Terrorism and international air travel

When a terrorist attack occurs in a country, it seems natural to expect that international tourists would be dissuaded from visiting that country. How big an effect does terrorism have in reducing international tourism arrivals? That is essentially the question addressed in this 2018 article by Devashish Mitra (Syracuse University), Cong Pham (Deakin University), and Subhayu Bandyopadhyay (Federal Reserve Bank of St. Louis), published in the journal The World Economy (ungated version here).

Their data covers the period 2000-2014, with bilateral air passenger travel between 58 source countries and 26 destination countries, drawn from the UN Service Trade Database, and terrorism data from the Global Terrorism Database maintained by the University of Maryland. They limit the terrorism data to:

...all non-state terrorist attacks that the GTD can classify without uncertainty as terrorist incidents to construct the measure of terrorism as our main explanatory variable of interest...

For the analysis, they employ a gravity model approach (which I have used in my own research, and previously described here and here). Mitra et al. find that:

...terrorism adversely and significantly impacts bilateral air passenger travel. What is the economic significance of our estimates? According to the results... a 10% increase in the number of terrorist incidents in the source country and the destination country results in a reduction in bilateral air passenger travel at least approximately by 1% annually for pairs with bilateral distances of 1,000 km or less... It equivalently means that an additional terrorist incident, which usually is of very small scale and non-fatal, can cause average bilateral air passenger travel between those source and destination countries to decrease by at least 1.3% or US$0.9 million approximately... Similarly, for pairs of countries with bilateral distance being 2,000 km or less an additional terrorist incident causes approximately a 0.82% decrease in their bilateral air passenger travel.

They also find that transnational terrorism has larger effects, and present a number of robustness checks of the results. However, I want to stop things right here, because there are two major problems with their analysis. First, their dependent variable is the dollar value of air passenger travel. That is problematic because the value of air passenger travel is made up of the number of air passengers multiplied by the cost of their travel. Theoretically, we might expect the number of air passengers to decrease due to terrorist attacks. However, the theoretical effect on the cost of travel is indeterminate. If the demand for air travel decreases, prices will decrease. However, if the supply of flights decreases, prices will increase. These effects offset each other. And besides that, I would argue that we should be more interested in the number of air passengers anyway, not the value of air passenger travel.

Second, the key explanatory variable (terrorism) is also problematic, because they measure it as the total number of terrorist attacks in both the origin (where the air passengers are coming from) and the destination (where the air passengers are going to). As noted at the start of this post, terrorism should dissuade international air travel. However, that applies to terrorist attacks at the destination. It doesn't apply to terrorist attacks at the origin. In fact, you could argue that terrorist attacks at the origin should increase international air travel, as people try to escape the risk of terrorism. So, again, the effect of terrorism as Mitra et al. measure it on air passenger travel is theoretically indeterminate.

Combining those two problems, I think the analysis doesn't really tell us much at all about how terrorism affects international air travel, because both the dependent variable and the key explanatory variable are both mis-measured. However, there is clearly an opportunity for some follow-up work by a good Honours or Masters student, using more appropriate data to explore the same research question.

Adam Smith on the gravity model of trade

I've written a number of posts that reference gravity models, including a lot of my own research. For example, see here for a post about my own research using the gravity model of internal migration flows, or here for a post about joint work with one of my PhD students on using the gravity model of international trade flows. Essentially, a gravity model suggests that the flow (of people in a migration model, or goods and services in a trade model) between two regions is positively related to the 'economic mass' (usually measured as population in a migration model, or economic output or GDP in a trade model) of the origin and the 'economics mass' of the destination, and negatively related to the distance between the two places. This idea of gravity models in migration goes back to work by Ernst Georg Ravenstein in the 1880s, and then developed mathematically by George Kingsley Zipf in the 1940s. In trade, the mathematical gravity model is usually attributed to Walter Isard in the 1950s.

So, the gravity model is old. But, as the inside joke among economists goes, there's been nothing new in economics since Adam Smith. And it turns out that Smith had a number of things to say in his most famous 1776 book The Wealth of Nations, as related by Bruce Elmslie (University of New Hampshire, and previously a visitor at Waikato) in this 2018 article published in the Journal of Economic Perspectives (open access). Elmslie notes a number of places where Smith alludes to gravity in the context of international trade, in Book IV, Chapter III, Part II of the Wealth of Nations. Elmslie writes that:

Smith (1776, pp. 624–25; emphases added) compares the trade that could take place between England and France if his system of natural liberty prevailed versus the forced, policy-driven trade between England and the North American colonies, and between France and its colonies:

