The Administrative Law Judge Test

At the beginning of this research, it was expected that the ALJ factor would be vital to the decision making process because each ALJ could be assumed (for the sake of argument) to have his or her own idea about patent law and remedy. The Commissioners also have to make their decisions based on the ALJ’s findings.

Chi-square tests seem to reject the initial ideas about ALJ’s personal preference and influence. On the other hand, if the Level of Significance were lowered to 20%, the ALJ factor would have passed all three tests.

ALJ Duvall Saxon Luckern Harris OTHERS Total Observed Frequencies

GEO 3 11 9 4 7 34

LEO 3 10 25 8 5 51

Expected Frequencies

GEO 2.4 8.4 13.6 4.8 4.8 34

LEO 3.6 12.6 20.4 7.2 7.2 51

Total 6 21 34 12 12 85

Observed Frequencies / Expected Frequencies

GEO 125% 131% 66% 83% 146%

LEO 83% 79% 123% 111% 69%

TABLE 4-8 TEST 1, ALL ALJS (p-value = 19%)

ALJ Duvall Saxon Luckern Harris OTHERS Total Observed Frequencies

VIO 6 26 41 19 15 107

SET 9 30 36 46 32 153

NVIO 4 14 17 21 13 69

Expected Frequencies

VIO 6.2 22.8 30.6 28.0 19.5 107

SET 8.8 32.6 43.7 40.0 27.9 153

NVIO 4.0 14.7 19.7 18.0 12.6 69

Total 19 70 94 86 60 329

Observed Frequencies / Expected Frequencies

VIO 97% 114% 134% 68% 77%

SET 102% 92% 82% 115% 115%

NVIO 100% 95% 86% 116% 103%

TABLE 4-9 TEST 2, ALL ALJS (p-value = 15%)

Beyond this somewhat discouraging finding, further examination revealed some even more interesting findings:

• ALJs possibly have limited influence over the choice of exclusion order.

• Earlier ALJs were quite consistent with their patent violation findings (less freedom).

• More recent ALJs are less consistent with their patent violation findings (more freedom).

The third point seems to support many USITC litigators’ concern that the assignment of ALJ could affect an investigation’s outcome. At least this research shows the assignment of ALJ may affect violation findings.

Generally, respondents may care less about the choice between general and limited relief (Test 1) (see Sections 1.5 and 4.1.3.3). They may care more about the violation finding (Test 2). Most of the times, fairness for the type of exclusion order relief is only questioned and challenged by U.S. trade partner governments. The power to decide exclusion order type has always been exercised by the USITC Commissioners. ALJs can only recommend.

After the 1980s, international pressure forced the U.S. to adopt a less U.S.-centered trade policy. The percentages of general relief declined as a result. However, in areas concerned only by individuals, the USITC seemed to allow more freedom for the ALJs.

4.2.5.1 ALJs Then and Now

The ALJs have limited power to determine the type of exclusion order relief (Test 1 questions). They make recommendations but Commissioners have the final say. As shown by the following tables, ALJs possibly used to be more influential in the determination of exclusion orders. This is no longer true.

Also based on the analysis to the same dataset, this research found earlier ALJs had much limited discretion over patent violation issues (Test 2 questions). In contrast, recent ALJs seemed to have more freedom in this area.²⁴²

ALJ Duvall Saxon OTHERS1 Total Observed Frequencies

GEO 3 11 5 19

LEO 3 10 0 13

Expected Frequencies

GEO 3.6 12.5 3.0 19

LEO 2.4 8.5 2.0 13

Total 6 21 5 32

Observed Frequencies / Expected Frequencies

GEO 84% 88% 168%

LEO 123% 117% 0%

TABLE 4-10 TEST 1, EARLIER ALJS (p-value = 13%)

242 Earlier OTHERS1 ALJs include: Ablondi, Leonard, Mathias, Minchew, Renick, Renwick, Timony and some Commissioners-decided earliest investigations. Recent OTHERS2 ALJs include: Barton, Bullock, Morriss and Terrill.

ALJ Luckern Harris OTHERS2 Total

Observed Frequencies / Expected Frequencies

GEO 94% 118% 101%

LEO 103% 93% 100%

TABLE 4-11 TEST 1, RECENT ALJS (p-value = 90%)

ALJ Duvall Saxon OTHERS1 Total Observed Frequencies

Observed Frequencies / Expected Frequencies

VIO 90% 106% 84%

SET 109% 99% 95%

NVIO 97% 92% 136%

TABLE 4-12 TEST 2, EARLIER ALJS (p-value = 91%)

Luckern Harris OTHERS2 Total Observed Frequencies Observed Frequencies / Expected Frequencies

VIO 139% 70% 74%

SET 80% 111% 121%

NVIO 88% 118% 90%

TABLE 4-13 TEST 2, RECENT ALJS (p-value = 2%)

p-Value Level of Significance

5% 10% 20% 50%

It is interesting to see Section 337 investigation’s “quality control” experienced a clearly shown groundbreaking change after the rise of international pressure:

• Regarding the type of exclusion order (Test 1):

o Earlier ALJs possibly had some control over it.

o Recent ALJs have very little control.

