How to Save Democracy From Technology
Ending Big Tech’s Information Monopoly
These days, people for the most part believe that governments should try to promote the general welfare of the populations they serve. The disagreements come over how to do that—what goals to focus on, what policies to adopt, and so on. These questions are usually approached through broad intellectual frameworks, such as political ideology or religion, and much time is spent debating the finer points of various doctrines. Often overlooked, however, is a simple and easy way to make lives better: use routine cost-benefit analysis to compare the expected returns from alternative policies and then choose the more effective ones.
Effectiveness sounds dull. But what if an extra dollar or rupee in a budget could feed ten people instead of one? Or if $100,000 of international aid spending could be tweaked so it would save ten times as many lives? When the stakes are this high, efficiency in spending becomes a moral imperative. Moreover, unlike debates over ideology or religion, debates over efficiency can actually get somewhere, because there is a straightforward mechanism for resolving them: compare the predictable costs and benefits of different courses of action and see which yields more bang for the buck.
Surely, this is just common sense, one might say, and governments must do it all time. Maybe they should, but in the real world, they rarely do—partly because this analysis involves a lot of work, but mostly because the results can be inconvenient, showing that a preferred policy is inefficient or even that elements of existing government bureaucracy may be unnecessary. Unsurprisingly, nobody wants to be the superfluous official—whether in a government, an international organization, a nongovernmental organization, or even a private philanthropy.
This means that decisions are affected by other factors. One town in rural Virginia, for example, holds an annual fair to support local charities. Each year, an animal-rescue organization brings a bald eagle to its booth as a prop, and each year, it receives more donations than other groups—which have a harder time using stagecraft to promote the virtues of, say, being a foster parent or working with at-risk youth. This sort of thing happens everywhere, and everybody knows it. Marketing and politics shape policy selection at least as much as technical merit, and the public suffers as a result.
The difference can be considerable: the philosopher Toby Ord analyzed 108 health interventions from the Disease Control Priorities Project, identifying the number of additional years of healthy life gained from spending the same amount on each. The most effective interventions were at least hundreds of times as powerful as the least. Moving $50 million from the bottom to the top of the list could save 1,000 lives instead of one. Likewise, extensive research on the United Nation’s Sustainable Development Goals reveals a similar pattern: the most efficient interventions aren’t just good; they’re remarkably better than the middle-of-the-road ones—and it’s likely that such a pattern holds true for spending by governments and development agencies in any country.
But just because inefficiency is common doesn’t mean it’s inevitable. Governments and other service providers can do better, even within their existing budgets, simply by disciplining themselves to embrace best practices across all their operations and by shifting time, effort, and resources from inefficient programs to efficient ones. Recently, my think tank, the Copenhagen Consensus Center, worked with the government of Bangladesh, as well as an extensive list of public- and private-sector organizations and Bangladeshi media, to find out how to improve the efficiency of development efforts in the country, and the lessons we learned in the process are applicable to other nations trying to improve their performance.
THE BANGLADESHI EXPERIMENT
The Bangladesh Priorities project has been funded by the C&A Foundation, an affiliate of the Dutch fashion company C&A, with help from the Swedish International Development Cooperaton Agency and the Danish embassy in Dhaka. We worked with all the major players in Bangladesh to assess what kinds of spending (both for the government’s $30 billion annual budget and for the $3 billion in development aid given by outside organizations) would do the most good for the country. The results were startling: they showed that major gains in national well-being could be achieved simply by rearranging budgets to favor policies with high returns on investment.
We began with the country’s latest five-year plan, which shapes most conversations about national development. Partnering with BRAC, the world’s largest nongovernmental development organization, we took each of the plan’s 20 topic areas, from gender equality to urbanization, and noted all the associated policies. Then we invited several hundred thought leaders from government, the academy, nongovernmental organizations, donors, and the private sector to add their own recommendations. This ultimately yielded 1,000 proposals, about half overlapping with those in the plan, on topics as varied as infrastructure, tax reform, public health, and more.
What if an extra dollar or rupee in a budget could feed ten people instead of one?
In 20 roundtables, we asked Bangladeshi experts to look at all the proposals and rank them—specifically identifying which ideas had the most potential or were likely to be politically popular, and also which had enough empirical data available to make a thorough examination possible. That whittled the list down to 76 proposals. Then, 30 teams of local and foreign economists estimated the costs and benefits of all 76 proposals. Most of the costs were monetary, but the benefits included several noneconomic ones as well.
Take a proposal to promote wetland conservation in the Sundarbans, a vast mangrove forest on the coast of the Bay of Bengal. It would help address climate change, enhance biodiversity, and create opportunities for fishing and tourism. The projected benefits added up to almost $4 billion, for a cost of $1.4 billion, generating a predicted nearly $3 of benefits for every $1 spent.
