I think there's strong evidence that there is in fact a far weaker "bench" of substitute capital in the Phillipines. Pre nurse-glut, 58% of nurses passed the registration exam, whereas post nurse-glut that dropped to 38%.
I agree that top talent loss is likely far less of a big deal in the Phillipines but for very different reasons - mainly because only 1 in 8 (or similar) new staff left, which should mean enough strong staff remain, even with the clear drop in average nurse quality.
Not sure if we disagree here. Of course I'd expect the average nurse quality to go down as the workforce increases by 9x. Rather, the claim about weaker substitutes in Nigeria was about explaining whyPhilippines nursing supply might be more price elastic than Nigerian supply. Specifically, since literacy, numeracy and high school graduation rates are likely significantly higher in Philippines than Nigeria, there's a larger share of the population that could plausibly respond to the migration demand shock by acquiring the relevant training.[1]
Phillipines and Nigerian CBAs are likely be wildly different, even just based off the one datapoint that Nigeria has lost tens of thousands of net nurses through emigration while Phillipines gained.
Agreed if we conducted the CBA today. However, as stated in the original comment, we want to be careful about lags here. Even in the Philippines, the migration increase started in 2000 when the US policy changed (Figure 3) and peaked in ~2006. While the enrollment rate in nursing programs did start increasing in 2000 itself (Figure 4, Panel A), the increase in the nurse graduation rate (i.e., the trained workforce) only started in 2004 (Figure 4, Panel B), and only hit it's peak in 2010, 10 years after the migration began. If we were looking at the change in Philippines' nurse workforce from 2000-2004, I think we might've concluded that they'd lost nurses and that the migration was a net-loss for them. Now, as we've discussed, there are reasons to believe that Nigerian nursing supply may not be as elastic as Philippines nursing supply, but I just wanted to emphasize that the current net-loss of nurses in Nigeria doesn't yet give us strong evidence that the CBAs will be wildly different.
Agreed that we should consider the broader set of costs/benefits you list! The top talent loss cost could be an especially a big deal in Nigeria, where I'd expect a weaker "bench" of substitute human capital than the Philippines (both for new potential nurses and for those who would train the new nurses/found new private colleges).
My (unquantified) view is that the CBA still looks pretty one-sided in the Philippines context, but I'd love to see what a formal modeling exercise produces (and if the conclusions are different for Nigeria or other Sub Saharan African countries).
TLDR: The immediate drop in Nigerian nursing workforce isn't indicative of the long-run effect of rich country nursing visas. Prior evidence suggests that the increasing in nursing supply will be lagged, but big enough swamp the effect of outmigration.
Hi Nick, thanks for this! Point well taken on the misleading numbers in the CGDev article. I share your skepticism that the UK policy could have caused an uptick in the Nigerian nurse population within 2 years of its introduction. In the first couple of years of the policy, the outmigration is plausibly a net-loss for Nigeria before we consider the effects of remittances.
That said, the crux of the cost-benefit calculation (bracketing remittances for now) is the effect of UK nurse recruitment on the Nigerian nurse workforce over the longer run. In particular, we have evidence (open access) from the Philippines[1] showing that expansions in US nursing visas caused an additonal 9 non-migrant nurses to be licensed for each nurse who migrated to the US. We can think of the big surge in nursing supply as being driven by the hopes of winning the "migration lottery" to the US, with most lottery entrants losing.
Crucially, the increase in nurse licensing only showed up after at least a 4-year lag (the time it took to complete a nursing degree). This increase involved both increasing enrollment at existing institutions, but also an increase in the number of nurse-training programs, which would show up with even further lags. If the Nigerian case mirrors that of the Philippines, we wouldn't have expected the increase to be visible yet. I think you made a similar point when you said :
There would be a long lag time (3-5 years of training) before we would see any response to a new policy which took in more nurses from another country.
The Philippines nursing example is the closest analog we have in the research, but we see a similar mechanism at work in the boom in engineering/CS skills in India in response to changes in the US H1-B cap (Khanna and Morales, 2021).[2]
Now, there may be institutional differences between the Filipino/Indian and Nigerian cases that cause there not to be similar effects. For example, there could be bottlenecks to expanding existing nursing colleges or setting up new ones. Alternatively, all the increase in nursing supply could be absorbed by other countries. I'm not familiar enough with the country to have a view on this, so I'd be curious if you think the Nigeria is particularly well or badly suited to expand nursing supply.
