Wednesday, May 14, 2014

What does migration to the United Arab Emirates tell us about labor mobility?

Summary: Some notes on migration to the United Arab Emirates (UAE). As in some other Gulf oil states, e.g. Qatar, almost the entire UAE private sector workforce is composed of foreign guest workers. The ratio of foreign workers to natives is high enough that if achieved by all developed countries it could absorb the labor force of the developing countries. The distribution is dominated by less skilled workers and workers from poor countries, who enjoy much higher wages than at home, but much lower than in countries such as the United States. Emirati tolerance of extremely high immigration may be related to the almost complete insulation of Emirati nationals private labor markets, and the exclusion of migrants from citizenship and access to government revenues. In Dubai, the native population primarily subsists on taxes on the foreign-dominated private sector, enjoying an extremely prosperous standard of living. The UAE shows that truly massive guest worker programs can greatly benefit migrants and natives when politically feasible, and could eventually eliminate most global poverty if broadly imitated. 

The extraordinary level of migration to the United Arab Emirates
In a previous post I discussed Singapore's exceptional openness to immigration, with half of the population foreign-born, including a large contingent of unskilled workers, greatly surpassing the OECD countries. However, even that high migrant share does not correspond to the levels implied by estimates that most workers in poor countries would migrate to developed countries under open borders. On that dimension, the closest models may be some of the Gulf oil states, including Qatar and the United Arab Emirates (UAE), where almost the whole private workforce is composed of migrant workers. I will focus on the UAE, but a largely similar story applies to Qatar.

In GDP per capita (PPP) per capita the UAE is at near European levels. IMF and CIA estimates put per capita GDP at around $30,000, while the 2012 World Bank estimate was over $40,000. Migrants to the UAE are entering a rich country, and they do so in extraordinary numbers. The government does not have exact statistics on the size of the foreign population, estimates all agree that the Emirati citizen population is a small minority. The increasing migrant share of the population has been central to a population explosion, as in these UN Population Prospects figures:

Historical population
YearPop.  ±%  

The UAE government's own mid-2010 population estimate put the total population at 8.264 million, and the Emirati nationality population at 11.5% of the total. Other estimates put the Emirati share slightly higher. In any case, the migrant share is enough to be relevant for evaluating concerns about a native population greatly outnumbered by migrants, or migrant populations that might result the OECD countries all opened their borders.

Michael Clemens (2013) has highlighted the scale of migration flows in the Gulf countries, and the UAE specifically in calling for further research:

By any of several measures, labor migration to the Gulf is a phenomenon of vast
importance. First, it is large relative to the size of the destination countries. The population of the GCC as a whole is approximately 41 percent foreign-born (EIU 2009). This rate is much higher than that seen even in high-immigration OECD countries such as the United States and France (both 13 percent), Canada (21 percent), and Australia (27 percent) (OECD 2013, p. 25). In all GCC countries except Saudi Arabia, foreign workers make up around 90 percent or more of the entire private sector labor force (Fasano and Goyal 2004). The large majority are from developing countries, principally South Asia and the Philippines. financial flows arising from Gulf migration are vast. In 2010, South Asia
received about US$83 billion in remittances from migrants in all countries—compared to
about US$56 billion in Foreign Direct Investment (FDI) and foreign aid combined (World

...Nonetheless, researchers know very little about the effects of labor migration to the Gulf. Those effects are very difficult to measure accurately. What has been the effect of low-skill  foreign workers on the economic productivity of GCC capital and GCC labor? This requires estimating what would happen to GCC economies with a smaller or different labor force. 
What has been the effect of GCC-bound migration on India or Pakistan? This requires
estimating what would happen to South Asian societies and economies without the jobs,
remittances, work experience, and time abroad that come through jobs in the GCC. What is the effect of work in the GCC on a particular migrant or that migrant’s family? This requires estimating what would have happened to the worker or family had migration not occurred. In short, measuring effects on countries or households requires estimating things that cannot be directly observed because they did not happen; economists call these things a “counterfactual.” 

