Jump to main content.

7.2. Hahn's Review

Quick Links

Tell me About

Media

Subject

Geographic Area

Index

Table of Contents

powered by Google

In his paper Regulatory Reform: Assessing the Government's Numbers, Hahn has produced a study of regulation as viewed through RIAs. He is interested primarily in the regulatory result rather than in the quality of the studies or the procedure by which they were developed although those matters receive some attention. That is, for the most part, he uses the RIAs to evaluate the resulting regulations. The core of his analysis is a review of RIAs, especially the benefit-cost tests they contain, to see if the resulting regulations add to or detract from the welfare of society. His conclusion is mixed.

Hahn reviewed 168 rules promulgated from 1982 to mid-1996 including 115 from EPA. This seems to include the complete set subject to the Executive Orders mandating benefit-cost studies. He summarizes the results of those studies in his Table 1, which is reproduced below without the columns for non-EPA agencies.

Table 7.2.1: Regulatory Scorecard, 1982 to Mid-1996

Total
EPA
Number of rules
168
115
Costs/savings assessed
164
111
98%
97%
Benefits or cost savings assessed
146
95
87%
83%
Human impacts estimated
94
49
56%
43%
Benefits monetized
44
26
26%
23%
Agency found that B>C
39
19
23%
17%

Because EPA dominates the totals, Hahn provides additional detail on its regulations classified by statute in his Table 2, which is reproduced below.

Table 2: Regulatory Scorecard: Breakdown for Environmental Statutes, 1982 to Mid-1996
EPA
CAA
ERCL
CWA
FIFRA
RCRA
SDWA
TSCA
Number of rules
115
62
5
14
2
19
8
5
Costs/savings
111
61
5
13
2
17
8
5
assessed
97%
98%
100%
93%
100%
89%
100%
100%
Benefits or cost
95
55
0
12
1
17
7
3
savings assessed
83%
89%
0%
86%
50%
89%
88%
60%
Pollution reductions
63
54
0
8
0
1
0
0
quantified
55%
87%
0%
57%
0%
5%
0%
0%
Human impacts
49
20
0
6
1
13
7
2
estimated
43%
32%
0%
43%
50%
68%
88%
40%
Benefits
26
13
0
7
0
2
3
1
monetized
23%
21%
0%
50%
0%
11%
38%
20%
Agency found
19
10
0
3
0
3
2
1
that B>C
17%
16%
0%
21%
0%
16%
25%
20%
Notes: CAA = Clean Air Act; ERCL = Comprehensive Environmental Response, Compensation, and Liability Act; CWA = Clean Water Act; FIFRA = Federal Insecticide, Fungicide, and Rodenticide Act; RCRA = Resource Conservation and Recovery Act; SDWA = Safe Drinking Water Act; TSCA = Toxic Substances Control Act

Tables 1 and 2 aggregate by count but Hahn also wishes to aggregate by value. For that, he needs to put values in the studies on comparable terms. Dollar figures were adjusted using implicit price deflators and expressed in 1995 dollars. A discount rate of 5% was used. Lives were valued at $5 million and plausible values for pollution decrements were used. Hahn, op. cit., Table 3. He displays the results in his Table 4, which is reproduced below without the non-EPA agency columns. The table covers proposed and final regulations.

Table 4: Net Benefits of Regulations, 1982 to Mid-1996
(present value in billions of 1995 dollars)
TOTAL
EPA
FINAL
Number of regulations
106
70
Monetized benefits exceed costs
43%
31%
Gross cost
601.9
444.0
Net cost
395.8
283.6
Benefits
1,948.6
764.2
Net benefits
1,552.8
480.7
Benefit to cost ratio
3.6
2.1
PROPOSED
Number of regulations
30
19
Monetized benefits exceed costs
43%
42%
Gross cost
100.8
84.7
Net cost
37.3
65.1
Benefits
362.0
85.0
Net benefits
324.8
19.9
Benefit to cost ratio
4.2
1.2
Note: Net benefits are calculated by subtracting net costs from benefits.

The three tables tell the main part of Hahn’s story:

Focusing on final rules [Table 4], I find that 46 of 106, or 43 percent, would pass such a test. For proposed rules, 13 of 30, or 43 percent, would pass such a test. (p. 15)

The net benefits for final regulations promulgated from 1984 [1982?] to mid-1996 approach $1.6 trillion in present value [Table 4]. Second, less than half the rules pass a benefit-cost test. Thus, even using agency numbers, we find that net benefits would increase substantially, by almost $280 billion, if agencies rejected rules that did not pass a benefit-cost test. (p. 32)

Hahn’s quoted remarks pertain to his Table 4 which resulted from his extensive reworking of the data as described earlier. However, Table 1 is even less flattering and there Hahn indicates that he merely reports agency results: only 23% of all regulations and 17% of EPA regulations pass a benefit-cost test. He partially reconciles the difference between the tables by stating:

The higher number of rules that pass when using my analysis is largely due to the fact I monetize benefits in several cases where the agencies do not. (p. 15)

He also acknowledges problems in preparing Table 4:

Table 4 covers thirty-two fewer regulations than table 1. I excluded some rules to represent more accurately the impact of regulations over the time period. (p. 13)

I did not include disease and injury that did not fit into fatality index categories. Thus, benefits may be understated. (footnote 48)

Although included in table 1, … [table 4] does not explore two rules on stratospheric ozone that, according to the EPA, have net benefits in the trillions of dollars. While those rules likely resulted in net benefits, the EPA’s estimates are likely to overstate the actual benefits significantly. (footnote 52)

However, these reservations do not extirpate the fundamental conclusion buttressed by Table 1 that most regulations fail a benefit-cost test but that in the aggregate, they produce net benefits. It is clear that a minority of rules are carrying the day:

… just two rules – NHTSA’s automatic restraints in cars and the EPA’s lead phasedown in gasoline – account for just over 70 percent of total net benefits. (p. 15)

One plausible explanation for the adoption of numerous inefficient regulations is that the statutes left the agencies no choice. They prepared the RIAs to comply with Executive Orders and then did what they had to despite discouraging findings. Hahn considers this possibility. He also notes that some statutes permit benefit-cost analysis but with limitations and some statutes are vague. His analysis is summarized in his table 8 which is excerpted below.

Table 8: Statutory Limitations on Balancing
(Percent of regulations passing benefit-cost test)
Agency analyses
% where B>C
Author’s calculations
% where B>C
Statutes without limitations48%50%
Other statutes18%42%

Using either agency data or Hahn’s transformation of it, no more than half of the regulations pass a benefit-cost test even when the agency is unconstrained. Hahn’s data show that the constraints have fairly little effect whereas agency data show a substantial impact. Hahn’s conclusions on this issue follow:

This preliminary analysis suggests that statutory limitations appear to have little effect on the efficiency of rules. Those results are counterintuitive to the common belief that statutory restrictions on balancing constrain an agency’s ability to promulgate economically efficient rules. Note, however, that my analysis uses very crude measures for the degree to which statutes constrain agencies. Moreover, isolating the impact of any single factor in the complex regulatory process is difficult. Finally, my analysis does not attempt to measure the efficiency gains associated with alternatives not chosen or even considered owing to such restrictions. (p. 29)

It would be interesting to have an analysis of the RIAs underlying rejected regulations but Hahn does not provide that.


Local Navigation


Jump to main content.