Information is a public good: Designing experiments to improve government


 * This essay is on Wikiversity to encourage a wide discussion of the issues it raises moderated by the Wikimedia rules that invite contributors to “be bold but not reckless,” contributing revisions written from a neutral point of view, citing credible sources -- and raising other questions and concerns on the associated '“Discuss”' page.


 * This article uses ISO 8601 dates except for References, which are controlled by standard Wikidata formatting, and direct quotes. In the initial author's experience, ISO 8601 dates seem to make it easier to remember dates and to compute differences between them.

Abstract
This article reviews literature relevant to the claim that "information is a public good" and recommends experiments to quantify the impact of news on society. We propose randomized controlled trials to evaluate the relative effectiveness of alternative interventions on the lethality of conflict and broadly shared economic growth. Experimental units would be polities in conflict or with incomes (nominal Gross Domestic Products, GDPs or gross local products) small enough so competitive local news outlets could be funded by philanthropies or organizations like the World Bank but large enough that their political economies have been tracked with sufficient accuracy to allow them to be considered in such experiments. One factor in such experiments would be subsidies for local journalism, perhaps distributed to local news outlets on the basis of local elections, similar to the proposal of McChesney and Nichols (2021, 2022).

Introduction

 * Information is a public good.
 * Misinformation is a public nuisance.
 * Disinformation is a public evil.

Public goods
In economics, a public good is a good (or service) that is both non-rivalrous and non-excludable. Non-rivalrous means that we can all consume it at the same time. An apple is rivalrous, because if I eat an apple, you cannot eat the same apple.

A printed newspaper may be rivalrous, because it may not be easy for you and me to hold the same sheet of paper and read it at the same time. However, the news itself is non-rivalrous, because both of us and anyone else can consume the same news at the same time, once it is produced, especially if it's published openly on the Internet or broadcasted on radio or television.

Non-excludable means that once the good is produced, anyone can use it without paying for it. Information is non-excludable, because everyone can consume it at the same time once it becomes available. Copyright law does not apply to information: It applies to expression. Stiglitz (1999) said that Thomas Jefferson anticipated the modern concept of information as a public good by saying, "He who receives an idea from me, receives instruction himself without lessening mine; as he who lights his taper at mine, receives light without darkening me." Stiglitz distinguished between "push and pull mechanisms" to promote innovation and creative work: "Push" mechanisms pay for work upfront, hoping that it will achieve a desired outcome, like citizen-directed subsidies for newspapers. "Pull" mechanisms set a target and then reward those who reach the target, like copyrights and patents.

Lindahl (1919, 1958) recommended taxing people for public goods in proportion to the benefits they receive. For subsidies for news, especially citizen-directed, this would mean taxing primarily the poor and middle class to fund this. If better news translates into productivity improvements whose benefits are broadly shared, as claimed in the literature cited in this article, the benefits that poor people receive would soon exceed the taxes they pay for it. Then the news subsidies would effectively be free in perpetuity, paid by benefits the poor would not have without those subsidies. If Piketty (2021, cited below with Figure 1) is correct, the ultra-wealthy would likely also benefit in absolute terms, though the relative distinction between them and the poor might be reduced.

This article recommends randomized controlled trials to quantify the extent to which experimental interventions benefit the public by modifying socio-political environment(s), including information environment(s), in ways that (a) reduce political polarization and any accompanying violence and (b) improve broadly shared peace and prosperity for the long term.

Sharing increases the value
The logic behind claiming that "information is a public good" can be easily understood as follows:


 * If I know the best solution to any major societal problem, it will not help anyone unless a critical mass of some body politic shares that perception. Conversely, if a critical mass of a body politic believes in the need to implement a certain reform, it will happen, even if I am ignorant of it or completely opposed to it.

We can extend this analysis to our worst enemies:


 * It is in our best interest to help people supporting our worst enemies get information they want, independent of controls that people with power exercise over nearly all major media today: If our actions reduce the ability of their leaders to censor their media (and of our political and economic leaders to censor ours), the information everyone gets should make it harder for leaders to convince others to support measures contrary to nearly everyone's best interests.

What kinds of data can we collect and analyze to evaluate who benefits and who loses from alternative interventions attempting to improve the media? See below.

World Bank on the value of information
In 2002 the President of the World Bank, James Wolfensohn, wrote, "[A] free press is not a luxury. It is at the core of equitable development. The media can expose corruption. ... They can facilitate trade [and bring] health and education information to remote villages ... . But ... the independence of the media can be fragile and easily compromised. All too often governments shackle the media. Sometimes control by powerful private interests restricts reporting. ... [T]o support development, media need the right environment—in terms of freedoms, capacities, and checks and balances."

This article includes proposals for evaluating alternative ways of improving the media and circumstances under which they may or may not be effective.

US Postal Service Act of 1792: a natural experiment
McChesney and Nichols (2010, 2016) suggested that the US Postal Service Act of 1792 made a major contribution to making the US what it is today. Under that act, newspapers were delivered up to 100 miles for a penny, when first class postage was between 6 and 25 cents depending on distance. McChesney and Nichols estimated that between 1840 and 1844, the US postal subsidy was 0.211% of GDP with federal printing subsidies adding another 0.005%, totaling 0.216% of GDP. That percent of GDP would be roughly $140 per person per year in 2019. We use 2019 dollars here to make it easy to compare with Rolnik et al. (2019), who recommended $50 per adult per year, which is roughly 0.06% of US GDP. Rolnik et al. added that the level of subsidies would require "extensive deliberation and experimentation". More recently McChesney and Nichols have recommended 0.15% of GDP ($98 per person per year in 2019), considering the fact that the advent of the Internet has nearly eliminated the costs of printing and distribution.

Tocqueville, who visited the US in 1831, observed the following:
 * [T]he liberty of the press does not affect political opinion alone, but extends to all the opinions of men, and modifies customs as well as laws. ... I approve of it from a consideration more of the evils it prevents, than of the advantages it insures.
 * The liberty of writing ... is most formidable when it is a novelty; for a people who have never been accustomed to hear state affairs discussed before them, place implicit confidence in the first tribune who presents himself. The Anglo-Americans have enjoyed this liberty ever since the foundation of the Colonies ... .  A glance at a French and an American newspaper is sufficient to show the difference ... .  In France, the space allotted to commercial advertisements is very limited, and the news-intelligence is not considerable; but the essential part of the journal is the discussion of the politics of the day.  In America, three-quarters of the enormous sheet are filled with advertisements, and the remainder is frequently occupied by political intelligence or trivial anecdotes:  it is only from time to time that one finds a corner devoted to passionate discussions, like those which the journalists of France every day give to their readers.
 * It has been demonstrated by observation, and discovered by the sure instinct even of the pettiest despots, that the influence of a power is increased in proportion as its direction is centralized.
 * [T]he number of periodical and semi-periodical publications in the United States is almost incredibly large. In America there is scarcely a hamlet which does not have its newspaper.
 * In the United States, each separate journal exercises but little authority; but the power of the periodical press is second only to that of the people ....



To what extent was Tocqueville's "incredibly large" "number of periodical and semi-periodical publications in the United States" due to the US Postal Service Act of 1792? To what extent did that "incredibly large" number of publications encourage literacy, limit political corruption, and help the US of that day remain together and grow both in land area and economically while contemporary New Spain, then Mexico, fractured, shrank, and stagnated economically? To what extent does the enormous power of the US today rest on the economic growth of that period and its impact on the political culture of that day continuing to the present? That growth transformed the US into the world leader that it is today; see Figure 1. In the process, it generated new technologies that benefit the vast majority of the world's population alive today. If the newspapers Tocqueville read made any substantive contribution to the growth summarized in Figure 1, the information in those newspapers were public goods potentially benefiting the vast majority of humanity (people who never read those newspapers and were unaware even of their existence) to the end of human civilization.

