Federal regulators scrambled to issue last-minute rules during the “midnight” period between Election Day and Inauguration Day. Now that Donald TrumpDonald TrumpMaxine Waters: ‘You can’t trust this president’ Obama shamefully lines pockets with 0K for Wall Street speech Dem senator fears Russian election interference could be ‘normalized’ MORE has taken the oath of office, his administration is scrambling to undo the Obama administration’s midnight rules. This familiar ritual has occurred at the end of most presidential terms since 1948.

These last minute regulations are a special example of regulations that get pushed through the regulatory process for political reasons. Their genesis creates self-imposed deadlines that favor action over accuracy. As a result, agencies’ analysis accompanying rushed midnight regulations is typically less thorough and less transparent than the analysis accompanying non-midnight regulations. And because midnight regulations may be less carefully thought out than other regulations, they are less likely to solve real problems at a reasonable cost.

Consider, for example, one of the Bush administration’s rushed midnight regulations in 2008, which required federal contractors to use the electronic e-verify system to see if prospective employees could be hired legally. The accompanying analysis provided no estimate of the regulation’s benefits, asserted that the regulation would reduce employers’ costs, then also made the contradictory assertion that no employer would use the system unless required to do so by regulation. The regulation hardly reflects a reasoned comparison of benefits with costs.

But many regulations issued outside the midnight period have the same types of problems. Midnight regulations should be considered the poster children for comprehensive regulatory process reform—not just a punctuated problem that periodically pesters politicians.

For example, the Agriculture Department finalized a new catfish inspection regulation in March 2016, even though the Government Accountability Office concluded three years earlier that the inspections would duplicate existing federal programs while doing nothing to improve food safety. Similarly, in 2015 the Food and Drug Administration issued a rule requiring that animal feed must be as safe as human food, even though a narrower rule focused on pet food would have produced the same public health benefits for humans at a fraction of the cost. In both cases, the agencies adopted unnecessary or overly broad regulations because they failed to conduct an evidence-based assessment of the problems they sought to solve.

These are not isolated examples. The Mercatus Center’s Regulatory Report Card assessed the quality of analysis accompanying 130 major, non-budget regulations proposed by executive branch agencies between 2008 and 2013. Only half of these regulations are even accompanied by evidence that the regulation addresses a significant problem. This suggests that even in the best of times executive branch agencies routinely ignore the instructions in President Clinton’s Executive Order 12866, which lays out requirements for regulatory impact analysis.

While it’s theoretically possible that regulatory agencies stumble upon the best regulatory design in the dark, it seems that performing robust impact analysis as instructed in executive order 12866 would significantly increase the possibility.

To generate better regulatory outcomes, effective regulatory reform would write analytical requirements into law, require agencies to seek public comment on alternatives before they decide how to regulate, and allow courts to review agency analysis to ensure that it meets some minimal standards. Effective regulatory reform would also require regular retrospective analysis, so that agencies would assess the actual effects of regulations after they are adopted and learn from the results to improve future regulations.

Incoming presidents always try to curb the prior administration’s midnight regulations, and Congress can invalidate them using an expedited process under the Congressional Review Act. But if policymakers want to prevent poorly designed regulations year-round, they should require agencies to analyze, act, and then account for the results.

Jerry Ellig is a senior research fellow with the Mercatus Center at George Mason University. Andrew Baxter is a first year MA student in the economics department at George Mason University.

The views expressed by authors are their own and not the views of The Hill.