Or take an early childhood education program that would help kids overcome setbacks from stunting. Stunting is caused by poor nutrition or repeated early infections, and its effects can last for many decades, with afflicted children earning less than their peers ever after. The program would bring specialists to work with stunted children and their parents to improve the children’s development skills, and the evidence shows that such efforts can boost the children’s lifetime earnings by 25 percent, completely eliminating the stunting effect. In Bangladesh, such a program would cost about $160 per child and increase each child’s future earnings by $2,884. So every $1 invested would bring an $18 return.
The Bangladesh Priorities project has generated more than 1,150 pages of peer-reviewed studies, available for free online and to be published in a two-volume book. Changes in spending require public support, so we published more than 40 articles on the research results in the largest Bengali and English newspapers, with a combined readership of more than ten million people. To help spread the message even more, we combined all the results in one chart, showing the bang for the buck of all 76 policies evaluated: the longer the line, the greater the multiplier effect.
TB OR NOT TB
To compare the policies fairly against one another, we had to translate all their impacts into a single ultimate scale of value, using common assumptions and calculations. For example, across all the studies, we used a standard figure for the economic value of a year of life and standard discount rates to calculate the value of future costs and benefits. Even so, the figures can obviously be only rough estimates, because of the inherent uncertainties involved in many of the projections.
Moreover, efficiency is not the only important value; governments need to consider other factors as well, such as justice, equality, and political sustainability. So we built in additional rounds of discussion in which the calculations and rankings could be challenged, including having a special panel of top economists scrutinize all the findings, make sure all variables were considered, and adjust the rankings as appropriate.
For example, microfinance programs have a relatively low economic return, but they promote equality and often benefit the poorest of the poor and so have more going for them than one might assume at first glance. A similar effect is true for family subsidies designed to prevent child marriage. The educational benefits of a delay in marriage are well established, but the broader social and health benefits are challenging to study. A simple economic cost-benefit analysis underestimates these and overlooks the moral benefits of deterring child marriage.
Each $1 spent on further trade liberalization would bring the country $10 in benefits.
In the end, the project’s most important finding related to the treatment of tuberculosis. It turns out that one in every 11 deaths in Bangladesh is caused by tuberculosis and that virtually all of those deaths are preventable. Today, proven treatments can cure TB patients for about $100 each. And yet nine Bangladeshis die from the disease every hour nevertheless. Why? Because only half of those who need treatment get it, thanks to the limited reach of Bangladesh’s health-care system, popular ignorance about how the disease is transmitted, and the shame and stigma associated with diagnosis.
Treating all TB patients in Bangladesh appropriately would not be easy or free. Identifying and treating people with the disease would require extensive outreach initiatives, costing $402 per death avoided. The value of the average life gained from those efforts, however, would be $8,503. So every $1 spent on treating TB—one of the country’s crucial problems—would produce an impressive $21 of benefits.
But wait, there’s more! Most of the benefits would go to the poorest of the poor, and curtailing TB would prevent all the disruption and tragedies stemming from the death of adults in their prime. Putting everything together, therefore, the expert panel decided that increasing expenditures on TB treatment was the single most effective way to improve life in Bangladesh. “For many years, [it has] been difficult to get enough attention and funding for TB,” according to BRAC’s Md. Akramul Islam. He has found that the results of our study are increasing the visibility of and funding for this neglected disease.
Perhaps more surprising, the second-biggest finding concerned expanding e-procurement. Reforming government purchasing procedures is about as unsexy a topic as one could imagine, but it turns out that it is extremely important in practice, particularly for a developing country such as Bangladesh. The government there spends more than $9 billion on procurement annually, on everything from roads to office buildings to pencils. There are opportunities for corruption at every step along the way: contractors have to hand in their proposals in person, and companies with political connections have been known to hire goons to physically block competitors from submitting bids. This leads to higher prices and sometimes subpar output.
Our research showed that switching to a digital procurement system would increase competition, reduce corruption (by an estimated 12 percent), and save money (up to $700 million annually). The practical requirements would involve little more than buying computers and educating staff—and for each $1 spent on such efforts, the return would be a whopping $663.
Doing this sort of exercise properly enables policymakers to see whether familiar nostrums live up to their billing. Bangladesh, and especially its garment industry, has benefited from trade liberalization, for example. But by how much? Now we can say that each $1 spent on further trade liberalization would bring the country $10 in benefits. Bangladesh has battled naturally occurring arsenic in its groundwater for decades, to cite another example. Now we can say that every extra $1 invested in fixing the problem for the worst-affected households would return $17 worth of benefits.
Even more important, this sort of process enables previously obscure ideas to get the audition and acceptance they deserve. Take a policy to counter malnutrition by providing small children with micronutrient supplements, including iodized salt, vitamin A, and zinc. Delivering the supplements would cost roughly $125 per child in need—in return for which the child would be healthier, do better in school, and have higher lifetime earnings. The result? For every $1 spent, the supplements program would generate $19 in benefits.
