Excellent comment—strongly upvoted for engaging with the data.
how did you calculate the median figure for Vietnam that you reference in section 4 ($6,914 GDP per capita)?
The sheet where we calculated the median growth episode within the spreadsheet is here:
Source: Pritchett, Labor p23
Vietnam was just the median of these selected growth episodes- because Pritchett in his example uses quite a big growth episode. Pritchett calculates the NPV gain from growth acceleration per person from this median case as $6,914. This is for illustrative purposes, picking Vietnam has no special significance here. “To be affected by a think tank” also has no special significance, we didn’t check whether this growth episode was likely affected by a think tank.
These are selected by Pritchett:
“These are a list selected the largest episodes of growth acceleration. Source: Selected episodes. Author’s estimates from estimates in Pritchett, Sen, Kar, and Raihan 2016”
so… re your question:
When I look at the those figures in Appendix A, though, it seems like the median growth episode calculated using PRM (without reference to dollar size) is somewhere around Ecuador’s negative growth in 1978, which doesn’t seem like it would line up even with the conversion to $PPP.
Yes, this is likely largely due to Vietnam having a roughly ~10x higher population and being 10x poorer back then.
I think it is okay to use, as Pritchett does, these selected growth episodes, because if one wants to maximize effectiveness using policy one can strategically only look at big poor countries. One could further look at only those countries where growth is sluggish and perhaps where economic policy is particularly bad.
I write about this in the appendix:
Because effective altruism often tries to focus on the poorest countries, where a dollar goes 100x further than in rich countries, there is perhaps most hope for growth diagnostics.
So perhaps Duflo is right in that “Growth is likely to slow, at least in China and India, and there may be very little that anyone can do about it.” And this is actually born out in China’s and India’s performance on the World Bank’s Doing Business indicators, where they score 63th and 31st out of 189 countries, though being relatively poor. Thus, there seem no low hanging fruit to improve their economic policy.
However, below I show a table where I multiply population size of every country by their poverty multiplier (i.e. $1 is worth x times more going to this country than to the richest country in the sample. See appendix 2 of this doc for more info). This can then be ordered by the utility created by increasing GDP per capita by $1. India comes out on top because of its large population (1.3bn) and relatively low GDP per capita ($6,574). China comes 3rd, because though it has a large population, it is already relatively rich ($15,531). Recall that the problem is that we might not know how to increase growth in India and China.
However, there are many smaller very poor countries in the top 10 sample such as DRC and Ethiopia—very poor countries with 100 million population. This can then also multiplied further by neglectedness/tractability criteria. For instance, in a country’s ranking on the WB Doing Business ranking divided by GDP. There one can see that, relative to its GDP per capita, China already does quite well on the Doing Business ranking. However, the DRC and Ethiopia do poorly on the doing business ranking, even relative to their GDP. These countries could be most cost-effective for economic policy assistance.
I think this view would probably be endorsed by many prominent development economists. But I concede that there are also development economists who believe that health and education is very important.
When I first read about Rodrik’s theory of development, I updated in the direction that health and education are not that important for growth at least for very poor countries, even though it’s quite unintuitive.
From the appendix doc:
Historically, almost all non-poor countries have grown their economies in three steps:
Rural to urban migration: Unskilled (subsistence) farmers migrate to cities and start working in factories. Over night, this increases their productivity many times over.
Manufacturing absorbs vast amounts of unskilled labor: These workers need very little human capital: they do not need to be educated because work in factories is very simple. Population health does not prevent growth either, because there are enough to replace sick workers.
Manufacturing exports niche products to the world market: The factories find their niche product (e.g. initially often garments) and export to the world market, which can absorb large amounts of the same good (e.g. billions of shoes)
Again quoting Weil’s review of “Health and growth” (emphasis mine):
As is often the case in economics, the observation that income and health are correlated, is only the beginning of the discussion. Such a correlation can be induced by causation running in either direction, as well as by the effects of some third factor. A priori, there are good reasons to think that all of these are possibilities. People who are healthier can work harder and learn more in school; and where people live longer they will be incentivized to invest more in education.Thus, we would expect better health to cause economic growth. On the other hand, higher income allows individuals or governments to make investments that yield better health. Finally, differences in the quality of institutions (looking across countries), in human capital (looking across individuals), or in the level of technology (looking over time) can induce correlated movements in health and income.”
re: Nunn: I’m not ruling out that invariant geographical factors influence economic development by way of health. But it’s a different question on whether we can do anything about that by ramping up health spending and ameliorate these differences and whether that’s important for growth.
