Based on Dunn's chapter 4 Forecasting Expected Policy Outcomes
From
Public Policy Analysis: An Integrated Approach, Sixth Edition by William N. DUNN (2014)
Introduction to ArtificiaI Intelligence in Higher Education
Policy prediction concise version
1. KAN YUENYONG, Thursday, November 26, 2020.
Thailand Policy Foresight in Covid-19 Era
Forecasting Expected Policy Outcomes
DA8020: Policy Studies and AnalysisGSPA
NIDA
2. Based on Chapter 4
Forecasting Expected Policy Outcomes
From
Public Policy Analysis: An Integrated Approach,
Sixth Edition by William N. DUNN (2014)
“We will begin with a speculation and end with the similar speculation”
3. Three Principle Forms of Forecasting
Can we foresee the future, how?
• Extrapolation (past historical trajectory)
• Prediction (based on known causal models)
• Expert Judgement (ask experts)
• Which one is the best method? Why?
A (Paul) Klee painting named ‘Angelus Novus’ shows an angel looking as though he is about to move away from something he is fixedly contemplating. His eyes are
staring, his mouth is open, his wings are spread. This is how one pictures the angel of history. His face is turned toward the past. Where we perceive a chain of
events, he sees one single catastrophe which keeps piling wreckage and hurls it in front of his feet. The angel would like to stay, awaken the dead, and make whole
what has been smashed. But a storm is blowing in from Paradise; it has got caught in his wings with such a violence that the angel can no longer close them. The storm
irresistibly propels him into the future to which his back is turned, while the pile of debris before him grows skyward. This storm is what we call progress.
— Walter Benjamin, “On the Concept of History” (1947, 1968)
5. Thai-related strain Nonthaburi/61/2020 is Thailand's patient zero also the first case outside China, who is a Chinese citizen from Wuhan
traveling to Thailand at January 8 2020, and Thailand/NIH-15/2020, and Thailand's the second patient is also a Chinese citizen who
traveled to Thailand at January 13, 2020. https://nextstrain.org/ncov/global
8. Compartmental models in epidemiology: The SIR Model
Facility
• Airborne Infection Isolation Room (AIIR) : Bangkok = 136, Upcountry = 1,042; Total AIIR = 1,178
• Cohort Ward : Bangkok = 143, Upcountry = 3,061; Total Cohort Ward = 3,204
• Isolation Room : Bangkok 237, Upcountry = 2,444; Total Isolation Room = 2,681
Human Resources
• Doctor : Public Hospitals = 29,449; Private Hospitals = 7,711; Total = 37,160
• Nurse : Public Hospitals = 126,666; Private Hospitals = 24,905; Total = 151,571
• Policy Recommendation (PROSPECTIVE) = Struck a quasi-lockdown to mitigate infectious rate, in order to satisfy medical supplies (pull down the curve)
9. 1
2
From number to policy intervention
1. Quasi-lockdown
2. Travel restriction
3. Quarantine
4. Contact tracing
3
Integrated Administrative Body
1. Emergency Decree to replace COMMUNICABLE DISEASES ACT, B.E. 2558 (2015)
2. The Center for COVID-19 Situation Administration (CCSA) ศบค
3. The Center for Economic Administration (and Relief and Recovery – Post COVID-19) ศบศ
4. Policy Assets Availability? (i.e. labs, medical resources, rt-PCR kit, etc)
4
12. Decision over Economic Recession VS Health Security trade-off: Reproduction number or R0 (R-naught) vs Fatality ratio. See US Presidential Election 2020.
But we need vaccine as a final solution. https://link.springer.com/chapter/10.1007/978-981-15-4814-7_14
Safe
Dangerous
?
[This chart is based on known pathogens]
What if, we face a different kind of
viral structure? Do we have any
assurance to counter such situation?
Evaluation
13. Moderna, Pfizer & Biontech and Oxford U & AztraZeneca
https://www.r-bloggers.com/2020/11/final-moderna-pfizer-vaccine-efficacy-update/,
https://www.biospace.com/article/comparing-covid-19-vaccines-pfizer-biontech-moderna-astrazeneca-oxford-j-and-j-russia-s-sputnik-v/
14. Accuracy
Q: Which one is better between qualitative vs quantitative? or mixed method?
