In his State of the Union address in January 1954, President Dwight D. Eisenhower, chief of state of the most prosperous economy in the world, called on Americans to support an ambitious plan to build a world-class highway system linking the nation as surely as the transcontinental railroads did a century before. Doing so, he said, would enhance both the security and the long-term prosperity of the nation — the $25 billion investment would yield untold returns for generations to come.

We all know the Federal Aid Highway Act, eventually passed in 1956, did just that. Now imagine if Eisenhower were to stand up today and call for the same. The response from Congress would almost certainly shout from the headlines: “President’s bill would add billions to deficits.”

Although this year’s edition of debt-ceiling drama lacked the Armageddon punch of prior years, it underscored once again how broken the budgetary process has become for the nation’s government. Less obvious as part of that dysfunction is the outsize role played by the Congressional Budget Office.

The agency has been much in the news these past weeks. Its estimates about the possible negative effects of raising the minimum wage were seized on by partisans as proof that the idea is a bad one. Before that, its analysis of the long-term employment effects of the Affordable Care Act also garnered the Washington spotlight and dominated the public discussion.

Until relatively recently, however, none of these debates would have been shaped by the CBO. Forty years ago, it didn’t exist. For most of American history, the government considered bills, spent or did not, without reference to this obscure but increasingly powerful agency. Now, it has become the gantlet all bills must run through, and where many go to die.

As all Washington insiders know all too well, before any spending bill passes, it must be “scored” by the CBO for its effects on long-term budgets and future deficits. To assess that, the CBO relies entirely on a set of government statistics. The primary ones are gross domestic product growth, inflation and tax receipts (which is largely a product of the first two). Those numbers constrain what the government can spend, to the point where passing anything remotely resembling a long-term investment bill such as the interstate highway (absent a crisis such as 2008-09) is impossible.

That process and those numbers stop us from channeling spending toward needed public works, long-term investments and innovations whose future outcomes are, by virtue of being new and untested, uncertain and unpredictable.

This isn’t about big or small government per se. Eisenhower was no apostle for an expansive role for the federal government. He was, instead, an advocate for a responsible role wherein government was the provider for common, public goods for which no one in the private industry and no state alone can provide.

Today, with the CBO as a gatekeeper and rigid numbers as its tools, we have reached a point where we can maintain spending on entitlements and defense but have no creative or dynamic means to invest in the long-term future of the country at a national level at precisely a time when we need it dearly and when societies around the world are doing so heavily and actively.

An agency’s mixed record

It’s not exactly the fault of the CBO, however. The agency exists within a web of legislation. It was created in 1974 by the Congressional Budget Act signed by President Richard Nixon. The original goal had little to do with deficits, which were not an issue in the 1970s, and everything to do with the relative balance of power between Congress and the executive branch and the desire of Congress to reassert its authority over federal spending.

In the 1980s, budgets regularly began to exceed tax revenue and deficits became a national issue. The Gramm-Rudman-Hollings Act of 1985 and subsequent deficit deduction acts in 1990 essentially made the CBO the official traffic cop of government spending. So in the past twenty years, the CBO has “scored” all legislation based on its effects on the long-term deficit of the U.S. government. That scoring system began with good and necessary intentions: to keep spending within certain boundaries by offering a neutral, nonpartisan analysis of how spending bills would play out over time and effect the fiscal balance of the federal government.

The CBO performed that role. In many instances, it prevented Congress and the White House from soft-pedaling the negative effects of their actions. One of the earliest examples was during the savings-and-loan crisis in the late 1980s. Given how complicit the government had been in creating the conditions for that debacle, it was very much in the interest of both branches of government to play down just how much money the crisis would cost the American taxpayers. The CBO made clear that the amount was in the hundreds of billions of dollars, and it squarely rejected efforts by the Reagan White House to make such payments external to the annual budget.

But the record has also been mixed when it comes to long-term projections.

The initial cost projections for Medicare Part D, for instance, proved wildly inaccurate. When the bill was passed in 2003, the CBO assumed that the bill would cost about $550 billion between 2004 and 2013. The actual cost was about $375 billion, nearly a third less. Estimates of long-term costs of the Affordable Care Act have also vacillated, by $100 billion in the past two years, as new information about enrollments and overall health costs fluctuates.