In my opinion, randomistas do not focus on growth at all, be it level effects or growth effects.
Though to be fair there’s this short passage in Duflo’s new book on this:
while we do not know when the growth locomotive will start, if and when it does, the poor will be more likely to hop onto that train if they are in decent health, can read and write, and can think beyond their immediate circumstances. It may not be an accident that many of the winners of globalization were ex-communist countries that had invested heavily in the human capital of their populations in the communist years (China, Vietnam) or countries threatened with communism that had pursued similar policies for that reason (Taiwan, South Korea).
Also we do say that “we do not think that the things assessed by RD do not increase economic growth at all: indeed some RD health interventions increase earnings and consumption later in life, and thus do increase growth to an extent. However, evaluating whether the effect size is trivial or not should be a top priority for proponents of RD.”
Yes, interesting take.
Aside from risk aversion, in the appendix, I list some more cognitive biases that might be at play for why people prefer RCTs.
Relatedly, perhaps people sympathetic to long-termism might believe that speeding up growth might speed up GCRs from emerging technologies. And while it is unclear when growth will speed up x-risk at all (see for instance), I think that when it comes to differential technological development, not all growth is equal.
What speeds up risks from emerging technologies is mostly growth in highly technical sectors in high-income countries. Growth in low-income countries will not increase world growth much and is less likely to cause risks from emerging technologies.
Put simply: Burundi’s catch-up growth won’t speed up global growth by much, is unlikely to speed up risks from AI or bio any time soon. Growth has been argued to lead to “Greater opportunity, tolerance of diversity, social mobility, commitment to fairness, and dedication to democracy.” Perhaps growth in poor countries will actually increase stability and thus be good from a differential technological development point.
Lower skilled labor also competes with AI R&D and so increasing trade and migration decrease AI R&D (see “Why Are [Silicon Valley] Geniuses Destroying Jobs in Uganda?”.
But even if growth in poor countries will slightly increase x-risks, then it might still be optimal to support it and offset the x-risk increase through targeted interventions to decrease x-risks. This is because multiobjective optimization for both x-risk reduction and global poverty is likely harder than single objective optimization for the most effective interventions in each category separately.
The paper we cited is a comprehensive recent meta-analysis on the topic of health and growth that synthesizes the literature on this topic.
The paper concludes:
“If improving health leads to growth, this would be a reason, beyond the welfare gain from better health itself, that governments might want to make such investments. However, the evidence for such an effect of health on growth is relatively weak. Cross-country empirical analyses that find large effects for this causal channel tend to have serious identification problems. The few studies that use better identification find small or even negative effects. Theoretical and empirical analyses of the individual causal channels by which health should raise growth find positive effects, but again these tend to be fairly small. Putting the different channels together into a simulation model shows that potential growth effects of better health are only modest, and arrive with a significant delay.”
We did however acknowledge that this claim is controversial:
Moreover, and more controversially, we do not believe that health interventions (whether directly funded or implemented by the state) are the best way to increase growth in the poorest countries. Here, we want to start a discussion on what the most effective causes of growth are, given its huge importance.
This is a topic of ongoing debate in the literature—future research could look into this topic more and a starting point could be the citation trail from the study above.
Thanks (strongly upvoted for trying to falsify a central claim). All opinions are mine.
1. While the interesting paper you cite shows that policies bad for growth are at historic lows and argues that much progress has been made, 20% of all countries still have bad policies, and 25% of SSA countries. Given the potential very high effectiveness of growth policy, that we tried to demonstrate in the piece, the value of information of looking into this further is high.
2. I do cite Rodrik in the Appendix who argues that these days, “standard prescriptions” (i.e. Washington Consensus) might not work any longer and we should be skeptical of top-down, comprehensive, universal solutions (though perhaps there are some more generalizable policy prescriptions to be discovered with further research—Rodrik for instance expands the Washington consensus with an additional 10 policy prescriptions).
However, technical assistance by more specialized agencies (e.g. DFID, USAID, GIZ as well as the World Bank’s country offices), and also NGOs such as the International Growth Center, the Copenhagen Consensus, etc. might be able to do “growth diagnostics” to find out where growth is bottlenecked and then help with tailor-made policies on a country-by-country basis.
They might also help with implementation issues, and even indifference issues.
