What Is the Alignment Tax?

Robin Young

Home/Archive/arXiv

arXiv · 2603.00047

What Is the Alignment Tax?

Robin Young

View original source ↗Read source document ↗Export metadata ↓

Abstract

The alignment tax is widely discussed but has not been formally characterized. We provide a geometric theory of the alignment tax in representation space. Under linear representation assumptions, we define the alignment tax rate as the squared projection of the safety direction onto the capability subspace and derive the Pareto frontier governing safety-capability tradeoffs, parameterized by a single quantity of the principal angle between the safety and capability subspaces. We prove this frontier is tight and show it has a recursive structure. safety-safety tradeoffs under capability constraints are governed by the same equation, with the angle replaced by the partial correlation between safety objectives given capability directions. We derive a scaling law decomposing the alignment tax into an irreducible component determined by data structure and a packing residual that vanishes as $O(m'/d)$ with model dimension $d$, and establish conditions under which capability preservation mediates or resolves conflicts between safety objectives.

Explore related subjects

econ.EM cs.AI cs.LG math.OC

Keep this discovery

Explore connections, maps & timelines

BibTeX RIS

Robin Young. 2026-03-03. What Is the Alignment Tax?. https://arxiv.org/abs/2603.00047

Cite the original work for its findings. Save a collection to share your selection of sources.

KEEP EXPLORING

Related papers

arXiv2026

Testing for Monotone Equilibrium Strategies in Games of Incomplete Information

This paper develops a unified framework for testing monotonicity of Bayesian Nash equilibrium strategies in unobserved types in games of incomplete information. We show that, under symmetric independent private types, monotonicity of differentiable equilibrium strategies is equivalent to monotonicity of a quasi-inverse strategy identified from observed actions. This allows the problem to be reformulated as testing a countable set of moment inequalities involving unconditional expectations. We propose a Cramer-von Mises-type statistic with bootstrap critical values. The method accommodates covariates and game heterogeneity. Monte Carlo simulations demonstrate finite-sample performance, and an application to procurement auctions illustrates cartel detection.

econ.EM↗

arXiv2026

Ordinal Distributional Change and Conservative Transition Benchmarks: Measurement, Identification, and Inference

Repeated cross-sections reveal changes in ordinal distributions but not the transitions producing them. I axiomatically characterize a threshold-weighted probability metric for ordinal change from threshold-crossing principles. For any threshold-additive ordinal geometry, the discrepancy coincides with the Wasserstein--1 distance induced by that ground metric and measures minimum displacement; its optimizing plans define conservative transition benchmarks. With missing outcomes, I derive sharp identified sets for the discrepancy and endpoint-conditioned benchmark plans. I develop finite-sample-valid projection inference using randomized Monte Carlo calibration and global search with an almost-sure convergence guarantee. Applied to Arab Barometer data, the framework documents a robust shift toward broader and more regular remittance receipt in Lebanon. The discrepancy interval remains well separated from zero after allowing for item nonresponse and sampling uncertainty, while benchmark bounds provide strong numerical evidence that least-displacement restructuring excludes movement toward less frequent receipt and requires reassignment from nonreceipt to recurrent receipt.

econ.EM↗

arXiv2026

A Stochastic Nested Fixed Point Algorithm for Large-Scale BLP Estimation

We develop a stochastic nested fixed point (SNFP) estimator for random coefficients logit demand models that updates model parameters using stochastic gradients and performs demand inversion one market at a time. Relative to the conventional nested fixed point (NFP) estimator, SNFP substantially reduces memory requirements and computational cost, making estimation feasible in very large datasets. We establish the large-$T$ (number of markets) asymptotic properties of the estimator under regularity conditions. We also characterize the effect of sharing one block of simulation draws across markets and show how to correct for it. Monte Carlo simulations show that the SNFP estimator achieves statistical accuracy comparable to the NFP estimator, and in our benchmark a single online pass estimates a model with 100 million markets in about 5.5 hours. An empirical application using scanner data further demonstrates the practical advantages of SNFP for large-scale demand estimation.

econ.EM↗