Why economics is hard

This post is motivated as a response to a comment Justin Kraus left on my post titled "Trade" from a few days ago. He's not an economist so I think his confusion is a product of thinking economics is a simpler (and more scientific) subject than it is.


Economics is hard because, while we like to think of it as a science, it's very difficult to get good experimental data. And without experiments you can't falsify bad hypotheses. So in economics, more than in other sciences, stories or theories that make a lot of sense but, as it turns out, are wrong can hang around for a long time.

Experiments, however, are not the only way to know something fishy is going on. Sometimes economists tell multiple stories, each specific to explaining one aspect of the economy. When you put them together, though, there is something inconsistent about them. For instance an economist might have a theory that international trade (1) benefits the United States and (2) doesn't create net unemployment because (1) if people choose to trade and are rational they're benefiting and (2) if people lose their jobs because their products are now imported, the economy will find a new use for their labor.

But then that same economist might say that when, say, Kenya imports t-shirts it puts the local producers out business. This is a net negative for the local economy because, presumably, the local t-shirt manufacturers can't find anything else to do with their labor and go idle. Trade causes unemployment and a long-term decline in GDP.

The problem with these stories is that you're left to wonder why Kenya can't reallocate resources but the United States can. What are the frictions in the Kenyan economy that don't exist in the United States? And while in the U.S. trade necessarily lowers some prices and thus benefits Americans, in Kenya the economist glosses over the fact that those imported shirts have lowered the cost of t-shirts which should increase utility.

That is more or less what happened when Bill Easterly, who usually touts the efficiency free markets and the general rationality of consumers, praised importing cars from Japan on those grounds but, in the past, has seemed to agree with arguments condemning cheap (free) imports to Africa on the ground that they harm local industry. At the same time he's also attacked Dani Rodrik for making pro-industrial policy arguments that rely on the assumption he (as far as I can see) will need to explain the frictions that exist in Kenya and make imports bad but don't exist in the United States.

Paths to Development

I'll write more about this in the future. Below is my taxonomy for approaches to development. Each includes an example of an effective practitioner who takes (primarily) that approach. There are plenty of hybrids, so I'll list some of my favorite examples of those.

All the names are a little mocking and irreverent because I think setting a tone of being both unserious and critical is important. Too many people are critical (which is important), but dead serious and rigidly ideological, when talking about different approaches to doing a project.

1. The Beggar (Lobbyist)

The beggar lobbies governments for money and legislation. For example, the Live8 concerts, organized by Bob Geldof, lobbied the G8 countries to commit to doubling funding for foreign aid. The NGO ONE in the United States, and its celebrity spokesman Bono, lobby for legislation on debt relief (Jubilee Act), cutting farm subsidies (FRESH amendment), trade reform (GROWTH Act), and providing funding for the treatment of AIDS (PEPFAR).

Examples: Bono and Bob Geldof, ONE Campaign

The appeal of the beggar is plain enough: governments have a lot of money and can use it to solve social problems. We spend a lot of time campaigning for our political parties in the US because we believe their approach to health care or education will improve our systems here. Since many problems in the developing world are likewise problems of education and health care, doesn't the same approach makes sense?

The main criticism of the beggar is that he or she usually doesn't understand the complexity of the issues. Lobbying for bad policy can hurt on a massive scale the same way lobbying for good policy can help on on a huge scale. Also, many intellectuals in the countries receiving the aid the beggar lobbies for think that the aid is paternalistic and encouraging laziness and outright corruption in the receiving country.

2. The Policy Entrepreneur (Academic)

The policy entrepreneur is usually a professor or other academic working for a think tank. They argue for policy reforms based on their research. Esther Duflo, co-founder of the Poverty Action Lab at MIT, is a prime example of an effective policy entrepreneur. Her and her colleague's research has shown which interventions in education, water infrastructure, health and microfinance have the biggest impact. Based on that evidence they have worked to reform policy in both the developed world (foreign aid policy) and the developing world (domestic policy).

Examples: Esther Duflo and co., Santiago Levy and Jose Gomez de Leon, Michael Clemens

The appeal of the policy entrepreneur is straight forward: policy decisions have a much bigger impact than small projects, so making good policy decisions should do more good than doing good small projects. The problem is this can work in reverse. Some policy entrepreneurs turn out to be wrong about issues in the long run and thus each runs the risk of making policy worse, not better. Also, many policy entrepreneurs never have much impact on policy.