I’m not sure the ‘extreme scepticism’ (perhaps we could just call it scepticism?) argument is given a fair shake. Note that answering the question of what causes a country to grow is basically the big question of development economics, and as such it has received considerable attention from economists. In the Duflo and Banarjee piece, they argue that economists did find good low hanging fruit, notably misallocation of resources, but they argue this is reaching a point of diminishing returns. Economists are now struggling to find great opportunities in growth economics, and so there is a good case for looking at different approaches to development. This argument feels plausible to me, and it means you do not have to make the apparently crazy claim that economists never had significant influence on past effective growth policies.
Yes, I steelman this view in the Appendix (my view not necessarily John’s):
“Growth is not as neglected as RD, its low-hanging fruit have been picked, and the marginal dollar is not as effective”
“The evidence that macroeconomic policies, price distortions, financial policies, and trade openness have predictable, robust, and systematic effects on national growth rates is quite weak—except possibly in the extremes. Humongous fiscal deficits or autarkic trade policies can stifle economic growth, but moderate amounts of each are associated with widely varying economic outcomes.”
For instance, take the debate over trade liberalization. Recall that there was exceptionally weak global trade growth over recent years. Relatedly to the previous point, some argue that the “low-hanging fruit” of economic liberalization has already been picked. For instance, Weyl argues that in “Radical markets”:
“There is a consensus that the economic gains from further opening international trade in goods is minimal. Studies by the World Bank and prominent trade economists find that eliminating all remaining barriers to international trade in goods would increase global output by only a small amount, 0.3–4.1%. For global investment, the most optimistic estimate in the literature finds a 1.7% increase in global income from the elimination of barriers to capital mobility. Many believe that liberalization of international capital markets has gone too far. Three top IMF economists recently argued that even liberalization that has already taken place has brought limited gains to economies while generating inequality and instability.”
However, there is a debate about this and counterarguments:
Others argue that trade policy is still very relevant. Complete rich-country liberalization would, after a 15- year adjustment, increase income in developing countries by $100 billion per year, which is approximately twice current aid flows.
Also, guarding against protectionism and not losing the growth from trade might be very important: one study suggest that an “increase in tariffs to average bound rates of 44.7 percent in highly protectionist countries such as India, Bangladesh, Pakistan and Sri Lanka would translate into a decline in real income in South Asia by 4.2 percent or welfare losses of close to US$125 billion relative to the baseline by 2020”.
Pritchett too seems much more optimistic about growth diagnostics and believes that while we might not know everything, we generally have a reasonable understanding of what causes growth and can even influence it.
Pritchett has edited a whole volume on growth diagnostics, including on the causes of growth in India.
Generally, my take is that growth diagnostics might get harder the richer a country becomes, by virtue of there being less and less data from other countries on how they developed. Thus, for the poorest countries, growth diagnostics might be easiest because we can draw lessons from all other countries on they developed.
But in the Appendix I have an analysis where I multiply population size of every country by their poverty multiplier (i.e. $1 is worth x times more going to this country than to the richest country in the sample. See appendix 2 of this doc for more info). This can then be ordered by the utility created by increasing GDP per capita by $1. India comes out on top because of its large population (1.3bn) and relatively low GDP per capita ($6,574). China comes 3rd, because though it has a large population, it is already relatively rich ($15,531). Recall that the problem is that we might not know how to increase growth in India and China.
However, there are many very poor countries in the top 10 sample such as DRC, Bangladesh and Ethiopia—very poor countries with +100 million population. This can then also multiplied further by neglectedness/tractability criteria. For instance, in a country’s ranking on the WB Doing Business ranking divided by GDP. There one can see that, relative to its GDP per capita, China already does quite well on the Doing Business ranking. However, the DRC and Ethiopia do poorly on the doing business ranking, even relative to their GDP. These countries could be most cost-effective for economic policy assistance.
The Copenhagen Consensus Center is actually doing something along the lines of assisting countries / highlighting the need to improve their economic policies. For instance they are helping Bangladesh to improve its economy and prioritize which policies would have the highest social, economic and environmental benefits for every dollar spent. On top of their list is e-procurement across government and land records digitization—related to criteria used to rank countries on the WB Doing Business index.
Randomista is clearly not a neutral term, and I think constitutes a kind of name calling (e.g. Corbynista in the UK). Do proponents of RCT development use this term for themselves?
We did not use it in a name calling way but rather as a neutral term to describe the intellectual movement. The term is used by mainstream economists who are critical in a respectful way, but also by randomistas themselves (note for instance that Duflo or Blattman have used the term).
