Skip to Content View Previous Reports



By the Project for Excellence in Journalism and Rick Edmonds of The Poynter Institute

“Newspaper circulation is in decline,” the inaugural edition of this report declared a year ago. After the events of 2004, it’s clear that things are worse than people thought.

Figures for 2001 to 2003 suggested that a 30-year decline in circulation was at least slowing and might even be leveling off. The circulation decline for dailies had slowed to a mere .0015 in 2003.1 What’s more, much of the long slide came at evening papers, many of which simply folded. Morning circulation was near its historic peak, and one could argue that the surviving papers were slightly bigger on average, and stronger. Count in the steady growth of newspaper online sites over the last decade, and you could even make the case that total audience was growing rather than shrinking.

U.S. Daily Newspaper Circulation
Circulation in millions, weekday and Sunday editions, 1940 to 2003
Design Your Own Chart
Source: Editor and Publisher Yearbook data
Average Circulation of U.S. Daily Newspapers
Weekday and Sunday editions, 1990 to 2003
Design Your Own Chart
Source: Editor and Publisher Yearbook data

That promise, however, was thwarted in 2004. Save for online, the story of newspaper circulation was a return to accelerated decline. For the six months ending September 30, 2004, circulation at the 841 daily papers and 662 largest papers for which audited totals were available was down 0.9% daily and 1.5% Sunday.2

At both the June and December 2004 meetings of analysts and investors, the industry leader, Gannett, reported year-to-year daily circulation losses, at slightly less than 2% excluding USA Today.3 Gannett’s president, Gary Watson, cited two factors: a business decision to price aggressively and the added cost of selling subscription starts now that the federal do-not-call registry limits telemarketing. At USA Today, which has been growing robustly and has a comfortable circulation lead among the national papers, Gannett raised the cover price from 50 to 75 cents in September, likely to lead to at least some circulation drop.4

The September results suggested that many other companies were following Gannett’s lead, essentially concluding that propping up circulation with expensive new pricing and determining churn needed to be scaled back.

It would have been a tough year for circulation in any case. And the audited September 2004 totals do not include purging the books of almost 250,000 phantom readers claimed by the four metropolitan papers (the Chicago Sun-Times, Newsday, Hoy and The Dallas Morning News) implicated in a circulation scandal. Long story short, each revealed that they had been bolstering their numbers the Enron way – by faking them, particularly by claiming single-copy street sales for papers that were literally dumped.

The succession of embarrassing announcements over the summer had several common threads. The cheating was extensive – overstatements from 50,000 to more than 100,000 – and had been going on for years.5 All four papers were in big cities and owned by public companies, which took large charges against earnings, almost $100 million for the Tribune Company’s Newsday and Hoy and tens of millions for the other two.6 (Circulation directors in Dallas and publishers at Hoy and Newsday departed. At the Sun-Times, a new management team had discovered practices of its predecessors).

By the summer’s end, aggressive New York prosecutors were looking to build a criminal case in the Newsday/Hoy situation. The Securities and Exchange Commission had launched its own investigation.

If it could happen at those four newspapers, two of them often rated among the nation’s 15 best, were there more shoes to drop? So far not. A couple of the larger companies, Gannett and Knight Ridder, said they were confident their own controls would prevent such fraud. E.W. Scripps hired a director of circulation compliance to double-check the work of its papers. Other companies simply announced they had charged publishers with making sure their numbers were clean.

As unraveling fraud schemes tend to do, this one revealed a number of unpleasant truths:

The ABC has already tightened up some rules, for instance halving the number of days that can be excluded from the audit count because of holidays or bad weather. Other rules are under review. Papers may also feel pressure now to tighten up on loophole categories on their own.10 Those things add to the difficulty and expense for newspapers if they seek to stabilize circulation numbers.

Shaky numbers and quality of circulation matter greatly to a business that draws about 75% of its revenues from advertisers.11 If there is damage, it will probably play out in a resistance to rate increases in 2005, which will in turn form the base for rate increases in years to follow. With many other factors at play, the damage could be substantial, yet also impossible to quantify exactly. It is part of what analysts see as soft industry prospects.

At the affected papers, the scandal is also translating into reduced news investments. The Morning News announced plans in September to eliminate 250 positions.12 The company didn’t specify how many of those would be in the newsroom, but other sources indicated news-editorial would take a hit of more than 10%, about 70 jobs. Newsday announced cuts of the same magnitude weeks later.

Silver linings were hard to find, but there was one: advertisers may demand rebates and will exert some leverage in pricing, but they still want to be in the paper. That played out at Newsday, where a group of car dealers suing the paper were banned for several months from placing ads. They protested, and Newsday relented.

It is important to note exceptions to the downward circulation trend. McClatchy newspapers, which have long emphasized circulation growth, recorded 0.2% growth both daily and Sunday in the six months ending in March 2004. The chain has recorded 20 consecutive years of circulation growth. Lee Newspapers were up 0.5% Sunday.13 The Boston Globe was up despite a price increase. In the longer run, national papers – The New York Times, The Wall Street Journal and USA Today – have all grown, The Times steadily substituting national for regional circulation, and at a premium price. Not coincidentally, all these papers would be classified by most newspaper industry insiders either as excellent or improving editorially.14 It is also worth emphasizing that generally the papers with the best circulation performance tend to be those that are investing in news resources.


