit to multilateral agreements.

Yet by the same token, one might see it as a sovereign right to choose to enter agreements with other countries. Countries that signed Kyoto did so (presumably) because they figured that, if everyone was bound to cut emissions, progress would be better. If we thought it was a good idea, it was in our prerogative to sign up.

At this point, I’m playing with semantics. What does it mean to be a sovereign country? The problem with the former definition is that it literally prohibits all international agreements. While there probably exist some who would agree with that, I doubt that it describes most TPP opponents. The latter definition, though, renders the concept of sovereignty somewhat useless. By that conception, what agreement would violate sovereignty? Does it require the agreement to be somehow revocable?

Use better arguments
You could dig into this as deep as you like. But if you question an opponent hard enough, no-one actually cares about sovereignty, per se. They either care about specific provisions that it would be stupid to commit to (and make no mistake, in its current form there are quite a few), or they generally disagree with market liberalisation. Neither of those stances are unreasonable, and there’s nothing wrong with some healthy scrutiny. It would just be nice if we stuck to arguments that matter.

Picture credit: NZ Herald

On inequality and education

If inequality is the “real cause” for poor education, does that mean poor education doesn’t drive inequality?

It is the opposition’s job, I guess, to oppose. Sometimes, though, the efforts baffle me. Such an instance occurred after Prime Minister John Key made a major education policy announcement on Thursday. The policy, which would create new positions in which top teachers and principals share best practice with other teachers and schools, was met with uncharacteristic approval from the education sector, including teachers’ unions. But it’s completely misguided, say the Greens:

This poorly thought out policy assumes that a possible improvement in teaching practice will address the driver of declining standards, inequality. […] The policy is not a blueprint to address the real needs of kids in lower decile schools to help them learn.

It’s worth taking a moment to consider the corollaries of the Greens’ position. If they believe that the only way to address poor educational performance is to address social inequality, it must follow that any education policy aimed at the education system, as opposed to the economy or welfare system, is a pointless exercise.

Which is a perfectly acceptable stance to take. But it’s hard to keep consistent. For starters, it requires them not to propose any education policy aimed at lifting achievement, for the precise reason that they don’t believe they work: the “real” cause is inequality, not the education system.

But moreover, it requires them to believe that you can’t address inequality through education. Their claim is that only by addressing inequality directly can you hope to improve educational outcomes. So they must not believe even in the potential for education to give poor children the chance to become richer adults. They must not believe that giving schools in poorer areas more funding (i.e., the decile system) will do anything to help those students. Their rhetoric is based on kids being “sick and hungry”, and “living in poverty”—not what teachers and schools must do to help break the cycle.

It might feel like I’m taking their views to the extreme—except that I’m not. Their comments are unambiguous: National’s education policy is bad because it fails to address inequality. “The best teachers and principals in the world can’t feed or heal the hungry and sick kids that show up to school each day.” My general bias is to try to read statements in the best light possible. I’m finding better interpretations hard to find. If they thought there exists a better education policy that is not welfare or economic policy, why didn’t they hint at that?

To be fair, poverty is indeed an excellent predictor for educational underachievement. And the Greens are justified in campaigning to reduce it, just because it’s generally bad. What is confusing is that they would criticise a policy aimed at doing exactly that, for trying to do so by improving education.

There is plenty that could go wrong with this policy. As with any novelty, there are bound to be teething issues; any good opposition will make the most of them. It could yet prove impractical to ferry teachers between schools on a daily basis. Labour at least reiterated its belief in “encouraging quality teaching” and “collaboration with teachers”. The Greens seemed determined to say nothing even neutral, and ended up advocating a principle that will soon prove to be a straitjacket. Perhaps they are excited about becoming the lead opposition party, but they should be more careful to leave themselves room to make policies of their own.

On the Snapchat user info leak/hack

Here’s an attempt to explain what the leak was, what it wasn’t and what it means, without getting carried away.