Hybrid: Jeff Sachs is both a grade A beggar and policy entrepeneur

3. The Mad Scientist (Inventor)

The mad scientist is focused on technology. There are a lot of mad scientists at engineering colleges such as MIT and in universities with Engineers without Borders chapters. The mad scientist thinks primarily on a small scale, hoping to invent neat new technologies that solve basic problems at low cost. Many of these technologies are directed at income generation, but some are also useful for improving health (e.g. delivering vaccines more effectively).

Examples: Amy Smith

The biggest problem for the mad scientist is that their technologies rarely reach the field after endless prototyping. Economists also often question the utility of this approach by noting that all the technology needed for development already exists and arguing that the problem is how the resources are utilized and distributed.

4. The Dentist (Behavior Change)

The dentist focuses on behavior change. In many cases people in the developing world have poor health or education outcomes because they don't make much effort in school or practice basic sanitation practices. The work of the dentist is difficult--like pulling teeth--hence the name. Particular examples of dentistry are total community-led sanitation, which encourages villagers in rural Asia, to take responsibility for their community's sanitation; hand-washing programs world-wide which encourage people to wash their hands; similar problems encourage the use of condoms and boiling water before drinking it.

Examples: TCLS, Public-Private Partnership for Handwashing with Soap

The main complaints about dentistry are that the dentists are trying to tell people what to do with their lives and that their projects tend to be very slow moving, with only incremental successes. The later tends to be valid while the former is not necessarily a fair representation of dentistry, see TCLS.

Hybrid: Paul Polack of IDE is a Dentist mixed with a Mad Scientist. His organization develops technologies for use by rural farmers, but spends at least as much time convincing farmers to take a risk and buy them.

5. The Santa Claus (Charity)

The Santa Claus has the most basic approach: hand out gifts. This approach is almost too simple to comment on, so I'll just throw in that it's my favorite. The Measles Initiative, which vaccinated millions of kids against measles (and other diseases); Partners in Health, which provides free medical care; and Nothing But Nets, which hands out bed-nets, are three of my favorite development projects.

Examples: Paul Farmer

The Santa Claus, while often the most effective development practitioner, and in rare cases the most celebrated (Paul Farmer), is far and away the most criticized. The Santa Claus is accused to eliminating intrinsic motivation, creating learned helplessness and a culture of dependency, and in general being ineffective as resources are (according to one theory) better allocated using the price system. For people who see development as capacity-building the Santa Claus is repulsive for ideological reasons, whereas pragmatists focused on improving health and quality of life, tend to appreciate some aspects of the Santa Claus approach.

Hybrid: Nearly every development practitioner plays the Santa Claus from time to time, often without noticing it.

6. The Petty Bourgeois (Small Business)

The petty bourgeois are businesspeople who espouse the virtues of microfinance and other private sector (but small scale) interventions. They tend to like projects that start restaurants, small service businesses or cottage industry manufacturing.

Examples: Jacqueline Novogratz, Iqbal Quadir

The main criticisms leveled at the petty bourgeois is that they aim too low. If you help someone making $500 a year to make $550 how much good does that do? What does that person spend the money on? Booze, parties or education? Economists have pointed out that because the poor tend to have little human capital and can only acquire small amounts of physical capital with small loans, they will never be able to earn large profits. Economists have also noted that many people who extend microcredit to the poor have enabled the poor to get deep into debt without necessarily raising their income, which might be a bad thing.

Dangers of Obscuring the Point

Sometimes people make arguments in what I call "shotgun" fashion. Instead of listing the key benefits of some policy or project they list all possible benefits in the hope that they'll appeal to a host of different interest and make the project sound like it makes a big difference--by killing two--or five--birds with one stone. I did this when I had to make a presentation about a project I did teaching West Africans to build cheap peanut shellers and I regret it.

I'll use a different example though because it's more illustrative. Take this pitch: "this new energy project lowers costs for individuals, creates jobs, and cuts down on deforestation. Oh, and kids benefit from the fact that the fuel is cleaning burning too!"