Thanks a lot for the detailed response Dean! The details on the motivational help that nurses provide make it clear that there's much less of an arbitrage opportunity/free lunch than I'd hoped (as is often the case with mere info).
Thanks again for all the great work (including on OD, I learned a lot from where India goes)!
Of the two components of KMC, breastfeeding assistance seems to me much more bottlenecked by nurses than skin to skin contact. That is, while breastfeeding assistance might need a nurse to provide bespoke information to each mother in the moment, skin-to-skin contact seems less individually specific and an easier piece of advice to share impersonally and by non-experts.
Two questions about this:
Is the distinction I drew above directionally correct, or does skin to skin contact require as much in person expert attention as breastfeeding assistance?
If the distinction is directionally correct, might it be possible to scale the provision of the skin to skin contact advice for much cheaper than it would take to hire a lot more nurses (some kind of information provision/belief change intervention in econ jargon)?
This could look like some kind of door-to-door campaign by community health workers, or a video version of text-message reminders for vaccines (though an internet requirement might screen out some of the households we care most about).
Are the two parts of KMC strong complements in a way that would make the provision of just one of them much less effective?
1. What do you think of the Coasian solution to Punjab/Haryana's paddy burning where Delhi (mostly) pays for the machines required to prevent the burning? Per the estimates Shruti cites here, the benefits would be more than 10x the cost. Are the main barriers to this political (it would be a bad look for the Delhi govt to pay Punjabi farmers), or something else?
2. The Delhi metro with an average daily ridership of over 5.5 million trips likely prevents a massive amount of tailpipe pollution. A larger, better run bus fleet would likely have a similar effect. Even if this is distinctly not a neglected area, influencing city-wide policy has potentially massive scale. Are there any cost-effective opportunities for philanthropists speed to speed up the development/improve the efficacy of public transport systems? Maybe in smaller cities with less expertise?
Here's the website, and here's their 501(c)3 approval letter. They seem to have a page describing their maternal health research, but not their involvement on the ground with the KMC program.
EDIT: the maternal health page has the following paragraph:
"In recent months, r.i.c.e. researchers are pursuing a project to promote Breastfeeding and Care Practices for Newborn and Low Birthweight Babies in Uttar Pradesh. The program is based on our earlier work. The program will test messages for breastfeeding and newborn care counseling to family members of newborns. It also intends to identify the existing systemic gaps in an adequate newborn care in hospitals and at home, and how can practitioners and caregivers overcome those."
Not sure if we disagree here. Of course I'd expect the average nurse quality to go down as the workforce increases by 9x. Rather, the claim about weaker substitutes in Nigeria was about explaining why Philippines nursing supply might be more price elastic than Nigerian supply. Specifically, since literacy, numeracy and high school graduation rates are likely significantly higher in Philippines than Nigeria, there's a larger share of the population that could plausibly respond to the migration demand shock by acquiring the relevant training.[1]
Agreed if we conducted the CBA today. However, as stated in the original comment, we want to be careful about lags here. Even in the Philippines, the migration increase started in 2000 when the US policy changed (Figure 3) and peaked in ~2006. While the enrollment rate in nursing programs did start increasing in 2000 itself (Figure 4, Panel A), the increase in the nurse graduation rate (i.e., the trained workforce) only started in 2004 (Figure 4, Panel B), and only hit it's peak in 2010, 10 years after the migration began. If we were looking at the change in Philippines' nurse workforce from 2000-2004, I think we might've concluded that they'd lost nurses and that the migration was a net-loss for them. Now, as we've discussed, there are reasons to believe that Nigerian nursing supply may not be as elastic as Philippines nursing supply, but I just wanted to emphasize that the current net-loss of nurses in Nigeria doesn't yet give us strong evidence that the CBAs will be wildly different.
Implicit here is that basic numeracy, literacy and high school graduation are pre-requisites for acquiring nurse training.
Agreed that we should consider the broader set of costs/benefits you list! The top talent loss cost could be an especially a big deal in Nigeria, where I'd expect a weaker "bench" of substitute human capital than the Philippines (both for new potential nurses and for those who would train the new nurses/found new private colleges).