However, it is true that
“the term was first used by another Nobel laureate and a long time and fierce critic of RCTs, Angus Deaton. According to Andrew Leigh, the author of a recent book titled, Randomistas: How Radical Researchers Are Changing Our World (2018), the term was meant mostly as an abuse which Leigh turned into a compliment. Leigh defined a Randomista as ‘’someone who believes we can find answers to important questions by tossing a coin and putting people into a treatment and control group, comparing the outcome, and then using the randomization to get a true causal effect.” (Social Science Space, 2018).”
Nice, you found another blunder in the literature!
“First, and classically, rating agencies’ fees tend to be high. The revenues of rating agencies come from new ratings and from the reexamination of former ones, as it is very difficult for a company, once it has been rated, to withdraw its rating from the market. It means the operational risk of rating agencies is quite low, just as the volatility of their revenues. We don’t know much about the prices of ratings and the profits of agencies. Nevertheless, in 2011, the operational profit of Standard and Poor’s and Moody’s was about 40 %; and Fitch’s was 31 %. For the first nine months of 2011, the revenue of Standard and Poor’s reached US$ 1.3 trillion for about 1,400 analysts. The figures for Moody’s were US$ 1.2 trillion for 1,300 analysts. These figures make for an annual revenue per analyst higher than US$ 1 million, which is quite high.”
from this paper on reforming rating agencies: https://sci-hub.tw/https://link.springer.com/chapter/10.1007/978-3-319-44287-7_12
So this should be billions, not trillions.
I had actually interpreted the figure differently and thought that rating agencies analysts rate trillions in value or something.
Have deleted these from the dataset.
 Pritchett, ‘Randomizing Development: Method or Madness?’ (2019), p. 23-24. see:
Thanks—yes, John Halstead and I co-authored this post.
(We’ll use the forum’s new co-authoring function, this is why I accidentally omitted the authors when first posting, but it will take a little time reflect it, so I’ve fixed this provisionally).
done! thanks for the suggestion
Thanks—I fixed the global debt in the non-financial sector figure!
And yes, you’re right that notionals need to be interpreted carefully—I initially had a paragraph in my post that notionals should be interpreted carefully, but then I cut it out. Your example is a good one and shows that, in theory, a world with a high notional value of derivatives trading can be one with a stable financial system.
However, I disagree that it is a “totally irrelevant number” and that the in practise notional total volume might be (a not entirely very bad) proxy measure for economic stability.
“To give an idea of the size of the derivative market, The Economist has reported that as of June 2011, the over-the-counter (OTC) derivatives market amounted to approximately $700 trillion, and the size of the market traded on exchanges totaled an additional $83 trillion. For the fourth quarter 2017 the European Securities Market Authority estimated the size of European derivatives market at a size of €660 trillion with 74 million outstanding contracts.
However, these are “notional” values, and some economists say that these aggregated values greatly exaggerate the market value and the true credit risk faced by the parties involved. For example, in 2010, while the aggregate of OTC derivatives exceeded $600 trillion, the value of the market was estimated to be much lower, at $21 trillion. The credit-risk equivalent of the derivative contracts was estimated at $3.3 trillion.
Still, even these scaled-down figures represent huge amounts of money. For perspective, the budget for total expenditure of the United States government during 2012 was $3.5 trillion, and the total current value of the U.S. stock market is an estimated $23 trillion. Meanwhile, the world annual Gross Domestic Product is about $65 trillion.
At least for one type of derivative, Credit Default Swaps (CDS), for which the inherent risk is considered high[by whom?], the higher, nominal value remains relevant. It was this type of derivative that investment magnate Warren Buffett referred to in his famous 2002 speech in which he warned against “financial weapons of mass destruction”. CDS notional value in early 2012 amounted to $25.5 trillion, down from $55 trillion in 2008.”
Yes, excellent question.
This is a really hard analysis to do because it’s very hard to assess what the money would have been spent on counterfactually- see my comment above to Khorton.
My subjective impression is that the $75k for the Better Science campaign was heavily skewed towards EA donors and would have gone to EA causes anyway. However, assuming returns to research this might have still improved the quality of donation within the EA community, which counterintuitively can sometimes be more effective than growing the pie.
However, the $200k raised for the climate change campaign was heavily skewed towards non-EA donor and perhaps the counterfactual here were less effective charities or even conspicuous consumption.
In government, we’d consider the money you crowdfunded a cost to society as well.