When I hear presentations like that is that it makes me wonder if any of those benefits are large. Are they listing many benefits because none of them alone is impressive? And doesn't having a long list obscure the original point of the project?

In the case of the energy project, limiting deforestation probably isn't important in the long run. The impact will be minimal and the environmental consequences are (to my knowledge) paltry compared to the consequences of poverty. The "lower cost" could be (probably is) as little as $10 a year for people, not exactly a kick out of poverty. And it might "create" jobs, but how many? And how many people are displaced in the competing industry? In some ways this is just another argument for trying to be quantitative when talking about benefits and for remembering to be honest about costs.

What I think is particularly interesting in this case is, though, that the added side-benefit, the one that was thrown in at the end, may well be the most important. Indoor air pollution kills millions of kids each year. If they don't die from air pollution they might live to be, say, 50. That's 45 years-of-life added, which we value at $25,000 a piece in the U.S. (roughly). So saving one kid from reducing air pollution is worth (something like) the savings from lowering costs for 11,250 people.

The example that motivated this post was Cash for Clunkers. The program was a total disaster, but many people liked it because of it's laundry list of beneficiaries. It would be progressive (help people driving older cars), environmentally friendly (get dirty cars off the roads), provide a needed stimulus and help the ail car industry. But in reality none of those objectives were achieved efficiently. If we were honest about which were important from the start we could have just designed specific policies for each--at less cost.

Useless Research

There's a new book out on the Lord's Resistance Army (HT: Chris Blattman, who authored a chapter).

The book is meant to set the record straight about the Lord's Resistance Army (hence the subtitle), and based on the quote from Chris they're particularly taking aim at the coverage Invisible Children and CNN have done. (I've never seen anyone else do a story or documentary on the Acholi's plight.)

Personally, I find the tone kind of smug. Invisible Children does a lot to help people. It's possible their account is very misleading and I'd like to read the book to get the facts. But I don't think reading the book is going to give me or anyone else much insight into how to end the war. I've never seen any academic research on foreign policy (non-military strategy) that is in any way useful.

For that reason the tone is offputting. Invisible Children, even if its just a propaganda documentary, inspired thousands of (mostly) high schoolers to raise money to help thousands of (mostly) kids get a better education. It also entertained a lot of people (8.0 average rating on IMDb). From a utilitarian perspective it was an "ethical" movie, however misleading. This book on the other hand, has little ethical value if it can't help end the war, and probably won't even entertain many readers.

You can file this post under the theme "unimportance of fact/reality" which I consider a central theme on this blog.

Random Thought

Here's a thought on why it's so hard to think about trade-offs between health and happiness: people think about different risk very differently depending on the situation. An increased chance of death from 4% to 5% this year is very different from an increase from 1% to 3%. Or consider:

Fishermen (think "Deadliest Catch") have a 1.1% chance of dying on-the-job each year (granted that's not really an increased risk of dying by quite 1.1%) and they make about $13 an hour, which is next to no premium over many other unskilled laborers. But suppose they make $5 more per hour and work 2000 hours a year more than someone at Wal-Mart. That means they'd being compensated for the risk by $10,000.

Now consider that you've been impressed into the English army for an amphibious landing in France. Your risk is dying that day is 2% (ignore non-fatal injuries) and they say they'll let you go after they establish a beachhead. How much would you be willing to escape 2% chance of sudden death? If they above scenario is any guide (or this valuation from speed limits) you should be willing to pay about $20,000 but no more.

Quote of the Day

In only two years . . .  more than 122,000 Floridians have saved $5 million dollars in prescription drug costs.

That is Charlie Crist's webpage on his big achievement in health care: the Florida Discount Drug Card. Many seniors are using the $20.49 a year savings to fulfill lifelong ambitions like visiting Vatican City or attending a game at every MLB park.

Agricultural Development

I've never been that interested in agriculture in relation to development. But I started reading about the history of agricultural extension on Wikipedia and the World Bank's website--it's surprisingly interesting. The page introduces a framework toward the end that is useful for organizing thoughts about any kind of technology transfer/behavior change.

