With rigorous economic research and practical policy solutions, we focus on the issues and institutions that are critical to global development. Explore our core themes and topics to learn more about our work.
In timely and incisive analysis, our experts parse the latest development news and devise practical solutions to new and emerging challenges. Our events convene the top thinkers and doers in global development.
CGD’s work in technology and development focuses on the macroeconomic implications of technology change as well as technological applications for specific development challenges.
Technological advances are a driving force for development. But policy choices determine who benefits. CGD focuses on three key questions around innovation, growth, and inequality: How can governments use existing technologies to deliver services more effectively to citizens? How can international institutions help create and spread new technologies to tackle shared problems like climate change and pandemics? And how can policymakers ensure advances in artificial intelligence, automation, and communications bring shared benefits and not greater global inequality?
Earlier this month, the first analysis of countries’ progress towards attaining the health-related Sustainable Development Goals (SDGs) was published in the Lancet. The Institute for Health Metrics and Evaluation (IHME) used Global Burden of Disease Data (GBD 2016) to create an index for 37 (out of 50) health-related SDG indicators between 1990–2016, for a total of 188 countries. Based on the pace of change recorded over the past 25 years or so, the researchers then projected the indicators to 2030. The punchline: if past is prologue, the median number of SDG targets attained in 2030 will be five of the 24 defined targets currently measured. Not very inspiring.
This is an enormous and laudable effort. A wealth of information is now in the public domain, standardised and comparable between countries, and measurable against SDG attainment, past and forecast. It is clearly better than the UN’s usual glossy pamphlet, and it signals something important—under business-as-usual arrangements, not much of the SDG agenda in health will get done.
Given that scenario, GBD past and future only gets us part of the way towards our goal: using regular measurement and data to drive more and better progress, policy, and spending by governments, donors, employers, as well as healthcare and product firms and investors.
GBD helps us understand the problem we are trying to solve in macro terms—lots of preventable deaths, or inadequate prevention and management of cardiovascular disease. And GBD in the US has truly been a wake-up call—the finding that life expectancy was going down in many counties, and presenting the data in a way that attracted the attention of leaders, may eventually trigger policy action. And as self-help programs always tell us: recognizing you’ve got a problem is the first step towards a solution.
But to use GBD to effect change to business-as-usual what we need is the kind of hyperlocal data, analysis and relationships that public and private payers routinely use in developed economies to inform their own spending and make their investment cases to their countries’ Treasuries or insurers. This critical missing link can help make the causal connection between investment in health and better performance in attaining the SDGs, including Universal Healthcare Coverage (UHC).
Some key (and some surprising) findings
The researchers namecheck Turkey, Rwanda, China, Cambodia, Laos, and Equatorial Guinea as having made significant progress towards UHC (to measure this they devised a kind of meta-index combining “risk-standardised mortality rates from 32 causes from which death should not occur in the presence of high-quality health care with estimates of nine types of intervention coverage for infectious diseases and maternal and child health outcomes”). Lesotho and the Central African Republic are singled out as not making much progress at all towards UHC during that same period. Incidentally, Equatorial Guinea spends much less of its public budget on health than the African average, less than half of the Abuja target of 15% and roughly half of what Lesotho does as a proportion of its public expenditure, suggesting more money alone is not a predictor of success, but more on this later (see Fig 2 here). All in all, they find that whilst good progress is forecast for a majority of countries towards reducing under 5 and neonatal mortality and maternal mortality, other targets such as childhood obesity, TB and road traffic accidents are likely to have been met by fewer than 5 percent of countries come 2030.
A new UHC index is presented as well as violence and vaccine coverage indicators (and more work is promised by the researchers on the latter). And there is more to look forward to in GBD 2017, including health worker density and distribution, “proportion of people who feel safe walking alone around the area where they live”, a refined version of financial protection using household catastrophic spending, and coverage of treatment for substance abuse and sexual violence by non-intimate partners. Most of the above indicators, however, hinge on data availability beyond Europe and North America.