Or take retrofitting kilns. More than a thousand kilns across Dhaka manufacture four billion bricks each year, emitting so much pollution along the way that the city’s air quality is often 16 times as bad as international standards. This air pollution kills 2,000 people each year. Upgrading the kilns with improved technology would make them burn more cleanly and efficiently and decrease fuel consumption by a fifth. And every $1 spent would yield $8 in value.
This kind of exercise also enables policymakers to tell which celebrated programs aren’t particularly effective, especially on a comparative basis. Household solar projects, for example, are darlings of the development community, but analysis by the economist A.K. Enamul Haque—who also co-wrote the recent World Bank report on solar energy in Bangladesh—showed that the panels produce only $1.80 in benefits for every $1 spent on them. Why such a poor showing? Because solar panels are relatively expensive and deliver fairly little energy, available for only a few hours at night.
Haque noted that most rich Bangladeshis use diesel generators rather than solar panels to provide alternative electricity sources during power cuts. So he decided to test whether it made sense for poorer Bangladeshis to emulate their richer neighbors. And sure enough, if five households chipped in to split the cost of a diesel generator, each $1 spent would yield $25 of benefits—even after accounting for the harm of higher carbon dioxide emissions. People who care about eliminating energy poverty should follow the numbers.
Similarly, Bangladesh is famous for its experiments in microfinance. But extensive research in many countries over long periods has shown that microfinance is not a particularly powerful intervention, as these things go. It carries a significant initial cost and produces modest benefits that taper off after a few years; all told, it yields $2 in benefits for every $1 spent. That’s better than nothing, but much less efficient than many other ways of using the same aid dollars.
Surprisingly, one program very popular in some development circles—unconditional cash transfers—turned out to be one of the least effective, according to the economists. These programs give a one-time cash amount to ultra-poor recipients, often microentrepreneurs, without conditions on how the money can be used. Multiple randomized controlled trials—the gold standard in estimates of effect—showed little direct impact: just 80 cents for each $1 invested, while the long-term impacts are not well studied.
Extensive research has shown that microfinance is not a particularly powerful intervention.
Nor was cash what the ultra-poor themselves wanted most. In addition to asking experts for their recommendations, we also engaged many poor Bangladeshis in remote areas directly. Many of their priorities were similar to those identified by the experts, but there were some crucial differences, depending on their circumstances. What the ultra-poor wanted most was increased agricultural productivity. Research has shown that efforts in this area can be extremely valuable, and our calculations predicted that investing $9,000 per agricultural worker would increase Bangladeshi farming productivity by ten percent over two decades—yielding a $4 return on each $1 spent.
MOVING FROM INTERPRETATION TO CHANGE
As one might imagine, the results of our study were not always popular, particularly among advocates of programs that ranked poorly. Sketching what could be done was easy; translating the findings into practice will be hard. But already, the discourse in Dhaka has changed for the better. As an editorial in Prothom Alo, one of the country’s leading newspapers, recently observed, “It is clear that the research is having a real impact on guiding decisions on Bangladeshi priorities and promises to help even more into the future.”
The prime minister’s office is now incorporating cost-benefit analysis across all government ministries. The finance minister has promised to complete e-procurement in two years, and his new budget sets aside $12 million for the effort. And the recommendations on nutrition have already been incorporated into the National Plan of Action for Nutrition, helping the country spend $1.5 billion over ten years even better. “Policymakers prioritize between competing options many times every single day,” Tofail Ahmed, Bangladesh’s minister of commerce, observed. “This project will help us to take a step back and ask, where are the areas where we should focus more attention and resources?”
At this point, the Copenhagen Consensus Center is continuing to work with BRAC in Bangladesh, helping move the reforms from concept to implementation. And what of our own project’s cost-benefit ratio? One immediate result has been the government’s decision to rapidly scale up its e-procurement. This will cost some $60 million in total, but the benefits will run to about $700 million every year. The move would likely have happened eventually anyway, but even if we can claim responsibility for only half the benefits for just the first year, that still means that the $2.5 million project has generated $350 million in benefits for Bangladesh, or $140 back on the dollar. If we were to include the impact of the other 75 proposals, the benefits would be even higher.
There is nothing special about Bangladesh when it comes to the potential gains to be realized. Any country could do a project like this, and we’re currently working on similar efforts for Haiti and India. This type of project is not a panacea for all of the world’s problems, and it would be naive to expect most of the gains to be realized. Nevertheless, the scale of the possible upside is so vast as to be sobering. For example, we estimate that shifting a mere one percent of Bangladeshi government spending from mediocre programs to great ones could end up producing more than $35 billion worth of social benefits every five years—a whole additional government budget’s worth.
Too often, politicians, voters, and donors fall for the bald eagle at the charity fair, letting catchy marketing and heart-rending anecdotes capture their imaginations and their wallets. But cost-benefit analysis provides a powerful tool to see the true track records and potential benefits of the policy alternatives before us, helping more people live longer, healthier, better lives. The moral is simple: If you really want to make the world dance, don’t forget about the price tag. Check it very carefully.