How does partisan conflict (PC) relate to economic policy uncertainty (EPU)? When does uncertainty feed back into conflict, and how stable is that link? We examine these dynamics using a time-varying Granger causality framework applied to United States monthly data from 1985 to 2025, combining the PC Index with aggregate and disaggregated EPU subindices. PC consistently precedes spikes in fiscal, tax, spending, and health care uncertainty. In contrast, the predictive direction reverses mainly during large shocks, such as trade conflicts or financial regulation crises, and largely disappears during recessions. Generalised forecast error variance decompositions show that PC shocks account for a modest but non trivial share of forecast error variance in fiscal, tax, spending, and health care uncertainty, while policy shocks in trade and regulatory domains explain a much larger share of the variance in partisan conflict itself. These results reveal asymmetric, time-varying predictive linkages between political polarization and policy uncertainty. The findings, which are based on the decomposition of EPU into its components and time-varying Granger causality, contribute to theories of veto players, gridlock, and signaling by identifying the institutional conditions under which political conflict is associated with greater economic volatility and when policy uncertainty appears to reinforce partisan division.