Experiments of the type discussed below can help quantify the magnitude of these suggested benefits in contemporary settings.

Other economists
We cannot prove that the diversity of newspapers in the early US contributed to the economic growth it experienced. Banerjee and Duflo (2019) concluded that no one knows how to create economic growth. They won the 2019 Nobel Memorial Prize in Economics with Michael Kremer for their leadership in using randomized controlled trials to learn how to reduce global poverty. More recently, Wake et al. (2021) found evidence that the economic costs of curbing press freedom persist long after such freedoms have been restored. And Mohammadi et al. (2022) found that economic growth rates were impacted by civil liberties, economic and press freedom and the economic growth rates of neighbors (spacial autocorrelation) but not democracy. These findings of Mohammidi et al. (2022) and Wake et al. (2021) reinforce Thomas Jefferson's 1787 comment that, "were it left to me to decide whether we should have a government without newspapers, or newspapers without a government, I should not hesitate a moment to prefer the latter."

To what extent might experiments like those recommended in this article either reinforce or refute this claim of Jefferson from 1787?

Randomized controlled trials to quantify the value of information
This article suggests randomized controlled trials to quantify the impact of citizen-directed subsidies for journalism, roughly following the recommendations of McChesney and Nichols (2021, 2022) to distribute some small percentage of GDP to local news nonprofits via local elections. Philanthropies could fund such experiments for some of the smallest and poorest places in the world. Organizations like the World Bank could fund such experiments as adjuncts to a random selection from other interventions they fund, justified for the same reason that they would not consider funding anything without appropriate accounting and auditing of expenditures, as discussed further below.

Before making suggestions regarding experiments, we review previous research documenting how information might be a public good.

Previous research
Before considering optimal level of subsidies for news, it may be useful to consider the research for which Daniel Kahneman won the 2002 Nobel Memorial Prize in Economics. Most important for present purposes may be that virtually everyone:


 * thinks they know more than they do (Overconfidence), and
 * prefers information and sources consistent with preconceptions. (Confirmation bias).

To what extent do media organizations everywhere exploit the confirmation bias and overconfidence of their audience to please those who control most of the money for the media, and to what extent might this reduce broadly shared economic growth? The proposed experiments should include efforts to quantify this, measuring, e.g., local incomes, inequality, political polarization and the impact of interventions attempting to improve such.

Plous wrote, "No problem in judgment and decision making is more prevalent and more potentially catastrophic than overconfidence." It contributes to inordinate losses by all parties in negotiations of all kinds including lawsuits, strikes, financial market bubbles and crashes, and politics and international relations, including wars.

Might the frequency and expense of lawsuits, strikes, financial market volatility, political corruption and wars be reduced by encouraging people to get more curious and search more often for information that might contradict their preconceptions? Might such discussions be encouraged by interventions such as increasing the total funding for news through many small, independent, local news organizations? If yes, to what extent might such experimental interventions threaten the hegemony of major media everywhere while benefiting everyone, with the possible exception of those who benefit from current systems of political corruption?



One attempt to quantify this appears in Figure 2, which summarizes a natural experiment on the impact of government subsidies for public media on public knowledge of domestic and international politics: Around 2008 the governments of the US, UK, Denmark and Finland provided subsidies of $1.35, $80, $101 and $101 per person per year, respectively, for public media. A survey of public knowledge of domestic and international politics found that people with college degrees seemed to be comparably well informed in the different countries, but people with less education were better informed in the countries with higher public subsidies.

Kaviani et al. (2022) studied the impact of "the staggered expansion of Sinclair Broadcast Group: the largest conservative network in the U.S." They documented a decline in Corporate Social Responsibility (CSR) ratings of firms headquartered in Sinclair expansion areas. They also documented a "right-ward ideological shift" in coverage that was "nearly one standard deviation of the ideology distribution" as well as "substantial decreases in coverage of local politics substituted by increases in national politics." Ellison (2024) said that "Sinclair's recipe for TV news" includes an annual survey asking viewers, "What are you most afraid of?" Sinclair reportedly focuses on that while implying in their coverage "that America's cities, especially those run by Democratic politicians, are dangerous and dysfunctional."

Sources in France are concerned that billionaire Vincent Bolloré has purchased a substantial portion of French media and used it effectively to promote the French far right. Scheidler (2024a) reported that the concentration of ownership the German media "has not yet reached the extreme forms observed in France, the United Kingdom or the United States, but the process of consolidation initiated several decades ago has transformed a landscape renowned for its decentralization." Scheidler (2024b) reported that there still exists a wide range of constructive media criticism in Germany, but it gets less coverage than before in the increasingly consolidated major media. This has driven many who are not happy with these changes to alternative media such as Die Tageszeitung, founded in 1978. In Israel, a 2022 survey found that the leading newspaper was Israel Hayom, founded in 2007 by billionaire Sheldon Adelson and distributed for free allegedly to skirt Israel's campaign finance laws. Various sources have suggested that it has encouraged Israeli policies that have driven many Palestinians to support Hamas, leading to the current Israel-Hamas war, discussed further below.

Benton wrote that past research has shown that strong local newspapers "increase voter turnout, reduce government corruption, make cities financially healthier, make citizens more knowledgeable about politics and more likely to engage with local government, force local TV to raise its game, encourage split-ticket (and thus less uniformly partisan) voting, make elected officials more responsive and efficient ... And ... you get to reap the benefits of all those positive outcomes even if you don’t read them yourself."

We feel a need to repeat that last comment: Benton says that we all benefit from others consuming news that we do not, because they become less likely to be stampeded into voting contrary to their best interests—and ours—and more likely to lobby effectively against questionable favors to major political campaign contributors or other people with power, underreported by major media that have conflicts of interest in balanced coverage of anything that might offend people with substantive control of their funding. That suggests that everyone might benefit from subsidizing a broad variety of independent local news outlets consumed by others.

Part of the mechanism here was documented by Trexler and Mullen (2024). They found that, "a few extra paragraphs of context increased support for spending and increased voters' willingness to hold local politicians accountable for infrastructure neglect by voting them out of office. ... With fewer reporters staffing newsrooms, the depth of reporting on invisible infrastructure declines."

Experiments along the lines discussed below could attempt to evaluate these claims and estimate their magnitudes.

How fair is the US tax system?
How fair is the US tax system? It depends on who is asked and how fairness is defined.



The Tax Foundation computed the effective tax rate in different portions of the distribution of Adjusted Gross Income (AGI), plotted in Figure 3. They noted that,"half of taxpayers paid 99.7 percent of federal income taxes". The effective tax rate on the 1% highest adjusted gross incomes (AGIs) was 26%, almost double (1.91 times) the average, while the effective tax rate for the bottom half was 3.1%, only 23% of the average.

The Tax Foundation did not mention that we get a very different perspective from considering gross income rather than AGI. Leiserson and Yagan (2021) estimated that the average effective federal individual income tax rate paid by America’s 400 wealthiest families was between 6 and 12% with the most likely number being 8.2%. The difference comes in the adjustments, while the uncertainty comes primarily from appreciation in the value of unsold stock, which is taxed at a maximum of 20% when sold and never taxed if passed as inheritance.

Divergent claims about business taxes can similarly be found. Watson (2022) claimed that, "Corporate taxes are one of the most economically damaging ways to raise revenue and are a promising area of reform for states to increase competitiveness and promote economic growth, benefiting both companies and workers." This "economically damaging" claim seems contradicted by evidence documented with Figure 1 above: When the top tax rate was cut in half, the rate of economic growth in the US fell by half, according to Piketty (2021, p. 139).