[T]he commerce of France might be more advantageous to Great Britain than that of any other country, and for the same reason that of Great Britain to France. France is the nearest neighbor to Great Britain. In the trade between the southern coast of England and the northern and north-western coasts of France, the returns might be expected, in the same manner as in the inland trade, four, five, or six times in the year. The capital, therefore, employed in this trade, could in each of the two countries keep in motion four, five, or six times the quantity of industry, and afford employment and subsistence to four, five, or six times the number of people, which an equal capital could do in the greater part of the other branches of foreign trade. . . . It would be, at least, three times more advantageous, than the boasted trade with our North American colonies . . . France, besides, is supposed to contain twenty-four millions of inhabitants. Our North American colonies were never supposed to contain more than three millions: And France is a much richer country than North America; . . . France therefore could afford a market at least eight times more extensive, and, on account of the superior frequency of the returns, four-and-twenty times more advantageous, than that which our North American colonies ever afforded. The trade of Great Britain would be just as advantageous to France, and, in proportion to the wealth, population and proximity of the respective countries...

Thus, Smith holds that if the trade volume between two countries is determined by each country’s consideration of “their real interest, without either mercantile jealousy or national animosity” (p. 624), it will be in relation to the size of the national produce of each country and the distance or proximity between them.

Notice that the three passages that Elmslie has emphasised in italics in the quote from Adam Smith all seem to relate to the gravity model of trade. How did Smith come upon the gravity model? Interestingly, Elmslie notes that:

Smith did not use the “gravity” terminology explicitly, but it is intriguing that for the determinants of the volume of trade Smith emphasized mass and distance, which is of course similar to Isaac Newton’s theory of gravity. Smith gives no direct indication that he had Newton’s gravity model in mind, but a connection is not implausible. Newton’s work had a significant impact on Smith’s methodology in the Wealth of Nations... In an earlier work written prior to 1758, Smith (1795) calls Newton’s theory of gravity “the greatest discovery that ever was made by man”...

So, there you have it. Adam Smith may actually have laid some of the initial foundations for the gravity model of trade (and migration), having been inspired by the work of Sir Isaac Newton.

The gravity model and cultural trade in restaurant meals

One of my favourite empirical models to work with is the gravity model. It is an extremely high performing (in terms of both in-sample and out-of-sample forecast accuracy) model when used in migration and trade contexts, and quite intuitive. Essentially, in a gravity model the flow (of goods and services, or people) from area i to area j is negatively related to the distance between i and j (so, if i and j are further apart, the flows are smaller, most likely because it costs more to move from i to j), and are positively related to the 'economic mass' of i and j (so, if i and/or j is larger, the flows from i to j will be larger).

I have used the gravity model myself (e.g. see this post), and so have my students (e.g. see this post). I particularly like it when I find examples of unexpected uses of the gravity model. For instance, there was this paper on running the gravity model in reverse to find lost ancient cities (which I blogged about here). 

Most of the time, a gravity model of trade involves goods and services that cross borders. However, that is not the case in this recent article by Joel Waldfogel (University of Minnesota), published in the Journal of Cultural Economics (appears to be open access, but just in case there is an ungated earlier version here). Waldfogel is probably best known for his work on the deadweight loss of Christmas (see also here), but in this research he looks at the cultural trade in restaurant dining.

The interesting thing about this article is that the data isn't really trade data at all. Restaurant meals don't cross borders. Instead, it is the intellectual property that is crossing borders, which is why this article relates to the literature on cultural economics. Waldfogel uses:

...Euromonitor data on aggregate and fast-food restaurant expenditure by country, along with TripAdvisor and Euromonitor data on the distribution of restaurants by cuisine.

He links each cuisine to an origin country (i in the description of the gravity model above), and the country location of the restaurant as the destination country (j in the gravity model description). He then calculates measures of 'trade flows' in restaurant meals, both including and excluding fast food, and runs a gravity model using those data. He finds that:

As in many models of trade, distance matters: a 1% increase in distance reduce trade by about 1%... Common language and common colonial heritage also matter.

Those are pretty standard results in the trade literature using gravity models. Then:

Which cuisines are most appealing after accounting for rudimentary gravity factors?... Excluding fast food, the ten most appealing origins are Italy, China, and Japan, which all have similar levels of appeal, followed by the USA, India, France, Mexico, Thailand, Spain, and Turkey. When fast food is included, the USA rises to the top, and the others remain in the same order.

Finally, on the balance of trade in restaurant meals, he finds that, of 44 selected countries:

...three are substantial net exporters: Italy (with net exports of $158 billion), Japan ($44 billion), and Mexico ($17 billion). Substantial net importers include the USA ($134 billion), Brazil ($39 billion), the UK ($20 billion), and Spain ($20 billion).