That said, I'm still skeptical of claims like this (quote from this paper):

The role of agriculture in sustainable development and poverty reduction for the vast majority of developing countries cannot be overemphasized. Forty-five percent of the developing world's population lives in households involved in agriculture, and twenty-seven percent in smallholder households, and most depend on agriculture for their livelihoods. The agricultural sector generates on average twenty-nine percent of gross domestic product (GDP), employs sixty-five percent of the labor force in agriculture-based countries ...

It seems obvious, right. Most poor people are farmers, so the key to reducing poverty is generating more income on farms. The problem with that reasoning is that assumption that people can't shift what kind of work they do. In fact, people can and are very rapidly moving from agriculture to manufacturing all over the world. Most of the poverty reduction world-wide can be attributed to growth in China and (without looking up he statistics) I'm fairly certain very little of China's growth has been driven by agriculture. The fact that "industrialized" and "developed" are nearly synonymous is telling.

Another problem is the potential for immizerating growth. Say farmers in Africa expand the supply of cotton dramatically. That supply shift should lower the price for cotton (e.g. harm Ghana and Togo's terms of trade). If the price effect is large enough it could wipe out most, all, or more than the gains from increasing quantity.

That said, agriculture is still crucial to improving people's quality of life. Millions of kids around the world are malnourished and a bumper crop can be used for, among other things, free school lunches. And for small-scale projects, you don't need to worry about general equilibrium effects, thus agricultural investment will often be a good candidate for a "project."

Update: Nature weighs in with an editorial. It's interesting that they think growth and trade won't solve most of the problem. By 2050 the average American family will have an income close to $100,000 and poor countries like China and Mexico should have incomes comparable to Europe today. Do they really think there will be that many poor people to feed? (And what, beyond supply and demand, caused the crisis in 2008?)

More illegal immigration

I read this in The New Yorker. It's well written and, as the headline suggested, it does provide "the real numbers on illegal immigration."

But I dislike this kind of writing. The author takes a very serious tone and, early on, demonstrates command of the facts. You start to trust him. If you don't know any better, you continue to.

Then he claims that deporting illegal aliens would be "economic suicide . . . since they are, for a start, essential to large sectors of the economy, beginning with the food supply." But that is nonsense. If illegal aliens didn't pick fruit then farmers would have to pay for citizens to do it, or buy machines. Someone would have to produce the machines. Suggest that cheap labor is irreplaceable is irresponsible.

He also writes that "the Fourteenth amendment . . . guarantees citizenship to any child born on these shores." That is one standard interpretation of the Fourteenth amendment, similar to the interpretation that the 2nd amendment allows for possession of any type of firearm without restriction. Both of those are questionable, certainly not the intent of the authors, and probably not the interpretation you'd get from some random person in Wal-Mart.

The article ends with a series of attacks against "demagogues" after briefly dealing with one of the three hard questions in immigration reform, "don't illegal immigrants drive down wages for Americans?" That is what supply and demand predict, but (1) that doesn't tell you the magnitude of the effect and (2) things are more complicated. There is a large and growing literature of good research on the topic (see here). It would have been nice if he cut the attack on the nebulous "right wingers" and dedicated a few paragraphs to this issue. He could have also addressed these hard questions:

1. what penalty should they pay to gain citizenship? what is fair?
2. why should Mexicans get preferential treatment in immigration?

That said the first few paragraphs are worth reading.

Letting Go

"Letting Go" by Atul Gawande is a certified Instant Classic essay.

This one isn't a masterpiece in terms of the writing, though Gwande is a good writer, nor does it really resolve the question it deals with. He just weaves a few interesting facts together with a number of very engaging stories. And that's enough.

A few thoughts: In one of his books, either Cosmopolitanism or The End of Ethics, Kwame Appiah mentions that he doesn't like the idea of rules in ethics. He doesn't think there is an algorithm for figuring out what the right thing to do is. Instead we should just try to understand stories, context, and feelings. We should read novels for ethical enlightenment.

This essay demonstrates the strengths of that kind of thinking. I do think we need to think about rules of thumb for when to stop treatment and we need to look more carefully at why hospice patients live so long--something that doesn't surprise me given what I know from positive psychology).  But stories are a start.

Tradeoffs

I'm an unabashed utilitarian--of sorts. I think that all there really is to life is having a happy life, and I'm partial to definition of happy that focus on moment to moment feelings, as opposed to satisfaction when you reflect on memories. (More on the two types in this old post.)

