In his First Year essay, “The hundred days myth,” historian David Greenberg offers the incoming Trump Administration some sage advice: “Pay a bit less attention to the hype about the first hundred days. Historically, that arbitrary benchmark has rarely correlated with the subsequent success or failure of a president’s time in office.”

How true! With the obvious exception of Franklin D. Roosevelt, those remembered as our “greatest” presidents did not enjoy exceptional success during their first three-and-a-half months.

Abraham Lincoln certainly did not—in the week following his inauguration, the Congress of the Confederate States adopted a permanent constitution. The next month, the Civil War commenced at Fort Sumter. And by the end of Lincoln’s third month in office, four more states had seceded. No major legislation was passed during that time. Yet Lincoln is remembered as the greatest of them all.

So what, then, is “correlated with the subsequent success or failure of a president’s time in office?”

Greenberg suggests that real success has little to do with the president’s ability to rack up a list of accomplishments in the first 100 days, and much more to do with whether the president can “set a new ideological direction for the country.”

Therefore, presidents would do well, he says, to stay focused on “the work that goes on in forging the economic direction of the country,” such as developing smart budgets and economic programs. This work will influence “what writers will say not in one hundred days, nor even one thousand, but over the lifetime of the administration and afterward.”

This is clearly good advice. Presidents should certainly set smart priorities and work to develop forward-looking economic plans. But even when they do, being remembered as “successful” or “great” will not always, or even very often, be within the president’s control. The president’s capacity to bring about the changes they seek is severely constrained by structural forces and contextual conditions about which they can do little.

For starters, presidents are highly constrained by the constitutional system and the limited powers of the presidency, especially relative to the Congress, but also to the Courts and the states. The framers of the Constitution, of course, limited presidential power and established a “separated system” on purpose. George W. Bush once quipped, “If this were a dictatorship, it would be a heck of a lot easier.”

To be sure, there is still a lot that presidents can get done unilaterally. After Democrats lost control of both houses of Congress in 2014, Obama told his Cabinet that even without the support of Congress, “I’ve got a pen, and I’ve got a phone,” indicating he would circumvent Congress by signing executive orders and rallying outside groups to help advance his agenda in non-legislative ways. But the limited ability of those strategies to actually accomplish Obama’s grand ambitions only reinforced the point. Presidents need the support of Congress to accomplish their major goals.

Yet Congress consists of 535 individual members, each of whom has different constituencies, interests, time horizons, and policy goals. This means that structurally speaking, the incentives of the president and members of Congress will not always, or even very often, align. Even when the president’s party controls both houses of Congress, support for the president’s initiatives is never guaranteed.

Indeed, the president’s prospects for legislative success depend to a large extent on their party’s competitive standing in Congress. How many seats do their fellow partisans hold in the House of Representatives and in the Senate? This, too, is almost entirely out of their control. But it matters more than just about everything else.

Most importantly, the partisan makeup of government shapes the content of legislation. Scholars have found that presidents who have faced a Congress controlled by the opposition party have actually signed “significant” legislation into law at a healthy rate. But the content of that legislation is almost always inferior, from the president’s perspective, to what it would have looked like under unified party government.

But even when presidents face unified government—as President Obama did in 2009-2010 and Trump will in 2017-2018—they often find that their ability to influence their fellow partisans is limited. In fact, the larger the party’s majorities, the more heterogeneous it tends to be, and the more difficult it is to get all factions to work together in concert. Trying to “herd the cats” is a major challenge that any president faces, whatever their natural political talents and skills.

A long tradition of scholarship has defined presidential leadership in precisely this way, as the struggle to enhance the president’s “personal capacity to influence the conduct of the men who make up government” (Neustadt 1960). Despite enjoying many years of unified party government during the 20th century, Democratic presidents usually found their party’s friendly majorities less friendly, and unified government less unified, than they had hoped. Such is the nature of political parties in a separated system of governing.

Moreover, there is little evidence presidents can do much to affect the partisan complexion of Congress. Given structural factors like congressional districts drawn for partisan purposes, unequal representation in the Senate, and staggered elections for members of Congress, presidents can try, but rarely succeed, in significantly altering the balance of party power in the legislative branch.

Given all of this, it is clear that even the most ambitious and highly skilled presidents will face formidable structural and institutional constraints. Try as they might to prioritize the right issues and alter the direction of the country through sound public policy, the constitutional system and the nature of political parties makes it extremely difficult for presidents to achieve their goals even in the “best” of circumstances.

(That is, speaking from the president’s perspective. It is worth remembering that the Founding Fathers were less concerned with the president’s ability to rack up legislative accomplishments in the first 100 days, or at any point for that matter, than with the stability and durability of the entire system—an objective they thought better advanced by limiting the president’s powers).

Finally, it is worthwhile to contemplate the limited tools presidents have to affect the direction of the economy. While the public certainly expects the president to deliver economic growth and prosperity for all, their ability to alter the trajectory of the economy is quite limited.

They can use budget proposals to set taxing and spending priorities, appoint members of the Federal Reserve Board, negotiate trade deals, and make other marginal interventions that can affect growth and unemployment rates. But altering the shape and direction of the economy depends far more on the sustained, coordinated efforts of myriad interests, organized groups, and political actors over long periods of time. Steering the national economy is a multifaceted, collective effort in which the president is only one—albeit important—player.

All of which returns us to the challenges all presidents face in trying to orchestrate concerted, collective action in a government and society as sprawling as ours, with structural features designed to thwart centralized control.

We know that presidents are notorious for their oversized ambitions. Their sights are set higher, and their purviews far wider than other political actors. They want to have a good first 100 days, but more than that, they want to have a lasting impact on that system. They do not just want to ride the currents of history, they want to steer them as well: they want to establish a historical legacy of their own designs.

But in many ways, they are just along for the ride. They will try their hardest and exploit the resources at their disposal to make an impact, but most presidents most of the time will find themselves affecting change, as political scientist George C. Edwards III has aptly put it, only “at the margins.”