What is most exciting, perhaps, about future work, is the researchers’ commitment to attempt to attribute likelihood of SDG achievement to targeted investment for meeting the SDGs, through tackling underlying risk factors, expanding coverage to existing or new interventions we know work, or increasing development assistance in some way. It would be interesting to see how such attribution could be empirically backed through global level analyses, given the myriad of policies, technologies and fluctuating funding streams which affect healthcare systems’ outcomes and spending. Indeed, most economists have not found a good way to cope with the endogeneity problem in these sorts of analyses (a good attempt using UK programme budgeting data can be found here).
GBD is highly visible in countries too
It is unclear what the impact of GBD analyses has been at country level, but the visibility of the work in the press is impressive. Hours from the publication of the latest wave of findings, in the national and regional press, India lamented the fact it lags so far behind countries with similar characteristics such as its fellow BRICs (see here and here), whilst Singapore celebrated securing the top slot. A 2010 edition of GBD, comparing countries on their performance in public health, triggered a rare policy response by the Secretary of State in the UK, who called on payers of services to better incentivise tackling risk factors for common NCDs, in particular cardiovascular disease through national health checks, better data collection and care integration (see here). Almost five years on, with unprecedented budgetary pressures on the NHS, and most of the agencies supposed to lead on these policies renamed or abolished, there is no follow up on the impact of the measures.
What will it take to use GBD and other data to change business-as-usual SDG trajectories?
Were such analyses able to link investment scenarios (e.g. more of a certain kind of aid, or better even, specific technology adoption and scale up decisions on the part of national health payers) to progress towards SDGs, as suggested by the researchers, then a stronger case could perhaps be made to national treasuries for more and ideally better targeted spending on health. To have local credibility, such work cannot rely solely on running regressions using country panel data. After all, we know that more money for health does not lead to better health outcomes (as shown in Figures 12 and 13 of this WHO report).
GBD rankings can serve as a useful source of information for prioritising those broader diseases and conditions warranting further attention in terms of interventions in the form of more funding; better targeted, outcomes-linked funding (as the UK minister urged his commissioners based on GBD 2010 data); and technology adoption or scale up including population level and behavioural interventions for tackling underlying risk factors. But when it comes to informing specific investment cases within these broader priorities, GBD data are not enough to allow consideration of trade-offs and of opportunity costs of alternative investment choices addressing the same problem. And this is precisely what is needed both in order meaningfully to attribute improvements in performance to investments in technologies or disease areas and to allow domestic payers and national treasuries to make confident allocation decisions including directing more domestic funds towards their own healthcare spending.
Germans have given Chancellor Angela Merkel a fourth term as chancellor, but once again without a parliamentary majority. It seems likely that Merkel will now try to negotiate a black-green-yellow “Jamaica coalition” (referring to the parties’ colors) with the Greens and the pro-business Liberals replacing the Social Democrats as coalition partners. Despite the gain in vote for nationalists, our analysis suggests the Jamaica coalition could actually strengthen Germany’s role in accelerating global development, as well as benefitting Germany.
In this blog, we look at the what the Jamaica coalition means using the framework of our Commitment to Development Index—which ranks rich countries on aid, migration, technology, environment, trade, finance, and security.
Germany’s starting point on Commitment to Development
Overall, Germany ranked fifth (out of 27 countries that we assess) and first on migration, largely because it has accepted so many refugees in recent years. We counted migrants as “1” when they came from the poorest country (Democratic Republic of Congo) and “0” when coming from the richest country (Norway). This method quantified that Germany lifted the equivalent of “880,000 poverty weighted migrants” out of extreme poverty last year! But a ratio of one new migrant for every 92 Germans, contributed to the rise of the far right nationalists (AfD) who have become the third largest party in parliament. Regardless of the election results, mounting public pressure will reduce migration. But a poll of economists thinks the Jamaica coalition is actually more migration-friendly than a continuation of the previous grand coalition would have been.
On aid, Germany met the international commitment of 0.7 percent of national income (GNI) on aid (overseas development assistance) for the first time in 2016. This included high expenditure on hosting refugees—but to maintain 0.7 when fewer refugees arrive, overseas development assistance would have to ramp up quickly.