Following the recent news about Snapchat users’ account info being leaked, I figure it’s worth sifting through the media drama for what appears actually to have happened. It’s not as bad as the headlines make it out to be, though depending on how protective you are of your mobile number and how you chose your username, you might reasonably be concerned.

In case this post propagates further than I intend, readers should be aware that I don’t claim specific expertise in this topic, I have not examined Snapchat myself (except to the extent detailed in an earlier post), and that by training I’m an electrical engineer, not a computer scientist (though I’ve done some work with software). That’s my disclaimer. Obviously I have a reasonable degree of confidence in what I’m about to write, otherwise I wouldn’t be writing it, but I caution against taking any of this as authoritative, and I’ll make technical corrections if alerted to them.

1. Accounts weren’t compromised. The “hack” was just what anyone can do if they have copious amounts of time.

The leak was the association of mobile numbers with usernames. It uses the same mechanism that you can use through the app to add the people in your mobile’s contacts list. When Snapchat “finds” your friends, it just sends your contacts’ numbers to the Snapchat servers, and the servers send back the usernames associated with them. So if you wanted to find out lots of username–number associations, say to build a database of them, you could (as Snapchat admits) just add every phone number in existence to your contacts list, then “find your friends”.

Of course, that would take forever. But if you can figure out how to pretend to be the Snapchat app, then you can write a script to do basically the same thing, much faster. You don’t need to hack anyone’s account to do this (you can just use your own); you just need to figure out how to trick the servers into thinking that you are the app. That’s what Gibson Security figured out how to do: they reverse-engineered Snapchat’s API and figured out the keys that the app uses to “prove” to the server that it is the app. (It should be noted that GibsonSec says they did not mine the data. Someone else did.)

The flaw is not so much that this is possible (finding your friends from their numbers is a legitimate function) as it is that doing it en masse is possible. Snapchat (according to GibsonSec) hasn’t done rate-limiting on its servers for this function, so you can send millions of requests in quick succession. They shouldn’t really allow this—a normal user should never need to ask for so many.

But it’s also not the end of the world. Specifically, to my knowledge (and I only know what I have read on the internet and seen on TV), the leak doesn’t involve your real name, passwords, snaps (expired or otherwise), or the ability to use your account without authorisation. At least not yet. But GibsonSec’s disclosure (as far as I can tell) doesn’t give any mechanisms for doing so.

So how bad is it? If you’re private about your mobile number, and one can take a reasonable guess at your username from your identity (not guess your identity from your username—guessing your username is probably harder than it sounds, depending on how you picked it), then you will probably be worried that someone can infer your mobile number. If your username isn’t very guessable, isn’t used anywhere else, or if you don’t care about people knowing your mobile number, it’s probably not such a big deal.

2. If you’re outside America and Canada, you’re probably safe, at least for now.

This is just what they said, here and here. Presumably the leakers were too lazy to check numbers other than North American ones. For example, users with a New Zealand mobile number aren’t affected. But if you want to check if your number was affected by the leak, do so here.

Of course, until Snapchat applies a fix, someone else could come along and run the exercise for other numbers. If you’re paranoid about it, just remove your mobile number from your account in the app (this might involve uninstalling and reinstalling the app). Snapchat doesn’t require your number for the app to work; it just means your friends can’t find you unless you tell them your username or add them first, and when they do it doesn’t associate a display name with you (i.e., lots of inconvenience for your friends and none for you, unless your friends do the same thing). You shouldn’t, in my opinion, feel like you need to delete your account.


When I wrote about Decipher Forensics in May, I criticised them for over-hyping the impressiveness of their claims. No such criticism applies here: Gibson Security’s reverse engineering effort is admirable and their description is reasonably clear. They expressly pinpoint the flaw to a lack of rate-limiting in their disclosure and don’t (as far as I can tell, unlike Decipher) make any unwarranted jumps to conclusions. Also, Snapchat obviously didn’t intend for the secret keys to be found. The keys aren’t user-specific; they’re universal, solely to prevent third parties from pretending to be the app. I presume they’re probably hardcoded into the app somewhere, though I wouldn’t know for sure. Snapchat was perhaps naïve to assume they wouldn’t be found eventually.

