Decisions, Decisions....
NATURE OF LOCATIONAL UNCERTAINTY; GAME THEORY;
PROBABILITIES; BEHAVIORAL AMBIGUITIES
(http://faculty.washington.edu/krumme/450/uncertain.html)
Supporting Pages:
1. Uncertainty: Definition
 Uncertainty: reference to the lack of knowledge (ignorance)
about the state of the environment or other decision variables.
 Uncertainty about the future vs. uncertainty about other places,
events, structures and other phenomena
 Risk: reference to the consequences of uncertainty, such as regret
often expressed in terms of probabilities of such consequences
uncertainty then reduced to uncertainty about actual occurrence and,
possibly, about the reliability of probabilities.
2. The roots of "relevant uncertainty"

 complex patterns and lack of visibility and control
 diversity of environments (with its impact on the amount of
required information)
 distance to environments (and its impact on the visibility and
cost of collecting information)
 directness or indirectness of links to relevant environments (and
its impact on visibility and cost of collecting information, as well as
the likelihood that different actors have different (and not easily known)
access to intermittent nodes thereby creating difficult to understand
differences in competitiveness...
 interconnectedness (and hostility) of environments.
Interdependence and its associated redundancies and multicausalities
makes the tracing of causal structures and processes difficult. Of
particular interest might be the possibilities of coalitions and collusive
structures forming in the environment which may become hostile ("talks
behind one's back"; "ganging up" etc. as frequently found as part of
unstable oligopolies.
Extreme uncertainties have often been associated with the term
"turbulence" which in turn has been most extensively conceptualized by
Emery & Trist (1965) and subsequently by
Terreberry (1968). In their systemstheoretical
view, complexity and dynamic change together tend to create
"turbulent
fields" which may become unpredictable due to pervasive and
significant "relevant uncertainties". Often, such situations are described
as environments where "the ground is in motion", where there are few if
any stable structures which could potentially be the basis for
understanding and projections of consequences of actions.
(See also
discussion in this online paper)
3. Causes of Uncertainty [Lawrence & Lorsch (1967)]
 quality and clarity of information
 lack of understanding of causal relationships between variables
 time required to receive clear information about consequences of
actions (leads and lags): gestation periods, environmental feedback.
4. Uncertainty and the classical location model
 price uncertainty: (see Day &
Tinney)
 locational uncertainty in duopoly: Hotelling model
 climatic uncertainty: "Man against Environment"
(Gould,
1963);
 more complex uncertainty situations
5. Uncertainty and the Nature of Spatial "Decisions"
 "decision" without uncertainty is "empty"
(Shackle)
 y = f(x; z); Payoff = f(controllable and uncontrollable
variables)
6. GameTheoretic Responses to Uncertainty: (see Walker et al.)
also: (Dean & Carroll) for a plant location example

What is Game Theory?
[Some References]