On environmental policies, high emissions per capita mean Germany might not meet the Paris agreement commitment to reduce emissions by 40 percent by 2020. The global poor will suffer the consequences: climate change might push 100 million people back into poverty by 2030. This is partly due to Germany’s poor policy choices, like burning and subsidising fossil fuels. Both the Greens and Liberals want to phase out these subsidies.
On technology more widely, there has been an increase in overall R&D spending to 0.88 per cent of GDP, but this is still lower than in many other countries. Spending more to create new technologies like mobile phones or biometric IDs can transform development and is a perfect example of investing in global public goods. All major parties want to increase R&D spending to 3.5 per cent of GDP by 2025—a “Jamaica coalition” will not change anything significantly here but this is a positive direction for development.
Germany’s trade policies have a significant impact on developing countries. Free trade agreements such as the EU’s “everything but arms” initiative give poor countries tariff-free access and have the potential to dramatically reduce poverty. For instance, a recent natural experiment suggests trade deals such as these can lower infant mortality by about 9 per cent.
On security policy, Germany has been criticized by the US for failing to spend 2 per cent of GDP on defence. This figure includes spending on UN peacekeeping, for which Germany spends only 0.03 per cent of GDP—less than the OECD average, and this at a time when the UN peacekeeping budget is facing deep cuts. This is a matter of real concern because security and development are closely interlinked—for instance, one study suggests that civil wars decrease GDP per capita by 17.5 percent. Merkel’s conservatives want to double defence spending to reach 2 percent of GDP by 2024. The Liberals also want to increase defence spending, unlike the Greens, who want to specifically focus on increasing support for UN peacekeeping.
Overall then, taking the policy commitments of the Liberals and Greens and adding them to Merkel’s conservative bloc in a “Jamaica coalition” could bode well both for Germany, and development beyond aid.
Under the international regulatory framework for anti-money laundering and countering the financing of terrorism (AML/CFT), banks are assigned significant responsibilities for detecting and preventing illicit financial flows. These responsibilities include performing due diligence on their customers, monitoring accounts and transactions for suspicious activity, and reporting suspicious activities to the government.
The “de-risking” problem
In recent years, regulators have raised their expectations for what counts as adequate AML/CFT compliance. At the same time, they have cracked down on institutions that have fallen short. While arguably necessary, this more stringent enforcement has produced some unintended side effects. In particular, it has put pressure on banks’ ability and willingness to deliver certain types of services, notably correspondent banking services.
Correspondent banking—the provision of financial services by one bank (the correspondent bank) to another bank (the respondent bank)—is vulnerable to illicit finance abuse. A correspondent bank generally does not have a direct relationship with the respondent bank’s customers. Often, the only information it has access to is the originator and beneficiary information contained in the payments messages themselves. Therefore, it can be a challenge for the correspondent bank to properly assess the illicit finance risk that such transactions pose. While regulators have clarified that as a general rule, banks are not expected to know their customers’ customers (KYCC), many correspondent banks nonetheless find these types of risk difficult to manage in a cost-effective way. In addition, correspondent banking has traditionally been a high-volume, low-margin business.
However, there are two new technologies that may help to solve the problem: big data and machine learning.
Big data refers to datasets that are high volume, high velocity, and high variety. These datasets necessitate different hardware, software, and analytical solutions than do traditional data sets. Banks generate enormous volumes of data in a wide variety of formats. Big data systems can help banks to analyze these data in order to identify abuse while preserving relationships with trustworthy customers.
Big data systems can help compliance staff organize and make sense of large volumes of information. Banks’ compliance staff can utilize data from a wide variety of internal and external sources such as transactions metadata, open-source information (such as negative news stories), and government information (such as sanctions lists, arrest warrants, and so on). Traditionally, these data have been siloed and consequently hard to access. Big data systems can reduce the time compliance staff spend searching for and consolidating information. These systems are typically paired with advanced analytics engines, such as machine learning algorithms, which can help identify patterns and relationships in the data that might have otherwise gone undetected by human investigators.