One reference on the difference between "adjusted" and "gross" income is US federal tax code and regulations, which grew from 1.4 million words in 1955 to over 10 million in 2015, averaging 145,000 additional words each year; see Figure 4.

How does this relate to media?

How do media organizations make money?
Media organizations everywhere sell changes in audience behaviors to the people who give them money. If they do not have an audience, they have nothing to sell. If they sufficiently offend their funders, they will not get the revenue needed to produce content.

The major media in the US have conflicts of interest in honestly reporting on discussions in congress on copyright law or on anything that might impact a major advertiser or might make it easier for politicians to get elected by spending less money on advertising. McChesney (2015) insisted that the major media are not interested in providing information that people want: They are interested in making money and protecting the interests of the ultra-wealthy, who control the largest advertising budgets. For example, media coverage of the roughly 40,000 people who came to Seattle in 1999 to protest the WTO Ministerial Conference there and the 10,000 - 15,000 who came to Washington, DC, the following year to protest the International Monetary Fund and the World Bank, included "some outstanding pieces produced by the corporate media, but those were exceptions to the rule. ... [T]he closer a story gets to corporate power and corporate domination of our society, the less reliable the corporate news media is."

Aaron (2021) said, "Bob McChesney ... taught me [to] look at ... the stories that are cheap to cover." Between around 1975 and 2000, the major commercial broadcasters in the US fired nearly all their investigative journalists and replaced them with the police blotter. It's easy and cheap to repeat what the police say. A news outlet can do that without seriously risking loss of revenue. In addition, poor defendants who may not have money for legal defense rarely have money to sue a media outlet for defamation. By contrast, a news report on questionable activities by a major funder risks both direct loss of advertising revenue and being sued. These risks impose a higher standard of journalism (and additional costs) when reporting on questionable activities by people with power than when reporting on poor people. This is a much bigger problem in countries where libel is a criminal rather than a civil offense or where truth is not a defense for libel.



After about 1975 television audiences in the US noticed increased coverage of crime in the broadcast news and concluded that crime was out of control, when there had been no substantive change in crime. They voted in a generation of politicians, who promised to get tough on crime. The incarceration rate in the US went from 0.1% to 0.5% in the span of roughly 25 years, after having been fairly stable for the previous 50 years; see Figure 5.



Around that same time, income inequality in the US began to rise; see Figure 6. To what extent might that increase in inequality be due to the structure of the major media?

To what extent might you and I benefit from making it easier for millions of others to research different aspects of government policies including the "adjustments" in the US tax system embedded in the over 10 million words of US federal tax code and regulations documented with Figure 4 above, encouraging them to lobby the US Congress against the special favors granted to major political campaign contributors against the general welfare of everyone else? Everyone except possibly the beneficiaries of such political corruption would likely benefit from the news that helps concerned citizens lobby effectively against such corruption, even if we did not participate in such citizen lobbying efforts and were completely ignorant of them.

Media and war

 * "You've got to be taught to hate and fear. ... It's got to be drummed in your dear little ear."
 * -- Lt. Cable in the 1949 Rodgers and Hammerstein musical South Pacific.

To what extent is it accurate to say


 * before anyone is killed in armed hostilities, the different parties to the conflict are polarized by the different media the different parties find credible?

This might seem obvious, but how can we quantify political polarization in a way that allows us to (a) model its relationship with the severity of conflicts and (b) use it to evaluate the effectiveness of efforts to reduce polarization?

The International Panel on the Information Environment (IPIE) is a consortium of over 250 global experts developing tools to combat political polarization driven by the structure of the Internet. The US Institute of Peace (2016) discusses "Tools for Improving Media Interventions in Conflict Zones". Previous research in this area was summarized by Arsenault et al. (2011). One such tool might be video games.

We suggest experimenting with interventions designed to reduce political polarization with some of the smallest but most intense conflicts: Interventions that require money could be more effectively tested with smaller, high intensity conflicts. It might be easier to measure a reduction from a higher-intensity conflict. And a given budget applied to a smaller population would mean more money per capita, which might produce a larger effect.

The Armed Conflict Location and Event Data Project (ACLED) tracks politically relevant violent and nonviolent events by a range of state and non-state actors. Their data can help identify countries or geographic regions in conflict as entities to be assigned randomly to experimental and control groups, whose comparison can provide high quality data to help evaluate the impact of any intervention. Might it be feasible to involve anti-war groups in the US in generating increased supply and demand for news regarding a random sample of places where the US is "invisibly" making war, according to Solomon (2023), to see if that can change the dynamics enough to reduce the level and lethality in experimental vs. control conflicts? Initial experiments of this nature might be done with a modest budget by working with organizations advocating nonviolence and with religious groups to recruit diaspora communities to do things recommended by experts in IPIE while also lobbying governments for funding. (One non-financial obstacle to such experiments is the US Supreme Court decision in Holder v. Humanitarian Law Project, which concluded that providing training to anyone designated as a "terrorist" by the US State Department is "providing material support to terrorism", a crime under the US Patriot Act.)

Any success might be leveraged into changes in foreign and military policies to make the world safer for all, though it could also increase resistance from people who may believe they benefit from the turmoil and oppose conflict resolution, often clandestinely.

Before discussing such experiments further, we consider a few case studies.

Russo-Ukrainian War, the US Civil War, World War I
In the Russo-Ukrainian War, Halimi and Rimbert (2023) describe "Western media as cheerleaders for war". Stiglitz (2002) noted this was a general phenomenon: "In periods of perceived conflict ... a combination of self-censorship and reader censorship may also undermine the ability of a supposedly free press to ensure democratic transparency and openness." Media organizations do not always do this solely to please their funders. Reporters are killed or jailed and news outlets closed to prevent them from disseminating information that people with power do not want distributed. Early in the Civil War in the US (1861-1865), some newspapers in the North said the US should let the South secede, because that would be preferable to war. Angry mobs destroyed some offices and printing presses. One editor "was forcibly taken from his house by an excited mob, ... covered with a coat of tar and feathers, and ridden on a rail through the town." Others changed their policies "voluntarily", recognizing threats to their lives or property or to a loss of audience.

Hochschild (2022) reported a substantial shift in the media ecology in the US following the approval of the Espionage Act of 1917. That act gave the Postmaster General the authority to declare as "unmailable" any publication that he decided might interfere with the armed forces of the US. He used this to effectively terminate many publications, because there was no other way to distribute publications nationally at that time (p. 61). The lack of broad discourse in the media allegedly amplified war hysteria, under which many people were persecuted, beaten, robbed, incarcerated, and even killed with impunity for peaceably assembling and petitioning for better wages and working conditions, or for speaking German. That media bias reportedly continued after the war to help major capitalists suppress labor organizers.

Stalin and Putin


A 2017 survey asked Russians to name 10 of the world’s most prominent personalities. The top two were Joseph Stalin and Vladimir Putin with 38% and 34%, respectively. When the study was redone in 2021, Putin had slipped from number 2 to number 5. Stalin still led with 39% followed by Vladimir Lenin with 30%, Poet Alexander Pushkin and tsar Peter the Great with 23% and 19% each, then Putin with 15%.

It may be difficult for some people in the West to understand how Stalin and Putin could be so popular, given the way they have been typically described in the mainstream Western media. However, this is relatively easy to understand just by looking at the accompanying plot (Figure 7) of average annual income in that part of the world between 1860 and 2022: Both Stalin and Putin inherited economies that had fallen dramatically in the previous years and had supervised dramatic improvements. Putin's decline between 2017 and 2021 may also be understood from this plot, because it shows how the dramatic growth that began around the time that Putin became acting President of Russia has slowed substantially since 2012.

