Politics / Regulation / U.S. Domestic Policy

The CRA, Regulation Day, and the Pathway for Regulatory Repeal

capitol1The phenomenon of “Midnight Regulations” — the high volume of rules issued as a lame duck President’s term winds down — has been well documented by the American Action Forum, as have Congressional attempts to curb the Midnight trend. Building on that analysis, we decided to investigate the effect of a 1996 law that was designed to provide Congress with greater regulatory oversight capabilities: The Congressional Review Act, or CRA.

The CRA mandates that federal agencies submit a summary of all final rules to Congress, and provides Members with 60 legislative days to review a rule’s report. During this review period, Members are permitted to introduce a joint resolution of disapproval for the regulation, which, if passed in both houses and signed by the President, would negate the proposed rule.

In reality, the power granted to Congress by the CRA does not often pose a viable threat to a regulation. Even if a joint resolution of disapproval is passed in both chambers by a majority vote, it would still presumably fail, as the President is unlikely to sign a resolution for the repeal of a rule his own administration created. Thus, for most of a President’s term, the CRA effectively necessitates a Congressional override in order to turn over a regulation that lawmakers find reprehensible.

However, one instance in 2001 highlights certain circumstances under which the CRA might provide Congress with some clout: when regulations are issued at the end of a President’s term.

Section 801(d) of the CRA promulgates that the review period for a regulation submitted to Congress with fewer than 60 legislative days remaining in a term will be treated as if it were submitted on the 15th legislative day of the subsequent Congressional term. So, when the 107th Congress convened in March 2001, they had an opportunity to review a controversial Occupational Safety and Health Administration (OSHA) rule regarding workplace injuries that had been published in the waning days of the Clinton administration. Both houses ended up passing a joint resolution of disapproval, and the repeal of the regulation was finalized with the signature of the newly inaugurated George W. Bush.

Because of Section 801(d), the CRA poses a threat to any regulation proposed late in an administration’s term — including Midnight Regulations — and the last legislative day a regulation could be proposed without its review period being extended to the next Congress is now known as “Regulation Day.” The prevailing wisdom is that an election-year administration would deluge OIRA with regulatory submissions in the weeks prior to Regulation Day in order to avoid the possibility of repeal later on.

But is that wisdom supported by empirical evidence? To answer that question, we compared the number of pre-Regulation Day submissions by election-year administrations with the number of submissions by administrations in non-election years.

In order to make this comparison, we first tallied the number of regulatory submissions in the four weeks prior to the CRA deadline every year since 1996, taking note of both the total number and the number of “major” (likely to have an annual effect on the economy of $100 million or more) rules submitted. We then categorized this data as “Election-Year” or “Non-Election-Year,” and ran a t-test to determine whether there was a significant difference in the means of these two groups. The data is shown below.

Screen Shot 2016-06-29 at 5.36.11 PMScreen Shot 2016-06-29 at 5.44.42 PM


As we can see, the mean number of submissions in Election Years is higher than the mean of Non-Election-Years for both total and major regulations. However, the large p-values resulting from our t-test indicate that the difference between the mean submissions in Election Years and Non-Election-Years is not statistically significant, and could be due to chance.

Thus, despite the prevailing wisdom regarding Regulation Day, there is not sufficient statistical evidence to conclude that election-year administrations exhibit significantly more regulatory activity 60 legislative days prior to the end of a Congressional term than they do in non-election-years.

One shortcoming of our analysis is that it relies on small sample sizes. There have not been many years since the passage of the CRA to use as data points for our t-test, especially when one considers the relatively low number of major regulations issued per month. This issue may muddle any existing statistical trend, and perhaps in future years, it would be helpful to re-try this exercise with more data.

Another point to consider is that we may not be viewing regulatory activity prior to Regulation Day through the proper lens. In 2001, it was not the volume of regulations issued that prompted Congress to issue a CRA joint resolution, but rather the controversial nature of the OSHA rule. Thus, future analysis might focus on quantifying whether rules submitted prior to Regulation Day have tended to be more controversial than those submitted after that date.