Place premium: migrants wages much higher than at home, much lower than in low-migration developed countries
Michael Clemens, who coined the term "place premium" has done research on the economic benefits to migrants moving to the UAE by looking at people offered jobs at a particular UAE construction firm. He finds a substantial place premium for Indian workers migrating to the UAE for construction jobs, although this premium is a lot smaller than those reported for the U.S.

Three key results emerge from analysis of these data. First, the economic benefit to migrant workers is extraordinarily and systematically large: migration to the UAE for basic construction work causes their daily wage to rise by a factor of five, and causes employment to rise by at least 20 percentage points. Second, there is no sign that many of the commonly-mentioned costs of migration are systematically experienced by migrants' households; migration to the Gulf causes the fraction of households in debt to sharply decline, and there is no evidence of labor force entry by school-age children or labor-force exit by adult family members. Third, households are generally well-informed about working and living conditions in the UAE, and there is no evidence that they enter into migration systematically overestimating the benefits. Households with migrants give estimates of migrants' income that closely reflect true income in UAE administrative records...
These are remarkable income gains, even in a country with a very large unskilled construction workforce. Some of those wages must be spent on plane tickets, increased cost of living, and fees, but the average benefit should still be substantial, even after taking into account the psychological costs of separation from family.

The point about expectations should be taken with a grain of salt: the fact that this particular larger firm did not take passports, pay less than promised, or otherwise cheat migrants means that this firm was honest, but only provides modest evidence. But it is a corrective to the overwhelming focus of media and scholarly attention on Gulf and UAE migration on cheating or mistreatment of workers (and more dubious downsides, such as the badness of having a poor person become less poor by working near rich people), which Clemens critiques:
The principal focus of scholarly and popular writings on labor migration to the Gulf is theharms and abuse perceived to be associated with the phenomenon, not the benefits to
workers or countries. The foremost theme of scholarly research on Gulf migration, Gardner (2012) writes, is “the problematic and exploitative labor relations that seemingly characterize the experiences of many of the poorest transnational labor migrants who spend time in the Gulf states.” To take two of numerous examples of academic research on foreign workers in the UAE, Keane and McGeehan (2008) describe “appalling” conditions in “a form of slavery,” while Zachariah et al. (2003) find that “nearly one-fifth of the Indian migrants have not received the same job, wages, and non-wage benefits as stipulated in their work contracts.” 
An even greater focus on perceived harms to workers emerges in more popular writings on migration to the GCC. For example, Human Rights Watch (2006) describes workers in the UAE as subject to “wage exploitation, indebtedness to unscrupulous recruiters, and working conditions that are hazardous to the point of being deadly.” Of all internet pages in English that mention migrant workers in Dubai, almost one-third contain the words “slave” or “slavery.”
...the migration of workers from South Asia to the UAE has been an important force for employment and poverty reduction for South Asians. It generates employment for hundreds of thousands who would otherwise be without work and creates billions of dollars in earnings each year for low-income South Asian workers... 

To look a bit closer at construction salaries, I will draw on Tong (2010), an excellent resource which I will refer to again and again. Tong combines data from a household survey (excluding poor migrant workers living in collective work camps) and from a government mandated payment system (which only includes private sector data, and is almost wholly drawn from firms with over 50 employees, biased towards low-skill construction). In combination, their weaknesses do not completely cancel out, but they provide a great start.

The payment system data puts construction workers at an annual income of 14,211 UAE dirhams. The dirham is pegged to the US dollar at about $0.27, so this gives an income of slightly under $4,000

This would put hourly wages for construction workers in the vicinity of $1-2, but these are augmented by non-cash benefits such as worker dormitories, and reduced by various costs. These construction wages are in the same ballpark as those for low-skill migrant construction workers in Singapore, after taxes (although that means the pre-tax productivity looks higher in Singapore). These wages may be small compared to migrants in the United States making $10 or more per hour in construction, but they show that even with massive migration wages can remain high enough to deliver substantial financial benefit.