Hitler
Fulda (2009) studied the co-evolution of newspapers and party politics in Germany, focusing primarily on Berlin, 1924-1933. During that period, the Nazis (NSDAP, National Socialist German Workers' Party, Nationalsozialistische Deutsche Arbeiterpartei) grew from 2.6% of the votes for the Reichstag (German parliament) in 1928 to 44% in 1933. Fulda described exaggerations in the tabloid press of an indecisive government incapable of managing either the economy or the increasing political violence, blamed excessively on Communists, and the potential for civil war. This turned the Nazis into an attractive choice for voters desperate for decisive action.

After the 1933 elections, the Reichstag passed the Enabling Act of 1933, which gave Hitler's cabinet the right to enact laws without the consent of parliament. The Nazis then began full censorship of the newspapers, physically beating, imprisoning and in some cases killing journalists, as the leading publishers acquiesced. The primary sources of news during that period were newspapers; radio was relatively new in Germany and carried very little news. Many newspapers were tabloids, interested in either making money or promoting a party line with minimal regard for fact checking. A big loser in this was the right‐wing press magnate Alfred Hugenberg, whose political mismanagement led to the substantial demise of his German National People's Party (Deutschnationale Volkspartei, DNVP), mostly benefitting Hitler.

This suggests the need for a counterfactual analysis of this period, asking what kinds of changes in the structure of the media ecology might have prevented the rise of the Nazis? In particular, to what extent might a more diverse local news environment supported by citizen-directed subsidies as suggested herein have reduced the risk of a demise of democracy? And might some sort of fairness doctrine have helped? And how might different rules for distributing different levels of funding to local news outlets impact the level of democratization? (Threats to democracy include legislation like the German Enabling Act of 1933 and other situations that allow an executive to successfully ignore the will of an otherwise democratic legislature, a self-coup, as well as a military coup.)

Iraq and the Islamic State
In 2014 in Mosul, two Iraqi army divisions totaling 30,000 and another 30,000 federal police were overwhelmed in six days by roughly 1,500 committed Jihadists, according to some accounts. Four months later, Reuters reported that, "there were supposed to be close to 25,000 soldiers and police in the city; the reality ... was at best 10,000." Many of the missing 15,000 were "ghost soldiers" kicking back half their salaries to their officers. Also, "[i]nfantry, armor and tanks had been shifted to Anbar, where more than 6,000 soldiers had been killed and another 12,000 had deserted."

To what extent might the political corruption and low morale documented in that Reuters report have been allowed to grow to that magnitude if Iraq had had a vigorous adversarial press, as discussed in this article? Instead, Paul Bremer, who was appointed as the Provisional coalition administrator of Iraq just over a week after President George W. Bush's Mission Accomplished speech of 2003-05-01, imposed strict press censorship. McChesney and Nichols contrasted this with General Eisenhower, who "called in German reporters [after the official surrender of Nazi Germany in WW II] and told them he wanted a free press. If he made decisions that they disagreed with, he wanted them to say so in print."

Israel-Palestine

 * Those who make peaceful revolution impossible will make violent revolution inevitable.
 * -- John F. Kennedy (1962)

To what extent is the Israeli–Palestinian conflict driven by differences in the media consumed by the different parties to that conflict?


 * To what extent are the supporters of Israel aware of violent acts committed by Palestinians but are unaware of the actions by Israelis that may have motivated those violent acts?
 * Similarly, to what extent are the supporters of Palestinians unaware of how Palestinian violence, even throwing rocks, may have reinforced long-standing fears among supporters of Israel and contributed to the repression the Palestinians have experienced?

To what extent are these differences in perceptions between supporters of Israel and supporters of Palestinians driven by differences in the media each find credible?

What can be done to bridge these gaps?

Gene Sharp, Mubarak Awad, and other supporters of nonviolence have suggested that when nonviolent direct action works, it does so by exposing a gap between the rhetoric [supported by the major media] and the reality of their opposition. Over time, this gap erodes pillars of support of the opposition. One example was the nonviolence of the First Intifada (1987-1993), which were protests against "beatings, shootings, killings, house demolitions, uprooting of trees, deportations, extended imprisonments, and detentions without trial." During that campaign, Israel got substantial negative international press for mistreating nonviolent protesters, killing some, breaking bones of others. Several thousand Israeli military refused to serve in the occupied territories. A hundred were court-martialed and imprisoned. Yitzhak Rabin, Israeli Defense Minister when the Intifada began, reportedly felt that the large number of military refusing to follow orders was threatening the ability of the Israeli Defense Forces to function. He resigned as Defense Minister and ran for Prime Minister on a platform of negotiating with Palestinians. His victory and subsequent negotiations led to the Oslo Accords and the joint recognition of each other by the states of Israel and Palestine. The West Bank and Gaza have continued under Israeli occupation since then with some services provided by the official government of Palestine.

During the Intifada, Israel tried to infiltrate the protesters with agents provocateurs in Palestinian garb. They were exposed and neutralized until Israel deported 481 people and imprisoned tens of thousands of others suspected of organizing the nonviolence. Finally, they got the violence needed to justify a massively violent repression of the Intifada.

The general thrust of this current analysis suggests a two pronged intervention to reduce the risk of a continuation of the violence that has marked Israel-Palestine since at least 1948:
 * 1) Offer nonviolence training to all Palestininans, Israelis and supporters of either interested in the topic. This is the opposite of the policies Israel pursued during the First Intifada, at least according to the references cited in this discussion of that campaign.
 * 2) Provide citizen-direct subsidies to local news nonprofits in the West Bank and Gaza at, e.g., 0.15% of GDP, as recommended by McChesney and Nichols, cited above and discussed further below.

How can we evaluate the budget required for such an experiment? The nominal GDP of the State of Palestine in 2021 was estimated at $18 billion; 0.15% of that is $27 million. Add 10% for research to get $30 million per year. That annual cost for the media component of this proposed intervention is 12% of the billion Israeli shekels ($246 million) that the Gaza war was costing Israel each day in the early days of the Israel-Hamas war, according to the Israeli Finance Minister on 2023-10-25. As this is being written, that war has continued for over 100 days. If the average daily cost of that war to Israel during that period has been $246 million, then that war will have already cost Israel over $24.6 billion. And that does not count the loss of lives and the destruction of property.

How much would training in nonviolence cost? That question would require more research, but if it were effective, the budget would seem to be quite modest compared to the cost of war, even if it were several times the budget for citizen-directed subsidies for local news in Palestine as just suggested.

The decline of newspapers


McChesney and Nichols (2022) noted that US newspaper revenue as a percent of GDP fell from over 1% in 1956 to less than 0.1% in 2020; see Figure 8. Abernathy (2020) noted that the US lost more than half of all newspaper journalists between 2008 and 2018. A quarter of US newspapers closed between 2004 and 2020, and many that still survive are publishing less, creating "news deserts" and "ghost newspapers", some with no local journalists on staff. Abernathy (2020) documented the problem of increasing "news deserts and ghost newspapers" in the US. A local jurisdiction without a local news outlet has been called a "news desert". She uses the term "ghost newspapers" to describe outlets "with depleted newsrooms that are only a shadow of their former selves." Some “ghost newspapers” continue to publish with zero local journalists, produced by reporters and editors that don't live there. One example is the Salinas Californian, a 125-year-old newspaper in Salinas, California, which lost its last paid journalist 2022-12, according to the Los Angeles Times (2023). They continue to publish, though "The only original content from Salinas comes in the form of paid obituaries, making death virtually the only sign of life at an institution once considered a must-read by many Salinans." A leading profiteer in this downward spiral is reportedly hedge fund Alden Global Capital. Threisman (2021) reported that, "When this hedge fund buys local newspapers, democracy suffers". And Benton (2021) said, "The vulture is hungry again: Alden Global Capital wants to buy a few hundred more newspapers". Hightower (2023) describes two organizations fighting this trend. One is National Trust for Local News, a nonprofit that recently bought several local papers and "is turning each publication over to local non-profit owners and helping them find ways to become sustainable." The other is CherryRoad Media, which "bought 77 rural papers in 17 states, most from the predatory Gannett conglomerate that wanted to dump them", and is working to "return editorial decision-making to local people and journalists ... and ... reinvest profits in real local journalism that advances democracy." News outlets acquired by something like the National Trust for Local News should be eligible for citizen-directed subsidies for local news, as discussed below, after their ownership was officially transferred to local humans. Outlets acquired by organizations like CheeryRoad Media would not be eligible as long as they remained subsidiaries.