But it would help for readers to understand exactly what the “hack” is. Fundamentally, it is the same as adding lots and lots of mobile numbers to your phone and finding all those “friends” on Snapchat—the hackers just found a more efficient way to do it. You should judge the severity of the breach on that basis.

“Resounding”, “emphatic”, “decisive”, “overwhelming”

Bluntly, the result of the asset sales referendum could have been better for its proponents.

These are the adjectives that have been used to describe the result of the asset sales referendum: “resounding”, “emphatic”, “decisive” (twice), “overwhelming” (twice), “great”. Sixty-seven percent is of course a reasonable margin. But in the context of citizens’ initiated referenda (CIRs), it’s not that impressive.

Stack it against CIRs of the past: New Zealand has had four since they were introduced in 1993. In each of those, the proponents claimed victories of 87.8%, 81.5%, 91.8% and 87.9%, respectively (ref). The asset sales referendum missed the lowest victory yet by more than 14 points. That’s more than the entire range of results before it.

The CIR votes all look like decent margins—why? It’s not surprising. It takes a mammoth effort to start one—petitioners must collect the signatures of 10% of all eligible voters in one year, more than 300,000. People don’t sign a petition just because they think the public should “have a say”. They sign it when they oppose the status quo and want that to be known. So unless you have a critical mass that both wants change and cares enough to petition, you won’t hit the threshold.

Finding one in every ten voters up and down the country is hard; many have tried and failed. Now imagine finding one in every nine, or one in every five, if you factor in turnout to the smacking referendum. Presumably actually reaching everyone to sign the petition is not that feasible. So by the time you’ve hit the 10% threshold, you should have a really good victory in store.

Of course, should is the keyword in that sentence. My criticisms of statistical fallacies make it churlish for me to pretend what I’m saying is known to be empirically true. There are probably reasonable narratives the other way; without comprehensive (and mandatory) surveys, all of it is guesswork. It’s tempting, even for me, to try to draw inferences about what the result means. My suspicion is the (relatively) low outcome just indicates that the issue is more polarized (or organized) that most, since probably a greater number of opponents actually signed the petition.

But consider the big picture. There are, at a basic level, three outputs: turnout, informal votes and the actual result. (Okay, maybe the electorate data is useful too.) There are lots more variables that go into voting. Desire to vote, hassle, stance, strength of stance, interaction with “meta-opinions” like the right to govern, opinion on the fact that the referendum is being held, and the impact of each of those on action taken. We can make inferences, but they inevitably require guesses (even if sensible ones) about how those things fit together.

This is not to say that the left should not dance in victory; 67% is not bad at all. And it is true that a majority of New Zealanders oppose asset sales—we knew that long before the referendum. But in the context of New Zealand CIRs, it’s hard to see why they seem over the moon. Compared to other petitioners in recent history, they’ve underperformed. Still a good result, but not as “emphatic”, “overwhelming” or “decisive” as they claim it to be.

Follow the majority, except when they disagree with me

It’s really easy to make up principles of democracy to suit your opposition. It’s far harder to keep them consistent.

The opposition hasn’t relented on their claim that the government lacks a “mandate” on asset sales. The majority of New Zealanders oppose it—even the government acknowledges so—so the government shouldn’t do it. In populist politics, this claim is appealing. As a principled claim, it deserves more examination.

The Greens have shown as much already when pressed on how this applied to their support of the anti-smacking bill. As I said at the time, I think the smacking referendum question had logical issues. But empirically, it was still clear that the vast majority of Kiwis wanted the law repealed. The Greens say that it was correct to ignore this because the bill had an overwhelming majority in Parliament. This is rather farcical: essentially, direct democracy is called on only when representatives are evenly divided. I’m not a fan of direct democracy, but insofar as one likes it, that principle is very belittling of it.

