Impact of Higgs-boson measurements on SMEFT fits

TL;DR

This work assesses how Higgs-boson measurements constrain dimension-6 SMEFT operators under two flavour-symmetric NP scenarios, using a LO RG framework to evolve Wilson coefficients from a high NP scale $\\Lambda$ down to the electroweak scale and performing a Bayesian global fit with the HEPfit tool. The analysis shows that Higgs observables significantly tighten bounds on several bosonic operators and, in the $U(2)^5$ case, on third-generation fermionic and dipole operators as well, pushing the NP scale upper reach toward $\\sim 15$–$20$ TeV in favorable fits. Results are compared with recent literature and shown to be broadly consistent, despite differing flavour assumptions and RG treatments. The study also outlines future directions, including NLO SMEFT renormalization-group improvements, broader observable sets, and possible extensions beyond dimension-6, to sharpen NP constraints and enable robust model discrimination.

Abstract

We present current bounds on SMEFT operators that are mainly constrained by Higgs-boson observables, under different assumptions for the flavour structure of the UV theory. We investigate how the accuracy reached through a dedicated Higgs-boson precision physics program is starting to have a major impact in setting a lower bound on the scale of new physics, and we discuss the relevance of considering the scale evolution of the SMEFT coefficients in this context. We compare our results with the literature, pointing out the consistency of the results in spite of the different assumptions adopted in each analysis, and we discuss future steps aimed at improving the accuracy of the fit.

Figures (2)

• Figure 1: Comparison of constraints from a full fit and a fit that does not include Higgs-boson observables, for the $U(2)^5$ case. The bounds on the bosonic-operator Wilson coefficients also apply to the $U(3)^5$ case. The scale of NP has been set to $\Lambda=3$ TeV. The limits shown correspond to the 95% HPDI. Results are presented with and without the RGE effects (the latter also adjusted to a value of $\Lambda=3$ TeV for this comparison), following the colour scheme indicated in the legend.
• Figure 2: Results from individual fits in the $U(2)^5$ flavour symmetric SMEFT. For each coefficient $C_i$, the top panel shows the scale of NP allowed by the data at 95% probability (normalized by the square root of the maximum of the 95% HPDI interval for $|C_i|$). The bottom panel shows the width of the 95% probability range divided by two. Both panels show results for 1) the full fit with RGE (for $\Lambda=3$ TeV) and without RGE (for $\Lambda=1$ TeV), and 2) the fit that does not include Higgs-boson observables. The color code is as explained in the legend. The horizontal lines indicate the maximum value allowed for each Wilson coefficient in the fit, corresponding to the perturbativity limit $4\pi$ (see Section \ref{['sec:framework']}). The cases in which the $95\%$ HPDI interval touches the prior's edges, indicated in red in Table \ref{['tab:u2-ind']}, are hatched with red diagonal lines. When the posterior distribution of a coefficient is completely flat the $95\%$ HPDI interval is hatched with diagonal white lines.