The concerning trend documented by Abernathy continues. The Fall 2023 issue of Columbia Journalism Review reported that 2023 "has become media’s worst year on record for job losses". Substantial advertising revenue has shifted to the "click economy", where advertisers pay for clicks, especially on social media.

Newspapers in other parts of the world have also experienced substantial declines in revenue. Some legislative bodies have attempted to tax Internet companies to replace some of the ad revenue lost by newspapers. In 2013 German law was changed to include "Ancillary copyright for press publishers", also called a "link tax". However, this law was declared invalid in 2019 by the European Court of Justice (ECJ), because it had not been submitted in advance to the EU Commission, as required. Before that ECJ decision, Google had removed newspapers from Google News in Germany. German publishers then reached an agreement with Google after traffic to their websites plummeted. Building on that and similar experience in Spain, in 2019 the European Union adopted a Directive on Copyright in the Digital Single Market. A similar link tax proposal in Canada led Meta, the parent company of Facebook, to withdraw news from Canada, and Google agreed to 'pay about $100 million a year into a new fund to support "news"' in Canada. As of 2023-11-30, California was still considering a link tax.

In Germany, Flößer (2024) found that the right-wing populist, political party in Germany, Alternative for Germany (Alternative für Deutschland, AfD), did better in places with no local newspaper.

Threats from social media
The growth of social media has been wonderful and terrible. It has been wonderful in making it easier for people to maintain friendships and family ties across distances.

But it has also been terrible as "antisocial media" have been implicated in the relatively recent rise in dysfunctional and counterproductive political polarization and violence. Ding et al. (2023) document, "Same words, different meanings" in their use by CNN and Fox News and how that has interacted with word usage on Twitter between 2010 and 2020 to increase political polarization, "impeding rather than supporting online democratic discourse."

Extreme examples of this increase have included violent efforts to prevent peaceful transitions of power in the US and Brazil. These changes even threaten the national security of the US and its allies, according to H. R. McMaster, President Trump's second national security advisor.

Various responses to these concerns have been suggested, beyond the recommendations of McChensey and Nichols. These include the following: :


 * Make internet companies liable for defamation in advertisements, similar to print media and broadcasting.
 * Tax advertising revenue received by large internet companies and use that to fund more local media.
 * Replace advertising as the source of funding for social media with subscriptions.

To these suggestions, we add the following:


 * Allow some of but not all citizen-directed subsidies for news to go to social media outlets, as suggested below.
 * Require all organizations whose income depends on promoting or "boosting" content, whether in advertisements or "underwriting spots" or clickbait, to provide copies of the ads, underwriting spots and clickbait to a central repository like the Internet Archive.
 * Use advertising to discuss overconfidence and encourage people to talk politics with humility and respect, recognizing that the primary differences they have with others may be the media they consume.

How to counter political polarization

 * "Never attribute to malice or stupidity that which can be explained by moderately rational individuals following incentives in a complex system."
 * -- Hubbard's corollary to Hanlon's razor.

More research is needed on how to counter the relatively recent increases in political polarization. For example, might some form of fairness doctrine help reduce political polarization? Conservative leaders in the US are vehemently opposed, insisting it would be an attack on First Amendment rights. However, as noted above, the tabloid media of Germany reportedly contributed to Hitler's rise to power between 1924 and 1933. How is the increase in political polarization since 1987 and 2004 different from the disregard for fairness of the news media that helped bring Hitler to power? Shouldn't conservatives and liberals be able to negotiate a fairness doctrine that targets unfair liberal media as well as unfair conservative media?

One example: The lawsuit Dominion Voting Systems v. Fox News Network was settled with Fox agreeing to pay Dominion $787.5 million while acknowledging that Fox had knowingly and intentionally made false and defamatory statements about Dominion to avoid losing audience to media outlets that continued to claim fraudulently that Donald Trump not Joe Biden had won the 2020 US presidential election. The settlement permitted Fox to avoid apologizing publicly, which could have threatened their audience share. That settlement was less than 6% of Fox's 2022 revenue of $14 billion. Evidently, if that decision made a difference of 6% in their audience ratings, Fox made money from defaming Dominion even after paying them $787.5 million. If so, it was a good business decision, especially since they did not have to publicly apologize. Fraud can be good business. Media executives could be fired if they lost money trying to protect democracy.

To what extent did Fox's lies about Dominion contribute to the mob attacks on the US Capitol on January 6, 2021, trying to prevent the US Congress from officially declaring that Joe Biden had won the 2020 elections? Experiments such as those proposed in this article would obviously not provide unequivocal answers to these specific questions, but they could help us understand the phenomena with quantified imprecision. Such results could help build a consensus for reforms that show a high probability of reducing the threats of political violence and increasing the prospects for broadly shared peace and prosperity for the long term.

McChesney and Nichols' Local Journalism Initiative
As noted above, McChesney and Nichols (2021, 2022) propose a "Local Journalism Initiative", distributing 0.15% of GDP to local news nonprofits via local elections. They based this partly on their earlier work suggesting that subsidies for newspapers in the US in 1840 was around 0.2% of GDP.

McChesney and Nichols' eligibility criteria
To be eligible, McChesney and Nichols say the recipient of such funds should satisfy the following:
 * Be a local nonprofit with at least six months of history, so voters could know their work.
 * Be locally based with at least 75% of salaries going to local residents.
 * Be completely independent, not a subsidiary of a larger organization.
 * Produce and publish original material at least five days per week on their website for free, explicitly in the public domain.
 * Each voter is asked to vote for at least three different local news outlets to support diversity.
 * No single news outlet should get more than 25% of that jurisdiction's annual budget for local news subsidies.
 * Each recipient of these subsidies should get at least 1% of the vote to qualify, or 0.5% of the vote in political jurisdictions with over 1 million people. Diversity and competition are crucial.
 * There will be no content monitoring: Government bureaucrats will not be allowed to decide what is "good journalism".  That's up to the voters.
 * Voting would be limited to those 18 years and older.

Alternatives
Some aspects of this might be relaxed for at least some political jurisdictions included in an experiment. For example, might it be appropriate to allow for-profit news outlets to compete for these subsidies as long as they meet the other criteria? However, we prefer to retain the rules requiring recipients to be local and completely independent, at least for many experimental jurisdictions. If citizen-directed subsidies for local news go to for-profit organizations, to what extent should their finances be transparent, e.g., otherwise complying with rules like those for 501(c)(3) nonprofits in the US? Might it also be appropriate to allow some portion of these funds to be distributed to noncommercial social media outlets that submitted, e.g., ads, underwriting, and click bait to a public, searchable database like the Internet Archive?

News written by people paid with these subsidies should be available under a free license like Creative Commons Attribution-ShareAlike (CC BY-SA) 4.0 international license but not necessarily in the public domain: Other media outlets should be free to further disseminate the news while giving credit to the organization that produced it.

