As my students are back in lecture halls this fall, they are being reminded that it is more important to complete their assignments properly, rather than quickly.  Mistakes at the beginning invariably lead to disappointing conclusions. That’s a lesson worth repeating to regulators in Washington too.

Take, for example, the Department of Labor’s fiduciary rule, which went into partial effect in June with the objective of improving the efficiencies in the market for retirement services upon which Americans rely. The rule’s regulatory impact analysis, produced in April 2016, found that imposing a fiduciary standard on brokers would save Americans billions.

However, as my new white paper shows, the Labor Department’s projections were based on incorrect interpretations of academic research. For example, the Labor Department’s analysis produces quantified benefits of approximately $32 billion, which are netted against quantified costs of $16 billion. My analysis concludes that a corrected estimate results in no quantified benefits, suggesting that instead of helping American retirees, the rule will impose net quantified costs of at least $16 billion.

What should be done in light of these findings? Although parts of the regulation were recently delayed until July 1, 2019, the Labor Department should revisit the rule in its entirety and conduct a de novo economic analysis. In doing so, it should carefully examine costs and benefits not just of the rule’s delayed and in-effect provisions, but also of potential alternatives to the rule that could do more to help American households save for retirement.

Fortunately, the Labor Department has a great playbook to follow in producing such an analysis: the law. President Bill Clinton’s Executive Order 12866 – left in place by Presidents Donald Trump, Barack Obama and George W. Bush – requires executive agencies to identify the need for a proposed regulatory action, assess costs and benefits of available regulatory alternatives, and select approaches that maximize net benefits.

Unfortunately, the Labor Department’s 2016 regulatory impact analysis fails to comprehensively address these requirements or those established by Trump’s January executive order on regulation.

For example, the regulatory impact analysis does not adequately establish that there is, in fact, a need for the fiduciary rule. It argues that brokers compensated by a commission for selling retirement products are more inclined to provide investment advice counter to their clients’ best interest, but largely offers only anecdotal and indirect evidence rather than empirical research in support.

In addition, the benefits computation in the analysis relies on a range of studies mostly focusing only on domestic equity funds and using pre-2005 data to show that mutual funds charging load fees (used to pay brokers’ commissions) systematically underperform other funds. By selecting this group of studies, the regulatory impact analysis fails to account for a wide range of fund strategies and the substantial post-2010 decline in mutual fund fees. Notably, research using more recent data shows the return gap between load fee-charging funds and others is immaterial.

Similarly, in the regulatory impact analysis, the benefits are skewed upwards based on an incorrect interpretation of a single study that examines whether fund performance is associated with load fees. This study finds that funds charging higher than average load fees underperform other funds, suggesting that firms charging lower than average load fees outperform other funds.

In its misinterpretation of the study, however, the Labor Department implicitly assumes that most funds charge unrealistically high load fees. By failing to accurately account for the better-than-average performance earned by funds that charge below-average load fees, the Labor Department estimates cannot be interpreted to reflect the effects of excess load fee amounts on fund performance broadly.

The costs could be even larger. For example, while the regulatory impact analysis acknowledges that the rule may make investment advice less accessible for those with modest retirement savings, it does not provide any estimate of the costs of reduced access. It also does not adequately address existing research showing that investment advice leads to increased savings and that lack of advice is associated with smaller retirement accounts.

The regulatory impact analysis also does not comprehensively analyze potential alternatives to the fiduciary rule or establish that the selected regulatory approach maximizes net benefits. For example, it does not quantify the costs of waiting for the Securities and Exchange Commission to act on this issue by implementing a uniform fiduciary standard, nor does it adequately address the costs that could result from two conflicting regulatory standards for U.S. brokers if the SEC chooses to exercise its authority after the Labor Department rule is implemented.

In light of these shortcomings, the Labor Department should delay the portions of the rule not yet in effect until a new regulatory analysis is complete. The revised analysis should transparently analyze costs and benefits of these provisions, as well as those that went live in June, and plausible regulatory alternatives to the rule, including waiting to see if the SEC adopts a uniform fiduciary standard so that the two regulatory frameworks can be aligned.

In fact, earlier this month SEC Chairman Jay Clayton told Congress that his agency is working on its own version of a fiduciary rule proposal. Clayton seems to believe in the need to get it done right when he said, “We’re going to work with the Department of Labor. If this were easy, it would already have been fixed.”

As students return to the lessons and rhythms of school this fall, so too should the Labor Department recall the reminders from professors to be thorough and the wisdom imparted by shop teachers throughout the decades: measure twice, cut once. American retirees are counting on it.

Craig Lewis, a senior adviser with Patomak Global Partners, is also an economics professor at Vanderbilt University and the former chief economist at the Securities and Exchange Commission.