One tradeoff that can comes up for utilitarians is between length and life and happiness during it. Is a better life one with more total happiness? More average happiness? We see this trade-off come up with sinful goods, like drugs, alcohol, and high-fat foods. Suppose you really would be happier if you drank 5 drinks every night. Would you be willing to trade ten years of your life (live to 67 instead of 77) so you could be 1 unit happier on a ten point scale?

That is a bad example because it rests on a mistaken premise. Becoming an alcoholic almost certainly isn't a road to bliss, nor are drug abuse or gratuitous gluttony.

But are there any good, relevant examples of a tradeoff where your risking your health for a clear gain in happiness?

The one that got me thinking about this is not fit for print.

Jeff Sachs on aid

I almost didn't read this column because Sachs often sounds like a broken record. But it turned out to be great.

Hosting this year’s G-8 summit reportedly cost Canada a fortune, despite the absence of any significant results. The estimated cost . . . reportedly came to more than $1 billion. This is essentially the same amount that the G-8 leaders pledged to give each year to the world’s poorest countries to support maternal and child health.

If only more economists had that kind of passion.

Of all of the G-8’s promises . . . the most important was made to the world’s poorest people at the 2005 G-8 Gleneagles Summit . . . The G-8 promised that, by this year, it would increase annual development assistance to the world’s poor by $50 billion relative to 2004. Half of the increase, or $25 billion per year, would go to Africa.

The G-8 fell far short of this goal, especially with respect to Africa. Total aid went up by around $40 billion rather than $50 billion, and aid to Africa rose by $10-$15 billion per year rather than $25 billion.

He can't say this, but look at those numbers. Aid to Africa increased by $10-15 billion. Millions of bed nets were distributed, measles vaccines injected, Rwanda has universal health insurance, and treatment for HIV is up 10,000% over the decade.

Would that have happened without the concerts, columns, and citizenship?

Awesome Links

1. Michael Bay has an awesome sense of humor

2. Do we need a Secular Religion? No. But he's right that

. . . religion [is] not just a matter of belief, but that it sat upon a welter of concerns . . . and that by getting rid of God, one would also be dispensing with a whole raft of very useful, if often peculiar and sometimes retrograde, notions that had held societies together since the beginning of time.

Religion makes people happy. We do need something to fill the void.

3. Uwe Reinhardt, a very good health economist, on insurance mandates

4. A slightly bloated but good edition of Why U.S. Food Aid Policy Sucks

5. Health halo effects. I think some of this reporting is misguided. First, lab experiments don't necessarily translate well into "the field." Second, most people probably have more to gain by finding a better mix of calories than cutting down on calories. Subway does provide a better balance than McDonald's even if you'll eat more calories there.

Whither Growth?

Probably not from an army of small-scale entrepreneurs. Chris Blattman notes these good points from Abhijit Banerjee:

1. The poor are not particularly well-suited to be entrepreneurs: They neither have the risk bearing capacity nor the human capital

2. Nor will anyone give them enough capital to really grow the businesses;

3. This suggests that the main source of dynamism has to be growth of medium to large firms.

Do voluntourists do more harm than good?

I've heard a lot of people throw fits about wasted aid. It would be nice if development projects were more successful or even if even knew which ones worked (see "Taking evidence seriously").

What I don't understand is why people insist it's easy for someone to do more harm than good? Does anyone really believe that in expectation people going on short-term development trips do more harm than good?

There are obvious benefits to the trips. The people who go are probably more likely to stay in the field of international development. They may be more likely to donate money in the future and have a network effect on their peers. They spend money in-country which increases aggregate demand in the economy, surely in good thing for most developing countries (esp. during the recession). How valuable are these actions? Probably, on average, not that significant, but they are benefits?

What are the costs associated with voluntourism? People have provided plenty of examples of how aid projects can not live up to their potential. But that is different from doing harm. I'm stunned that so few people can tell the difference between tripping someone and watching them trip when you might have prevented it. As far as I can tell (this may be revised over time) voluntourist's are bad because they might annoy some people who don't want them meddling in their community and they might make those people feel they can't help themselves. Those don't sound like big costs and it's hard to understand why, if someone is such a nuisance, people just let them hang around.