I’d argue mine is more appropriate for a fundraising charity, but at least I understand the difference now!
Yes, I agree that for all non-profits or public benefit companies a net present value analysis from a societal perspective would be optimal and what we ultimately care about. However, I feel like my analysis is good approximation: of course, the money I crowdfund is a cost to society given the opportunity costs, but the implicit assumption here is that the donor’s money would counterfactually be spent on conspicuous consumption or perhaps ineffective charities. If this is the case, then because the value of increasing consumption in advanced economies is comparatively small and the cost-effectiveness analysis is relatively insensitive to whether we count this, then the business analysis is a good approximation for the societal value and it’s ok to leave it out for simplicity’s sake.
Fundraising charities routinely use “fundraising ratios”, which are benefit-cost ratios and similar to net values, so this seems standard practise.
In the future, I might look more into the true counterfactual societal net value and see whether some of the money donated would have gone to similarly effective charities in the future.
I think EAF did a good job at this where they estimated the money they raised that would not have been donated otherwise.
One year ago I’ve gotten $40k from the EA community, these are the cost to society, because I’ve spent this money on Let’s Fund operations.
Then, on average roughly 1 year later, I’ve crowdfunded $300k for my campaigns (these are the benefits).
The benefits minus the costs are the also called the net value, which are then simply $260k. But if you discount the costs at 5% with the formula:
then you get a net present value of 234.
Does that make sense?
Really interesting questions—thank you!
How confident in your analysis and conclusion do you have to be in order to publish a recommendation? For example, do you believe “better wrong than vague”?
I’m very confident in the conclusions of the research for campaigns and the bar for publication is substantially higher than for what I post on the EA forum. I usually also ask many people to review my research for campaigns (see acknowledgment sections in the reports).
On the EA forum, I sometimes don’t excessively hedge my claims for clarity’s sake. And I have sometimes epistemic status disclaimers which you’re referring (‘better wrong than vague’, “say wrong things”, “Big, if true”, “Strong stances’)
Do you try to caveat to show your degree of confidence?
Yes, I use sensitivity analysis and careful language throughout.
For instance, in my cost-effectiveness analysis I caveat:
“Below we present a very rough, simple, back-on-the-envelope cost-effectiveness analysis (“Fermi estimate”). This model is crude and should not be taken literally. Rather than leaving our assumptions unarticulated and fuzzy, we think it is better to be wrong than vague. By stating assumptions explicitly that can be questioned and falsified (as the common aphorisms in statistics go: “Truth will sooner come out of error than from confusion” and “All models are wrong, but some are useful”). It also helps us think through relevant considerations and formalizes our intuitions. If you disagree with any of the inputs to our model, then you can create a copy of our spreadsheet and plug in your own parameters.”
I also use the word “might” about 90 times in the Clean Energy campaign.
But there are some statements even in the report where I’m intentionally wrong for clarity’s sake. For instance, when I write:
“The focus of advanced economies like EU countries to prioritize reducing their own domestic emissions is a natural impulse (‘clean up your own backyard first’). But 75% of all emissions will come from emerging economies such as China and India by 2040. Only if advanced economies’ climate policies reduce emissions in all countries, will we prevent dangerous climate change. We call this the cool rule: only if all countries reduce their emissions will the planet stay cool.”
The bolded sentence is clearly wrong on some level, because we can perhaps use geoengineering to cool the planet or maybe emerging economies such as China will solve the issue. However, I feel this is less important to emphasize because it’s somewhat unlikely and writing all that out would distract from the central message. By making strong statements such as “Only if” you’re making your writing and central claims really clear so that they can be more easily falsified. But some people might disagree and like to hedge more.
How easy would it be to find a demonstrably incorrect statement or paragraph in your work?
I’m quite careful I think but given the length of the report I cannot rule out that there are errors in somewhere. But I’d be somewhat surprised if they were easy to find. So I’ll pay a bug bounty of $20 for any statement that is demonstrably incorrect.
However, the central claims I’m very confident in, because I’m trying to triangulate with multiple lines of evidence, so that the conclusions do not depend on a single piece of data (https://blog.givewell.org/2014/06/10/sequence-thinking-vs-cluster-thinking/ ).
If the recent Bill Gates documentary on Netflix is to be believed, then Gates first became seriously aware of the problem of diarrhea in the developing world thanks to a 1998 column by Nicholas Kristof. It’s hard to assess the counterfactual here (would Gates have encountered the issue in a different context? Would he have taken the steps he ultimately did after reading the Kristof piece?) but it seems plausible that Kristof’s article constitutes a cost-effective intervention in its own right (if a not particularly targeted one).