• Predictive forecasts using causal models are approximately three times more
accurate than simple extrapolative forecasts.
• Forecasts based on expert judgment that employed multiple rounds (Delphi
technique) improved accuracy in about 70 percent of cases, as compared with
methods using one-time expert judgment as part of a largely unstructured
process, which is the standard procedure used with panels of experts in areas
such as health and medicine, the environment, and the economy.
• Predictive forecasts based on theoretical models to not appear to be more
accurate than forecasts based on expert judgment, although forecasts based
on theoretical models plus expert judgment outperform extrapolative forecasts.
Dunn (2014)
15. The conclusions reached by a particular school of modern philosophy —
logical positivism — will be accepted as a starting point.
— Herbert A Simon, Administrative Behavior (1947)
It is necessary to deny … that empiricism is the essence of science.
— Dwight Waldo, The Administrative State (1948).
(quantitative, post-positivist, and value-laden perspective)
C. Dwight Waldo
Herbert A. Simon
The greatest unending debate in Public Administration field, and both challenged traditional PA both in research
methodology and politics & administrative dichotomy. Waldo staged the Minnowbrook I (separated from ASPA annual
meeting) on 1968 and led to the birth of National Association of Schools of Public Affairs and Administration (NASPAA)
and NPM movement (solidified later by the Ostroms).
Accuracy -> Why Forecasting!
How to blend quantitative vs qualitative research altogether? How to compensate strength & weakness of each other?
GSPA
NIDA
17. Accuracy: Blending experience (from season expert) vs empirical data (RETROSPECTIVE)
Adjusting the model to conform the reality, from SIR to SEIR with vital dynamics. From single
shot to multiple shot. https://docs.idmod.org/projects/emod-hiv/en/latest/model-seir.html
18. CONPLAN 8888: COUNTER-ZOMBIE DOMINANCE http://i2.cdn.turner.com/cnn/2014/images/05/16/dod.zombie.apocalypse.plan.pdf
Unknown Unknowns
or The Black Swan
“There are known knowns; there are things we know we know. We also know
there are known unknowns; that is to say we know there are some things we
do not know. But there are also unknown unknowns” — Donald Rumsfeld
“Most of us view the world as more benign than it really is, our own
attributes as more favorable than they truly are, and the goals we adopt as
more achievable than they are likely to be. We also tend to exaggerate our
ability to forecast the future, which fosters overconfidence.” – Daniel
Kahneman (With Amos Tversky and others, Kahneman established a
cognitive basis for common human errors that arise from heuristics and
biases)
Correction: The Tenth Man Rule (in WWZ), or Devil Advocate, or “Red Team”
Problem in “assumption drag,” that is, a tendency among forecasters to cling
to questionable or plainly implausible assumptions despite contrary evidence
(Dunn, 2014).
20. Possible futures: Future that may occur differently from what eventually do occur.
Plausible futures: Future that tend to occur, based on theory, if policymakers do not intervene to redirect the course of events.
Normative futures: Futures that are consistent with future value.
What to forecast?
• The consequence of existing policies
• The consequence of new policies
• The content of new policies
• The behavior of policy stakeholders
Club of Rome report states that inequality reduction and new economic models are necessary for long-term economic and planetary stability
https://www.stockholmresilience.org/research/research-news/2018-10-17-a-smarter-scenario.html
24. https://www.youtube.com/watch?v=zWA43-6taNA
The classic Holzinger and Swineford
(1939) dataset consists of mental
ability test scores of seventh- and
eighth-grade children from two
different schools (Pasteur and Grant-
White). In the original dataset
(available in the MBESS package),
there are scores for 26 tests. However,
a smaller subset with 9 variables is
more widely used in the literature (for
example in Joreskog's 1969 paper,
which also uses the 145 subjects from
the Grant-White school only).
25. • Extrapolation -> Big Data (i.e.
GDELT), ML & AI
• Causal Model -> Bayesian SEM,
Complex system causal model,
Graph theory
• Expert Judgement -> Transformative
Scenario Planning
• Framework -> Phenomenology,
Mixed Method
• How to simplify, combine and
economize it all?
• Differentiate between policy
forecasting [known (un)knowns] and
future foresight (unknown unknowns)