The History of Game Theory [Timeline]
[http://williamking.www.drexel.edu/top/class/histf.html]
 Minimax/ Maximin: "Wald Criterion"
 Minimizing maximum regret strategy (Savage)
 Savage's regretbased decisionmodel avoids the extreme conservatism
of the Wald model. Based on a "regret matrix" which compares (subtracts)
the highest outcomes of each strategy from other outcomes. The new matrix
shows the extent to which a decisionmaker could have (expost, i.e. based
on hindsight) done better (opportunity loss). The Wald solution rule
(maximin/minimax) is applied to this new matrix to gain the minimax
regret solution.
 Equal likelihood principle (LaPlace)
 the Laplace model can be interpreted as a transition model between
the
probability/risk model of decision theory and game theory in that it
suggests that in the
absence of any probabilities which could potentially differentiate the
payoffs, equal probabilities should be assigned. "If one is 'completely
ignorant' as to the state of nature which will occur, he should behave as
if all states are equally likely." (Walker et al, p.170)
 Optimism (Hurwicz criterion)
 This decision criterion involves the identification of the worst and
best outcomes for each strategy. An optimism coefficient (a) and a
pessimism
coefficient (1a) are
then determined (a represents a given individual's optimism that a
positive outcome will occur) and multiplied by the (best+worst) outcomes.
The sum of the resulting coefficients is the optimism index. The strategy
with
the highest index will be followed. One should not interpret as the
likelyhood of states of nature. Instead, it is the decisionmaker's
optimism index relating to the occurrence of an outcome. "When a=0, the
Hurwicz solution is the same as the pure Wald solution." (Walker et al.,
p.171)
7. Probabilistic Approaches
 Probability: quantitative expression for the likelihood
of an event
 Hotelling model with discrete locations and probabilities:
"expected value
maximization"
The "expected" is applied here in the usual statistical sense. Expected
values are weighted averages found by multiplying the value (payoff)
associated with each environmental/competitive state (location of the
other duopolist) by the probability of
incurring
it and then adding these products. The weights (probabilities) sum to 1.
 Decision trees for decision making (John Magee)
 Subjective probabilities
 Improving probabilities through learning (gathering additional
evidence)

Bayesian Statistical Sciences
 Morris: "Management Science: A Bayesian Introduction"
 Expected value of perfect information: Upper limit of costs expended
for additional information.
Expected cost associated with acting under uncertainty = Value of perfect
information  (minus) Payoff of best present action under uncertainty
 Limitations of expected value/utility maximization model
(How (in)appropriate is the model for the locational discourse?)
8. Other responses to uncertainty (overview & transition)
 Passive vs. active approaches
 Passive: accepting the uncertainty and e.g. select a maximin strategy
or do nothing
 Active: reducing the uncertainty through information searches and
learning activities
 Building responsiveness into a project, policy or strategy
 Taking out insurance
 Collecting more information
 Hiring consultants
 Opening a research department
 etc.
 Procrastinating, waiting for the future to arrive, waiting for better
weather, riding out periods of particularly high levels of relevant
uncertainty
Brian Goodall, Dictionary of Human Geography (key words):
Behavioral environment; Behavioral matrix;
Game Theory (incl. minimax; payoff matrix; saddle point); Information field;
Probabilistic; probability; Activity space;
Subjective probability;
Internet Sites:
Newspaper Clippings:

Vulnerable companies bet billions on
weather; Seattle Times,
Sunday, July 12, 1998 by Sharon Walsh, The Washington Post
"Global warming, greenhouse gases and
the most extreme El Nino since the 1800s are contributing
to severe swings in climate... companies scurrying to
find ways to protect themselves against the vagaries of Mother Nature.
For years, businesses have hedged against
economic losses with financial products based on interest rates,
currencies, commodities... These
products ... are called derivatives
because they are derived from different bets on the future
movement of just about anything that can't be predicted with certainty.
And, what's more unpredictable than the weather?
So some companies are sheltering themselves with
something new: weather derivatives..."
Literature:
Lists:
G. A. Bradshaw and Jeffrey G. Borchers
Uncertainty as Information: Narrowing the Sciencepolicy Gap
[National Center for Ecological Analysis and Synthesis (NCEAS) and USDA
Forest Service; 2Department of Forest Science, Oregon State
University]
Chichilnisky. Graciela, Economics
of Uncertainty. August 1997 (online)
Downs, George W. and David M. Rocke, Optimal Imperfection? Domestic
Uncertainty and Institutions in International Relations. Princeton
Univ.Press 1995. [JX1395. D69]
Duncan, Robert B., "Characteristics of Organizational Environments and
Perceived Environmental Uncertainty," Administrative Science Quarterly
September 1972, 31327.
Ekinsmyth, C. Hallsworth, A. Leonard, S. Taylor, M., Stability and
instability: The uncertainty of economic
geography. AREA. Vol. 27, no. 4, DEC 1995, p.289299.
F.E.Emery & E.L.Trist (1965)
Hardaker, J. Brian, Coping with Risk in Agriculture. CAB International/
Oxford UP, 1997. (288pp.)
Forges, F. and JF. Thisse, Game Theory and Industrial Economics, in:
Norman, George and M. La Manna, in: The New Industrial Economics. 1992
pp. 1247. [pp.33ff. The LocationPrice Model]
[HD 2326.N42, 1992]
Goodchild, Michael F.,
UNCERTAINTY IN GEOSPATIAL INFORMATION REPRESENTATION, ANALYSIS
AND DECISION SUPPORT
[FY97 NURI RESEARCH PROPOSAL; SUBMITTED BY THE NATIONAL CENTER FOR
GEOGRAPHIC
INFORMATION AND ANALYSIS AT THE UNIVERSITY OF
CALIFORNIA, SANTA BARBARA AND THE UNIVERSITY OF
MAINE]
Gould, Peter, Man against his Environment: A GameTheoretic Framework,
Annals (AAG), 53(3), 1963, 2907.
reprinted in:
 Smith, Taaffe & King, Readings in Economic Geography, Chicago, 1968.
Gould, Peter, Wheat on Kilimanjaro: The Perception of Choice Within Game
and Learning Model Frameworks, in: Ludwig von Bertalanffy and Anatol
Rapoport, eds., Yearbook of the Society for General Systems Research,
Vol.X, 1965, pp. 157ff.
David F. Hendry and Neil R. Ericsson, eds., Understanding Economic
Forecasts (MIT Press, 2001)
[HB3730.U49.2001/Suzz]
[
PDF files (prepublication) "Forecasting Uncertainty in Economic
Modeling"]
Hey, John D., Uncertainty in Microeconomics. New York: NYU Press, 1979.
Kahneman, Daniel; Paul Slovic and Amos Tversky, eds., Judgment under
Uncertainty: Heuristics and Biases. Cambridge: Cambridge University Press,
1982.
King, L.J. and R.G.Golledge, "Bayesian Analysis and Models in Geographic
Research," in Univ. of Iowa Geography Dept. Discussion Papers No.12, 1969,
pp.1545 [Geographical Essays Commemorating the Retirement of Harold H.
McCarty]
Rabin, Matthew. "Psychology and Economics," Journ. of Econ.Lit. 36(1)
March 1998, 1146 (Review Paper)
Sandmo, Agnar,
"On the Theory of the Competitive Firm under Price
Uncertainty," American Economic Review Vol. 61, No. 1, Mar., 1971,
Savage, L.J., The Foundations of Statistics, New York, 1954.
Thrift, Nigel. The Geography of International Economic Disorder. In:
Johnston & Taylor, eds., A World in Crisis? 2nd ed., Oxford: Blackwell,
1989, 1678.
Thrift, Nigel. A Hyperactive World. In: Johnston, Taylor and Watts, eds.,
Geographies of Global Change: Remapping the World in the Late 20th
Century. Oxford: Blackwell, 1995.
Wald, A., Statistical Decision Functions, New York, 1950.
Walker, Warren E., Uncertainty : the challenge for policy analysis
in the 21st century. Santa Monica, Calif. : RAND, [2001]
[H97 W3395 2001]
Webber, Michael. Impact of Uncertainty on Location. Cambridge:
M.I.T.Press, 1972.
Winberg, Alan R., Managing Risk and Uncertainty in International Trade:
Canada's Natural Gas Exports. Boulder: Westview, 1987, Ch.6, pp.113ff.
("Uncertainty") [HD61.W56 1987, Suz]
Zhang, Jingxiong & Michael Goodchild,
Uncertainty in Geographical Information
Taylor & Francis (Routledge) 2002 [ISBN: 0415243343]
Conferences:
"Prediction is very difficult, especially about the future" (Niels Bohr,
quantum physicist)
Return to Econ & Bus Geography
2003 [
econgeog@u.washington.edu]