Push vs. Pull: asylum seeker numbers and statistics

Round 42 in the never ending war against ratbaggery brings us to the argument of what drives asylum seeker numbers – push factors vs. pull factors.

Proponents of the push factor view claim that the numbers of people attempting to enter Australia to seek refugee status – including those that arrive by boat – are primarily driven by events outside of Australia’s direct control. They argue that events like war, political unrest and other causes of human displacement and general misery around the world create a supply of asylum seekers that spread throughout the globe seeking sanctuary and a better life, and that the numbers coming to Australia are primarily a function of these events rather than domestic Australian policy.

The Pull Factor school of thought on the other hand claims that it is primarily Australian domestic policy responses that define the number of people seeking asylum in Australia. They argue that there is always a large supply of those seeking asylum around the world, and that marginal changes in Australian domestic policy lead to large changes in the proportion of that global pool that will choose to seek asylum in Australia rather than alternative destinations. The Pull Factor school ultimately argues that marginal changes in the deterrence level of Australian policy is the difference between pushing asylum seekers away to be someone else’s problem, or pulling them toward Australia to become our problem.

Fortuitously, we have a handy little natural experiment available to test the broad accuracy of the Pull Factor school. Firstly, Australia and New Zealand exist in the same part of the world, meaning that we would expect to experience the same regional dynamics when it comes to localised asylum seeker numbers. Secondly, we are both relatively desirable destinations with a western orientation and free from any internal political persecution of minorities. Thirdly, and most importantly, over the last 15 years or so Australian and New Zealand border protection policy became sharply divergent. From the end of 2001 Australia implemented the Pacific Solution while New Zealand policy has remained fairly consistent over the entire period.

If the Pull Factor school of thought was accurate – if pull factors really do dominate asylum seeker numbers – then we would expect to see very little correlation in total asylum seeker application numbers between Australia and New Zealand – afterall, our respective policies are different and during the Pacific Solution period were vastly different.

If we take the total asylum seeker application numbers for both Australia and New Zealand over the period of 1994-2008, we can run a scatter plot and regression line to see if there is any correlation.

This tells us that those carping on about Pull Factors as being the dominant effect, are engaging in a few pull factors of their own. The Australian and New Zealand experiences are highly correlated in a very strong statistically significant way. This is the exact opposite of what would occur were our respective domestic policies the dominant influence on our respective asylum seeker numbers.

Something else is driving our numbers together – giving us this high correlation.

We don’t have good numbers for total global asylum seeker applications, but we do have good data on the following 38 developed and borderline developed nations: Australia, Austria, Belgium, Bulgaria, Canada, Cyprus, Czech Rep, Denmark, Estonia, Finland, France, Germany, Greece, Hungary, Iceland, Ireland, Italy, Japan, Latvia, Liechtenstein, Lithuania, Luxembourg, Malta, Netherlands, New Zealand, Norway, Poland, Portugal, Rep. of Korea, Romania, Slovakia, Slovenia, Spain, Sweden, Switzerland, Turkey, United Kingdom and the United States

If we compare the combined Australia and New Zealand numbers against this global 38 group, we get:

Again, a strong correlation. Not as strong as that between New Zealand and Australia which share the same regional dynamics, but strong none-the-less, suggesting that even regional differences get swamped by larger trends in global asylum seeker supply numbers.

Those folks promoting “Pull Factors” as being the dominant influence of total asylum seeker numbers are, quite simply, wrong.

Where it does become more complicated however is if we ignore total asylum seeker numbers and just look at those asylum seekers arriving by boat. Yet, by doing so, we are effectively ignoring the majority of asylum seeker claims in Australia – sometimes ignoring nearly all asylum seekers in Australia. If we look at just the total amount of people arriving by boat each year as a proportion of total Australian asylum seeker applications, it tells a complicated story:

The years 1999 to 2001 saw the dreaded ‘boat people’ make up a relatively high proportion of our total asylum seeker numbers– but it’s also worth noting at this point that those three years make up 3 of the top 4 years in total asylum application numbers in our collection of 38 nations.

When those proponents of “Pull Factors” that are slightly more nuanced in their approach (more nuanced than, say, boneheads like Sharman Stone and Andrew Bolt for example) limit their argument to just “boat people” – they are still left arguing at best about a minority of asylum seekers, at worst about virtually none. However, they do have some evidence backing their argument – although evidence of high uncertainty.

If we rank the years 1994 to 2008 by numbers of ‘boat people’ arriving in Australia from lowest to highest (so the lowest year gets a score of 1, while the year with the highest number gets a score of 15), and do the same for total global asylum application numbers from our 38 nations (1 being the year with the lowest number and 15 being the year with the highest number) – we can compare the relative strength of boat people arrivals in Australia with the relative strength of broader global asylum seeker trends. When we run a scatter and regression with the two sets of ranks, we find ourselves with an interesting outlier.

If we ignore the 2002 result as an outlier and use the other 14 years to build our regression line, it tracks well – suggesting that the relative patterns through time of boat arrivals in Australia, is itself a function of broader global asylum seeker trends. The reason 2002 is an outlier (and also a dodgy stat that we’ll get to in a tick) comes from it being the starting point of the Pacific Solution – the year when 1 billion dollars worth of forward expenditure began to be implemented. It did reduce numbers – by redefining parts of Australia as not actually being Australia. Any boat people that happened to land were conveniently excluded from the statistics by an act of definition.

Let’s be clear – this is what the Pacific Solution did – it diddled the stats by redefinition. Boats still made the attempt to enter Australia – which is a point worth noting as many of the proponents of Pull Factors cite reducing the risk of death from reducing the number of people attempting the voyage by boat, as one of their key rationales. Yet we know that SIEV(s) 5,7,11 and 12 in 2002 attempted to make the journey and were returned to Indonesia while SIEV(s) 4,6 and 10 actually sank. That was in very late 2001 through late 2002. In 2003 we know that boats were still attempting to make the voyage such as SIEV 14, but were again towed back from whence they came.

The UNHCR estimates that 1600 people were diverted throughout the time of the Pacific Solution, but hard numbers are difficult to come by.

What that figure doesn’t take into account are the numbers that attempted the voyage but were turned back – nor those that sank or were suspected of being lost at sea.

If we adjust the 2002 number to account for boats that not only attempted to make the voyage, but ended up detained within the Pacific Solution system using the numbers provided by the Select Committee for an inquiry into a certain maritime incident 2002, we can add 1546 to the 2002 number. Redoing the same chart as above now gives us:

Which paints a somewhat different picture.

So while the more nuanced argument of Pull Factors – that they reduce the number of boats attempting to make the voyage – may have some force of weight on paper as evidenced by the statistics of boat arrivals following the Pacific Solution, it still relies on a set of very dubious numbers by ignoring the reality of voyages still being undertaken, but conveniently redefined as somebody else’s problem. While Pull Factors most likely have some relatively small effect on boat numbers, they are simply swamped – overwhelmingly swamped – by Push Factors.

UPDATE:

Just a quick update – I forgot to put links in for where the actual data came. For asylum seeker application numbers you can go here and here at the UNHCR, while the boat arrival data came from the parliamentary library here (now moved here)