The Evolution of American Government Implications for Business and Industry

1
The Evolution of American GovernmentImplications
for Business and Industry

• Frank R. Baumgartner
• Professor and Head
• The Pennsylvania State University
• Management Strategy and the Business Environment
• The Wharton School, University of Pennsylvania
• March 27, 2004

2
Outline of Todays Talk

• The Policy Agendas Project
• With Bryan D. Jones
• The Policy Advocacy Project
• With Berry, Hojnacki, Leech, and Kimball
• The Evolution of American Government
• Book manuscript with Jones, in progress
• Drawing Lobbyists To Washington
• Lobby Disclosure Reports data linked to Agendas
  Project data
• With Beth Leech (Rutgers)
• My focus today

3
Intellectual Goals

• Develop Punctuated Equilibrium Theory of
  Government
• Encourage attention to forces of stability and
  forces of change, simultaneously
• Encourage larger-scale empirical work
• Often in collaboration
• Orders of magnitude larger then typical work
• Provide infrastructure for teaching and research

4
The Policy Agendas Project

• Http//www.policyagendas.org/
• With Bryan D. Jones
• Interactive web site allows analysis as well as
  dataset retrieval
• Research, archive, teaching uses

NSF grants SBR9320922 and SBR0111611
5
Policy Agendas - Datasets

• Congressional Hearings (72,000 cases)
• Statutes (12,000)
• Stories in CQ Almanac (18,000)
• Sample of NYTimes Abstracts (36,000)
• OMB Budget, by 62 categories (3,000)
• Gallup Most Important Problem data
• All are comprehensive, from 1947-

6
Policy Agendas In Progress

• Bill Introductions (400,000 cases)
• Supreme Court Decisions
• Executive Orders of the President
• State of the Union Speeches
• Association Activities, from the Encyclopedia of
  Associations, from 1959 to present (with John
  McCarthy)
• 25,000 associations in 2002
• 6,000 associations in 1959
• Other data sources to be added in the future

7
Agendas Project - Organization

• Everything coded according to
• 21 Major Topics of Policy Activity
• Macro-economics
• Health
• Agriculture
• Etc.
• 226 subtopics
• Each major topic further subdivided
• Consistently coded over time, allows time-series
  analysis
• Facilitates systematic historical work, often by
  adding new coding from same sources.
• Sophisticated index system for public sources.
• Can be used alone, or as a base for more detailed
  analysis.

8
The Policy Advocacy Project

• Http//lobby.la.psu.edu
• Frank Baumgartner (PSU)
• Jeff Berry (Tufts)
• Marie Hojnacki (PSU)
• Beth Leech (Do you know her?)
• David Kimball (Missouri, St Louis)
• Grads Christine Mahoney and Tim LaPira

NSF grants SBR9905195 and SBR0111224
9
The Policy Advocacy Project

• Sample of lobbyists, weighted by level of
  activity with the federal government
• Could you take the most recent issue you've been
  spending time on and describe what you're trying
  to accomplish on this issue and what type of
  action are you taking to make that happen?
• Then snow-ball sample of others involved in that
  issue.
• Result A random sample of objects of lobbying.

10
The Policy Advocacy Project

• Random sample of 102 issues, based on 350
  interviews, extensive internet research on each
  case as well, all archived on our web site.
• Focus on issue-definition, advocacy efforts
• Micro-approach on advocacy and policy
  definitions, how change or stability occur, in a
  cross-section.

11
The Combination

• Macro-level attention to characteristics of the
  American political system over 50 years, evidence
  for p-e theory of policy change.
• Micro-level attention to the struggles of policy
  advocates and government officials to work within
  this environment.
• Theory testing but also infrastructure
  development for political science large scope.

12
The Evolution of American Politics

• Important changes from 1947 to present in
  structure and functions of government.
• Implications of rise of new issues in politics,
  complexity of the governmental agenda.
• Information-based theory of government attention.
• Manuscript forthcoming, University of Chicago
  Press, revisions to be completed in next several
  months
• (Comments therefore welcome and timely)