Many countries have some form of community radio. Some of those radio stations include what they call news and / or public affairs, and some of those are made available as podcasts via the Internet. If their "news & public affairs" programs are subsequently posted to a website as podcasts, preferably accompanied by some text if not complete transcripts, under a license no more restrictive than CC BY-SA, that should make them eligible for subsidies under the criteria mentioned above if they add at least one new podcast of that nature five days per week. If the programming of this nature that they produce is not available on the web under an appropriate license, part of any experiments as discussed here might include offers to help such radio stations become eligible.

Might it be wise to allow children to vote for news organizations they like? Ryan Sorrell, founder and publisher of the Kansas City Defender, insists that, "young people ... are very interested in news. It just has to be produced and packaged the right way for them to be interested in consuming it". The French-language Topo presents news and complex issues in comic strip format. Their co-editor in chief insists, "there are plenty of ways to get young people interested in current affairs". Might allowing children to vote for news outlets increase their interest in learning and in civic engagement among both children and their caregivers? Might this translate into increased civic engagement after they become eligible to vote as adults, leading to reductions in political corruption and improvements in government long term? Should this be tested in some experimental jurisdictions?

Some of the money may go to media outlets that seem wacko to many voters. However, how different might that be from the current situation? Most importantly, if these subsidies have the effect that Tocqueville reported from 1831, they should be good for democracy and for broadly shared peace and prosperity for the long term: They could stimulate public debate, and wacko media might have less power than they currently do, with "each separate journal exercis[ing] but little authority; but the power of the periodical press [being] second only to that of the people."

Tocqueville's comparison of newspapers in France and the US in 1831 is echoed in Cagé's (2022) concern about "the Fox News effect" in the US and that of Bolloré in France. She cites research claiming that biases in Fox News made major contributions to electing Republicans in the US since 2000. These shifts, including changes by the conservative-leaning broadcasting company, Sinclair Broadcast Group, reportedly made a substantive contribution to the election of Donald Trump as US President in 2016, while a comparable estimate of the impact of changes in MSNBC "is an imprecise zero." In France, she provides documentation claiming that the media empire of French billionaire Vincent Bolloré has made a major contribution to the rise of far-right politician Éric Zemmour and is buying media in Spain. The pattern is simple: Fire journalists and replace them with talk shows, which are cheaper to produce and are popular, evidently exploiting confirmation bias and the overconfidence effect, as discussed above.

To what extent is the increase in political polarization since 1987 and 2004 due to increased concentration of ownership of both traditional and social media (and how those organizations make money selling changes in audience behaviors to the people who give them money)? To the extent that this increase in polarization has been driven by those changes in the media, citizen-directed subsidies for diverse news should reverse that trend. This hypothesis can be tested by experiments like those proposed herein.

Roadmap for local news
Green et al. (2023) describe "an emerging approach to meeting civic information needs" in a "Roadmap for local news". This report insists that society needs "civic information", not merely "news". It summarizes interviews with 51 leaders from nonprofit and commercial media across all forms of distribution (print, radio, broadcast, digital, SMS) in member organizations, news networks, news funders and researchers. They say that, "Rampant disinformation is being weaponized by extremists", and "Democratic participation and representation are under threat." They recommend four strategies to address "this escalating information crisis":


 * 1) Coordinate work around the goal of expanding “civic information,” not saving the news business;
 * 2) Directly invest in the production of civic information;
 * 3) Invest in shared services to sustain new and emerging civic information networks; and
 * 4) Cultivate and pass public policies that support the expansion of civic information while maintaining editorial independence.

Part of the motivation for this article on "Information is a public good" is the belief that solid research on the value of such interventions should both (a) make it easier to get the funding needed, and (b) help direct the funding to interventions that seem to make the maximum contributions to improving broadly shared peace and prosperity for the long term at minimum cost.

Budgets for experiments
What factors should be considered in evaluating budgets for experiments to estimate the impact of citizen-directed subsidies for news?



Rolnik et al. (2019) suggested that $50 per adult, roughly 0.06% of US GDP, might be enough. However, that's a pittance compared to the revenue lost by newspapers in the US since 1955, as documented in Figure 8 above. It's also a pittance compared to the money spent on advertising (see Figure 9): Can we really expect local media funded with only 0.06% or 0.15% of GDP to compete with media funded by 2% of GDP? Maybe, but that's far from obvious.

Might it be prudent to fund local journalism in some experimental jurisdictions at levels exceeding the money spent on advertising, i.e., at roughly 2% of GDP or more? If information is a public good, as suggested by the research summarized here, then such high subsidies would be needed in some experimental jurisdictions, because the maximum of anything (including net benefits = benefits minus costs) cannot be confidently identified without conducting some experiments beyond the point of diminishing returns.



Also, news might serve a roughly comparable function to accounting and auditing, as both help reduce losses due to incompetence, malfeasance and fraud. Two points on this:


 * 1) CONTROL FRAUDS:  Black (2013) noted that many heads of organizations can find accountants and auditors willing to certify accounting reports they know to be fraudulent. For example, on 2024-05-03 BF Borgers and its owner, Benjamin Borgers, agreed to pay a $14 million fine and stop working as an accountant for having committed “massive” fraud in more than 1,500 SEC filings involving over 500 public companies, including Former President Trump's Truth Social. Black calls such executives "control frauds." Primary protections against these kinds of problems are vigorous, independent journalists and more money spent on independent evaluations beyond the control of such executives.  In this regard, we note two major differences between the Savings & Loan scandal of the late 1980s and early 1990s and the international financial crisis of 2007-2008: First the major banks by 2007 were much bigger and controlled much larger advertising budgets than the Saving & Loan industry did 15-20 years earlier.  This gave major media a much bigger conflict of interest in honestly reporting on questionable activities of these major accounts. Second, major banks had made substantial political campaign contributions to much larger portions of both the US House and Senate. Might the larger campaign contributions have been as effective if the major media had not had such conflicts of interest? To what extent might this corrosive impact be estimated in experimental polities?
 * 2) ADEQUATE RESEARCH OF OUTCOMES:  Many nonprofits and governmental agencies officially have outcome measures, but many of those measures tend to be relatively superficial like the number of people served.  It's much harder to evaluate the actual benefits to the people served and to society.  For example, the Perry Preschool and Abecedarian programs divided poor children and caregivers into experimental and control groups and followed them for decades to establish that their interventions were enormously effective. Meanwhile, US President Lyndon Johnson's Great Society programs, and Head Start in particular, did not invest as heavily in research.  That lack of documentation of results made them relatively easy targets for political opponents claiming that government is the problem, not the solution. These counter arguments were popularized by US President Ronald Reagan and UK Prime Minister Margaret Thatcher to justify reducing or eliminating government funds for many such programs. Banerjee and Duflo (2019) summarized relevant research in this area by saying that the programs were not the disaster that Reagan, Thatcher, and others claimed, but they were also not as efficient and effective as they could have been, because many local implementations were underfunded, poorly managed and poorly evaluated.   Bedasso (2021) analyzed World Bank projects completed from 2009 to 2020, concluding that high quality monitoring and evaluation on average made a major contribution to the positive results from the successful projects studied. To what extent might citizen-directed subsidies for local media as suggested here improve the demand for (and the supply of) better evaluations, leading to better results?  To what extent might these effects be estimated using randomized controlled trials comparing different jurisdictions, analogous to the research for which Banerjee, Duflo, and Kramer won the 2019 Nobel Memorial Prize in Economics?