I was a little surprised that Spain was such a net importer of cuisine from other countries. I guess that reflects that Spanish cuisine isn't as available outside of Spain as many other European cuisines are. The US and UK being large net importers is not a surprise though.

The results are mostly uncontroversial. However, I did take issue with some of the choices. Waldfogel codes all "pizza" restaurants as Italian. I'm not convinced that Pizza Hutt or Domino's count as Italian food - more like generic fast food, most of which was coded to the US. It would be interesting to see whether re-coding pizza would make any difference to the results - possibly not, as Waldfogel does test for the impact of coding "fried chicken" as either domestic or US and that appears to make little difference.

The gravity model is clearly very versatile, and deserves much greater attention in research than it currently receives. This research demonstrates a slightly new direction for it.

[HT: Offsetting Behaviour, last year]

Are international trade and migration complements or substitutes?

As migrants move from their origin country to a destination country, does that result in increased trade as well as migration? Intuitively, it seems like it would. In the simplest sense, those migrants might send goods back to family and friends at home in the origin country, and they might import goods from the home country to their destination. Causality need not run from migration to trade though. People are more likely to migrate to places that they are more familiar with, and having experienced goods from a country might increase familiarity with it - a mechanism leading from trade to migration. Either way, those explanations would suggest that international trade and migration are complements - an increase in one is associated with an increase in the other. However, the international literature has been inconsistent in its findings. Some studies find that trade and migration are complements, while other studies find that they are substitutes - an increase in trade is associated with a decrease in migration, and vice versa.

In a new working paper, my PhD student Rosmaiza Abdul Ghani and I, along with Bill Cochrane (University of Waikato) and Matthew Roskruge (Massey University) use a newly available migration dataset, along with longstanding trade flows data, to investigate these relationships. Most studies of migration and trade limit themselves to a few countries, or use migrant stocks (the number of migrants living in a particular country) as a proxy for migration. However, this dataset by Nikola Sander and Guy Abel covers migration flows between over 240 countries. And it comes with cool graphics (try them at this link).

Anyway, that data allows us to investigate the relationships between trade and migration more thoroughly than previous studies. We make use of seemingly unrelated regression, which is a technique that allows us to simultaneously model the relationships that run in both directions. We found that:

...trade and migration have positive coefficients in all of the specifications except for the fixed effects model (where, as noted above, the interpretation of the coefficients is challenging). That is, trade and migration are complements. In our preferred PPML-SUR specification, an additional migrant from country i to country j is associated with 1.7 percent higher trade flows from country i to country j, while an additional USD1000 in trade flows from country i to country j is associated with 25.4 percent higher migration flows from country i to country j.

The second of those coefficients seems a little large, but it starts from a very low base - perhaps we should have re-centered the data. In any case, the results support the story I noted at the beginning of this post - international trade and migration are positively related, so they are complements. This analysis is correlational though - we haven't established any causality here. That is the subject of the second paper contributing to Rosmaiza's PhD, which I'll blog about in a future post.

How the internet affects international migration decisions

In the 1960s, Everett Lee came up with a model of migration decisions (ungated version here) that remains the most widely used theory of what determines peoples' migration decisions. Lee emphasised that there are: (1) push factors - things in the origin that cause people to want to move away, like low wages or high unemployment; and (2) pull factors - things in the destination that cause people to want to move there, like high wages or low unemployment. Push factors can be positive (they make you want to move away), or negative (they make you want to stay), and likewise pull factors can be positive (they make you want to move there) or negative (they make you not want to move there). A third factor has been added to these push and pull factors - facilitating factors, or things that make migration easier, like a simplified visa process.

With the theory out of the way, we can now consider how the internet might affect international migration decisions. The internet facilitates improved communication, including between diasporas and those living in their home country. So, perhaps greater internet access might make migration easier, by increasing the flow of information about how or where to migrate to.

Alternatively, maybe the internet facilitates outsourcing of jobs to previously lower-income countries, increasing job prospects and incomes in the origin, and reducing migration (a negative push factor), or allowing people to work for a firm in the destination country without having to leave the origin country (a negative pull factor). Or maybe the internet facilitates access to goods or services (e.g. Netflix) that previously weren't available in the origin country, improving the quality of life there (again, a negative push factor).