13
The Evolution of American Government Information,
Attention, and Policy Punctuations Chapter 1
Introduction Information Processing and the
Evolution of American Government Chapter 2
American Government, Then and Now Part I Signal
Detection and Public Policy Chapter 3 Information
Processing and Policy Evolution Chapter 4 The
Inefficiencies of Attention Allocation Chapter 5
Is Agenda Setting Event Driven? Chapter 6
Representation and Agenda Setting Part II Policy
Punctuations and Macro Policy Chapter 7 Issue
Intrusion and Policy Punctuations Chapter 8 The
Pace of Legislation Chapter 9 How the US
Government Budget Evolved Chapter 10
Understandable Complexity in the Budget Chapter
11 Policy Punctuations Chapter 12 Institutional
Friction and Policy Punctuations Part III Issues
and Structures of Government Chapter 13
Co-evolution of Issues and Structures in Congress
Chapter 14 Entropy, Institutional Overlap, and
Politics Chapter 15 The Punctuation Entropy
Paradox Chapter 16 Information Processing and
Governance
14
Inefficiency in Attention-Shifting

• Given multi-dimensional issues, how does
  attention shift from one dimension to another?
• Individuals attention-allocation
• Organizations agenda-setting
• Agenda-setting the first step in decision-making

15
A Model of Decision-Making
16
Two approaches to theory building in public
policy

• One case-study-at-a-time
• State of the art in the policy sciences
• Value in understanding particular policies
• Post-hoc and ad-hoc approach
• No ability to predict when shifts occur
• Theories often dont apply from one policy area
  to another
• Broad theoretical breakthroughs not likely with
  this approach
• (Agendas data can be used for this)

17
Two approaches (cont.)

• Stochastic modeling
• Measure policy changes across scores of policies
  across scores of years
• Combine policy change indicators
• Analyze the resulting distribution of changes
• No ability to discuss particulars of cases
• More analogous to biology than to physics
• Attention is on probabilities, behavior of the
  population, not of individual units in the
  population
• Hopefully, greater theoretical possibilities even
  if less useful for studying particular policies

18
Explaining a Punctuated Equilibrium Model

• Attention to equilibrium periods is just as
  important to the punctuations
• A focus on the causes and inevitability of
  discontinuities and inefficiencies in
  decision-making.
• Cascades, Sieves, and Friction as sources of
  inefficiency in decision-making

19
Studying Policy Change Stochastically

• Indicators of Change
• Change in attention
• Media Attention
• Congressional Hearings
• Change in Levels of Activity
• Lawmaking
• Major Laws
• Change in Policy
• Budgets

20
What is Incrementalism?

• Pt Pt-1 et
• Policy today is a function of fluctuations of
  policies yesterday.
• Rearranging the terms yields (Pt - Pt-1) et
• That is, policy is a random walk in time.
• The resulting distribution of policy changes
  Normal Distribution.
• This is logical if one thinks of efficient
  institutions responding to stochastic (random)
  inputs
• Easily testable by taking all annual change
  measures across all years and all policies.

21
Is Normality Expected in a P-E Model?

• Under bounded rationality, choices are
  inefficient as attention moves from one issue to
  another.
• Over-attention to previously defined attributes
  no constant weighting to all relevant dimensions
  zero-weights to many potentially relevant issues.
• Friction, resistance to change, inefficient
  reaction to new information.
• Individuals ideology, cognitive limits
• Organizations missions and focus on particular
  solutions rather than (inherently
  multidimensional) problems
• Both under- and over-reactions expected.

22
Three Explanations of Inefficiency Cascades,
Sieves, and Friction

• Cascades imitation, cue-taking, threshold models
• Sieves informational short-cuts, satisficing
• Friction zero attention to particular elements
  until they are forced on the agenda. New
  institutions (agencies) may be created, suddenly
  focusing large amounts of attention (and effort)
  on a new dimension.
• All imply lurching rather than smooth responses
  to change

23
Fig. 11-2. Response with Interactive Costs
24
Disproportionate Response to Inputs

• Reaction to low levels of input close to zero
• Reaction to medium levels close to zero
• Reaction to sufficient levels extreme
• Should be significantly different from Normal,
  specifically leptokurtic high peak, low
  shoulders, high extremes

25
Leptokurtosis A Signature Characteristic of
Friction-Heavy Processes

• Friction and inertia lead to much greater than
  normal incrementalism reduced variance seems
  apparent.
• However, this is combined with occasional bursts
  of extreme change. Variance is actually quite
  significant, but skewed.
• Moderate change is lower than expected.
• Kurtosis, a measure of peakedness (k0 for Normal
  curve

