After the midterms, Congress needs to take a hard look at the Pentagon’s leadership shake-up. Secretary of War Pete Hegseth is replacing senior commanders while pursuing a 20% reduction in general and admiral positions by January 1.

On October 28, Hegseth removed Lt. General Brett Sylvia, the Joint Staff’s director for strategy, plans and policy, amid allegations of leaking classified information. His departure follows a string of dismissals and pressured retirements.

Some of those changes may be justified. But will they produce the leadership America needs as artificial intelligence, drones, missile attacks, and cyber-warfare make military operations more demanding?

Before settling on a number of generals and admirals, the Pentagon needs to examine the missions, headquarters, decision authority, and supporting staffs those officers oversee.

Sylvia’s case has gone to the Justice Department, and the allegations require evidence and a fair process. That investigation should be judged separately from the need to reduce senior positions. A necessary headquarters can have an unsuitable leader. An excellent officer can occupy a position the military no longer needs.

The J5’s responsibility is to help the chairman connect military options to national objectives and assess strategic risk. Whatever the outcome of Sylvia’s case, that work remains.

I spent 24 years in uniform and more than two decades working as a civilian strategist inside the Pentagon. I support reducing unnecessary senior positions and giving capable colonels and Navy captains greater responsibility. Changing the rank above an office door, however, accomplishes little if the paperwork, overlapping responsibilities, and approval requirements remain.

The military has roughly 850 active-duty generals and admirals for approximately 1.3 million personnel. An analysis published by National Defense University Press estimated more than 2,000 for over 12 million during World War II. Today’s proportion is roughly four times higher.

The ratio invites scrutiny, but it does not tell us which positions to cut. Wartime mobilization expanded the force dramatically, and nuclear, space, and cyber responsibilities cannot be measured simply by counting troops.

General George Marshall, the Army chief of staff, reorganized the service during a far larger war. In early 1942, administrative demands and competing agencies were consuming time he needed for strategy. The March 1942 reorganization consolidated responsibilities under three major commands: Army Ground Forces prepared ground units for combat; Army Air Forces managed the air arm; and Services of Supply handled supply and support. It abolished General Headquarters and trimmed several General Staff divisions, leaving Marshall freer to direct the war.

Marshall also pushed work and authority down to the new commands and enlarged the Operations Division that directed the war. He cut some staffs while strengthening the functions most important to winning.

That selective approach matters as America confronts Iran, supports Ukraine, deters China, expands its Arctic presence, and protects its interests in space and cyberspace. The new Greenland agreement expands Pituffik Space Base and adds defense areas at Narsarsuaq and Mestersvig. Such missions demand Arctic expertise. They do not automatically justify another layer of Washington staff.

Artificial intelligence may reduce routine reporting and analysis. But commanders will still have to determine whether machine recommendations are reliable, targets are lawful, and attacks could widen a war. We need officers who understand the technology well enough to challenge it and who have the authority to intervene when it fails. Faster analysis could eliminate some headquarters layers, while more complex operations may require stronger leadership in particular missions. The number and rank of officers should follow those responsibilities.

Golden Dome, Trump’s proposed layered missile defense for the homeland, illustrates the challenge. Protecting Americans requires linking sensors, interceptors, communications, and commanders across services. Another headquarters helps only if it clarifies responsibilities and improves performance in the field.

The same test applies to Hegseth’s planned four-star Autonomous Warfare Command, intended to accelerate drones and robotic systems across the services. The Army is also creating its own autonomous systems command. The Pentagon aims to establish the joint command in October 2027. Congress will need to examine its funding and authorities, while the Pentagon explains how it will work with existing commands and service programs. A new four-star headquarters should identify which responsibilities it will assume and which existing structures can then shrink or disappear.

Then there is personnel policy. Must exceptional officers become generals to keep serving effectively?

Hyman Rickover, the nuclear propulsion pioneer, was passed over twice before selection for rear admiral in 1953. His experience warns that conventional promotion systems can misjudge unusual talent. But it does not prove that every valuable specialist needs a star.

The services should examine longer specialist careers, meaningful assignments, and targeted compensation for exceptional colonels and captains. Congress can change career and compensation rules where necessary. Promotion should reflect responsibilities requiring higher rank; retaining expertise should not depend on inventing those responsibilities.

The same scrutiny should extend to headquarters personnel rosters. Determine which duties require commissioned officers, which warrant officers or enlisted leaders can perform, and which belong with civilian specialists. Include contractors, and eliminate unnecessary work before shifting its burden elsewhere.

The Goldwater-Nichols reforms of 1986 strengthened joint leadership and clarified command relationships. Who decides can matter as much as how many people are involved.

Hegseth and the chairman should require every major headquarters to justify its mission, decision authority, senior ranks, and supporting staff against warfighting requirements. Support functions, though indispensable, must show how they sustain the force.

The lame-duck defense debate gives Congress the chance to demand that accounting. The Senate Armed Services Committee’s bill would authorize roughly $1.1 trillion in fiscal 2027, less than the administration’s $1.5 trillion request. But authorization is not appropriation. Lawmakers must align assigned missions with the funding and legal authorities needed to accomplish them.

The Pentagon should then test its proposed structure against simultaneous crises, disrupted communications and mobilization, measuring costs, decision speed, and operational performance, including civilian and contractor support.

Officers who undermine lawful civilian direction or substitute political fashion for readiness should be replaced. But officers also owe civilian leaders candid professional advice, even when it is unwelcome. Each personnel decision should rest on demonstrated conduct and competence.

Hegseth may find that fewer generals can lead a stronger military. The troops deserve clear orders, competent commanders, and headquarters that help them fight. Congress should judge the overhaul by those results. Readiness matters more than any personnel deadline.