Machine learning is a type of artificial intelligence which enables computers to learn without being explicitly programmed. There are two broad types of machine learning—supervised learning and unsupervised learning. With supervised learning, the machine learning algorithm analyzes a dataset in order to build a model that predicts a pre-defined output. For example, a supervised machine learning algorithm may be presented with transactions labeled “suspicious” and “not suspicious” and instructed to develop a model that best categorizes transactions as one or the other, based on the available data. With unsupervised learning, the machine learning algorithm explores the features of a dataset, looking for patterns and relationships without attempting to predict a pre-defined output.
Machine learning algorithms are already being used to tackle money laundering and the financing of terror. The application of machine learning to customer segmentation and transactions monitoring has the potential to greatly reduce both false negatives and false positives in identifying suspicious activities. Clustering, a type of unsupervised learning, can be used to develop much more sophisticated customer typologies than traditional methods. This can help banks to gain a better understanding of their customers’ financial behavior. In addition, classification algorithms, a type of supervised learning, can be used to identify suspicious transactions. These algorithms can be trained so that, over time, their accuracy improves. Recently, HSBC has partnered with Silicon Valley-based artificial intelligence firm Ayasdi to automate some of its compliance functions. Another American company, QuantaVerse, is helping several large international banks to fight money laundering and other financial crimes.
In early October, we will be discussing the scope of new technologies to address de-risking at RegTech 2017 in Brooklyn, NY. We will also be publishing a report, Technology Innovations to Address De-Risking, in which we will examine this topic in detail. While there have been many positive actions on the regulatory side, our view is that technologies that have emerged over the past few years present very real opportunities to solve the complex problem of de-risking.
Gender data are essential. How else are we going to monitor progress in the wellbeing of women and girls?
Kudos to the Bill and Melinda Gates Foundation for their strong commitment to the importance of metrics and their new user-friendly and visually stunning data publication, Goalkeepers:The Stories Behind the Data 2017, which they have promised to make an annual publication.
This year’s report tracks progress in women’s reproductive health by measuring progress in access to family planning and in reducing maternal mortality, but most of the 18 indicators included in the report are aggregate measures, representing averages of men’s and women’s outcomes. As we are only too aware, there is a dearth of gender data. In the future, we hope that better gender data and better coverage will allow for tracking of women’s work—both paid and unpaid—and their contributions to the economy, their participation in politics and public policy, and their wealth and well-being as individuals, not just as members of households. All are dimensions that go beyond women’s traditional roles as child bearers and mothers, and provide a more well -rounded picture of their lives. They are critical components for measuring progress, but these data are not there right now. Indeed, Goalkeepers found so little data for its one sex-disaggregated indicator—a measure of ownership of agricultural land—that they couldn’t report results.
Below, we provide background on data shortage, and outline three key steps that countries and the international community should take in order to produce better data and close gender gaps.
The holes in gender data
The need for more finely disaggregated data to monitor the progress of excluded groups is one of the statistical challenges of the Sustainable Development Goals, which remind us that we must “leave no one behind” in our efforts to achieve the 2030 development goals. The SDGs also call for disaggregation (“where relevant”) by sex, and it is a glaring irony that data on women and girls, who, after all, constitute the majority of human beings, are absent or infrequently recorded in the national and international databases used to measure the SDGs.
The holes in gender data happen because there are no measures for many events in women’s lives, because the measures that exist are bad (read biased) or not well defined, or because they are measured infrequently or not at all. Of the 232 indicators included in the SDG framework, 53 refer explicitly to women or girls or specify disaggregation by sex. Among these, only 15 are well defined and generally available (classified as Tier I).
Significant investments are needed in gender data—both in producing more and better data, and in using these data for policymaking. This is a tall order, and one that may need quite a bit of heavy lifting, especially in poor countries with few resources for statistics and data. But there are easier, less resource-intensive initial steps to be taken.