This discussion makes us wonder if better research and better news might deliver dramatically more benefits than costs in reducing money wasted on both funding wasteful programs and on failing to fund effective ones? In particular, might society benefit from matching the 1% of the workforce occupied by accountants and auditors with better research and citizen-directed subsidies for news (see Figure 10)? If, for example, 1% or 2% of GDP distributed to local news nonprofits via local elections, as described above, increased the average rate of economic growth in GDP per capita by 0.1 percentage point per year, that increase would accumulate over time, so that after 10 or 20 years, the news would in effect become free, paid by money that implementing political jurisdictions would not have without those subsidies. Moreover those accumulations might remain as long as they were not wiped out by events comparable to the economic disasters documented above in discussing "Stalin and Putin" -- and maybe not even then as suggested by the economic rebounds documented in Figure 7.

Other recommendations and natural experiments
Table 1 compares the recommendations of McChesney and Nichols (2021, 2022) and Rolnik et al. (2019) with other possible points of reference. Karr (2019) and Karr and Aaron (2019) recommend "a 2 percent ad tax on all online enterprises that in 2018 earned more than $200 million in annual digital-ad revenues". They claim that this "would yield more than $1.8 billion a year", which is very roughly 0.008% of GDP, $5 per person per year; Google has negotiated agreements similar to this with the governments of Australia and Canada. Other points of reference include the percent of GDP devoted to accounting and auditing and advertising. As displayed in Figure 10, accountants and auditors are roughly 1% of the workforce in the US. It's not clear how to translate that into a percent of GDP, but 2% seems like a reasonable approximation, if we assume that the average income of accountants and auditors is a little above the national norm and overhead is not quite double their salaries; this may be conservative, because many accountants and auditors have support staff, who are not accountants but support their work.

Another point of reference is the average annual growth rate in GDP per capita in the US since World War II: A subsidy of 2% of GDP would be roughly one year's increase in average annual income since World War II, as noted with Figure 1 above. More precisely, the US economy (GDP per capita adjusted for inflation) grew at 2.3% per year between 1950 and 1990 but only 1.3% between 2008 and 2023. Inequality expert Thomas Piketty attributed that slowing in the rate of economic growth to the increase in income inequality in the US since 1975, documented in Figure 6 above. Whether Piketty is correct or not, if 2% per year subsidies for journalism closes the gap between 1.3% and 2.2% per year, those media subsidies would effectively become free after a little more than two years, paid out of income the US would not have without them. This reinforces the main point of this essay regarding the need for randomized controlled trials on any intervention with a credible claim to improving the prospects for broadly shared economic growth for the long term. This table includes other interventions for which humanity would benefit from better evaluations of their impacts. This includes Seattle's "Democracy Voucher" program, which gives each registered voter four $25 vouchers, totaling $100, which they could give to eligible candidates running for municipal office. However, only the first 47,000 were honored; this limited the city's commitment to $4.7 million every other year. If Seattle can afford $100 per registered voter, many other governmental entities can afford something very roughly comparable for each adult in their jurisdiction. Seattle's "democracy vouchers" are used to fund political campaigns, not local media; they are mentioned here as a point of comparison.

Other interventions that deserve more research than we've seen are the New Jersey Civic Information Consortium (NJCIC) and a program in California to improve local news in communities in dire need of strong local journalism. The NJCIC was initially funded at $500,000, which is only 0.00008% of New Jersey's 2020 economy (GDP) of $630 billion.

In 2022, the state of California authorized $25 million for up to 40 Berkeley local news fellowships offering "a $50,000 annual stipend [for 3 years] to supplement their salaries while they work in California newsrooms covering communities in dire need of strong local journalism." This Berkeley program is roughly $0.21 per person per year, roughly 0.0002% of the Gross State Product. A similar project in Indiana funded by philanthropies began as the Indiana Local News Initiative and has morphed into Free Press Indiana.

Some local Leagues of Women Voters have all-volunteer teams who observe official meetings of local governmental bodies and write reports. The City Bureau nonprofit news organization in Chicago, Illinous, "trains and pays community members to attend local government meetings and report back on them." The program has been so successful, it has expanded to other cities.

For an international comparison, we include amaBhungane, whose investigative journalism exposed a corruption scandal that helped force South African President Jacob Zuma to resign in 2018; amaBhungane's budget is very roughly one penny US per person per year in South Africa, 0.0002% of GDP. To the extent that this essay provides a fair and balanced account of the impact of journalism on political economy, South African and the rest of the world would likely benefit from more funding for amaBhungane and other comparable investigative journalism organizations. This could initially include randomized controlled trials involving citizen-directed subsidies for local news outlets in poor communities in South Africa and elsewhere, as we discuss further in the rest of this essay. Without such experiments, we are asking for funds based more on faith than science.

Table 1. Media subsidies and other points of reference. At the low end, political corruption exposed in part by amaBhungane forced the resignation in 2018 of South African President Zuma on a budget that's very roughly one penny US per person per year. If much higher subsidies of 1% of GDP restored an annual growth rate of 2.2% per year to the US from the more recent 1.3% discussed with Figure 1 above, those subsidies would pay for themselves from slightly more than one year's growth that the US would not otherwise have.

Other factors
We feel a need here to suggest other issues to consider in designing experiments to improve the political economy: education, empowering women, free speech, free press, peaceful assembly, and reducing political polarization.

EDUCATION: Modern research suggests that society might have lower crime and faster rates of economic growth with better funding for and better research on quality child care from pregnancy through age 17.

EMPOWERING WOMEN: Might the best known way to limit and reverse population growth be to empower women and girls? Without that, might the human population continue to grow until some major disaster reduces that population dramatically?

FREE SPEECH, FREE PRESS, PEACEFUL ASSEMBLY: Verbitsky said, "Journalism is disseminating information that someone does not want known; the rest is propaganda." Discussion of threats, arrests, kidnappings, and murders of journalists and violent suppression of peaceful assemblies encourages us to consider the potential utility of efforts to improve local news, as noted by contributors to Islam et al., eds. (2002), cited above. Data on such problems should be considered in selecting sites for experiments with citizen-directed subsidies for journalism and in analyzing the results from such experiments. Such data should include the incidence of legal proceedings against journalists and publishers as well as threats, murders, etc., in jurisdictions comparable to experimental jurisdictions. Before providing external funding to improve local journalism, it may be wise to review with government officials in candidate polities the history of attacks on journalists in their jurisdictions, continue collecting data on that during interventions, and use such data in analyzing the results.

REDUCING POLITICAL POLARIZATION: What interventions might be tested that would attempt to reduce political polarization while also experimenting with increasing funding for news through small, diverse news organizations? For example, might an ad campaign feature someone saying, "We don't talk politics", with a reply, "We have to talk politics with humility and mutual respect, because the alternative is killing people over misunderstandings"? Might another ad say, "Don't get angry: Get curious"? What can be done to encourage people to get curious rather than angry when they hear something that contradicts their preconceptions? How can people be encouraged to talk politics with humility and respect for others, understanding that everyone can be misinformed and others might have useful information?

FOCUS ON POLITICIANS: Mansuri et al. (2023) randomly assigned presidents of village governments in the state of Tamil Nadu in India to one of three groups with (1) a financial incentive or (2) a certificate with an information campaign (without a financial incentive) for better government or (3) a control group. They found that the public benefitted from both the financial and non-financial incentives, and the non-financial incentives were more cost effective. Might it make sense in some experimental jurisdictions to structure the subsidies for local news by asking voters to allocate, e.g., half their votes for local news to outlet(s) that they think provide the best information about politicians with the other half based on "general news"?