Who migrates to the UAE?
The expatriate population in the UAE is drawn from many countries, but half or more stems from IndiaPakistan, Nepal, Sri Lanka, and Bangladesh. Several million are drawn from other counties in the Middle East, North Africa, and Muslim-majority countries around the world. These migrant sources on average have much lower income levels than the UAE. There are also perhaps half a million Westerners present in the country, mostly British and American. The skew towards less developed countries may make the UAE a better model for open borders than, e.g. the common labor market in the European Union. The household survey data show the prominence of developed countries, even excluding the masses of poor workers in construction labor camps:

The Ministry of Labor payment data show South Asians making up most of the private workforce:

 It is also worth considering human capital. How does the education level stack up to sending countries? If migrants to the UAE have on average been much more skilled than the average in potential sending countries (weighted by migrant supply), then its experience might not scale: if many developed countries adopted policies on the UAE they might wind up with less relevant human capital and reduced productivity, especially of low-skilled labor. Household and payment data

This educational distribution is substantially enriched compared to the secondary and tertiary education levels in the labor force of poor migrant-sending countries, but skewed low compared to developed countries, such as the U.S.:

Educational attainment in the United States, Age 25 and Over (2013)
High school graduate88.15%
Some college58.33%
Associate's and/or Bachelor's degree41.50%
Bachelor's degree31.66%
Master's and/or Doctorate and/or professional degree11.57%
Doctorate and/or professional degree3.16%
The UAE is able to maintain as high a level of human capital as it does in part by attracting many highly skilled workers from developed countries, partly by only admitting those with job offers and a few attached individuals (excluding the unskilled unemployed), and it seems in part by attracting and admitting more skilled workers from poor countries.

Overall, the UAE's migration system is probably substantially more selective than open borders but not overwhelmingly so.

The UAE seeks temporary workers, not new citizens
To enter the UAE, a foreign worker must be sponsored by an Emirati national or a business with Emirati national involvement (as a partner, board member, etc). If that employment is terminated, the worker must leave the country, and visas are temporary and must be regularly renewed. Workers may not bring their family members to reside with them in the UAE unless they meet minimum salary requirements, and family sponsorship rules discriminate on the basis of sex.

As a result, the adult sex ratio is extraordinarily skewed by male migrant workers who have left their families at home. Wikipedia gives a 2010 estimated sex ratio of about 3 males:1 female among those age 15-64. With a 91% male labor force participation rate according to the World Bank, hours worked per person are very high, which helps to keep the UAE's reported GDP per capita at European levels, even as average wages are much lower.

Keeping the families of low-skill workers out encourages guest workers to leave rather than staying illegally, and shields the UAE from responsibility for their welfare, crime, or political pressure to grant migrants the rights of Emirati nationals. Any such problems from workers result in rapid deportation or imprisonment. Migrants do not receive government benefits and preferential access to government jobs funded by oil wealth, and the military and security forces are securely in Emirati national hands. The UAE is not a democracy, and migrants do not gain electoral influence (which might let them appropriate oil and real estate wealth to themselves).

The government actively cultivates a distinctive Emirati national identity, which helps to bolster the social separation and unequal rights and status of Emiratis vs poor migrants. Discrimination based on race and national origin is permitted and widespread, as in Singapore.

This illiberal system complicates attempts to generalize from the experience of the UAE. While these factors may make migration more politically feasible for the UAE, they draw intense criticism in developed liberal democracies. This may mean that the UAE is not a good model for the political feasibility and subsequent political and fiscal effects of extreme openness to migration in developed liberal democracies.

Emirati citizens have unusually little to lose and more to gain from migrants
Are there other reasons why the Emirati government permits so much more migration per native than does Singapore?

One obvious possible explanation is oil and gas revenue. When the native population owns abundant oil and gas resources (which can be extracted using the aid of foreign energy companies and immigrants), this makes it easy to buy tradable goods, but to use oil wealth to buy non-tradable services (construction, domestic workers, electricians, etc) requires either huge increases in wages or the use of migrant labor to bring trade to the supposedly non-tradable sector.