It’s no surprise that projections often fail to match outcomes. No one can predict the future, and it is not the fault of the large staff of economists and nonpartisan brains at the CBO that they cannot know future growth patterns, future costs and the entire fluid set of variables that shape economic life. No government agency, for instance, is held accountable for specific predictions about political outcomes. Even the CIA is not expected to have a crystal ball about long-term political outcomes in other countries. A good analysis of likely outcomes and possibilities, yes, but not specifics.

Yet that is precisely the mandate of the CBO — to calculate long-term numbers. The only way it can do so, however, is to use current numbers and patterns and make assumptions. This is what economists and Wall Street institutions do all the time, and the legacy is at best spotty and at worst strikingly wrong.

The belief that economic systems are like machines or science experiments pervades academic economics and policymaking. While the staff of the CBO most certainly recognizes that economic systems are complex, the presumption is nonetheless that long-term projections are meaningful and feasible. That has led to a political culture of elected and unelected officials making promises and forecasts about the effects of legislation:The 2012 budget proposal by Rep. Paul Ryan (R-Wis.) was replete with figures projecting the federal deficits a decade or more from now. Bills are constantly debated in terms of what the CBO assesses as the multiyear consequences.

The problem is that projections can be bound only by what is known at present. To make any projections, you need to decide which variables matter and how they will shape one another. For instance, you need an assumption of the growth of GDP, the amount of tax revenue that will be collected, the rate of inflation, and incomes and jobs. These assumptions are based, as they must be, entirely on past patterns. Yet the numbers themselves are only marginally older than the CBO. GDP only replaced gross national product in the 1990s, and GNP was only regularly calculated starting in the late 1940s. Inflation, employment and other key indicators have been around for only about 60 years — a while, yes, but not nearly enough for statistical calculations that have an acceptable margin of error. Given that, the projections made by the CBO are statistically bound to be inaccurate.

Finally , and most crucially, there is nothing in the CBO’s mandate that would allow it to game-out scenarios for changes in the law. It can offer outcomes based on different rates of long-term growth, but it cannot and will not offer “what-if” scenarios such as “if Congress changes the corporate income tax rate” or “if the U.S. become a net energy exporter” or “if the United States invades Iran.”

Variables change — constantly, dynamically and unpredictably.

Understandably, if any scenario could be contemplated, then partisans of one bill or another could just invent a scenario that suited their interests. Let’s spend a trillion dollars on energy independence because we could then generate enough revenue on energy taxes plus enough savings from not importing oil to more than offset the costs. Let’s cut Social Security payments in half because the amount of debt that would reduce would allow companies and individuals to invest the money and more than make up the shortfall. Everyone would naturally assume whatever is best for their goals.

Even more problematic, the current framework is incapable of assessing the savings and other potential benefits of certain types of spending. That makes political sense. If any official could justify a bill that spends hugely now by promising that it will have long-term benefits that cannot be substantiated until it happens, then every bill would come with iron-clad promises of long-term benefits. Yes, this legislation will cost $100 billion, but think of what it will save and yield!

Yet major benefits of U.S. government programs over the past decades have come precisely because current investments yielded future returns: the Interstate Highway System; medical and scientific breakthroughs from the National Science Foundation and the National Institutes of Health; the commercial applications of defense spending, including the Internet; the GI Bill funding higher education; the Fair Housing Act. The list is long, and the positive consequences for our world today are manifest. And none of those could survive the scoring system we have today.

This is a call for removing a straitjacket that we have only recently donned. Government and societies around the world are actively investing public funds in much-needed public goods — China, of course, but also dozens of others. Infrastructure is the most obvious, and one that is sorely lacking in the United States. So, too, are public-private partnerships that split the costs of job training, skills education and public works, from an efficient energy grid to better transportation networks to high-speed Internet and cellphone coverage.

In key respects, we are living off the investments of earlier generations. Those generations did not need to filter all investments through a scoring system that treats economic statistics as facts and future outcomes as predictable. We should rigorously test all of our spending assumptions, but not at the cost of investing for the future.

Karabell is head of global strategy for Envestnet. His latest book is “The Leading Indicators: A Short History of the Numbers That Rule Our World.”