Of course, there are also plenty more circumstances where Parliament’s been evenly divided, that haven’t been subject to “mandate” criticisms. So more generally, under what circumstances should the government abandon its own election policy in favour of topic-specific polls? The left has been shy on exactly what their criteria are, so it’s hard to tell. So I admit this is reading between the lines. But here’s my best effort at inference so far:

  • It’s fine when much more than 50 per cent (say, maybe 70 per cent) of Parliament supports it.
  • It’s fine when the amount of public opposition isn’t much more than 50 per cent (say, less than 60 per cent).
  • It’s fine when it doesn’t concern a long-term economic decision.
  • It’s fine when it’s not something “built up by generations of New Zealanders“, or I suppose something with a similarly long history.
  • It’s fine when the public isn’t very vocal about opposition. By “the public”, of course, I just mean political parties with lots of media time.

The weird thing about these principles is that they carve out criteria for a lack of a mandate. Conversely, to have a mandate, only one of the criteria has to hold. Logically, it is necessary to do this to avoid the inconsistencies inherent in the Greens’ and Labour’s record.

Intuitively, it’s stupid. Each of those planks implies a different principle. It’s impossible, in this framework, to devise a single consistent conception of what “mandate” means. It’s hard to escape the feeling that the left has been making up as narrow a set of exceptions as they need to, so they can harp on about democracy for this one issue.

Let me suggest a simpler principle: The government should follow the majority view when and only when it happens to agree with my view. This is convenient: you are free to oppose whatever you are opposed to. It’s also empirically true. Being in the minority has never stopped anyone from arguing that a policy shouldn’t be enacted. It certainly hasn’t stopped the Greens, who are in the minority for most of their core policies.

Of course, the problem with it is that it doesn’t sound very nice. But when would this principle ever give a different outcome?

If the fact that a view is held by a majority can’t change your view on that policy, then it’s not a very useful discriminant and it shouldn’t be used as if it is.

I can’t decide how to vote in this referendum

I don’t know… I just really don’t know.

I was never convinced by the case National made in 2011 for selling 49% stakes in certain state assets. I didn’t really understand what they were trying to achieve. First, they would use the proceeds to pay off public debt. Then, they would use the proceeds to build schools and hospitals. I realise you can do both, but the more you do of one, the less you can of the other. Clearly debt wasn’t that pressing an issue. Then, they would give “mums and dads” a chance to invest in some of New Zealand’s best performing assets. So they obviously thought they were high-value, presumably because they brought good dividends. But they also thought a good way to balance books was to sell them. Which explanation flies?

At the same time, Labour’s opposition tactic baffled me. Something emotional about “owning our future”, whatever that means. Something nationalistic about being “tenants in our own country”, as if I felt ownership because my government owned it. Something about profits going elsewhere, when presumably the sale price reflects anticipated profits. It amazed me how a technical economic topic could become as emotion-driven as marriage equality. As unconvinced as I was about National’s case, I couldn’t see why it’d be a disaster, either.

Here’s what I never got. If you think the government will “win”, as in, get a good price for the shares, then you must also think that these New Zealand “mums and dads” will be paying a premium for them. Conversely, if you think that that it’s a real opportunity for investors, then you must also think that the government will get lacklustre proceeds from the sale. So either the government cheats itself, or it cheats its constituents.

Okay, not quite. Firstly, you might think that neither party will win or lose, and it all won’t be much different. But then, why the hype, the angst, the excitement? Secondly, you might think it is a win-win, like most economic transactions. But stocks don’t work like consumable goods: their value (in general) is in their capacity to make you money. Thirdly, you might take the ideological line, about states being bad conductors of business. But no-one made this argument. They couldn’t, because National was going to retain a controlling stake.

This tension might seem convenient to opponents, but a similar one strings them. If the assets are high-value, as Labour and the Greens campaigned in 2011, then selling them means giving up good dividend streams—but also getting a good price from buyers. If the assets are low-value, as Labour and the Greens think now, then selling them means getting disappointing sale proceeds—but also ridding its books of poor investments. Either way, the price will (presumably) reflect the value of the asset.

