Syndicate content

Stark evidence on the jobs quality-quantity trade-off: Evidence from migration

David McKenzie's picture

“More and better jobs” is a goal for many policymakers around the world (along with part of the title for a recent World Bank South Asia flagship report on employment). How to create “good jobs” is a key question that the next World Development Report is also expected to help answer. Few people are likely to argue against more good jobs, but one of the trickier policy questions is how much of a trade-off there may be between the number of jobs and how good each job is?

In a new working paper with Caroline Theoharides and Dean Yang, I find strong evidence for a jobs quantity-quality trade-off when it comes to migrant jobs. We start by noting that two key facts about international migration are:

1.       The same worker can earn very different wages depending on where they work. See for example my experimental evidence on Tongans, or Orley Ashenfelter’s new work on what McDonalds workers are paid for the same tasks in different countries (he finds wage rates can differ by a factor of ten).

2.       Economic shocks at destination are typically accompanied by large drops in migration numbers. For example, emigration to Ireland from the new EU fell 60% from 2008 to 2009 as Ireland’s economy tanked.

The main message of our paper is that these two facts are intimately related: the high vulnerability of migrant jobs to economic jobs is tied to the large wage gains migration brings. The reason is that various regulations (as well as possibly norms) keep migrant wages higher than market clearing levels, causing adjustment to shocks to occur through quantities rather than labor prices.

We carry out two types of analysis to show this. First, we use a new database on all new work contracts from the Philippines over the 1992-2009 period, which is unique in having information on which countries workers leave for, and the wages they are contracted at. We aggregate this information up to the destination country-year level, and regress the quantity of migrants going to a destination, and the wages paid, on GDP at destination, controlling for time and destination country fixed effects. We are then effectively asking whether Dubai gets more Filipino migrants when its economy is booming relative to other countries in the world, and whether it pays them more when its economy is booming than when it is doing relatively less well.

We find that

·         Migrant numbers are strongly significantly related to GDP shocks at destination: a 1% increase in GDP at destination leads to 1.5% more migrants, a 1% fall leads to 1.5% fewer migrants.

·         Migrant wages have a small and insignificant relationship to GDP shocks at destination: a 1% increase in GDP leads to an insignificant -0.06% change in wages.

We then do several robustness checks to show this is not being driven by changes in the occupational mix of workers migrating at different points in the business cycle, nor in changes in their observable characteristics, and that the pattern of adjustment through wages rather than quantities occurs for both high and low-skilled workers.

Impact evaluation of a minimum wage change

In the second part of the paper we provide further support for the idea that minimum wages bind by examining the impact of a minimum wage change. In 2006 the Philippines introduced a reform which increased the minimum wage that domestic workers (e.g. maids) contracted abroad could earn to US$400 per month. This applied only to domestic workers, and doubled the prevailing wage of $200 per month being paid to Filipino domestic workers in countries like Saudi Arabia, Singapore and the U.A.E.. We carry out difference-in-differences analysis to examine the change in quantities and wages of domestic workers going to destinations where pre-reform wages were below $400, relative to two comparison groups: (a) domestic workers going to destinations which pre-reform were already paying above $400; and (b) Filipino migrants in other occupations going to the destinations where pre-reform domestic worker wages were below $400, but who were not affected by the reform since it applied only to domestic workers.

Using either comparison group, the difference-in-differences analysis finds the minimum wage reform DID result in an increase in the wages domestic workers earned: we estimate a 27-46% increase in the average wage earned; but also resulted in a large (55-57%) fall in the number of Filipino domestic workers being hired. Since the quantity adjustment was even larger than the price adjustment, total wages paid to Filipino domestic workers going to these destinations fell, even though the workers who did go each got paid more.

These types of minimum wages are not just a feature of the international labor market for low-skilled workers. For example, many countries require that high-skilled workers be paid the “prevailing wage” for their occupation (e.g. the H1B program in the U.S. requires this), to prevent local workers getting their wages undercut by migrants. The consequence of this is that the few migrants who do get to migrate experience massive wage gains, but far fewer people from developing countries get to participate in migrant jobs than would be the case if these minimum wages didn’t bind.

Of course wages are only one element of job quality, and it may be that the workers hired at a higher wage are also treated better by their employers and that as a result, the overall welfare effect might still be positive. Yet Lant Pritchett has argued in his excellent book Let their people come that there is also a stark quantity-quality jobs trade-off when it comes to migrant rights, contrasting the migration of programs of countries which let few people in, but give them lots of rights and protections when they arrive, with those of other countries which provide far more job opportunities, but with more limited protections. As a result, policymakers may sometimes have to choose between “more jobs” and “better jobs”, with relatively little empirical evidence to date of where these choices arise and which policies can help manage these tradeoffs.