In recent months, we have seen a string of shocking headline figures based on 10-year budget projections.

For example, when the House’s American Health Care Act (AHCA) bill got scored by the Congressional Budget Office (CBO), we learned that 23 million fewer people would be insured, but that it would reduce the deficit by $119 billion. When Trump proposed his initial budget, Bloomberg said the plan included $3.6 trillion in budget cuts.

These numbers sound enormous. It’s estimated that there are 325 million total in the US, so the 23 million people would represent 7% of the entire current population. Likewise, total expenditures of the US government in FY 2016 came to $3.9 trillion. In the normal meaning of language, a $3.6 trillion budget cut would suggest that Trump was planning to eliminate nearly the entire federal government.

But of course, that’s not what he actually proposed–in fact, on net, Trump didn’t propose cutting the budget at all. His plan called for a 16 percent increase in the budget through 2020.

And the CBO analysis didn’t say the AHCA would cut the deficit or health insurance immediately. Rather, their analysis found that 10 years from now, the plan would result in 23 million fewer people having insurance and the cumulative deficit over 10 years, would be reduced by $119 billion relative to current law. That is, on average, the annual decrease to the deficit would be a mere $12 billion due to the AHCA.

This type of long-term analysis and language is commonplace in US national politics, but it achieves no useful outcome. It confuses far more than it clarifies. It does not provide accurate estimates of long-term results. It does not improve the average voter’s understanding of policy effects. And it gives politicians a means to claim they are being fiscally responsible without actually exercising any prudence whatsoever.

Missing the Mark

When the president prepares a budget plan, there should be little pretense about accuracy. It is fundamentally a political document prepared by the Executive Branch, and we should not be surprised if reality turns out somewhat less rosy than the future foreseen by the president’s sales pitch.

But when long-term projections are provided by the CBO, the figures are supposed to be much more credible. Indeed, in the media, the Congressional Budget Office is almost always described as the nonpartisan Congressional Budget Office, giving it an air of authority.

The problem is that the CBO estimates have proven to be consistently wrong as well. And in general, the CBO has been wrong in the same direction–it overstates economic growth, underestimates government outlays, and underestimates the effects of economic incentives. The net result is that the CBO projects lower debt and lower deficits than actually end up occurring.

Meanwhile, when it comes to healthcare, the Congressional Budget Office underestimated the job declines that would be produced by changing economic incentives, and it substantially overestimated how many people would gain healthcare coverage by about 11 million. Again, the bias points in the same unhelpful direction–the CBO underestimates the costs of government intervention and overestimates the benefits.

This bias is concerning, but it’s not really the main problem. The problem is that the CBO has been given an impossible task of projecting the performance of a massively complex economy over a very long time period. Even with the very best economists, that would not turn out well.

A Logical Contradiction

The CBO’s already herculean undertaking is made more difficult by the fact that they must make their projections as if current law will not change.

This is the logical and only conceivable assumption for them to make. (Clearly, they can’t conjure into existence entirely new policies for the purpose of their analysis.)

Yet, at the same time, this assumption is plainly absurd. In virtually every election cycle, politicians campaign on bringing new changes to Washington. Even incumbents run on making change, and national elections of some magnitude occur every two years.

So the CBO has to assume the law won’t change for ten years, even as most of the politicians that make the laws risk changing in two years–and all of them advocate for changes of one stripe or another while in office.

This combination is not going to produce an informative result.

Instead, the time horizon for analysis should take the US political cycle into account. Looking beyond one or two years is a fool’s errand.

Budget Cuts “Tomorrow”

Some bars have a sign in them that says “Free beer tomorrow”. The sign is always up and never changes. There will always be free beer tomorrow, but alas, tomorrow never comes.

Something similar happens with these 10-year budget projections. Congress has a knack for starting any proposed cuts in the later years, while letting spending run wild in the immediate future.

Since the headlines are often based on the full 10-year run, it doesn’t superficially matter when the cuts occur. A $200 billion cut in year 7 would look the same as a $200 billion cut in year 1 in much of the reporting. However, a $200 billion cut in year 1 has real world consequences that a $200 billion cut in year 7 does not–maybe taking away the money from a powerful special interests, maybe eliminating programs, etc. Even if these cuts are good policy, they will encounter resistance, and the resistance will be much stronger if the cut happens in the near term rather than in the hypothetical future.

The focus on 10-year projections thus gives politicians a way to have their cake and eat it too. In any given bill or budget, they can continue spending at current levels and rates of increase for next couple years–thereby avoiding hard decisions and difficult political fights. And then they can put the “draconian” cuts in the out years. Over the full 10-year cycle, they can claim to balance the budget or reduce the deficit and pretend to be fiscally responsible.

But in reality, they’ve done nothing at all. And when it comes time to implement the hard cuts needed, well, it will be time for a new bill and a new 10-year analysis. Rinse. Repeat.

In fact, this is the tack taken in the ostensibly fiscally conservative AHCA bill passed by the House. Above we noted that the bill will cut (merely) $119 billion from the cumulative deficit over 10 years. However, the deficit reduction doesn’t even start until 2021–or year 4 (see page 35). Before that time, the bill would actually increase the deficit over baseline estimates.

This approach should not come as a surprise. The focus on 10-year projections helps Congress avoid making hard choices. Instead of encouraging long-term fiscal discipline, it just distracts us from the reckless decisions being made in the here and now.

If Congress had a sign like the neighborhood bar, it might say “Balanced budget in Year 7”. But alas, Year 7 never comes.