Week 9: Modeling Data Policies

DSAN 5450: Data Ethics and Policy
Spring 2026, Georgetown University

Author

Affiliation

Jeff Jacobs

jj1088@georgetown.edu

Open slides in new window →

Schedule

Today’s Planned Schedule:

	Start	End	Topic
Lecture	3:30pm	3:45pm	Final Projects →
	3:45pm	4:15pm	Extended Recap →
	4:15pm	4:50pm	Contractual Power →
Break!	4:50pm	5:10pm
	5:10pm	6:00pm	The Power of Mechanism Design →

\[ \DeclareMathOperator*{\argmax}{argmax} \DeclareMathOperator*{\argmin}{argmin} \newcommand{\bigexp}[1]{\exp\mkern-4mu\left[ #1 \right]} \newcommand{\bigexpect}[1]{\mathbb{E}\mkern-4mu \left[ #1 \right]} \newcommand{\definedas}{\overset{\small\text{def}}{=}} \newcommand{\definedalign}{\overset{\phantom{\text{defn}}}{=}} \newcommand{\eqeventual}{\overset{\text{eventually}}{=}} \newcommand{\Err}{\text{Err}} \newcommand{\expect}[1]{\mathbb{E}[#1]} \newcommand{\expectsq}[1]{\mathbb{E}^2[#1]} \newcommand{\fw}[1]{\texttt{#1}} \newcommand{\given}{\mid} \newcommand{\green}[1]{\color{green}{#1}} \newcommand{\heads}{\outcome{heads}} \newcommand{\iid}{\overset{\text{\small{iid}}}{\sim}} \newcommand{\lik}{\mathcal{L}} \newcommand{\loglik}{\ell} \DeclareMathOperator*{\maximize}{maximize} \DeclareMathOperator*{\minimize}{minimize} \newcommand{\mle}{\textsf{ML}} \newcommand{\nimplies}{\;\not\!\!\!\!\implies} \newcommand{\orange}[1]{\color{orange}{#1}} \newcommand{\outcome}[1]{\textsf{#1}} \newcommand{\param}[1]{{\color{purple} #1}} \newcommand{\pgsamplespace}{\{\green{1},\green{2},\green{3},\purp{4},\purp{5},\purp{6}\}} \newcommand{\pedge}[2]{\require{enclose}\enclose{circle}{~{#1}~} \rightarrow \; \enclose{circle}{\kern.01em {#2}~\kern.01em}} \newcommand{\pnode}[1]{\require{enclose}\enclose{circle}{\kern.1em {#1} \kern.1em}} \newcommand{\ponode}[1]{\require{enclose}\enclose{box}[background=lightgray]{{#1}}} \newcommand{\pnodesp}[1]{\require{enclose}\enclose{circle}{~{#1}~}} \newcommand{\purp}[1]{\color{purple}{#1}} \newcommand{\sign}{\text{Sign}} \newcommand{\spacecap}{\; \cap \;} \newcommand{\spacewedge}{\; \wedge \;} \newcommand{\tails}{\outcome{tails}} \newcommand{\Var}[1]{\text{Var}[#1]} \newcommand{\bigVar}[1]{\text{Var}\mkern-4mu \left[ #1 \right]} \]

(Normative) Issues with Notice and Consent

The Crux of the Normative Issues

Does Reading = Understanding?

• Does reading \(\implies\) understanding implications / contingencies / ambiguities?
• NLP could (and should!) be helpful (“making privacy policies machine readable […] would help users match privacy preferences against policies offered by web services”), but mostly just reveals how bad the problem is:

Figure 15 from Wagner (2023). “Obfuscatory words” are words like acceptable, significant, mainly, or predominantly, interpretated at the discretion of companies rather than users (see next slide!)

The Fundamental Problem of Contracts

• Just as we can’t observe two simultaneous worlds \(W_{X = 0}\) and \(W_{X = 1}\) which differ only in the value of \(X\),
• We can’t foresee all possible contingencies that need to be included in a contract
  • (We can try, though! Hence use of obfuscatory words to minimize liability)
• So, when a situation arises which is not covered by a clause in the contract, what happens? What principle determines whose interpretation wins out?
  • (Hint: It is actually literally my legal middle name…)

…POWER!