For the rest of the industry, one alternative may be to find a new and better way to describe their business performance. And the events of 2004 gave a fresh push to the preference of many to make readership, rather than paid circulation, the standard measure of audience reach. They consider readership, defined as the number of people who actually read a newspaper, more meaningful than the paid circulation number (especially now that a bright light has shone on the ways the system can be abused and gamed). Readership has fallen through the years, but not as fast as circulation. There is evidence that more people are reading the newspaper at work or in settings like coffee shops and waiting rooms and that the demographic groups newspapers have a harder time reaching, like women and young people, are well represented among occasional readers of this kind.15

Commercial information services like Scarborough have been measuring readership in large markets for years, and it has been the standard measure in Canada for two decades. Since 1999, the ABC has begun offering “reader profile” reports, which not only estimate total readership but provide demographic information that helps advertisers design tailored multi-media campaigns for their clients. After a slow start, a majority of larger newspapers – 102 of the top 150 – have now done ABC profiles.16

This slow shift in emphasis to readership dovetails with the work of the Readership Institute at Northwestern University, a five-year-old, $10 million industry-sponsored research initiative aimed at helping papers halt the slide. The Institute prefers an even broader measure of readership than a number count – looking at the time a reader spends with the paper, how thoroughly the different sections are read and how many days a week. These factors are combined into a measure called Reader Behavior Score (RBS).17

The most current estimates of industry-wide readership are 78,285,000 (53% of the adult population) on an average weekday and 90,765,000 Sunday (61%). That is a slippage of about 1% daily since 1998 and 1% Sunday. By comparison, circulation had fallen 2% daily and 2.5% Sunday through 2003 (the latest year for which industry-wide estimates are available) with more decline coming in 2004.

Daily and Sunday Newspaper Readership
Percentage reading newspapers in an average week, 1999-2004
Design Your Own Chart
Source: Scarborough Research unpublished survey data,

Readership provides an apples-to-apples comparison for selling advertising against competing media like television and radio. But it is not risk-free. At least some advertisers may devalue pass-along readership as well as the deeply discounted elements of paid circulation. Even though it is more comparable to broadcast, some hesitate to embrace the readership standard since the traditional metric of paid circulation often leads to a higher cost per thousand (CPM). And the readership approach may also prove only a short-term solution. Were that new standard in place, stabilizing readership numbers over time would probably still be a challenge as fewer people turn to newspapers.

Number of Papers

Despite declines in circulation, the number of daily papers seems to be leveling. In the latest full year of data, 2003, there was a loss of only one paper, leaving 1,456 total daily papers in the U.S.18

Number of U.S. Daily Newspapers
Weekday and Sunday editions, yearly increments, 1990 to 2003
Design Your Own Chart
Source: Editor and Publisher Yearbook data

A shift from evening to morning editions has been going on for decades with the rise in evening television news and the decline in the number of cities with multiple newspapers (see 2004 State of the News Media Report). In the last year for which data are available, 2003, the number of morning papers rose by 10, to 787, while afternoon papers continued their three-decade decline by dropping 12, to 680. Circulation remains heavily tilted toward the morning, with 5.7 morning copies sold for each afternoon copy.19

In 2003, total Sunday circulation was down 0.5% from 2002, at 58,495,000. That is right in line with the annual decline of 0.5% in Sunday circulation since 1991.20

Strangely, the latest circulation decline came in a year when the total number of papers with Sunday editions rose by 4 to 917, tying 2000 for the most-ever Sunday papers. The combination of declining circulation and a rising number of Sunday papers means that the average Sunday circulation of 63,790 is the lowest ever.21


The majority of circulation is still clustered at the top, with the 12 largest papers (those over 500,000 circulation) accounting for 20% of all circulation. The smallest category, on the other hand (50,000 or less), comprises 1,239 papers and commands just 31% of overall circulation.22

Daily Newspapers by Circulation Category, 2003
Design Your Own Chart
Source: Editor and Publisher Yearbook 2004
Due to rounding, percents do not add up to 100.

Readership Trends

While readership calculations suggest that the picture may not be so bad, survey data offer more reason for concern. According to polling by the Pew Research Center, in April 2004, 60% of people said they read a daily newspaper “regularly,” the lowest number since Pew began asking the question in 1990. Through the 1990s, the number was stable at around 70%.23

Percent of People Reading a Newspaper “Regularly”
Design Your Own Chart
Source: Pew Research Center for the People and the Press, ’’Pew Research Center Biennial News Consumption Survey,’’ June 8, 2004
qu.:Do you happen to read any daily newspaper or newspapers regularly, or not?

