A firm's cost of capital is the market-value-weighted blend of what debt and equity holders require: WACC = w_d·k_d·(1 − T) + w_e·k_e. Equity is priced by the CAPM, k_e = r_f + β·ERP, while debt is taken after the interest tax shield, k_d·(1 − T). Only debt earns the shield, so it usually sits below equity in the blend.
Shift weight toward debt and WACC usually falls, because shielded debt is cheaper than equity. Raising the tax rate deepens the shield and lowers the after-tax cost of debt.
Note: only debt carries the (1 − T) tax shield, since interest is tax-deductible and dividends are not. The weights w_d and w_e should be market values of debt and equity, not book values, because WACC is the return the market requires today.
Reflect: if cheap shielded debt always lowers WACC, why not finance the firm almost entirely with debt? What does this model leave out about how k_d and k_e respond as leverage climbs?