Tax Migration as Social Contagion: A Tipping-Point Model with Application to the Scandinavian Wealth Tax Debate

Technical Summary: Tax Migration as Social Contagion

Problem Statement
The paper addresses the empirical and theoretical validity of extrapolating micro-level tax migration estimates to macroeconomic GDP impacts, specifically in the context of the 2026 Danish wealth tax debate. The central issue is the reliance on Blandhol (2025), which estimates that wealth-tax-induced emigration from Norway reduces long-run GDP by 1.3%. This figure was scaled linearly by Dansk Industri (DI) to argue that a Danish wealth tax would cost billions. The author challenges the "micro-to-macro" extrapolation underlying this figure, arguing that it rests on five identification conditions that are simultaneously violated. The paper posits that tax migration among the ultra-wealthy is not a smooth, independent elasticity but a social contagion process exhibiting tipping-point dynamics, rendering linear cross-country scaling invalid.

Methodology
The paper employs a hybrid theoretical and empirical approach:

1. Social Contagion Model: The author develops a scalar ordinary differential equation (ODE) for the emigrant fraction $n(t)$. The effective emigration rate $\lambda$ is modeled as a product of four components:

   • Baseline Rate: Driven by tax differentials, reference-dependent tax pressure (prospect theory), and policy regime hostility.
   • Social Contagion: An exponential multiplier $\Phi(\tilde{n}) = e^{\kappa \tilde{n}}$, where $\tilde{n}$ is a visibility-weighted fraction of prior emigrants (the "celebrity effect").
   • Exit-Tax Anticipation: A factor accelerating departure based on the subjective probability of future exit taxes.
   • Return Migration: A constant rate $\rho$.
     The model is analyzed for saddle-node bifurcations to identify tipping points where the system jumps discontinuously from a low-emigration to a high-emigration equilibrium.

2. Fokker–Planck Framework: The migration model is embedded within a Fokker–Planck framework for wealth distributions (drawing on companion papers by Frøseth, 2026b,d,c). This framework models wealth evolution as geometric Brownian motion with demographic turnover. It establishes the "drift-shift symmetry" of proportional wealth taxes and characterizes migration as a "permeable boundary" leakage. This allows for the derivation of the Pareto tail exponent and the analysis of how tax policy affects the wealth distribution's steady state.

3. Empirical Analysis (Kapital 400 Panel): The author constructs a panel of the 400 wealthiest Norwegians (2011–2025) from the Kapital magazine list. The data is used to:

   • Estimate the Pareto tail exponent ($\hat{\alpha}$) directly.
   • Identify specific emigrants within Blandhol's sample window (2016–2020).
   • Distinguish between "active-control" owners (who retain voting rights) and "passive" wealth holders (heirs with economic exposure only).
   • Document a "hidden channel" of heir-emigration where controlling owners remain in Norway while heirs emigrate with B-shares.

Key Contributions

1. Theoretical Mechanism (Tipping Points): The paper demonstrates that the emigration system admits a saddle-node bifurcation when the contagion strength $\kappa$ exceeds a critical threshold ($\kappa_{crit} = 4$). In this regime, two stable equilibria (low and high emigration) coexist. The Norwegian 2022 episode, where emigration rates jumped from ~0.2% to >2%, is presented as consistent with a discontinuous tip rather than a smooth response to tax changes.