• Examples from employment contracts:
• In a private, cooperatively-owned, democratic firm, outcome determined by conversation, majority vote, unanimity, etc.
  • These technically exist in the US! Employing 2,380 workers, \(\frac{2380}{127509000} \approx 0.0019\%\) of US workforce
• Otherwise, in a non-unionized private firm (94% of total), the outcome is determined by organizational hierarchy
  • This is the case for \(\frac{125000000}{127509000} \approx 98.03\%\) of US workforce

Descriptive and Normative Implications

• Who has power w.r.t. incompleteness of contracts?
• Who ought to have power w.r.t. incompleteness of contracts?
• Residual rights of control…

Hart’s Nobel Prize Speech

Complete contracts are contracts where everything that can ever happen is written into the contract. Actual contracts aren’t like this, as lawyers know. They’re poorly worded, ambiguous, leave out important things. They’re incomplete.

A critical question that arises with an incomplete contract is, who has the right to decide about the missing things? We called this right the residual control or decision right. The question is, who has it?

Further thought led us to the idea that this is what ownership is. The owner of an asset has the right to decide how the asset is used where the use is not contractually specified (Hart 2017)

Understanding Rights \(\leftrightarrow\) Fighting for Rights

• “Hohfeldian” framework (Hohfeld 1913)
• A right \(r_i\) granted to person \(i\) \(\implies\) A duty/obligation imposed on everyone in the world besides \(i\) (to respect \(r_i\))
• A duty or obligation \(d_i\) imposed on a person \(i\) \(\implies\) A right granted to everyone in the world besides \(i\) (to… be a potential beneficiary of \(d_i\))
• \(\implies\) rough measures of relative power in a contract:

\[ \frac{\text{rights}_i}{\text{rights}_j} = \frac{\text{obligations}_j}{\text{rights}_j} = \frac{\text{rights}_i}{\text{obligations}_j} = \frac{\text{obligations}_j}{\text{obligations}_i} \]

Descriptive vs. Normative

• Much of Part 1 has been adjusting to weirdness of normative reasoning

• Descriptive reasoning looks like [Rules of math \(\implies \theta^* = 2.5\)], but [rules of math] part isn’t mentioned bc extraneous

  • (Even if it was mentioned, intersubjective agreement not so hard, very few people fighting wars over “we should denote repeated addition with \(\otimes\) not \(\times\)!”)

• Normative reasoning looks like [Antecedent A \(\implies\) Answer 1 but Antecedent B \(\implies\) Answer 2], and people do fight wars over A vs. B (implicitly or explicitly)

• Part 2: Rapid cycling back and forth between normative and descriptive!

• One new aspect: “Descriptive Ethics” (How do people act, not how should people act) \(\leadsto\) Study of Power

  [What is] right, as the world goes, is only in question between equals in power; otherwise, the strong do as they please and the weak suffer what they must. (Thucydides 2013, chap. 411 BC)

Recap 1: Privacy Policies Take a Long Time to Read!

Recap 2: Reading Privacy Policies \(\neq\) Understanding Privacy Policies!

• Reading vs. understanding implications / contingencies / ambiguities…
• NLP could (and should!) be helpful (“making privacy policies machine readable […] would help users match privacy preferences against policies offered by web services”), but mostly just reveals how bad the problem is:

Figure 15 from Wagner (2023). “Obfuscatory words” are words like acceptable, significant, mainly, or predominantly, interpretated at the discretion of companies rather than users (see next slide!)

Conclusion

Real Conclusion(?)

Wars with One Side

It would be ideal except for the Porto Ricans [sic]. They are beyond doubt the dirtiest, laziest, most degenerate and thievish race of men ever inhabiting this sphere. It makes you sick to inhabit the same island with them. They are even lower than Italians. What the island needs is not public health work but a tidal wave or something to totally exterminate the population. It might then be livable. I have done my best to further the process of extermination by killing off 8 and transplanting cancer into several more. (Cornelius Rhoads)

By 1930, the police had files on at least 135,000 individuals (about 3 percent of the island) suspected of favoring independence. (Source)

Wars with One Side?

Georgetown Law School Center on Privacy and Technology, American Dragnet: Data-Driven Deportation in the 21st Century

Contracts Through a Game-Theoretic Lens: Mechanism Design

The Fundamental Problem of Contracts

• Just as we can’t observe two simultaneous worlds \(W_{X = 0}\) and \(W_{X = 1}\) which differ only in the value of \(X\),
• We can’t foresee all possible contingencies that need to be included in a contract
  • (Hence use of obfuscatory words to minimize liability)
• So, when a situation arises which is not covered by a clause in the contract, what happens? What principle determines whose interpretation wins out?
  • (Hint: It is actually literally my legal middle name…)

…POWER!

• Examples from employment contracts (tooting own horn):
• In a private, cooperatively-owned, democratic firm, outcome determined by conversation, majority vote, unanimity, etc.
  • These technically exist in the US! Employing 2,380 workers, \(\frac{2380}{127509000} \approx 0.0019\%\) of US workforce
• Otherwise, in a non-unionized private firm (94% of total), the outcome is determined by organizational hierarchy
  • This is the case for \(\frac{125000000}{127509000} \approx 98.03\%\) of US workforce

Descriptive and Normative Considerations

The combined effect of incomplete contracts and conflicts of interest is that the determination of outcomes depends on who exercises power in the transaction.

Power is generally exercised by those who hold the residual rights of control, meaning the right to determine what is not specified contractually (Bowles 2009)

• [Step 1: Empirically measurable given antecedents] Who has power w.r.t. a given incomplete contract?
• [Step 2: Up to you and your ethical axioms; e.g., efficiency] Who ought to have power w.r.t. incomplete contracts?

Working Definition of Power

Note Defining (Dyadic) Power (Bowles and Gintis 1992, 326–27)

For agent \(A\) to have power over agent \(B\) it is sufficient that, by imposing or threatening to impose sanctions on \(B\), \(A\) is capable of affecting \(B\)’s actions in ways that further \(A\)’s interests, while \(B\) lacks this capacity with respect to \(A\).

Mechanism Design

• Prisoner’s Dilemma
• Assurance Game
• Invisible Hand Game
• Mechanism Design = Creating incentives to push existing game from one form to another!
• Second Price Auctions…

Prisoners’ Dilemma

		\(B\)
		Silent	Snitch
\(A\)	Silent	\(-1,-1\)	\(-3,\phantom{-}0\)
	Snitch	\(\phantom{-}0, -3\)	\(-10,-10\)

Assurance Game

Palanpur, Gujarat, India

The farmers do not doubt that earlier planting would give them larger harvests, but no one, the farmer explained, is willing to be the first to plant, as the seeds on any lone plot would be quickly eaten by birds…

[What if you all organized to plant on the same day, to reap rewards of earlier planting while minimizing bird losses (dividing by \(N\) instead of \(1\))?]

“If we knew how to do that”, he said, looking up from his hoe at me, “we would not be poor.” (Bowles 2009)

Assurance Game in Normal Form

		\(B\)
		Early	Late
\(A\)	Early	\(4, 4\)	\(0, 3\)
	Late	\(3, 0\)	\(2, 2\)

Invisible Hand Game (Normal Form)

		\(B\)
		Corn	Tomatoes
\(A\)	Corn	\(2, 4\)	\(4, 3\)
	Tomatoes	\(5, 5\)	\(3, 2\)

The “Goal” of Policymaking!

		\(B\)
		Silent	Snitch
\(A\)	Silent	\(-1,-1\)	\(-3,\phantom{-}0\)
	Snitch	\(\phantom{-}0, -3\)	\(-10,-10\)

Figure 1: Prisoners’ Dilemma

\(\leadsto\)

		\(B\)
		Early	Late
\(A\)	Early	\(4, 4\)	\(0, 3\)
	Late	\(3, 0\)	\(2, 2\)

Figure 2: Assurance Game

 

		\(B\)
		Corn	Tomatoes
\(A\)	Corn	\(2, 4\)	\(4, 3\)
	Tomatoes	\(5, 5\)	\(3, 2\)

Figure 3: Invisible Hand Game

 

Prisoners’ Dilemma 😫 \(\prec\) Assurance Game 🤨 \(\prec\) Invisible Hand Game 🥳

Prisoners’ Dilemma (Fishers’ Dilemma)

• Single, unique Nash equilibrium, and it’s Pareto inferior

The Game

		\(j\)
		Fish 6 Hours	Fish 8 Hours
\(i\)	Fish 6 Hours	\(1.0,1.0\)	\(0.0,\boxed{1.2}\)
	Fish 8 Hours	\(\boxed{1.2}, 0.0\)	\(\boxed{0.4},\boxed{0.4}\)

• Boxes = Best Responses:
• \(\text{BR}_i(6\textrm{ hr}) = 8\textrm{ hr}\), \(\text{BR}_i(8\textrm{ hr}) = 8\textrm{ hr}\)
• \(\text{BR}_j(6\textrm{ hr}) = 8\textrm{ hr}\), \(\text{BR}_j(8\textrm{ hr}) = 8\textrm{ hr}\)

Pareto Dominance

Operationalizing Power

• Equally good outside options \(\implies\) can contract to Pareto-optimal point \(o^P\)
• \(i\) has better outside options \(\implies\) can make take it or leave it offer to \(j\):
  • “You (\(j\)) fish 6 hrs all the time. I (\(i\)) fish 6 hrs 41% of time, 8 hrs otherwise”

• Ever so slightly better for \(j\) \(\implies\) \(j\) accepts (Behavioral econ: \(j\) accepts if 41% meets subjective fairness threshold; observed across many many cultures!)
• Later / next week: observe policy with outcome \(o^{C}_{i \rightarrow j} \iff\) policy values \(i\)’s welfare more than \(j\)’s welfare (inferred social welfare weights \(\omega_i > \omega_j\))

Policy Interventions: Fish Dilemmas \(\mapsto\) Assurance Games

• Notice: To “escape” prisoners’ dilemma, we had to literally change the rules of the game (permanent intervention)
• Fishers’ Dilemma:
  • No institutions: \(a_i, a_j \in \{6\text{ hr}, 8\text{ hr}\}\)
  • Institutions (courts or social norms): \(\{\text{Accept}, \text{Reject}\}\)
• Driving “game”:
  • No institutions: \(a_i, a_j \in \{\text{Stop}, \text{Drive}\}\)
  • Institutions (stoplights installed by govt or community agreement): \(a_i, a_j \in \{\text{Obey Light}, \text{Run Light}\}\)
• Within assurance games, only need to nudge (one-time intervention) \(\leadsto\) new equilibrium (self-enforcing by definition)

Assurance Game

• Multiple equilibria; the particular outcome we observe is a function of history (path dependency)
• Drive-on-left vs. drive-on-right: Assurance game where neither equilibrium Pareto-dominates other option
  • Swedish Dagen H: Nudge from \(a^*_{\textsf{L}} = (\textsf{L},\textsf{L})\) to \(a^*_{\textsf{R}} = (\textsf{R},\textsf{R})\)
  • Either eq is self-reinforcing! (Unless you want to crash out)

• QWERTY vs. DVORAK / Palanpur farmers: Assurance game where observed equilibrium Pareto inferior

		\(j\)
		Early	Late
\(i\)	Early	\(\boxed{4},\boxed{4}\)	\(0, \, 3\)
	Late	\(3, \, 0\)	\(\boxed{2},\boxed{2}\)

Invisible Hand Game

• Single, unique Nash equilibrium, and it’s Pareto efficient
• \(\Rightarrow\) Acting in self interest \(\leadsto\) best possible outcome

It is not from the benevolence of the butcher, the brewer, or the baker that we expect our meal, but from their regard to their own interest (Smith 1776)

		\(j\)
		Corn	Tomato
\(i\)	Corn	\(2, \, 4\)	\(4, \, 3\)
	Tomato	\(\boxed{5}, \boxed{5}\)	\(3, \, 2\)

• Wealth of Nations SPOILER: The wealth comes from division of labor
  and also dumbleydore dies. semperus snake too. and even poor ron the weasel, who never deserved such a fate

An economic transaction is a solved political problem. Economics has gained the title “Queen of the Social Sciences” by choosing solved political problems as its domain. (Lerner 1972)

First Fundamental Theorem of Welfare Economics

Thm: [Antecedents (Coase Conditions)] \(\Rightarrow\) markets produce Pareto-optimal outcomes

• Even Jeff finds proof (and corollaries) compelling / convincing / empirically-supported
  • (It’s a full-on proof, in the mathematical sense, so doesn’t rly matter what I think; I just mean, imo, important and helpful to think through for class on policy!)
  • Ex: Conditional on antecedents [(Coase) minus (perfect competition) plus (thing must be allocated via markets)], \(\uparrow\) Competition \(\leadsto\) More efficient allocations
• Like how Gauss-Markov Assumptions \(\Rightarrow\) OLS is BLUE, yet our whole field (at least, a whole class, DSAN 5300) built on what to do when GM Assumptions don’t hold
• For policy development, helpful to think through
  • which cases “break” FFT (more honored in the breach)
  • How each violation might be “fixed” through policy
• Our violation: No externalities assumption
  • Possible policy “fixes”: property rights, market-socialist nationalization

Part 2 Suddenly Collides with Part 1: Property Rights

• Rawlsian Rights: Vetos on societal decisions; Constitution can make some inalienable (can’t sell self into slavery), some alienable
• Property rights: alienable. You can gift or sell the rights if you want (veto is over society just, like, taking your property if someone else would be happier with it)

Case : Society decides Right to Clean Air \(\prec\) Right to Smoke \(\Rightarrow\) Start at \(E\)

• \(A\) can pay \(B\) to alienate right (Pay $50/month, can smoke 5 ciggies) \(\leadsto\) \(X\)
• Movement along light blue curve: giving up \(x\) money for \(y\) smoke, equally happy. \(u_A(p)\) identical for \(p\) on curve
• Movement to higher light blue curve () \(\Rightarrow\) greater utility \(u_A' > u_A\)

Case Society decides Smoke \(\prec\) Clean Air \(\Rightarrow\) Repeat for \(E' \leadsto X'\)

“Edgeworth Box” for Right to Smoke vs. Right to Clean Air: \(A\), \(B\) are roommates; \(A\) loves smokin, \(B\) loves clean air

Externalities \(\Leftrightarrow\) Costs of Actions Paid by Someone Else!

• Firm \(S\) produces amount of steel \(s\), pollution \(x\)
• Firm \(F\) “produces” amount of fish \(f\)
• \(S\) optimizes

\[ s^*_{\text{Priv}}, x^*_{\text{Priv}} = \argmax_{s,x}\left[ p_s s - c_s(s, x) \right] \]

• While \(F\) optimizes

\[ f^*_{\text{Priv}} = \argmax_{f}\left[ p_f f - c_f(f, x) \right] \]

• If [Yugoslavia-style] nationalized, new optimization of joint steel-fish venture is

\[ s^*_{\text{Yugo}}, f^*_{\text{Yugo}}, x^*_{\text{Yugo}} = \argmax_{s, f, x}\left[ p_s s + p_f f - c_s(s, x) - c_f(f, x) \right] \]

• Can prove/“prove” that \(o(s^*_{\text{Yugo}}, f^*_{\text{Yugo}}, x^*_{\text{Yugo}})\) Pareto-dominates \(o(s^*_{\text{Priv}}, x^*_{\text{Priv}}, f^*_{\text{Priv}})\)

Social Welfare Functionals

• (Next week… let’s do this in-depth next week!)

Functionals?

• You probably know what a function \(f(x)\) is; a functional is a function of functions: \(\mathscr{G}(f)\)
• It’s from math, which is scary, but it’s just notation to remind us that we’re analyzing functions of functions
• In our case, they “work the same way” as regular functions, e.g., \(\mathscr{G}(f,g) = f^2 + g^2\), so \(f(x) = x, g(x) = 2x \Rightarrow \mathscr{G}(f,g)(x) = x^2 + 4x^2 = 5x^2\)

We Live In A Society, Part 2

• Utilitarianism, Kantian Ethics, and Rawls can all be modeled as Social Welfare Functionals

\[ W(\mathbf{u}) = W(u_1, \ldots, u_n) \Rightarrow W(\mathbf{u})(x) = W(u_1(x), \ldots, u_n(x)) \]

• \(u_i(x)\): Given bundle of resources \(x\), how much utility does \(i\) experience? \(u_i: \mathcal{X} \rightarrow \mathbb{R}\)
• \(W(\mathbf{u})\): Aggregates \(u_i(x)\) over all \(i\), to produce measure of overall welfare of society. \(W: (\mathcal{X} \rightarrow \mathbb{R})^N \rightarrow \mathbb{R}\).
• \(W(\mathbf{u}) = \sum_{i=1}^n \omega_iu_i(x)\). \(\omega_i\) is \(i\)’s welfare weight
• (Preview) Decomposition: evaluate policies by estimating marginal utility \(u'_i(x)\) compared to \(\omega_i\))

Alternative SWF Specifications

• Social values

\[ W(\underbrace{v_1, \ldots, v_n}_{\text{Values}})(x) \overset{\text{e.g.}}{=} \omega_1\underbrace{v_1(x)}_{\text{Privacy}} + \omega_2\underbrace{v_2(x)}_{\mathclap{\text{Public Health}}} \]

• Stakeholder Analysis

\[ W(\underbrace{s_1, \ldots, s_n}_{\text{Stakeholders}})(x) = \omega_1\underbrace{u_{s_1}(x)}_{\text{Teachers}} + \omega_2\underbrace{u_{s_2}(x)}_{\text{Parents}} + \omega_3\underbrace{u_{s_3}(x)}_{\text{Students}} + \omega_4\underbrace{u_{s_4}(x)}_{\mathclap{\text{Community}}} \]

• (Adapted from this great intro video!)

Utilitarian SWF

• Easy mode (possibly/probably your intuition?): Everyone’s welfare weight should be equal, \(\omega_i = \frac{1}{n}\)

\[ W(u_1, \ldots, u_n)(x) = \frac{1}{n}u_1(x) + \cdots + \frac{1}{n}u_n(x) \]

• \(\implies\) Utilitarian Social Welfare Functional!
• The Silly Problem of Utilitarian SWF: What if everyone is made happy by \(u_{\text{Jeef}} = -\infty\)?

The Hard Problem of Utilitarian SWF

While the rhetoric of “all men [sic] are born equal” is typically taken to be part and parcel of egalitarianism, the effect of ignoring the interpersonal variations can, in fact, be deeply inegalitarian, in hiding the fact that equal consideration for all may demand very unequal treatment in favour of the disadvantaged (Sen 1992)

• \(\implies\) “Equality of What?”
• What is the “thing” that egalitarianism obligates us to equalize (the equilisandum/equilisanda): Utility? Opportunity? Resources? Money? Freedom?

Policy Evaluation via Inverse Fairness

We Can Finally Understand This Image from Week 1!

References

Bowles, Samuel. 2009. Microeconomics: Behavior, Institutions, and Evolution. Princeton University Press. https://books.google.com?id=HAiMDU4qv0IC.

Bowles, Samuel, and Herbert Gintis. 1992. “Power and Wealth in a Competitive Capitalist Economy.” Philosophy & Public Affairs 21 (4): 324–53. https://www.jstor.org/stable/2265369.

Hart, Oliver. 2017. “Incomplete Contracts and Control.” American Economic Review 107 (7): 1731–52. https://doi.org/10.1257/aer.107.7.1731.

Hohfeld, Wesley Newcomb. 1913. “Some Fundamental Legal Conceptions as Applied in Judicial Reasoning.” Yale Law Journal 23. https://heinonline.org/HOL/Page?handle=hein.journals/ylr23&id=24&div=&collection=.

Lerner, Abba P. 1972. “The Economics and Politics of Consumer Sovereignty.” The American Economic Review 62 (1/2): 258–66. https://www.jstor.org/stable/1821551.

Sen, Amartya. 1992. Inequality Reexamined. Clarendon Press.

Smith, Adam. 1776. The Wealth of Nations. Random House Publishing Group. https://books.google.com?id=Anwp7vTR1usC.

Thucydides. 2013. The War of the Peloponnesians and the Athenians. Cambridge University Press. https://books.google.com?id=dSYnCTf3rSAC.

Wagner, Isabel. 2023. “Privacy Policies Across the Ages: Content of Privacy Policies 1996–2021.” ACM Transactions on Privacy and Security 26 (3): 32:1–32. https://doi.org/10.1145/3590152.