Financial Relativity: An Information-Geometric Interpretation of Asset Pricing

This paper introduces "Financial Relativity," an information-geometric framework that unifies asset pricing by treating physical and risk-neutral measures as equivalent probabilistic reference frames, thereby reinterpreting prices as geometric projections of terminal payoffs and deriving volatility as an endogenous manifestation of evolving posterior uncertainty.

The Gist

Imagine you are trying to predict the weather.

In the old way of thinking (Classical Asset Pricing), economists say: "The weather is what it is (the Physical Reality, or measure $P$), but because people are scared of getting wet, they pay extra for umbrellas. So, the price of an umbrella reflects both the real chance of rain and the fear of rain."

This paper, "Financial Relativity," by Lin Li, suggests we are looking at this wrong. It argues that we shouldn't think of prices as a mix of "reality" plus "fear." Instead, it suggests that prices are a geometric projection of reality, shaped by what we currently know.

Here is the core idea broken down with simple analogies:

1. The Two "Maps" of the World

Imagine you are navigating a city.

• Map A (The Flat Map): This is a blank map where every street looks equally likely to be the destination. It's the "default" setting when you know nothing. In the paper, this is called $P$ (the Physical Measure).
• Map B (The Curved Map): Now, imagine you get a text message saying, "Traffic is bad on 5th Avenue." Suddenly, your mental map changes. 5th Avenue looks "longer" or "harder" to traverse, while side streets look "shorter." You aren't changing the city; you are changing your geometry based on new information. In the paper, this is $Q$ (the Risk-Neutral Measure).

The Paper's Big Insight:
Traditional finance says Map B is just Map A distorted by human fear (risk aversion).
This paper says: Map B is the only map that matters right now. It's not "distorted"; it's just the map that fits the current information. There is no "true" map underneath; there is only the map that makes sense given what you know.

2. The "Shadow" Analogy (What is a Price?)

Imagine a 3D sculpture (the Terminal Payoff) sitting in a dark room. You can't see the whole sculpture yet.

• The Light Source: This is your Information.
• The Wall: This is the Market Price.

As you turn on more lights (get more information), the shadow of the sculpture on the wall changes shape.

• At the start (no info), the shadow is just a blurry blob.
• As you get more info, the shadow sharpens, revealing more details of the sculpture.

The Paper says: A stock price isn't a "value" floating in the air. It is simply the shadow (projection) of the final outcome, cast by the current level of information.

• If the shadow looks "expensive," it's not because people are greedy; it's because the geometry of the light (the information structure) is casting it that way.

3. The "Gravity" of Information

In physics, Einstein's General Relativity says that mass tells space how to curve, and space tells matter how to move.

• Mass = Structural Information (e.g., a company's debt, a government policy, a pending lawsuit).
• Curved Space = The Probability Geometry (The "Risk-Neutral" world).
• Moving Matter = The Stock Price.

The Analogy:
Think of a trampoline.

• If you put a bowling ball (a major piece of news) on the trampoline, the fabric curves.
• If you roll a marble (a stock price) across it, the marble doesn't move in a straight line; it follows the curve.
• Old View: "The marble is curving because someone is pushing it with a stick called 'Risk Premium'."
• New View (Financial Relativity): "The marble is moving in a straight line relative to the curved fabric. It only looks like it's curving because we are looking at it from a flat perspective."

The Takeaway: What we call "Risk Premium" (extra return for taking risk) might just be an optical illusion caused by looking at a curved world through a flat lens.

4. The "Flashlight" of Uncertainty

Imagine you are in a dark room with a flashlight (the market).

• Volatility (how much the price jumps around) isn't random.
• The paper suggests volatility is highest when the flashlight beam is uncertain.
• If you are 50% sure the light is on the left and 50% sure it's on the right, the beam wobbles wildly (High Volatility).
• If you are 99% sure it's on the left, the beam is steady (Low Volatility).

The paper provides a mathematical formula showing that volatility is directly determined by how much "uncertainty" remains in the geometry. It's not an external force; it's a natural consequence of the information geometry.

Summary: Why does this matter?

1. No More "Magic" Risk Premiums: Instead of guessing how much "fear" investors have, we can look at the structure of information. If the geometry is curved, prices move that way.
2. Unified Language: It connects things that usually seem separate: how we price options, how volatility behaves, and how news affects stocks. They are all just different views of the same "curved space."
3. Better Predictions: If we understand that prices are just "shadows" of information, we can build better models to see how prices will react when new "lights" (news) are turned on.

In one sentence:
This paper suggests that financial markets aren't driven by a mix of "reality" and "fear," but are instead a geometric dance where the shape of our knowledge (the geometry) dictates how prices move, just like gravity dictates how planets move.

Technical Summary

1. Problem Statement

The paper addresses a fundamental conceptual asymmetry in classical asset pricing theory. While the Fundamental Theorem of Asset Pricing (FTAP) establishes that in an arbitrage-free market, asset prices are the discounted conditional expectations of future payoffs under a risk-neutral measure ($Q$), the economic interpretation of $Q$ remains problematic.

• The Asymmetry: Pricing is governed by $Q$, but economic interpretation is anchored in the physical measure ($P$).
• The Conflict: $Q$ is often treated merely as a technical device or a preference-weighted transformation of $P$ (via a stochastic discount factor). This leads to multiple, sometimes conflicting, explanations for the same pricing relations depending on how the $P \to Q$ link is specified (e.g., via representative agent preferences, behavioral biases, or ambiguity aversion).
• The Core Question: Is $P$ inherently "more true" than $Q$, or are they simply different probabilistic reference frames corresponding to different informational constraints? The paper argues that the traditional view of $P$ as the "true" prior and $Q$ as a "distorted" posterior is an epistemic convention rather than a necessary market property.

2. Methodology: Financial Relativity

The author proposes a new framework termed Financial Relativity, which utilizes Information Geometry to reinterpret asset pricing. The methodology draws a structural analogy between General Relativity in physics and asset pricing in finance.

Core Conceptual Shift

• From Preferences to Geometry: Instead of deriving prices from utility functions and preferences, the framework posits that terminal structural information (constraints on the terminal state space) determines the curvature of the probability geometry.
• Reference Frames:
  • $P$ (Flat Background): Represents the "flat spacetime" or the least-committed, symmetric prior (maximum entropy) in the absence of structural constraints.
  • $Q$ (Curved Spacetime): Represents the "curved geometry" induced by terminal structural constraints (institutional, fundamental, or technological). It is the posterior geometry formed under these constraints.
• Price as Projection: Asset prices are not weighted averages but orthogonal projections of the terminal payoff vector ($X$) onto the current information subspace ($L^2(\mathcal{F}_t)$) within the geometry defined by $Q$.
  $$\tilde{S}_t = E^Q[X | \mathcal{F}_t] = P_{\mathcal{F}_t} X$$
  where $\tilde{S}_t$ is the discounted price.

Theoretical Tools

1. Geometric Potentials: The deviation of $Q$ from the flat prior $P$ is modeled via a geometric potential $\phi$, where $Q(w) \propto P(w)e^{\phi(w)}$.
2. Field Equations: The paper introduces "Financial Field Equations" to describe how structural sources ($I$) generate geometry ($G(Q)$), analogous to Einstein's field equations ($G_{\mu\nu} = 8\pi T_{\mu\nu}$).
3. Information Conservation: The framework utilizes entropy decomposition to show that price evolution is the gradual revelation of pre-existing terminal structural information, not the generation of new information.

3. Key Contributions

A. Reinterpretation of the Risk-Neutral Measure

The paper elevates $Q$ from a technical computational tool to the actual probability geometry employed by the market. $Q$ is not a distortion caused by risk aversion but the natural geometry resulting from structural constraints. This resolves the "Recovery Problem" (recovering $P$ from $Q$) by suggesting $P$ is not necessarily the fundamental object; rather, $Q$ is the operational geometry.

B. The Equivalence Principle in Finance

The paper establishes a financial version of the Equivalence Principle:

• Inertial Frame ($Q$): Under the correct geometric reference frame ($Q$), discounted asset prices follow a martingale (geodesic) with zero systematic drift.
• Apparent Drift ($P$): When observed from a "non-natural" frame (e.g., the flat prior $P$), the same price motion appears to have a drift (risk premium).
• Conclusion: Risk premia are not necessarily intrinsic forces compensating for risk but are apparent drifts arising from observing curved geometry through an incorrect (flat) reference frame.

C. Derivation of Financial Field Equations

The author derives specific equations for both discrete and continuous settings:

1. Discrete Field Equation:
   $$L\phi = \kappa \rho$$
   Where $L$ is a discrete Laplacian, $\phi$ is the geometric potential, and $\rho$ is the terminal structural source. This shows geometry is endogenously generated by structural sources.
2. Continuous Field Equation (Posterior Dynamics):
   $$dq_t(x) = \sigma(x - m_t)q_t(x) dW^Q_t$$
   This describes the stochastic evolution of the posterior density $q_t$ as structural constraints arrive continuously. It replaces the static "source" with a dynamic updating mechanism.

D. Endogenous Volatility

A critical result is the derivation of price dynamics where volatility is endogenously determined by posterior uncertainty (variance), rather than being an exogenous parameter.
$$dS_t = r_f S_t dt + \Sigma_t dW^Q_t$$
$$\Sigma_t \propto \text{Var}^Q(X | \mathcal{F}_t)$$
This implies that volatility peaks when uncertainty is highest (e.g., $\pi_t \approx 0.5$ in a binary outcome) and collapses as the geometry stabilizes (certainty approaches 0 or 1).

4. Results and Empirical Implications

• Unified Framework: The paper unifies no-arbitrage pricing, stochastic discount factors, filtering theory, and volatility modeling under a single "Information–Geometry–Motion" hierarchy.
• Event-Driven Dynamics: The framework explains the "rise-then-fall" volatility patterns around major events (e.g., earnings, regulatory decisions) as the endogenous adjustment of posterior geometry as constraints tighten, rather than exogenous shocks.
• Testable Hypotheses:
  1. Volatility-Variance Link: Price volatility should co-move with posterior uncertainty (variance).
  2. Geometric Reconfiguration: Major information events should induce systematic geometric shifts in implied distributions.
  3. Risk Premium Origin: Part of the observed risk premium is an artifact of the reference frame (observing curved geometry from a flat frame) rather than pure preference-based compensation.
  4. Information Efficiency: Market efficiency can be measured by the rate at which prices compress the state space (reducing conditional entropy) relative to the total terminal information.

5. Significance

• Theoretical Unification: It offers a coherent alternative to the "preference-based" paradigm, suggesting that market structure (geometry) is more fundamental than individual agent preferences. It bridges the gap between mathematical finance (martingales) and information economics (learning).
• Resolution of Ambiguity: By treating $P$ and $Q$ as symmetric reference frames defined by information constraints, it removes the ontological hierarchy that makes the "Recovery Problem" difficult.
• New Research Agenda: The paper proposes "Financial Relativity" as an extensible research program. It suggests new ways to model incomplete markets (where multiple $Q$ geometries exist), analyze market microstructure via information partitions, and develop operational measures of market efficiency based on information geometry.

In summary, Lin Li's paper argues that asset pricing is a geometric phenomenon. Prices are the observable manifestations of how terminal structural information curves the probability space, and market dynamics are the geodesic motion of these projections. This shifts the focus from "what agents prefer" to "how information structures the market's geometry."