Презентация на тему «Джордж кавассилас 2011»

The Market for Diversity in Television News Media Economics Workshop

New Economic School, Moscow October 28-29, 2011

Lisa George Hunter College and the Graduate Center City University of New York

Felix Oberholzer-Gee Harvard Business School

Question

How does competition among local television stations influence diversity in local news programming? Do viewers value this diversity? Implications Welfare? Political engagement? Policy?

What We Know

Theory Incentive to differentiate depends on relative importance of price competition & market cannibalization effects Empirical Evidence: Radio (Berry & Waldfogel, 2001; Sweeting 2010) Newspapers (George 2007) TV (Baker & George 2011) Business stealing matters, consumption effects less clear

What We Don’t Know

What dimensions of differentiation matter to consumers in news markets? Politics? Issues? Race? Localism?

Why Diversity Matters

Some illustrations Hispanic political participation Newspaper readership Political competition & turnout

Hispanic Political Participation

Source: Oberholzer-Gee & Waldfogel, AER 2009

Variety in Newspaper Markets

Per Capita Sales (ABC)

Per Capita Sales (ABC)

(1)

(2)

-0.0021

-0.0007

(1.83)+

(0.46)

-0.0024

(1.38)

0.0453

0.0483

(17.61)**

(14.27)**

498

498

249

249

Distance (DMA)

Unique Beats (DMA)

Topics Covered (DMA)

Topics Covered (DMA)

(1)

(3)

(5)

(6)

Owners

-0.01

-1.02

-1.84

-1.50

(1.86)+

(2.36)*

(4.43)**

(2.88)**

Papers

0.01

-0.23

-0.82

(2.30)*

(0.37)

(1.08)

Constant

0.20

16.80

35.33

38.87

(6.70)**

(5.34)**

(18.28)**

(10.22)**

N

534

534

534

534

DMA's

173

173

173

173

Source: Lisa M. George, Information Economics and Policy, 2007

Political Competition & Diversity

Source: Lisa M. George, Content in Campaigns, 2011

Why Study TV News

TV News remains the primary news source for US households.

9

Source: Pew Research Center for the People and the Press, Ideological News Sources: Who Watches and Why. September 10, 2010. Sample size about 1500.

Why TV is Hard to Study

Typical empirical strategy Identify changes in diversity measures from (exogenous) changes in market structure TV News Static markets with very limited entry & exit Regulated ownership Need to measure content directly or infer variety from demand

10

Empirical Approach

Demand-side diversity measures Supply-side diversity measures Issues, Politics, Race, Geography How do they relate – To each other? To news viewing? To ownership and other policy variables?

Data

Newsbank Transcripts 40 markets, 2006-2010 FCC Market Structure Data Number of stations and owners Nielsen Viewership One month (all 210 markets) in 2006, 2008, 2010 Total viewing by timeslot & program type Black & Hispanic viewing (not used here)

Demand-Side Diversity Measures

Lead-in effects are important in TV Cost of changing the channel Consumer tendency to switch channels from prime time programming to local news reveals programming differentiation. Two illustrations . . .

Local News Viewing

. . Lead-in Matters

Local News and Entertainment Viewing -- Station Shares by Network

Local News & Lag News Viewing Population Share

Local News Viewing

. . Loyalty Matters

Local News and Entertainment Market Shares by Day (2010)

Measuring Differentiation

Other

ABC Prime

NBC Prime

Other

NBC News

ABC News

Estimating Differentiation

Lagged Viewing (LV) coefficients measure diversity. Pairwise estimates are summed across competitors for a station-market-year measure. Alternatives & adjustments

I= Local News Indicator for S = {1,0} LV = Lagged Viewing for S S={ABC, CBS, NBC, FOX, Other}

m=market = 210 DMA’s, y=Year={2006, 2008, 2010} D= Day ={M, T, W, Th, F}, T= ? Hour Timeslot {9pm-12am}

Station Cross-Effects

Prime Time

Prime Time

Prime Time

Prime Time

Prime Time

ABC

CBS

NBC

FOX

Other

Local News

Local News

Local News

Local News

Local News

ABC

CBS

NBC

FOX

Other

Average effect of an additional prime-time viewer on local news viewing.

0.054

-0.030

-0.038

0.058

0.021

-0.006

0.076

-0.008

-0.054

-0.005

-0.048

-0.080

0.151

-0.110

0.019

0.013

-0.020

-0.020

0.033

-0.015

NA

NA

NA

NA

NA

Pairwise effects summed over competing (off-diagonal) stations.

Supply-Side Diversity Measures

Basic Measure – Word Counts Word share (keyword frequency/total words) Market Deviations in word shares Issues: “Policy Agenda Projects” categories Keywords matched to categories Inductive keywords (next round) Politics & Race Members of Congress by party & race (Shares & Totals) Geography & Localism Local place names & local titles

Issue Diversity Metrics

Mean Word Share (%) (N=1523)

Market St. Deviation (N=398)

Crime

0.976

0.205

Weather

0.615

0.187

Government

0.391

0.091

Business & Economics

0.226

0.045

Foreign Affairs & Trade

0.161

0.040

Education

0.152

0.031

Defense

0.138

0.035

TV & Media

0.112

0.031

Social Welfare

0.108

0.021

Health

0.097

0.029

Traffic

0.096

0.044

Infrastructure & Environment

0.087

0.022

Sports

0.069

0.018

Ideological Issues

0.062

0.017

Labor & Employment

0.054

0.016

Taxes

0.041

0.016

Agriculture

0.008

0.004

Death Notices

0.0003

0.0003

Political Diversity Metrics

Mean Word Share (%) N=1523

Market Standard Deviation N=398

Ethnicity & Race

All Minority

0.0019

0.0020

Non-Hispanic White

0.0320

0.0102

Party & Office

Democrats House

0.0062

0.0037

Republicans House

0.0039

0.0022

Democrats Senate

0.0111

0.0042

Republicans Senate

0.0100

0.0036

Politician Counts

Mean

St. Dev.

Total Covered Politicians

56

26

Average Share of Stations Covering

44%

5.9%

Local Diversity Metrics

Mean Word Share (%) N=1523

Market St. Deviation N=398

Place Coverage

0.592

0.114

Local Government Titles

0.086

0.025

Place Counts

Mean

St. Dev.

Total Place References

1152

304

Average Share of Stations Covering

61%

5%

Some Results

Diversity and market structure Diversity and content Issue Diversity Political Diversity Diversity and viewership

Differentiation & Market Structure

Larger markets, more differentiation. More stations (owners), less differentiation. No relationship between racial diversity and differentiation. Market-level measures do not get us far.

Differentiation & Coverage (Word Shares)

* * *

* * *

* * * *

*

All issues included in each specification, split table.

Differentiation & Coverage (Word Deviations)

* * *

* *

* *

All issues included in each specification, split table.

Differentiation & Politics

Specifications with network interactions indicate this is not a FOX effect, but rather a CBS effect.

Differentiation & Race

Covering minority politicians does not contribute to loyalty.

Differentiation & Viewing (All Stations)

Differentiation & Viewing (Most Differentiated Station)

Some Conclusions & Extensions

First attempt to measure diversity in local television news Initial evidence on points of differentiation + Ideology Government + Politicians Differentiation increases viewing, largest effect for stations Still to do – Inductive differentiation measures Adjusted loyalty measures Extensions Station loyalty & the market for advertising Minority coverage & viewership