2. Identification of Five Failures in Extrapolation: The paper systematically dismantles the Blandhol/DI extrapolation by showing five conditions required for validity are violated:

   • Wealth-Weighted Representativeness: The sample used for micro-estimates (mid-tier wealthy) does not represent the wealth-weighted population. The GDP impact is dominated by the top fortunes, none of whom emigrated during the sample window with active domestic control.
   • Time Horizon: The 5-year observation window is too short to separate structural drift from stochastic noise (diffusion), as the relaxation half-life of the wealth distribution is ~21 years.
   • Outcome Variable Confounding: Firm revenue (the outcome in Blandhol's study) conflates control effects with ownership effects. Since many emigrants are passive heirs, revenue declines may reflect generational transitions or industry shocks, not a "productivity haircut."
   • Tax-Wealth Mapping: The sample is selected based on taxable wealth (book value), while the GDP integral depends on market wealth. The nonlinear mapping between these (due to book-value assessment discounts) is ignored in the extrapolation.
   • Confounded Regime Shocks: The emigration wave coincided with the abolition of the five-year exit-tax lapse rule and the crystallization of deferred capital gains. The response cannot be isolated to the wealth tax alone.

3. The "Hidden Channel" of Heir-Emigration: Using the Kapital 400 data, the author identifies 36 recent heir-emigrants carrying approximately 127 billion NOK. In these cases, the controlling owner retains A-shares (voting rights) and domestic residency, while the heir emigrates with B-shares (economic rights). This results in a "productivity haircut" ($\eta$) of near zero, as the enterprise's management and human capital remain in Norway. This channel is invisible in standard panel data but dominates the wealth-weighted emigration volume.

4. Asymmetry of Assessment Regimes: The paper contrasts the Norwegian system (book-value assessment with statutory discounts, creating anisotropic drift and portfolio distortions) with the proposed Danish system (market-value assessment, preserving drift-shift symmetry). It argues that the behavioral response in Norway was driven partly by these specific distortions, which are absent in the Danish proposal, making linear scaling invalid.

Results

• Pareto Stability: The Pareto tail exponent for the ultra-wealthy in Norway is estimated at $\hat{\alpha} \approx 1.3$ and remains stable across the 2011–2025 period, including the post-reform years. This suggests the tax change did not alter the fundamental shape of the wealth distribution's tail.
• Visibility Weight: The ratio of visibility-weighted emigrants to head-count emigrants is estimated at $\hat{\xi} \approx 1.1$, indicating that high-wealth departures exert a disproportionately large influence on the contagion process.
• Non-Identifiability of Contagion Strength: While the data is consistent with a tipping event, the specific contagion strength parameter $\kappa$ cannot be identified from a single historical episode. However, the data rules out the assumption of independent, rational decision-making ($\kappa=0$) required by the DI scaling exercise.
• GDP Impact: The wealth-weighted integral determining GDP loss is controlled by individuals absent from the event-study sample or those whose emigration carries negligible productivity costs (passive heirs). Consequently, the 1.3% GDP loss estimate is uninformative regarding the true aggregate cost.

Significance and Claims

The paper claims that the Norwegian emigration wave is a non-scalable, path-dependent tipping event, not a smooth elasticity. Its significance lies in:

1. Methodological Correction: It argues that standard micro-elasticity tools break down when behavioral responses exhibit tipping-point dynamics. Policy analysis should focus on the distance to the tipping point (determined by network structure, regime hostility, and exit-tax rules) rather than attempting to scale elasticities across jurisdictions.
2. Policy Implications for Denmark: The paper concludes that the Danish proposal is fundamentally different from the Norwegian episode. Denmark lacks the specific combination of a regime shift, a closing exit-tax loophole, and the accumulated deferral wedge of the Norwegian fritaksmetoden. Furthermore, the Danish proposal's market-value base avoids the portfolio distortions that characterized the Norwegian system. Therefore, the DI extrapolation answers the wrong question: it estimates the cost of replicating the Norwegian episode, not the cost of introducing a wealth tax under Danish conditions.
3. Theoretical Insight: It provides a formal mechanism (social contagion with visibility weighting) for understanding why tax migration can cluster geographically and temporally in a discontinuous manner, consistent with the observed "rush-for-the-door" dynamics.

The paper remains modest regarding the precise quantification of the contagion strength, acknowledging that the available data cannot pin down $\kappa$ but is sufficient to invalidate the assumptions underpinning the linear extrapolation used in the Danish political debate.