When asked whether they had read a newspaper “yesterday,” a stricter measure, the numbers appeared more stable, though lower than for “regular” reading; 42% of people said yes, compared with 41% in April 2002. Readership also increases with age – 23% for people aged 18 to 29 compared to 60% for people 65 and older. As with other media, this is another worrisome trend, and the 2004 data suggest it has not abated.24

Given the problems of attracting the next generation of readers, it can be only temporary solace that newspapers still have a hold on what many advertisers consider attractive demographics. Readership among college graduates was high, 72%, as it was for families with incomes greater than $75,000 (74%).25

When it comes to how much time people spend with the paper, Pew found that among those who read a paper the day before, 62% spent half an hour or more with it. Another 12% spent less than 15 minutes, and 26% spent 15 to 29 minutes. Those who read regularly, in other words, made a significant time commitment. (By comparison, 85% of television news watchers spent more than half an hour with TV news, though that can be a more passive activity and involve doing other things at the same time.26 )

Readership by Party Affiliation

There was much discussion, too, of declining trust in the media on ideological grounds, particularly among Republicans. Did that affect readership? The broad answer is no.

In the spring of 2004, Republicans and Democrats were at nearly the same level for readership of a newspaper the day before at 45% and 46% respectively.27


While a realistic prognosis is that paid-circulation erosion will continue over the next decade, there is heightened focus on possible solutions. Embracing the readership standard as an alternative to the circulation standard is one. Broadly, companies seem to be choosing some mix of that and these three additional strategies:

So consider 2004 a watershed year for newspaper circulation. It brought a whiff of scandal to an industry long considered a model of business probity. That will harm at least a little – maybe more than a little – pricing power to advertisers. And two transitions speeded up: to valuing readership above paid circulation, and to increasingly running newspaper companies as providers of content across multiple platforms rather than newspapers first, foremost and forever.


1. 2004 Editor and Publisher Yearbook Online data, 1940-2003,

2. Newspaper Association of America press release, “Newspaper Readership Remains Strong in Top 50 Markets.” November 1, 2004. Available online at:

3. press release, December 8, 2004. Available on line at:

4. Associated Press Wire, “USA Today to raise its cover price to 75 cents,”, June 22, 2004. Available on line at:

5.Rick Edmonds, “Ignoring the Elephant in Newspaperland.” Poynteronline, December 3, 2004. Available on line at:

6. Ibid.

7.Jennifer Saba, “Tossing Out a Few Numbers. ” Editor & Publisher, August 1, 2004, pp. 36-41.

8. Tom McGinty et al., “Anatomy of a Circulation Fabrication.” Newsday, September 12, 2004.

9.Deutsche Bank Securities (New York, New York). “Circulation uncensored.” Paul Ginocchio et al. Published August 20, 2004. 10.3% is a weighted average of both daily and Sunday circulation.

10. Rick Edmonds, “EXTRA! EXTRA! Newspapers Cope With Basic Problems.” Poynteronline, December 16, 2004. Available on line at:

11.Merrill Lynch (New York, New York).”Newspaper Industry Primer (8th Edition).” Lauren Rich Fine. Published August 26, 2004.

12. Associated Press Wire, “Dallas Morning News to close Cuba news bureau.” October 29, 2004.

13. Data provided by Rick Edmonds’s attendance at company presentations.

14. Data provided by Rick Edmonds’s attendance at company presentations.
Columbia Journalism Review editors, “American’s Best Newspapers.” Columbia Journalism Review, November/December 1999. Available on line at:

15. Rick Edmonds, “Readers vs. Subscribers.” Poynteronline, October 21, 2002.

16. “400th Reader Profile Report Released: The Philadelphia Daily News, Audit Bureau of Circulations News bulletin on line, August 2004. Available on line at:

17. An extensive account of the Institute’s work is available at The gist of the complex series of studies is this: papers can improve how much time people spend reading only if they follow the entire set of recommendations, which includes reader-friendly content and formats, stronger in-paper promotion, attention to the content of advertising, and transforming a newsroom culture that the Institute found was usually resistant to change and experimentation. Papers that have done all of that for several years have shown RBS improvement. Those that pick only an idea or two don’t show much progress.

18. 2004 Editor and Publisher Yearbook Online data, 1940-2003,

19. Ibid

20. Ibid

21. Ibid

22. Ibid

23. The Pew Research Center for the People & the Press (Washington, D.C.). “Pew Research Center Biennial News Consumption Survey.” Published June 8, 2004.

24. Ibid

25. Ibid

26. Ibid

27. Ibid

28. MediaMark Research, Inc. (New York, New York). “Newspaper Section Readership 2004.” Published Spring 2003.

29 . Ibid

30. Rick Edmonds, “Improving Editorial Quality From the Top Down.” Poynteronline, February 23, 2004. Available on line at:

31. Stefanie Olsen, “Extra! Wall Street Journal gives away Web content.”, October 22, 2004. Available on line at:

32. press release, June 22, 2004. Available on line at:

33. GThis was discussed during a panel at the American Society of Newspaper Editors (ASNE) Annual Convention, April 20, 2004, most explicitly in a presentation by Diane McFarlin, past ASNE president and publisher of The Sarasota Herald-Tribune.