So, it isn't clear from the theory whether the internet would increase, or decrease, international migration. What do the data say? A 2017 article by Hernan Winkler (World Bank), published in the journal Applied Economics Letters (I don't see an ungated version, but it appears to be open access), provides some answer. Winkler estimated a gravity model (which I have previously discussed here) using data from 6072 origin-destination pairs of countries over the period 1990-2010. He found that:

...a 10% increase in internet penetration in the source country is accompanied by a 1% decrease in the stock of migrants born there.

That was based on the simplest model he reported, but other models supported that the internet reduced migration. He also found similar results using an instrumental variables approach, which implies that the results are causal - that is, the internet caused a reduction in international migration. Winkler concluded that:

...the internet may weaken the importance of push factors in the decision to migrate, and that these effects dominate any declines in mobility costs associated with this new technology.

This isn't the last word on this, but it is consistent with a story that the internet is associated with job outsourcing from high-income countries to low-income countries, and weakens the incentives for workers from low-income countries to migrate. Maybe, rather than building a wall, the U.S. should be building out internet infrastructure in low-income countries?

Running the gravity model in reverse to find lost ancient cities

The gravity model of trade or migration (which I have written about before here) must be one of the consistently best-performing empirical regression models (in terms of in-sample prediction, at least). The model is really simple too. In its simplest form, a gravity model suggests that the migration (or trade) flow between two regions is positively related to the 'economic mass' (proxied by population in the case of migration, or by GDP in the case of trade) of the origin and the economic mass of the destination, and negatively related to the distance between the two places. So really, you don't need a whole lot of data in order to run a gravity model (though you do need data on trade or migration flows).

The standard gravity model is based on known data such as the distances between countries (or regions, or cities). But what if you didn't know where the cities were (as might be the case for lost ancient cities), but you did know the size of the trade flows? Could you use the gravity model to triangulate the likely location of those lost cities, by estimating the distance from their trade partners? It turns out that yes, you can.

In what might be the most ingenious use of the gravity model I've ever seen, a new NBER Working Paper (ungated version here) by Gojko Barjamovic (Harvard), Thomas Chaney (Sciences Po), Kerem A. Coşar (University of Virginia), and Ali Hortaçsu (University of Chicago) does almost exactly that. The authors use a dataset of over 12,000 Assyrian tablets from 1930-1775 BCE, 2,806 of which contain mentions of mentions of multiple cities in Anatolia (modern-day Turkey). Of those tablets, 198 contain merchants' itineraries for 227 itineraries relating to travel between 26 cities (15 of which are known, and 11 of which are 'lost'). The authors explain the difference between known and lost cities, as:

‘Known’ cities are either cities for which a place name has been unambiguously associated with an archaeological site, or cities for which a strong consensus among historians exists, such that different historians agree on a likely set of locations that are very close to one another. ‘Lost’ cities on the other hand are identified in the corpus of texts, but their location remains uncertain, with no definitive answer from archaeological evidence. From the analysis of textual evidence and the topography of the region, historians have developed competing hypotheses for the potential location of some of those.

So, the authors use the data from the itineraries to construct a dataset of trade between known cities, and between known and lost cities. Using that dataset they then estimate a gravity model of migration, which provides an estimate of the distance elasticity of trade of about 3.8. That means that each 1 percent increase in distance between two cities reduced trade by about 3.8 percent. This is much higher than modern models of trade where the elasticity is usually about one, but given ancient trade was mostly by road (or by coastal shipping) and the roads were not high quality, that doesn't seem too unusual.

Next comes the really cool bit. They then use the distance elasticity measure to 'back out' estimates of the location of the lost cities. Their method even gives confidence bounds around the estimated point location of each lost city. They conclude that:

...[f]or a majority of the lost cities, our quantitative estimates come remarkably close to the qualitative conjectures produced by historians, corroborating both such historical models and our purely quantitative method. Moreover, in some cases where historians disagree on the likely location of a lost city, our quantitative method supports the conjecture of some historians and rejects that of others.

Eyeballing the results from the maps though, the estimated location of the lost cities doesn't appear (to me) to be particularly close to the historians' qualitative estimates. However in spite of that, this is a very cool paper using the gravity model in a very novel way. Hopefully we see more of this in the future.

[HT: Marginal Revolution]

Climate change won't much affect internal migration in NZ

Climate change is likely to be one of the key challenges facing humankind over the coming century (or more). We are likely facing increases in mean temperature, desertification, rising sea levels, and increasing frequency and intensity of extreme weather. But how big is the impact likely to be on a country like New Zealand, anyway?