PIGGYBACK ON COMMUNITY AND LOCAL DEVELOPMENT PROGRAMS: The World Bank (2023) notes that, "Experience has shown that when given clear and transparent rules, access to information, and appropriate technical and financial support, communities can effectively organize to identify community priorities ... . The World Bank recognizes that CLD [Community and Local Development]  approaches and actions are important elements of an effective poverty-reduction and sustainable development strategy." This suggests that experiments in citizen-directed subsidies for news might best be implemented as adjuncts to other CLD projects to improve "access to information" needed for success. Experiments like those described herein should provide an independent assessment of the value of monitoring and evaluation, which was "significantly and positively associated with project outcome as institutionally measured at the World Bank".

However, some potential recipients of CLD funding may be in news deserts or with "ghost newspapers", as mentioned above. Some may not have at least three local news outlets that have been publishing something they call news each workday for at least six months, as required for the local elections recommended by McChesney and Nichols (2021, 2022), outlined above. In such jurisdictions, the local consultations that identify community priorities for CLD funding should also include discussions of how to grow competitive local news outlets to help the community maximize the benefits they get from the project.

The need for "at least three local news outlets" is reinforced by the possibilities that two or three local news outlets may be an oligopoly, acting like a monopoly. This risk may be minimized by working to reduce barriers to entry and to encourage different news outlets to serve different segments of the market for news. The risks of oligopolistic behavior may be further reduced by requiring all recipients of citizen-directed subsidies to release their content under a free license like the Creative Commons Attribution-ShareAlike (CC BY-SA) 4.0 international license. This could push each independent local news outlet to spend part of their time reading each other's work while pursuing their own journalistic investigations, hoping for scoops that could attract a wider audience after their work was cited by other outlet(s).

This preference for at least three independent local news outlets in an experimental jurisdiction puts a lower bound on the size of jurisdictions to be included as experimental units, especially if we assume that the independent outlets should employ on average at least two journalists, giving a minimum of six journalists employed by local news outlets in an experimental jurisdiction. The discussions above suggested subsidies ranging from 0.06% to 2% or more. To get a lower bound for the size of experimental jurisdictions, we divide 6 by 0.06% and 2%: Six journalists would be 0.06% of a population of 10,000 and 2% of a population of 300.

Sampling units / experimental polities
Many local governments could fund local news nonprofits at 0.15% of GDP, because it would likely be comparable to what they currently spend on accounting, media and public relations. If the results of such funding are even a modest percent of the benefits claimed in the documents cited above, any jurisdiction that does that would likely obtain a handsome return on that investment.

Experimental jurisdictions might include some of the members of the United Nations with the smallest Gross Domestic Products (GDPs) or even some of the poorest census-designated places in a country like the US.

Alternatively, they might include areas with seemingly intractable cycles of violence like Israel and Palestine: The budget for interventions like those proposed herein are a fraction of what is being spent on defense and on violence challenging existing power structures. If interventions roughly comparable to those discussed herein seem to reduce the lethality of a conflict at a modest cost, it would have an incredible return on investment (ROI) with important lessons that might help reduce the lethality of other conflicts.

For illustration purposes only, Table 2 lists the six countries in the United Nations with the smallest GDPs in 2021 in US dollars at current prices according to the United Nations Statistics Division plus Palestine and Israel, along with their populations and GDP per capita plus the money required to fund citizen-directed subsidies at 0.15% of GDP, as recommended by McChesney and Nichols (2021, 2022). The rough budgets suggested here would likely be for news subsidy companions to Community and Local Development (CLD) projects.

Other factors should be considered in detailed planning. For example, the budget for such a project in Montserrat may need to be increased to support greater diversity in the local news outlets subsidized, because 0.15% of GDP would fund only 3.3 journalists. And a careful study of local culture in Kiribati may indicate that the suggested budget figure there may support substantially fewer than the 97 journalists suggested by the naive computations in this table. The key point, however, is that subsidies of this magnitude would be modest as a proportion of (a) many other projects funded by agencies like the World Bank or (b) the money spent on defense or war.

Table 2. Rough estimate of the budget for subsidies at 0.15% of GDP for the 6 smallest members of the UN plus Palestine and Israel. Population and GDP at current prices per United Nations Statistics Division (2023). (*) "Number of journalists" was computed assuming each journalist would cost twice the GDP / capita. For example, the GDP / capita for Tuvalu in this table is $5,370. Double that to get $10,740. Divide that into $90,000 to get 8.4.

Other possibilities for experimental units might be historically impoverished subnational groups like Native American jurisdictions in the United States. As of 2023-01-12 there were "574 Tribal entities recognized by and eligible for funding and services from the Bureau of Indian Affairs (BIA)", some of which have multiple subunits, e.g., populations in different counties or census-designated places. For example, the largest is the Navajo Nation Reservation that is split between Arizona, New Mexico, and Utah. Some of these subdivisions are too small to be suitable for experiments in citizen-directed subsidies for news. Others have subdivisions large enough so that some subdivisions might be in experimental group(s) with others as controls.

Supplement not replace other funding
The subsidies proposed here should supplement (not replace) other funding, similar to the subsidies under the US Postal Service Act of 1792. McChesney and Nichols recommended that an organization should be publishing something they call news five days per week for at least six months, so the voters would know what they are voting for. Those criteria might be modified, at least in some experimental jurisdictions, especially in news deserts, as something else is done to create local news organizations eligible to receive a portion of the experimental citizen-directed subsidies.

The Institute for Nonprofit News and Local Independent Online News (LION) Publishers help local news organizations get started and maintain themselves. Organizations like them might help new local news initiatives in experimental jurisdictions as discussed in this article.

Sources for funding research
We know of three plausible source of funding for research to quantify the value of local news:


 * World Bank: The World Bank has already discussed the value of news. We would expect that organizations that fund community and local development projects would also want to fund experiments in anything that seemed likely to increase the return on their investments in such projects.
 * Major philanthropies: Folkenflik (2023) wrote, "Some of the biggest names in American philanthropy have joined forces to spend at least $500 million over five years to revitalize the coverage of local news in places where it has waned." This group of philanthropic organizations includes the American Journalism Project, which says they "measure the impact of our philanthropic investments and venture support by evaluating our efficacy in catalyzing grantees’ organizational growth, sustainability and impact."
 * Governments of nation states for funding citizen-directed subsidies for local news nonprofits in conflict areas like Palestine and maybe Israel: If such interventions have the anticipated effect, the results would have profound implications for national security the world over.

If the claims made above for the value of news have much validity, then appropriate experiments should be able to quantify the groups who benefit from improving the news ecology, how much they benefit, which structures seem to work the best, and even the optimal level of funding.

Beyond that, we would hope that many others, including nation states, members of the United Nations, would like to fund similar interventions, especially in high conflict areas, because changing the media environment (including social media) should be substantially cheaper and more effective than investing in lethal high tech weapon systems.

Summary
This article has summarized numerous claims regarding different ways in which information may be a public good. Many such claims can be tested in experiments crudely similar to those for which Banerjee, Duflo, and Kremer won the 2019 Nobel Memorial Prize in Economics. We suggest funding such projects for high conflict zones and for relatively small poor communities and countries. For the latter, we suggest funding such projects as companions to Community and Local Development (CLD) projects. If the research cited above is replicable, the returns on such investments could be huge, increasing productivity and thereby delivering benefits to the end of human civilization. If this works as expected, it will benefit other economies in ways similar to how subsidies for newspapers published in the US in the early nineteenth century allegedly benefited the US. This period of US history included the development of technologies that benefit the vast majority of humanity the world over today.

Acknowledgements
Thanks especially to Bruce Preville who pushed for evidence supporting wide ranging claims of media influence in limiting progress against many societal ills. He also helped with the literature search. Thanks also to Dave Black for suggesting experimenting with Native American jurisdictions in the US and to Joy Ellsworth for describing the substantial cultural challenges that such interventions might face. Thanks also to Mark Hull for help improving the discussion of Hitler.