Ready to Measure
Where to begin? Some indicators are—or should be—“ready to measure.” Data2X, an initiative of the UN Foundation, along with Open Data Watch last year proposed a list of 20 such gender indicators. Sixteen of these indicators are the same or similar to SDG indicators; four more broaden the traditional view of women’s roles by adding measures of economic status measures. See the original Ready to Measure report for documentation on the sources of these indicators. The recently released Ready to Measure Phase II report compiles the data from international databases and household surveys. This database and all accompanying metadata are available online.
While all of these indicators are, indeed, ready to measure, large gaps remain. The table below (cited in R2M-II) summarizes the availability of data for the 16 R2M indicators found in international databases.
Table 1. Data Availability for 16 Ready to Measure Indicators, 2011-2015
Ready to Measure IndicatorNumber of countries with at least one observation 2011–2015Countries with 3+ obs. 2011–2015
1. Ratio of women to men living below the international and national poverty lines
2. Prevalence of stunting in children under 5 years of age by sex
3. Prevalence of anemia in women of reproductive age*
4. Maternal deaths per 100,000 live births*
5. Under-five mortality rate* by sex
6. HIV Prevalence* by sex
7. Adolescent birth rate*
8. Contraceptive prevalence, modern methods
10. Individuals using the Internet by sex
11. Women aged 20-24 who were married or in a union before age 18
12. Proportion of seats held by women in national parliaments
15. Employment to population ratio* by sex
16. Employed persons who are own-account workers* by sex
17. Non-agricultural wage employment* by sex
18. Women with an account at a financial institution
20. Children under 5 whose births have been registered with a civil authority by sex
*Modeled estimates by international organizations
The four remaining measures in the R2M report, which can be compiled from existing surveys, are:
completion rates of women by school stage
share of women among mobile telephone owners
share of young women who are not in school nor looking for work
growth rate in adult women’s share of household earned income among the bottom 40 percent of the population
The current state of gender data
The effort to gather the R2M indicators has revealed new or confirmed old observations on the state of development data:
Indicators with high coverage rates are usually the product of statistical models using limited direct observations, indirect observations (from recall or sisterhood methods), or covariates to extend or interpolate values from survey data.
Gender data collected through surveys sponsored by bilateral and multilateral agencies in developing countries, such as MICS, DHS, and LSMS, are more likely to be included in international databases. For example, coverage rates for data on anemia, HIV, and contraceptive prevalence are higher for low- and lower-middle-income countries than for upper-middle and high-income countries.
Many microdata sets from surveys are not publicly available, even from those sponsored by international organizations. Therefore, it is not possible to confirm the statistics derived from them or to use them to construct new indicators
Indicators on the use of new technologies have only recently appeared in public databases. For example, since 2014 only a single observation on use of the Internet by women is available for 83 countries
Data collection and publication schedules for most indicators are not known and their future continuity is uncertain.
We have identified at least three steps, all doable if countries and the international community are committed to “leaving no one behind” by 2030:
1) Commit to the R2M indicators and close key gender data gaps
A first next step is to help countries collect full information on the R2M indicators or a similar set on a regular basis. Concurrently, countries need to commit themselves to begin closing gender data gaps in key areas that are most important to them. The international community should support country efforts, particularly for the poorest countries.
2) Parity is not always equality
The Goalkeepers report shows substantial progress in reducing child mortality. The R2M report shows the most recent sex-disaggregated mortality rates and discusses how to read the data. In this case, parity does not mean that girls do better than boys. On the contrary—for biological reasons, infant boys have higher mortality rates and, therefore, the “normal” acceptable ratio in countries with no discrimination against girls is a ratio favorable to girls. A parity ratio of 1 (same number of infant boys and infant girls dying) means that girls are in trouble—dying at higher rates than they should. Research on what sex differentials mean and how they arise needs to go along with any effort to improve the collection of gender data.
3) Measure women’s poverty: A challenge long overdue
Lastly, the international community should invest in identifying, testing, and agreeing on robust measures for a few core gender data applicable globally. One outstanding problem is measuring women’s poverty, separate from household poverty. The often-cited claim that 70 percent of the poor are women is false, and, hopefully, well discredited by now. However, we do not have an adequate response to this very core question.