To clarify, the win here was not to influence Gates, because he is already very much on board with clean energy innovation agenda (though perhaps if he really read the article then it might be that it might have ever so slightly shifted his views towards the importance of government vs. private R&D which I feel he doesn’t focus on enough).
Rather that he has 50mn followers on Twitter and is considered a public intellectual / authority on climate change (publishing a book on climate change in 2020).
But yes your general point is good, because for instance Reid Hoffman and others retweeted the Gates tweet—so perhaps there might be a very small chance of a “Kristof’s effect”.
I’m curious if you consider the wide propagation of your research in the news media a “risky and very effective” project, and if your research products have been intentionally structured toward this end.
Yes, the research is intentionally structured for wide-ish dissemination. This manifests in several ways:
I feel there are relatively few and only modest downside risks to this project. For instance, there’s little information hazard, however there are some moral hazards. See the “Risks, reservations, drawbacks section of the report” where we write:
“Overall, what these quotes have in common are concerns about the moral hazard of spreading a meme like ‘breakthrough technology by itself will solve climate change’. Authors repeatedly caution that additional policies—especially carbon taxes—are needed. We think these concerns are warranted, but do not believe that this suggests that clean energy R&D should not be increased. We believe there is consensus amongst even these climate policy scholars that R&D levels must be increased substantially. Crucially, while the moral hazard aspect of clean energy R&D increases might drag out emission reduction, another aspect pushes more strongly in the other direction and make carbon taxes more likely. For instance, one economic model suggests that “if a carbon tax imposes a dollar of cost on the economy, induced innovation will end up reducing that cost to around 70 cents”. Given that political acceptability is mainly a function of cost, making clean energy cheaper might make carbon taxes more likely.”
The crowdfunding aspect of the campaign means that the campaign and its topic are at least a little bit optimized for being more readily understood by the wider public. This means that there are other harder to explain topics that are perhaps more neglected and thus might be more effective. For instance, in the report in the Section on Climate change is relatively non-neglected, we write:
“Climate change is a high-profile topic that many people work on. It is funded by both governments and big private foundations. Thus, even though clean energy innovation in particular has been relatively underfunded within the climate policy space, it is conceivable that in the future ITIF might receive grants for their clean energy innovation program from other funders, which lowers the counterfactual impact of donating to this project. In other words, comparatively, climate change is not very neglected. For instance, the risks and expected losses of pandemics are of a similar magnitude than those of climate change, yet the area is more neglected by other funders.”
Also, there the page is intentionally structured hierarchically going from less to more in-depth, with summaries at the top and then for people who really want to read all the details there’s heavily footnoted analysis further down on the page.
If you have some takeaways from your big success so far, it could be very helpful to post them here- widely taken-up tweaks to make research propagate more effectively through the media are marginal improvements with potentially very high value.
Generally with questions about success, there’s of course a lot of survivorship bias and a lot of it was perhaps just luck. Similarly, the first step to make research propagate widely is that you need to spend a lot of time and effort researching and editing until it’s the research not only really good but also very readable, which requires a lot of resources/privilege.
So perhaps take the following with a grain of salt- your mileage will vary.
If you want your research be covered by the media you of course need a good pitch and get in touch with a lot of relevant journalists (numbers game). You can have it peer-reviewed by people and say so and so has reviewed it and says it’s really good/interesting/novel research.
Then to push the coverage of your research you can find influencers for whom your content is highly relevant. I used social proof and had a few select academics and policy wonks I was connected to retweet/endorse the article because it was very much in their field of expertise even if they didn’t have very many followers. Then I used this to contact relevant influencers who in the past had tweeted about climate change and had also in the past retweeted Vox articles (aligned political leaning). You can tell them that so and so has already retweeted it as social proof, and ask if they could perhaps also retweet because it’s relevant to their audience.
Then there’s a technique called power mapping that I used, where you get in touch with people that are connected to even more influential people. You’re connected to many people through only very degrees of separation (small world phenomenon), so you perhaps know someone who knows someone who knows an “influencer”. You can see who for instance, Obama follows on twitter and then if you get to those people to to get say Obama to retweet the coverage of your research (because it’s on reputable site such as Vox).
Sorry if this was a bit rambly, but I hope you get the general idea.