So when asset sales became the defining issue of the election, I never got it.

After the election, the re-elected National government proceeded to do exactly as it said. And Labour and the Greens continually claimed that National didn’t have a “mandate” to run the sales. Why? Because polls showed that 70% of New Zealanders opposed them.

There is a huge meta-question behind that chain of logic. If the majority of a constituency opposes a policy, is the government obliged to follow them? How do you ascertain that for every single policy? Everyone made this issue the defining question of the election, and one side lost. Why does or doesn’t that constitute a mandate? What does that mean for the concept of representative democracy? When do we abandon it in favour of direct democracy?

Frustratingly, those questions got almost no attention. I don’t intend for this post to be an analysis of constitutional principles (hint: this means you should not comment giving or asking for an analysis of constitutional principles), but to provide some context, briefly: I think we elect governments to do a job, and then let them do their job. There are lots of nuances to this, and I think it’s important to hear opposition, but I don’t think polls should alone be enough to dictate government policy.

So when National proceeded, I wasn’t fussed. I don’t think it’s an affront to democracy, and I can’t see why it’ll be hugely either damaging or beneficial.

I actually abstained in the last citizens’ initiated referendum, because I thought the question was dumb. My threshold for boycotting a vote is very high: along the lines of, there is no correct way to answer the question. Otherwise, in general, I think people should vote in referenda. Tactically, there’s no way to distinguish between a boycott vote and an apathetic one. In principle, even if opposed to citizens’ initiated referenda, I think we should change the system, not try to undermine it.

So I’d like to vote. If someone can suggest to me why I should vote one way or the other, I’m happy to hear it. But please don’t revert to ideological references to the past, emotive nationalist ownership arguments, arguments that contradict the 51% stake thing, or other ridiculousness. And please try to be consistent. If you think a low price makes it a failure, I expect you to think that a high price makes it a success. Otherwise, find an argument that isn’t premised on the amount of proceeds or dividends.

On consensus and compromise

This is hard enough at a smaller scale. How could it work at a global one?

I’ve been intrigued by the accounts of a friend (and old flatmate) who’s at the COP19 talks in Warsaw. She makes a plethora of insightful observations, most of which this post is not about, but a theme of the conference—even more so in general accounts than hers—has been a lack of progress towards any sort of agreement.

The phenomenon isn’t new to Warsaw, of course. Copenhagen 2009 is widely remembered for its failure to produce an agreement, or in the eyes of many, any meaningful progress. The conventional wisdom is that everyone talks a lot about urgency, but no-one acts on it.

I’ve always found the idea that 198 countries could gather in a room and reach unanimous agreement fanciful. This isn’t unique to climate change talks. I wrote about last year’s UN telecommunications conference, which failed similarly. I don’t really keep up with international politics much, but I sort of get the feeling this generally runs through most attempts to thrash out a global consensus.

The thing is, most of the time, we have a hard enough time coming to a consensus domestically. I could point to the recent U.S. government shutdown, which is analogous not just for its lack of timely compromise, but also because everyone talked about the need to avert a greater calamity. But that’s an extreme example. It’s actually more difficult to think of counterexamples than it is to think of supporting examples. Extraordinarily few votes in national legislatures in most countries (or at least most democratic countries) are unanimous. Very few are even near-unanimous. Most votes dealing with major issues split relatively evenly, as in, between one- and two-thirds.

So what hope is there of achieving consensus among the entire world?

I could write for ages about potential solutions. Pretty much no country demands such agreement in its domestic political process, of course, but there are lots of reasons why that won’t work at an international level. Perhaps a game-theoretic approach would help the stalemate. Truthfully, I don’t really know what the answer is, so I’ll just avoid rambling aimlessly.


Get every new post delivered to your Inbox.

Join 1,329 other followers