As a candidate, President Trump said that he planned to “cancel every unconstitutional executive action, memorandum and order issued by President Obama.” According to a briefing from then-Vice President-elect Mike Pence, immediately after the inaugural parade, the new president would “go into the Oval Office with a stack of papers” and “start signing them to roll back” Obama’s executive actions. As of Jan. 21, the news media had reported the signing of only one such executive order (aside from some routine administrative actions), which affected the Affordable Care Act.

Presidential executive orders are a comparatively popular tool. If we tally them all from Harry Truman (1945-1953) up through Barack Obama, we get a total of 4,244. Different presidents, of course, leaned on them to different degrees: At the high end, Truman averaged 117 each year; Obama’s average was a mere 35 .

Presidential candidates often promise to change policy with the “stroke of a pen” on “day one” of a new administration. Such rhetoric implies that a flurry of executive orders issued immediately after inauguration — like the calendar leaves falling in an old movie — could quickly reverse the previous president’s course.

But how often do presidents reverse executive orders — especially in the first days of a new administration — after a party turnover?

Here’s how we checked.

We used the National Archives’ Executive Orders Disposition Tables to determine how many executive orders were explicitly “revoked” or “superseded” by each of the 12 presidents who have served since Franklin D. Roosevelt.  We counted reversals within newly elected presidents’ first 30 days in office. That count was quite low. Nine presidents reversed three or fewer in that time; Obama hit the high, reversing eight previous presidents’ executive orders within 30 days.

When we checked to see how many orders were reversed within a new administration’s first 60 days, the numbers were still relatively low. Six presidents still reversed three or fewer. Gerald Ford took second place, reversing 17, while Ronald Reagan took top place, reversing 22 of his predecessors’ executive orders.

What about over their entire terms? There George W. Bush takes top place, reversing 64 executive orders. The rest reversed 52 or fewer. Their annual averages ranged widely. Bill Clinton reversed 5.5 per year; John F. Kennedy reversed 10.3. Richard Nixon, Ford and George W. Bush each averaged eight or more reversals annually. You can see the range of our data in the table below.

Note that these are not the number of executive orders issued by presidents, but rather the number of previous executive orders reversed by each president; some executive orders reverse more than one previous executive order.

Let’s look at the totals.

Out of those 4,244 executive orders issued by Truman and presidents since, their White House successors reversed only 508, or about 12 percent of the total. Only 5 percent of those reversals came during the first 30 days of their terms, and only 14 percent during the first 60 days.

Wouldn’t you think that when the White House changed parties — when a Republican took over after a Democrat, or vice versa — the new occupant would reverse more executive orders? It’s not so. Interestingly, the annual average of reversals didn’t differ much after party turnover transitions (7.8 percent) or party continuity transitions (7.1 percent). Nor do presidential reversals vary much by party: Republicans revoked roughly seven orders on average, Democrats eight.  That may be because most executive orders are routine and administrative.

Only the rare executive order truly sets or changes policy — and those keep changing.

A few flip back and forth as presidents — and the political climate — changes.

For instance, in reaction to Nixon’s abuses of presidential power, Jimmy Carter replaced Nixon’s Executive Order 11652 on secrecy and classification by issuing his own Executive Order 12065, requiring more open access to government policies and documents. His successor, Reagan, then reversed Carter’s order with his own (12356), tightening secrecy again. Clinton revoked Reagan’s order with Executive Order 12958, pushing the government back toward openness. George W. Bush then issued Executive Order 13233 to assert more presidential control over the release of governmental records. And shortly after his inauguration, Obama revoked the Bush order and later replaced it with his own approach to classification (EO 13526).

There’s more to executive power than executive orders.

Of course, as Andrew Rudalevige regularly writes here at TMC, many unilateral actions can be taken without issuing executive orders. Memoranda (previously “letters”) have the same legal authority as executive orders, although they are often not listed in the Federal Register, as is required for executive orders.

For instance, Reagan’s “Mexico City Policy,” which denied U.S. government funding for nongovernmental organizations that provided “advice, counseling, or information regarding abortion,” was reversed and then reinstated in a series of memoranda by presidents Clinton, George W. Bush and Obama. Trump will likely make another reversal.

In addition, presidents have other ways of changing the direction of government. Even without issuing a formal directive by executive order or memorandum, presidents can institute important policies by, for example, instructing agency heads to issue guidelines. This is how the Department of Homeland Security implemented Obama’s instruction to carry out parts of the “Dream Act” without congressional approval, through guidances known as Delayed Action for Childhood Arrivals (DACA) and Deferred Action for Parents of Americans (DAPA). DHS then prioritized the implementation of deportation of undocumented persons, later reversed by the U.S. Court of Appeals for the 5th Circuit.

But to sum up, the rhetoric of sweeping, day-one change is almost always belied by the reality of presidential life. It will be interesting to see whether Trump meant his promises to reverse Obama policies literally — and will issue that flurry of executive orders — or whether he meant it seriously, and will reach his goals through other means. Or perhaps changing the direction of the ship of state will be more difficult than he anticipated.

James P. Pfiffner is University Professor of Public Policy at the Schar School of Policy and Government at George Mason University.

Joshua Lee is a doctoral student in public policy at the Schar School of Policy and Government at George Mason University.