1\section{Results II: who delegates, who owns, and the wealth distribution}2\label{sec:results_b}34Table~\ref{tab:b_scenarios} reports the stationary equilibria of Model B across the pre-registered scenarios and5Table~\ref{tab:b_decomp} (Appendix~\ref{app:extra}) the channel decomposition relative to the no-AI economy. The6no-AI economy is a standard Aiyagari calibration: $r=\numrnoai\%$, Gini \numgininoai, top-10\% share7\numtoptennoai. All numbers come from \texttt{scripts/20\_model\_b.py}.89\input{tables/model_b_scenarios}1011\subsection{The decision-quality channel concentrates wealth even without any production effect}1213The ``delegation only'' scenario switches on the information channel alone: no compute in production, no change14in factor prices other than through saving. The rich delegate first (Figure~\ref{fig:B_lorenz}(b)): with a fixed15component in the fee and a stake proportional to wealth, delegation is zero in the bottom deciles and universal at16the top, the cross-sectional counterpart of Proposition~\ref{prop:threshold}. Because delegated savings earn a17higher and less noisy return, the delegators accumulate faster: the Gini rises from \numgininoai\ to18\numginidelegationonly\ and the top-10\% share from \numtoptennoai\ to \numtoptendelegationonly, while output rises by19only \numdYdelegationonly\% (the resource cost of human mistakes is saved). This is H3 in a form that does not depend20on who owns the AI infrastructure: better decisions are a scale-dependent return, and scale-dependent returns21concentrate wealth.2223\begin{figure}[htbp]24\centering25\includegraphics[width=\textwidth]{figB_lorenz_delegation.pdf}26\caption{Model B. (a) Lorenz curves of wealth across scenarios. (b) Share of households delegating within each wealth27decile (households only; the owner class of the concentrated scenarios is excluded). Source:28\texttt{scripts/20\_model\_b.py}.}29\label{fig:B_lorenz}30\end{figure}3132\subsection{AI as a production input works in the opposite direction}3334Adding compute to production (``production only'' and the ``both'' scenarios) raises output by35\numdYfullbroad\% in the broad-ownership case, lowers the labour share from 0.67 to about 0.44, and lowers the36interest rate. With broad ownership this \emph{reduces} wealth inequality (Gini \numginifullbroad, top 10\%37\numtoptenfullbroad): wages rise, the return to wealth falls, and the compute income flows to everyone in proportion38to savings. The net effect of the two channels under broad ownership is therefore an equalisation---a result that39depends entirely on the ownership assumption.4041\subsection{Ownership}4243Under concentrated ownership (a measure-0.10 owner class holding the scarcity rents, Proposition~\ref{prop:ownership})44the same technology yields a Gini of \numginifullconcentrated\ and a top-10\% share of \numtoptenfullconcentrated,45against \numginifullbroad\ and \numtoptenfullbroad\ under broad ownership; the scarce-compute variants (fixed46stock, larger rents) show the same ordering. Figure~\ref{fig:B_shares} summarises: automation moves the factor47shares identically in both cases, and the distributional outcome is decided by the ownership structure of48algorithmic capital, which is H3 as stated. The interest rate is also higher under concentration because the rents are49not available to finance capital accumulation through the household fund.5051\begin{figure}[htbp]52\centering53\includegraphics[width=\textwidth]{figB_shares_concentration.pdf}54\caption{Model B. (a) Factor income shares (labour, capital, compute) by scenario. (b) Top-10\%, top-1\% wealth55shares and Gini. Source: \texttt{scripts/20\_model\_b.py}.}56\label{fig:B_shares}57\end{figure}5859\subsection{Welfare by wealth percentile and the controls}6061Figure~\ref{fig:B_welfare} converts the value functions at given wealth percentiles into consumption-equivalent62gains relative to the no-AI economy. Delegation alone benefits the top of the distribution and leaves the bottom63unchanged; AI in production benefits every percentile through wages, more so at the bottom where labour income64dominates; concentration shifts the gains upward. Two controls behave as pre-registered: when human mistakes are65switched off (``no mistakes''), nobody delegates ($D=0$, the $\tau_A=\tau_H$ falsification test of H1 in the66cross-section), and a cheaper fee raises adoption (\numDfullbroad\ to 0.87) and lowers inequality slightly because67delegation reaches further down the distribution. Removing borrowing (``no borrowing'') changes little, because the68borrowing limit binds for households who do not delegate in any case.6970\begin{figure}[htbp]71\centering72\includegraphics[width=0.7\textwidth]{figB_welfare_percentile.pdf}73\caption{Model B: consumption-equivalent welfare gain relative to the no-AI economy at selected wealth percentiles.74Source: \texttt{scripts/20\_model\_b.py}.}75\label{fig:B_welfare}76\end{figure}77