My (unquantified) view is that the CBA still looks pretty one-sided in the Philippines context, but I'd love to see what a formal modeling exercise produces (and if the conclusions are different for Nigeria or other Sub Saharan African countries).
TLDR: The immediate drop in Nigerian nursing workforce isn't indicative of the long-run effect of rich country nursing visas. Prior evidence suggests that the increasing in nursing supply will be lagged, but big enough swamp the effect of outmigration.
Hi Nick, thanks for this! Point well taken on the misleading numbers in the CGDev article. I share your skepticism that the UK policy could have caused an uptick in the Nigerian nurse population within 2 years of its introduction. In the first couple of years of the policy, the outmigration is plausibly a net-loss for Nigeria before we consider the effects of remittances.
That said, the crux of the cost-benefit calculation (bracketing remittances for now) is the effect of UK nurse recruitment on the Nigerian nurse workforce over the longer run. In particular, we have evidence (open access) from the Philippines[1] showing that expansions in US nursing visas caused an additonal 9 non-migrant nurses to be licensed for each nurse who migrated to the US. We can think of the big surge in nursing supply as being driven by the hopes of winning the "migration lottery" to the US, with most lottery entrants losing.
Crucially, the increase in nurse licensing only showed up after at least a 4-year lag (the time it took to complete a nursing degree). This increase involved both increasing enrollment at existing institutions, but also an increase in the number of nurse-training programs, which would show up with even further lags. If the Nigerian case mirrors that of the Philippines, we wouldn't have expected the increase to be visible yet. I think you made a similar point when you said :
The Philippines nursing example is the closest analog we have in the research, but we see a similar mechanism at work in the boom in engineering/CS skills in India in response to changes in the US H1-B cap (Khanna and Morales, 2021).[2]
Now, there may be institutional differences between the Filipino/Indian and Nigerian cases that cause there not to be similar effects. For example, there could be bottlenecks to expanding existing nursing colleges or setting up new ones. Alternatively, all the increase in nursing supply could be absorbed by other countries. I'm not familiar enough with the country to have a view on this, so I'd be curious if you think the Nigeria is particularly well or badly suited to expand nursing supply.
The CGDev authors cite this research, and I'm guessing this drives their views more than the spotty WB/WHO data.
And here's another paper with similar findings in the context of Fiji.
Thanks a lot for the detailed response Dean! The details on the motivational help that nurses provide make it clear that there's much less of an arbitrage opportunity/free lunch than I'd hoped (as is often the case with mere info).
Thanks again for all the great work (including on OD, I learned a lot from where India goes)!
Hi Dean!
Of the two components of KMC, breastfeeding assistance seems to me much more bottlenecked by nurses than skin to skin contact. That is, while breastfeeding assistance might need a nurse to provide bespoke information to each mother in the moment, skin-to-skin contact seems less individually specific and an easier piece of advice to share impersonally and by non-experts.
Two questions about this:
Two questions:
1. What do you think of the Coasian solution to Punjab/Haryana's paddy burning where Delhi (mostly) pays for the machines required to prevent the burning? Per the estimates Shruti cites here, the benefits would be more than 10x the cost. Are the main barriers to this political (it would be a bad look for the Delhi govt to pay Punjabi farmers), or something else?
2. The Delhi metro with an average daily ridership of over 5.5 million trips likely prevents a massive amount of tailpipe pollution. A larger, better run bus fleet would likely have a similar effect. Even if this is distinctly not a neglected area, influencing city-wide policy has potentially massive scale. Are there any cost-effective opportunities for philanthropists speed to speed up the development/improve the efficacy of public transport systems? Maybe in smaller cities with less expertise?
Here's the website, and here's their 501(c)3 approval letter. They seem to have a page describing their maternal health research,
but not their involvement on the ground with the KMC program.EDIT: the maternal health page has the following paragraph:
"In recent months, r.i.c.e. researchers are pursuing a project to promote Breastfeeding and Care Practices for Newborn and Low Birthweight Babies in Uttar Pradesh. The program is based on our earlier work. The program will test messages for breastfeeding and newborn care counseling to family members of newborns. It also intends to identify the existing systemic gaps in an adequate newborn care in hospitals and at home, and how can practitioners and caregivers overcome those."