For the UAE, the oil sector accounted for 29% of GDP in 2011 and for most government revenues when oil prices are high. Historically, the oil share was much higher, but oil spending and migrants have been used to diversify the economy. The oil is unevenly distributed among the seven constituent emirates, with Abu Dhabi delivering the overwhelming majority. Dubai's production has been declining while it has developed major industries, especially in tourism, trade, finance, and travel, so that its share of UAE oil revenues accounted for only 2% of Dubai's GDP in 2010 (when oil prices were lower, and down from half in the 1970s).

Presently, government revenues in Dubai, which has been disproportionately attracting migrants, are firmly dominated by fees on the foreign business sector, so that the native population is directly subsisting on taxation of migrants:

TAX SYSTEM: The UAE does not impose an income tax on wages and salaries. While each emirate has the authority to change this policy, none currently do so. Oil companies pay one of the highest tax rates in Dubai as they are subject to a 55% tax in addition to paying royalties. However, oil is a minor contributor to Dubai’s GDP, with total government revenues from oil and gas adding up to just over Dh5bn ($1.4bn) in 2011, according to the IMF. Foreign banks are also subject to a 20% corporate tax on profits earned in Dubai. Hotels and entertainment venues pay a 5% municipality tax for rooms, food and other services. Import taxes on luxury items, including those shipped through the free zones, is 10%, while all other imports are taxed at 4%. While they are not taxed, many government related entities also pay the government royalties. Finally, alcohol, which is not widely sold outside hotels, attracts a 30% sales tax in Dubai.
In total, revenues from these taxes were only Dh7bn ($1.9bn) of the government’s accounts in 2010. While this is a major increase from the Dh1.9bn ($517m) in 2003, it still represents only 17% of government income. Dubai therefore depends on other revenue streams to finance its annual budget. 
REVENUES: The 2013 budget, approved at the start of 2012, offers some insight into these revenue streams. Total public revenues for the year are expected to be around Dh32.6bn ($8.9bn). Fees to establish businesses and conduct trade through Dubai will come to Dh20.22bn ($5.5bn), or 62%, of the total revenues. This is a significant increase of almost 10% from 2012, even though the government is not raising its fees for these services, indicating that growth is in the number of businesses and the level of activity, as well as a greater variety of government services.
Much of the tourism trade in Dubai is drawn from beyond the Gulf:
In 2011, Saudi Arabia emerged as the top source market for Dubai’s tourism industry with 873,152 guests, followed by India (702,142), UK (643,196), Iran (476,708) and the US (462,653). Germany ranked sixth with 275,663 guests, while Kuwait and Russia scooped seventh and eighth positions with 273,253 and 255,746 guests, respectively. Oman emerged 9th on the list with 223,993 guests followed by Pakistan (221,374), China (193,791), Australia (179,214), France (152,439), Egypt (149,130), Philippines (125,408), Qatar (122,319), Italy (105,523), Jordan (95,818), Bahrain (91,238) and Lebanon (90,984).
Dubai's role as an open port and air travel hub reflects not only Gulf demand, but its central position between Europe and Asia. Manufacturing made up 16% of Dubai's GDP in mid-2013:

Inline image 1

Finance serves not only Dubai but also other countries in the region.

However, despite all this diversification, the occupational makeup of Emirati nationals remains dependent on the government jobs once financed with oil, and now increasingly with taxes on foreign workers. Essentially, the data show that the Emirati population doesn't work for private businesses. Emirati workers make up just over 0.3% of records in the private sector payment system database.

Partly this reflects low labor force participation, but the primary driver is the availability of lucrative public sector jobs, for which the government strongly favors Emirati nationals. The household survey shows government jobs superior pay, and the elevated incomes Emiratis enjoy relative to most of the population (exceeded only by mostly highly skilled Western expats).

Household income surveys also show Emiratis enjoying greatly elevated incomes. Emiratis also enjoy access to state benefits, favored treatment by the police and judiciary, and the economically valuable right to hire and sponsor migrants. With ready access to protected government sinecures, Emiratis benefit from low migrant wages as consumers without fear of private sector labor market competition.