In a new working paper, I evaluate the impact of climate change on internal migration in New Zealand, and what that means for the future spatial distribution of population. That is, which regions are likely to gain population from climate change, and which will lose population? I make use of a gravity modelling framework (which I have written about before). Essentially, a gravity model suggests that the migration flow between two regions is positively related to the population of the origin and the population of the destination, and negatively related to the distance between the two places. I tried out a bunch of climate variables from NIWA to find those that appeared to have the biggest impact on internal migration, using data on inter-regional migration from the last four Censuses (1991-2013).

Three climate variables are found to have statistically significant associations with internal migration: (1) mean sea level pressure in the destination; (2) surface radiation in the origin; and (3) wind speed at ten metres at the destination. The sign of the effects suggest that migrants attracted to areas with more settled weather (higher mean sea level pressure); migrants are less likely to move away from areas with more sunlight hours (but interestingly, don't move towards those areas); and migrants prefer to avoid moving to areas that are windier.

I then embedded the gravity model within a cohort-component population projection model, which is something that Jacques Poot and I have been working on for a number of years. I used the projections model to evaluate the effect of different climate change scenarios on regional populations out to a horizon of 2100.

Including the three climate variables in the population projection model makes a small difference to the regional population distribution. The inclusion of climate variables increases the projected populations of Northland, Bay of Plenty, Gisborne, Hawke’s Bay, Taranaki, and Nelson. The overall impact is quite small, as you can see from the diagram below for Northland. The orange line tracks the projected population of Northland excluding any impact of climate, while the grey line includes the impact of climate. Bear in mind that Northland shows the biggest effects in relative terms - the effects on other regions are smaller.

I also looked at the effect of different climate change scenarios, and the difference between different climate scenarios is negligible. The diagram below shows the projections under different climate scenarios for the Southland region. As you can see, there is little difference between them (and that result is similar for other regions as well).

Overall, the results suggest that, while statistically significant, climate change will have a negligible effect on the population distribution of New Zealand at the regional level. This is not to say that climate change will not have important and substantial effects at very localised levels, as a result of sea level rise, for instance. However, most if not all of the displacement of people will be within regions. For example, maybe those displaced by sea level rise simply move a little further inland, or we build walls to keep the sea at bay.

Read the full working paper here.

Distance matters less for students going to high quality universities

A new paper published in the journal Spatial Economic Analysis (sorry I don't see an ungated version anywhere) by John Cullinan and Jim Duggan (both National University of Ireland, Galway) uses gravity modelling to investigate the factors associated with student flows from secondary schools to tertiary institutions in Ireland. I really like gravity modelling, and it's an approach that I've been applying to internal migration flows in New Zealand in two MBIE-funded projects (see here and here), and a Marsden-funded project (see here). Cullinan and Duggan's paper was the first time I have seen it applied to student flows to universities (though it is not the first paper to do this).

Essentially, a gravity model suggests that the flow (of migrants, trade, or students) between two places is positively related to the 'economic' size of the origin (more potential migrants, or more things to trade), the economic size of the destination (more opportunities for migrants, or more demand for tradeable goods), and negatively related to the distance between the two places (since movements over longer distances are more costly). The gravity modelling approach does a pretty good job of explaining trade and migrant flows, even without including any other variables - whether they be factors in the origin that 'push' migrants out, or factors in the destination that 'pull' them in.

Cullinan and Duggan include both school-level variables (push factors) and tertiary-institution-level variables (pull factors) in their models. They find that:

...school size, girls-only schools, mixed-gender schools and Catholic schools are associated with higher students flows, all else being equal. On the other hand, schools with DEIS [disadvantaged] status are found to be associated with lower student flows. The effect for schools located in more deprived areas is somewhat surprising, suggesting flows are lower from schools located in more affluent areas.

Given that they have already controlled for schools that are considered disadvantaged, that last result might not be as surprising as it seems. The size of the effect is tiny. They also note that:

...HEIs [Higher Education Institutions] located in Dublin are found to have lower predicted student flows once the other factors are accounted for, an effect that may be driven by the higher costs of living in the capital city.

It might also be because the top two universities (Trinity College Dublin and University College Dublin) are both located there, and will be the most selective in terms of their student intakes. Finally, one other result I found particularly interesting (though not necessarily surprising):

...the average [distance] elasticity masks considerable variation both within and across HEI types, suggesting that students are much more willing to travel further to attend some HEIs than others.

Specifically, the distance elasticities were lowest for Trinity College Dublin and University College Dublin, as you might expect for the highest quality universities.

It would be really interesting to repeat a similar analysis for New Zealand, and I expect the IDI holds the necessary data (it certainly has tertiary data and secondary school data - the question is whether they are easily linked). This would be even better than the Irish study, as there are many years of data available, rather than just a single cross-section. There's definitely a potential future Honours or Masters project in that.