R2M includes data for Latin America on the proportions of women and men living in poor households (compared to the totals for each sex), and shows that women are overrepresented among those living in poverty in most countries. This gets the closest perhaps to an objective measure, with a big caveat that it does not capture inequalities within the household that most often (and across cultures) benefit male over female household members.
The World Bank, building on its household survey capacity and their work on a multidimensional poverty measure, should lead an international effort involving UNSD and other relevant UN agencies, as well as academics and women’s advocates, to devise robust, doable, and simple measures of women’s poverty.
We are now almost 20 percent of the way toward the SDG target date of 2030, still we lack regular observations on at least 70 percent of the proposed gender indicators. The Millennium Development Goals taught us two lessons: first, setting quantified targets and rigorously monitoring progress can galvanize action; second, databases do not build themselves. If we are to honor the SDG goal of achieving gender equality and empowering women and girls and keep our commitment to leave no one behind, then there must be an equally strong commitment to gathering and publishing the data needed to build effective programs and measure outcomes.
About the Authors
Mayra Buvinic is a senior fellow at both the Center for Global Development and the United Nations Foundation. Previously, she was director for gender and development at the World Bank.
Eric Swanson is director of research at Open Data Watch. Previously he was senior adviser in the Development Data Group of the World Bank.
CGD, in partnership with the World Bank Group, Bill and Melinda Gates Foundation, and Omidyar Network, is delighted to host Nandan Nilekani, the founding chairman of UIDAI (Aadhaar), the unique identification system of India, which has enrolled more than a billion people. Nilekani will speak on “Societal Platforms: A Cambrian Approach to Sustainable Development”—how we can distill principles from the unique architecture of Aadhaar to develop new platforms, like EkStep, that can enable people to access an increasingly wide array of transformative services.
For the world’s middle-income countries, the changes unleashed by automation, digital technologies, and the advent of increasingly more capable AI pose major challenges. They threaten to upend the few tried and tested development strategies.
How well do your country's policies make a positive difference for people in developing nations? That’s the question CGD seeks to answer each year in our Commitment to Development Index (CDI). It’s a ranking of 27 of the world’s richest nations based on seven policy areas: aid, finance, technology, environment, trade, security, and migration.
The team behind the CDI, deputy director of CGD Europe Ian Mitchell and policy analyst Anita Käppeli, join me this week on the CGD podcast to discuss why these rankings matter and how countries stack up.
In first place this year is Denmark, followed by Sweden, Finland, France, and Germany. Greece, Japan, and South Korea rank at the bottom—though South Korea actually ranks first on the technology component.
Among the countries in the middle are the UK, tying with the Netherlands for 7th place, and the US, all the way down at 23rd. In the future, how might these scores be impacted by the changing politics of the two nations?
“On Brexit, there’s real potential for this to affect the CDI score,” Mitchell tells me in the podcast. “The UK will take control of its own migration policy more fully and it will have its own trade policy and it will take control of agricultural policy form the EU. All of those things feature in the Commitment to Development Index.”
As for the the Trump Administration’s America-first approach, Mitchell says, “It’s surely in the interest of countries to see other countries developing to reduce the security risk, to make sure there’s lower risk of disease emerging . . . and the CDI is a framework for prioritizing action on that.”
Overall, Käppeli tells me, the CDI is a reminder to countries that “policy coherence is an issue; that they should not pursue policies in [only] one field—for instance, give a lot of aid, but then close the boarders for products from developing countries.”
“The CDI is holistic,” Mitchell adds, pointing out that the CDI’s focus on policy is “complementary” to the Sustainable Development Goals’ focus on outcomes: “If you think about how we’re going to achieve the SDGs, then looking at the CDI [is] a great way to do that.”
How well do your country's policies make a positive difference for people in developing nations? That’s the question CGD seeks to answer each year in our Commitment to Development Index (CDI). The team behind the CDI, deputy director of CGD Europe Ian Mitchell and policy analyst Anita Käppeli, join me to discuss why these rankings matter, how countries stack up, and how their scores may be impacted by the shifting political environment.