• Regarding the issue of violation (Test 2):

o Earlier ALJs were very consistent with their findings.

o Recent ALJs have much more freedom in decision making.

It may not be a good idea to call ALJs Luckern and Harris (retired in 2007) “recent”

ALJs because their long period of service overlaps with “earlier” ALJs, such as ALJ Saxon, for more a significant period of time (e.g., more than a decade). They are listed under

“recent” ALJs for purposes of this research for convenience and to emphasis their increased freedom in decision-making.

ALJ Earliest Latest

Duvall 1978 1985

Saxon 1978 1996

Luckern 1984 2005

Harris 1985 2005

OTHERS1 1984 1991 OTHERS2 1999 2005

TABLE 4-15 ALJS AND DATES OF THEIR EARLIEST AND LATEST INVESTIGATIONS

CHART 4.7 TERMS OF OFFICE OF THE ALJS

4.2.5.2 USITC’s Measures to Maintain Conformity

Numbers out of context (as well as labels) can mislead. It is noteworthy that former ALJ Harris, a widely known believer of free trade and free competition, had a higher-than-average rate for General Exclusion Orders (1 : 2).

For example, in the case of ALJ Harris, all of his General Exclusion Orders were issued in the beginning of his service.²⁴³ Many of his exclusion orders were issued only after Federal Circuit remanded the case back to him. He issued no General Exclusion Order after 1992. On the other hand, over a more recent period of time, ALJ Luckern is generally more willing to issue Limited Exclusion Orders.

ALJ Harris also has a significantly below average rate for patent infringement findings.

Possibly, he used a higher level of proof for patent infringement and his violation finding could have been mostly worst violations. As a result, these worst cases (e.g., low technological level and many obvious infringers) are more likely to require General Exclusion Order relief.

4.2.5.3 Notes on “Judge Shopping”

Before making a quick conclusion that the assignment of ALJ may not significantly affect an investigation’s outcome, one may need to know the USITC’s in-house rules to see how they may affect case assignment. From ALJ Renick (early 1978) to Duvall (1978-1985) to the retirement of Saxon (1979-1996), the USITC operated under a chief ALJ system.²⁴⁴ The chief ALJ had the power to assign cases to all ALJs.²⁴⁵ After the retirement of ALJ Saxon, the USITC practically abandoned the chief ALJ system. However, the Rules related to the chief ALJ system remain unchanged.²⁴⁶ During this phase, the chief ALJ decides which case goes to which ALJ.

243 337-TA-225 (1986/09/11), 337-TA-254 (1989/09/26; after remanded by Federal Circuit), 337-TA-287 (1989/09/28) and 337-TA-324 (1992/08/06).

244 See, e.g., U.S. Government Manual 731 (1995). (“Chief Administrative Law Judge: Janet D. Saxon”); U.S.

Government Manual 701 (1996). (“Chief Administrative Law Judge: (vacancy)”); U.S. Government Manual 699 (1997). (“Administrative Law Judges: Sidney Harris, Paul J. Luckern”).

245 See, e.g., 43 Fed. Reg. 4126 (Jan. 31, 1978). (Renick assigned 337-TA-37 and -39 to Duvall); 43 Fed. Reg.

31245 (July 20, 1978). (Duvall assigned 337-TA-52 to Saxon). 49 Fed. Reg. 34422, 23 (Aug. 30, 1984). (The USITC delegated the power to designate presiding officer to acting chief ALJ Saxon).

246 C.f., 19 C.F.R. § 210.15 (a)(1). (“During the period between the institution of an investigation and the assignment of the investigation to a presiding administrative law judge, all motions shall be addressed to the chief administrative law judge. …”).

There has been no chief ALJ at the USITC since the retirement of Saxon.²⁴⁷ It adopted a roster system that assigns cases to ALJs in an orderly fashion. In response to the change, some complainants try to gain an upper hand by filing the complaint at a calculated time when an ALJ perceived to be friendlier may be available.²⁴⁸ The two case assignment systems could have affected the distribution of outcomes. It may take detailed cross-analysis of case assignment records to determine whether non-random case assignment could be a factor statistically affecting the decision histories available for our review.

The issue of judge shopping and the history of USITC’s case assignment systems have been discussed by ALJ Harris in Synchronous DRAM, 337-TA-437. In that case, Complainant Rambus withdrew the complaint shortly after the complaint was assigned to ALJ Harris but before its publication in the Federal Register. In filing the motion to withdraw, Rambus argued, “because it was withdrawing its complaint prior to the formal institution of the investigation, there was no prejudice to the parties and an order terminating the investigation was appropriate.”²⁴⁹ ALJ Harris ruled “The circumstances of Rambus’ filing its complaint and then abruptly withdrawing it when the presiding judge was named creates a strong but rebuttable inference of judge shopping.”²⁵⁰