This situation is very different from Singapore, where the native population primarily depends on unsubsidized labor income for its standard of living, rather than the largesse of the state, and some older native workers (with lower skill and education levels from before Singapore's development) face direct competition from low-skill migrants. This means that the Singaporean government faces political tradeoffs that the UAE does not, and may help to explain the difference in permitted migration. In theory, increasing the Singaporean migrant population and migration tax revenue by many times would permit Singaporean citizens to live off migration taxes, and in the interim taxes on migrants could be used to provide wage subsidies to natives. However this has not been realized in practice, due to some combination of political opposition, real estate constraints, and other barriers.

Could UAE levels of migration be scaled in other developed countries?
The UAE seems to suggest that very high ratios of migrants to natives (near to the maximum possible if the developed world opened its borders and all workers in poor countries migrated) are possible while retaining a sizable place premium, albeit lower than in other developed countries with low migration.

One scaling issue might be the supply of capital, as the UAE was able to use oil revenues and foreign investment to keep the capital:labor ratio from getting too extreme, while the world as a whole faces global capital supply constraints. The exclusion of non-workers and the denial of political rights makes it difficult to evaluate how the fiscal and political effects would differ in the other developed countries.

The political conditions leading to such high immigration may be the most difficult to replicate. First, while UAE and Gulf migration policies have enormous humanitarian and poverty-reduction benefits, they seem to be enabled by a kind of callousness to the condition of foreign workers that is unconcerned by their low wages, restrictions on family sponsorship, ineligibility for government benefits, and lack of political rights. As Clemens notes, media coverage of the migration system in the UAE is overwhelmingly negative, despite its very positive humanitarian effects. Voters in rich liberal democracies may reject migration with citizenship out of nationalism, UAE policies out of revulsion, and wind up with the status quo policy of barring the poor outright. Second, it would be challenging to credibly mimic the UAE's insulation of natives from fear of migrant competition.

Nonetheless, the UAE provides an example of an almost entirely migrant workforce, relatively poor and less skilled, where the sky hasn't fallen, skilled workers do very well, and the global poor have achieved huge benefits. If the United States and European Union adopted UAE's policies they could eliminate most global poverty while gaining immense revenues to be distributed among the natives. While the UAE system treats migrants poorly compared to natives, it treats potential migrants far better on average than any rich liberal democracy, by allowing them to migrate at all and greatly improve their standard of living. So immense guest worker programs can work as a matter of policy, if not politics. I'd also note that while unlimited-scale guest worker programs are currently far from mainstream, Republican primary candidate Newt Gingrich and the Krieble Foundation have backed such a plan, the "Red Card."

Similar examples exist in other oil states
I focused on the UAE because I happened to have more relevant data, and because of the increasing diversification of the economy from oil, especially in Dubai. However, I might instead have focused on Qatar, which has similarly extreme demographics:

The 2010 census recorded the total population at 1,699,435.[2] In January 2013, the Qatar Statistics Authority estimated the country's population at 1,903,447, of which 1,405,164 were males and 498,283 females.[107] At the time of the first census, held in 1970, the population was 111,133.[106] The population has tripled in the decade to 2011, up from just over 600,000 people in 2001, leaving Qatari nationals as less than 15% of the total population.[108] The influx of male labourers has skewed the gender balance, and women are now just one-quarter of the population.[108]Non-Arabs make up the majority of Qatar's population, and government statistics refer to them as non-Qatari. As of 2013, the four largest ethnic groups are Arab 40%, Indian 18%, Pakistani 18%, and Iranian 10%.[109] Of the remaining 14%, the most prevalent ethnicities are Nepali, Filipino, and Sri Lankan; however, exact percentages are unavailable.[110]
Qatar also enjoys even higher GDP per capita, estimated at some $100,000 in 2011, although this is driven by a much larger per capita supply of fossil fuels, which make up more than 60% of GDP. Wages for foreign workers remain low by developed country standards, but appear better than in the UAE (as wages are dramatically higher in oil-rich Norway than neighboring Sweden).

The success of the extreme migrant worker model in multiple countries, albeit neighbors with shared cultural background, is a point in favor of its robustness or potential for further extension.

No comments: