Deep Projections and the Local Nature of the Cass Criterion

This paper introduces deep projections of indifference and offer hypersurfaces to quantify their relative curvature and applies these geometric insights to establish the local nature, sufficiency, and necessity of the Cass criterion in consumption-loan overlapping generations economies with unbounded demographic and endowment dynamics.

The Gist

Imagine you are a traveler standing at a crossroads in a vast, hilly landscape. This landscape represents the world of economics, specifically how people make choices about what to buy and sell over their lifetimes.

This paper, written by Leandro Lyra Braga Dognini, introduces a new way to map this terrain. Instead of just looking at the flat roads, the author invents a tool called "Deep Projections" to measure how much the hills curve.

Here is the breakdown of the paper's ideas using simple analogies:

1. The Three Main Landmarks

In the world of consumer theory (how people choose goods), there are three main "surfaces" or shapes that define a person's choices:

• The Trade Hyperplane (The Flat Road): Imagine a perfectly flat, straight road. This represents the budget line—the limit of what you can afford if prices are fixed. It's the "straight" version of reality.
• The Indifference Hypersurface (The Gentle Hill): This is a curved hill. Every point on this hill gives you the exact same level of happiness (utility). If you are on this hill, you don't care which specific spot you are on; you are equally happy.
• The Offer Hypersurface (The Steeper Hill): This is another curved hill, but it's shaped differently. It represents the best possible deals you can make given your preferences.

The Author's Insight:
Usually, economists look at how "curved" these hills are using complex math (like Gaussian curvature). Dognini says, "Let's look at this differently." He asks: If I start on the flat road, how far do I have to "dive" or project myself into the landscape to hit the Gentle Hill? And how far to hit the Steeper Hill?

He calls these dives "Deep Projections." They measure exactly how much "bent" or curved the hills are compared to the flat road.

2. The "Bend" Factor

The paper discovers a specific mathematical number (a formula involving a "Lagrange multiplier" and how spending changes with happiness) that acts like a curvature gauge.

• The Metaphor: Imagine the "Gentle Hill" (Indifference) is a slightly bent piece of paper. The "Steeper Hill" (Offer) is that same piece of paper, but bent even more sharply.
• The Result: The author proves that the "Steeper Hill" is always bent more than the "Gentle Hill." The formula he found tells you exactly how much more bent it is. It's like having a ruler that says, "This hill is 1.5 times curvier than that one."

3. The Big Question: Is the Economy Efficient?

The paper applies this new "curvature ruler" to a specific type of economy called an Overlapping Generations (OLG) model.

• The Scenario: Think of a society where people are born, live for two periods (young and old), and then die, while new generations are constantly being born. They trade with each other.
• The Problem: Sometimes, even if everyone is making the best choices they can, the economy as a whole might be "wasteful" or inefficient. The famous Cass Criterion is a rule used to check if an economy is efficient. It looks at the prices of goods over time. If the prices get too small too quickly, the economy might be broken.

4. The New, More Flexible Rule

For a long time, the Cass Criterion had strict rules. It assumed that the population size and how much money people had stayed roughly the same (bounded).

Dognini uses his "Deep Projections" to prove a more general version of the Cass Criterion.

• The Old Rule: "The economy is efficient only if prices don't drop too fast, assuming the population and wealth stay steady."
• The New Rule: "The economy is efficient if the prices don't drop too fast, even if the population explodes or shrinks, and even if people's wealth grows or shrinks wildly."

How he did it:
He showed that the condition for efficiency depends on the local shape of the hills (the indifference curves) right around where people are standing. He didn't need to know the shape of the entire mountain range (global information). He just needed to know how curved the ground is right under your feet.

5. The "Local Nature" Discovery

The most important takeaway is that the Cass Criterion is local.

• Analogy: Imagine you are trying to balance a ball on a hill. To know if it will roll away, you don't need to know the shape of the whole mountain. You only need to know if the ground immediately under the ball is flat, sloped, or curved.
• The Paper's Claim: The author proves that whether an economy is efficient or not depends entirely on these tiny, local curves (the Deep Projections) around each person's current situation. If the curves are "bent" enough in a specific way, the economy is efficient. If they are too flat, it might be inefficient.

Summary

This paper invents a new way to measure the "curvature" of economic choices. By using this new tool, the author updates a famous rule (the Cass Criterion) to work in much wilder, more unpredictable economic scenarios (where populations and wealth change drastically). The key discovery is that you don't need to see the whole picture to know if an economy is working well; you just need to measure the local "bend" of the choices people are making right now.

Technical Summary

Technical Summary: Deep Projections and the Local Nature of the Cass Criterion

Problem Statement
The paper addresses the characterization of Pareto efficiency in consumption-loan overlapping generations (OLG) economies. While it is established that all equilibria in these economies are weakly Pareto optimal, the First Welfare Theorem fails, meaning equilibria may be inefficient. The literature, following Cass (1972), has sought necessary and sufficient conditions for efficiency based on the asymptotic behavior of supporting price sequences. The standard Cass criterion, $\sum_{t=1}^\infty 1/\|p_t\| = \infty$, assumes bounded dynamics for demography and per capita endowments. Previous generalizations (e.g., Geanakoplos and Polemarchakis, 1991) relaxed demographic constraints but not necessarily endowment constraints. This paper aims to provide a more general statement of the Cass criterion that allows for unbounded dynamics in both demography and per capita endowments, while simultaneously revealing the "local nature" of the criterion's underlying assumptions.

Methodology
The paper employs differential topology and consumer theory to define and analyze "deep projections" of canonical manifolds.

1. Deep Projections: The author defines the trade hyperplane, indifference hypersurface, and offer hypersurface as canonical manifolds in the consumption set. The "deep projections" ($\pi_i$ and $\pi_o$) are introduced to quantify the curvature of the indifference and offer hypersurfaces relative to the trade hyperplane. These projections measure how far the manifolds "run deep" into the interior of strictly convex sets along the direction of the marginal utility vector.
2. Local vs. Pointwise Assumptions: The proof strategy distinguishes between:
   • Pointwise assumptions: Conditions depending solely on the allocation and the supporting price sequence.
   • Local assumptions: Conditions depending on the behavior of indifference hypersurfaces in a neighborhood of specific household allocations.
3. Step-by-Step Derivation: Following the reasoning of Benveniste (1976), the paper constructs a step-by-step proof for the sufficiency and necessity of the Cass criterion. It utilizes Taylor expansions and bounds on the deep projections (derived from the Hessian of the expenditure function) to establish inequalities relating price sequences to utility improvements.

Key Contributions and Results

1. Characterization of Curvature via Deep Projections:

   • The paper defines deep projections $\pi_i$ (indifference) and $\pi_o$ (offer) and derives their local behavior.
   • Theorem 2 and Corollary 3 establish that the factor $\lambda_n(p) \frac{\partial e(p, v_n(p))}{\partial u} > 0$ measures how much more "bent" the offer hypersurface is relative to the indifference hypersurface. Here, $\lambda_n$ is the Lagrange multiplier and $v_n$ is the indirect utility associated with normalized Walrasian demand.
   • Theorems 6 and 7 provide local quadratic lower and upper bounds for the deep projections of the indifference hypersurface, linking the geometry of the manifold to the consumption vectors themselves rather than just marginal rates.

2. Generalized Cass Criterion:
   The paper derives a generalized statement for the Cass criterion that accommodates unbounded dynamics for both demography ($H_t$) and per capita endowments. The condition for Pareto optimality is given by:
   $$\sum_{t=1}^\infty \frac{1}{\|p_t\| \sum_{h \in G_t} \|c^h_t\|} = \infty$$

   • Theorem 9 (Sufficiency): Under specific local and pointwise assumptions (Assumptions 1–4), if the series above diverges, the allocation is Pareto optimal.
   • Theorem 10 (Necessity): Under a complementary set of assumptions (Assumptions 5–9), if the series converges, the allocation is not Pareto optimal.

3. Revealing the Local Nature of the Criterion:
   A primary contribution is demonstrating that the validity of the Cass criterion relies on local properties of the indifference hypersurfaces (specifically, that they do not flatten out to the trade hyperplane in the limit).

   • Example 3 illustrates a case where the Cass criterion fails because the indifference curves become increasingly flat (approaching the trade hyperplane), violating the local assumptions regarding the deep projections. This highlights that the criterion's validity is not merely a global property of prices but depends on the local curvature of preferences.

Significance and Claims
The paper claims to provide a "relevant milestone" for the deployment of OLG models in diverse economic applications, including monetary theory, bubble literature, and social security design, by allowing for unbounded dynamics in demography and endowments.

Crucially, the paper asserts that the deep projections serve as a complementary analytical tool to traditional curvature measures (such as Gaussian curvature). The author claims these projections are "possibly more convenient for economic applications" because they offer a tractable method for measuring curvature that directly reveals the local nature of efficiency conditions without requiring global information about the indifference hypersurfaces. The results suggest that the Cass criterion is fundamentally a local phenomenon, contingent on the specific geometric behavior of preferences in the neighborhood of the equilibrium allocation.