26
Leptokurtosis (cont.)

• Reactions are not proportionate to inputs
• Either extreme resistance to change, or massive
  over-reactions
• Earthquakes, forest fires, many natural processes
  have these characteristics
• Difficulty to observe when N is low appears
  roughly Normal with a few outliers
• Easily apparent, and statistically discernable,
  when N is high enough, as in our large datasets

27
Two Normal Distributions
Two Normal distributions with mean zero and
standard deviation of 5 and 15.
28
A Normal and a Leptokurtic Distribution of
Similar Variance
29
Analyzing Policy Distributions

• Level of friction should not be the same across
  all policy distributions
• Efficient processes should not have the
  distinctive pattern we expect in the policy
  process as a whole.
• Institutions vary in their relative efficiencies
• Kurtosis a measure of deviation from Normal
• Kurtosis a measure of efficiency of decisions

30
From Inputs to Outputs Increasing Costs of
Decision-Making

• Normal distribution expected for problem inputs
  and low-cost decision-making
• Increased stickiness, costliness of decisions
  made in different institutional environments
  should yield increased kurtosis scores
• Stages of policy process higher kurtosis

31
Markets Not perfectly Normal
Figure 12.8. Dow-Jones Industrial Average, Daily
returns (percents), 1896-1996.
32
Election Results Relatively Efficient
Figure 12.7. Election to election change in House
vote margin by district, 1898-1992.
33
Hearings Moderately low-cost
Figure 12.3. Yearly percentage change in Senate
hearings by major topic, 1946-1999.
34
Lawmaking Moderately Costly
Figure 12.4. Yearly percentage change in statutes
by major topic, 1948-1999.
35
The US Budget since 1800 A High-Cost Policy
Process
Figure 12.6. Annual change in Real US Budget
Outlays, 1800-1994.
36
The Distribution of Annual Budget Changes,
1947-1999
37
Kurtosis Scores for 12 Processes
38
Implications

• The contingencies of the policy process mixture
  of positive and negative feedback processes.
• Politicians seek issues on which to make their
  names technical communities develop solutions to
  problems they may not even be working on
  advocates seek attention to pressing social
  concerns

39
Implications (cont.)

• Models of these complicated processes need to
  incorporate uncertainty
• A stochastic approach is not to supplant, but to
  supplement, traditional policy studies.
• Todays results show we should not underestimate
  the forces of stability or the opportunities for
  dramatic change.
• Smooth, efficient, reasonable reactions to
  changing inputs are missing, however.

40
Drawing Lobbyists to Washington

• 74 Issue Areas Defined in Lobby Disclosure
  Reports
• 56 of these overlap with areas defined by Policy
  Agendas Project
• 8 time periods of lobby registrations
• Data reported in paper some have read

41
Drawing Lobbyists to Washington

• Three hypotheses
• Government draws lobbyists, like it or not, to
  create and expand their government relations /
  lobbying / public affairs divisions
• Government spending draws lobbying by those
  groups seeking rents
• Economic activity creates more interests who
  naturally come to Washington, simply reflecting
  greater economic activity

42
Drawing Lobbyists to Washington

• Data
• Number of hearings, contemporaneous period
• Number of hearings, previous 10 years
• (Number of distinct committee venues also tested
  highly collinear with above)
• Federal spending (not available for all
  issue-areas)
• Number of firms in this area of the economy (not
  available for all issue-areas, or for all times)

43
Drawing Lobbyists to Washington

• Variable Coeff (St Err)
• Hearings, 0.009 (0.004)
• (previous 10 years)
• Federal spending 0.02 (0.003)
• (Billions)
• Firms .0000185 (.0000177)
• Organizations, t1 0.98 (0.03)
• Intercept -0.67 (0.67)
• R-2 0.93 N84 p

44
Drawing Lobbyists to Washington

• Size of Economy not Important
• Spending has an impact, no surprise
• Much governmental activity is unrelated to
  federal spending reflected however in hearings
  activity
• Trade issues
• Health and regulations
• These explain stronger pull factor of hearings.

45
Drawing Lobbyists to Washington

• Agendas Project documents
• Growth in Government
• Increased Diversity of Range of Government
  Activities
• This analysis explains
• Growth in Group System, increased political
  involvement of business as a result, not so much
  a cause of this previous trend

46
Conclusions

• Data resources may be of interest to many
• Government activity is an important variable in
  studying business activity
• Variability in government activity
• by economic sector
• over time
• These provide opportunities for academic study
  but also have great impacts on business and
  lobbying strategies. Much is reactive, not
  proactive.