Someone should put together a compelling case for why students shouldn't do their little touristy trips abroad that no one thinks helps much but I can't imagine, on average, do more harm than good.

Development Contradictions

All debates are plagued by some messy thinking here and there. But "development" debates, I've noticed, tend to have particularly murky thinking. I think this is rooted partly in the political and racial overtones and partly in the complexity of the issues (they tend to be multi-dsiciplinary). I've complied a list of my top six example of doublethink in development.

1. Some people are against farm subsidies because they lower food prices worldwide--Haiti was flooded with cheap "Miami rice," West Africans complain about cotton subsidies, etc.

But when food prices spiked in 2008 everyone was in outrage. We need to increase supply to ensure low enough prices for the urban poor. So which is it--are high prices good or low prices good? (There is a nuanced position here but I didn't hear anyone taking up that complexity.)

2. A lot of people are for a "human dignity"-based approach. We should listen to what people need and make the technologies they want. The patronizing attitude that we know what they need is wrong.

Then I hear this classic D-lab story: someone built a latrine with locks and the ownered ended up using it as a shed. This is explained as an example of a failed project. But wasn't the result good? The owner chose what they needed (a shed, not a latrine). There is also the malarial net version: they give out nets, people use them to fish, Easterly says this was a waste. But shouldn't some people say "well look they needed a fishing net more than a malarial net?" (There is again a nuanced position here that acknowledges on some issues, esp. health, we know best and should be patronizing.)

3. Some people love this idea of making everything a business. Sure, Nike pays low wages but it's not exploitation, it's the only way to develop. Micro-finance institutions charge usurous interest rates but it's better than the moneylenders.

But then when poor people contract with Millennium Villages to make money giving tours to American tourists in Rwanda, it's a bad thing. It's an affront to human dignity and exploitation. What?

4. Some people think we need to sell things, otherwise people won't understand the value of them. And in any event if you charge a price you ensure the people who need the limited resource are most likely to get it. Plus, it's bad to give things away because it promotes a culture of dependency.

But one tour of India let tourists donate goats to one village they toured. The "problem" was that they contracted with the only guy who could speak English in the village and he just took the goats and sold them. It was, of course, an outrage. It was fostering corruption in the village.

But doesn't it follow the principles we like--allocating the few goats efficiently, promoting a business, making sure people put the goats to use?

5. Daimbisa Moyo wants to end aid. But then says she only means development aid. Humanitiarian aid for health project is necessary and NGOs are ok in her book. But she dislikes PEPFAR. What gives? Is she for AIDS treatment or not? I have no clue after reading the whole debate and her book. You'd think the most influential book in development in several years could be clear on a basic point like that.

5.1 Moyo, like many others, notes that donations can be bad because they put Africans out of work. If we just donate bed nets then it will put the Africans who could have gone to work making the nets out of business. This simple story is not so obviously wrong. But it should be to economists like Moyo. It goes by the name broken window fallacy. I agree that a more complicate story, which emphasizes one of: the value of learning by doing, increasing returns to scale, and positive production spillovers. But people need to note those things if they're important, and explain which ones.

6. The biggest contradiction in development is between the constant exhorting that our goal is to help them help themselves and the reality that we don't. Paul Farmer makes this point--although people want to pretend they are in a position of equality with the people they help they are not. Your health care is better. You can go home. If you live in the community you'll never be treated as an equal, if you're not black. Some people try to become equals by living in the communities for extended periods of time. But how much can that help? Most people argue that the people are smart and are doing what they can for themselves. If you become one of them then you have nothing to do that they haven't tried. Also, the more context specific the project the less likely it is to scale.

If people really believed the goal was to help them develop themselves they'd spend their time on anti-corruption measures (these are complicated), lobbying to reform trade deals, including cutting farm subsidies, and raising funds for NGOs that employ local people to do the work. We'd also lobby for immigration to let more Africans come here, get skills, send remittances, and eventually go back with their human capital and be leaders, entrepreneurs etc. But I don't see a lot of people working on any of the above.

Taking evidence seriously

Banerjee (brief) on taking evidence seriously

Tim Harford's experimentalist manifesto