It seems plausible to me that the arbitrator gets a lot of business
from such agreements and therefore would have a lot of trouble
As an empirical matter this is correct. This has been demonstrated, for example, in securities arbitration and in consumer contract arbitration.
Links to a three part New York Times series in 2015 on unfairness in consumer arbitration, and a rebuttal op-ed in a Bloomberg article by a law professor, can be found here.
Also, Consumer arbitration in credit card agreements almost never actually happen.
The CFPB found that large banks are much more likely than small banks
to include arbitration clauses, but that because of their market
share, around 50% of credit card loans and 44% of insured checking
account deposits are covered by arbitration agreements. (The numbers
would be far higher but for the NAF settlement, under which many
issuers removed arbitration clauses from their contracts.) The
percentages are much higher for prepaid cards.
Ninety percent of the arbitration agreements studied include class
waivers. Most contain small-claims court carve outs. The banks are far
more likely than the consumers to go to small claims court. That makes
sense. For small debts, a collection action in a small claims court
will usually lead to a default judgment, which is then immediately
enforceable. Arbitration requires two steps, the arbitration
proceeding and then the filing of the award.
Out of these millions of agreements, only about 300 arbitration claims
have been filed by consumers per year over the last three years, and
they were all for high dollar-value claims (more than $1,000). . . .
the Bureau observed that almost no consumers filed arbitrations about
disputes under $1,000. For arbitration filings involving debt
disputes, the average amount of debt at issue was over $13,000. For
other arbitration filings, the average consumer claim was for over
A number of arbitration clauses allow a consumer, and sometimes the
company, to use small claims courts rather than arbitration for
dispute resolution. The CFPB’s preliminary analysis indicates that not
many consumers initiate small claims court cases in credit-card
disputes. Rather, the analysis shows that small claims court cases are
much more likely to be brought by banks than by consumers. In the
states and counties studied, the Bureau was able to identify at most
870 credit card cases brought by consumers in small claims court
against large credit card issuers, but more than 41,000 cases brought
by these banks against consumers in small claims court.
Got that, there were on average only about 100 consumer arbitration claims filed per year against banks, despite the fact that they cover half of all credit cards and more than 40% of deposit accounts. If people in Colorado file consumer arbitration claims in numbers proportional to the national average, that means there are only 1-2 consumer arbitration claims filed concerning credit card and account disputes per year in the entire state.
Basically, when compared to non-arbitration claim rates, the result of consumer arbitration clauses appears to be to deny consumers any meaningful remedy.
In a 2008 survey of securities arbitration participants:
Customers were much less likely than “non-customers” to perceive the
arbitration panel as open-minded (27.97% vs. 49.43%), much less likely
to perceive the arbitration panel as impartial (24.84% vs. 47.82%)
and, ultimately, much less likely to view the arbitration process as
fair (27.84% vs. 50.64%). Indeed, 60% of customers disagreed with the
statement “I have a favorable view of securities arbitration for
customer disputes,” and 61% disagreed with the statement that
“[a]rbitration was fair for all parties.” . . . Participants with
recent experience in a civil court case also were asked to compare the
fairness of securities arbitration with the fairness of the court
process. A whopping 75.55% of customers found that arbitration was
“very unfair” (62.96%) or “somewhat unfair” (12.59%) when compared
with their court experience.
Employees also fair poorly in arbitration:
[In American Arbitration Association] arbitrations that occurred as a
result of clauses in employer-promulgated agreements . . . employees
won only 19.7% of their cases. Employees did even worse when they
faced employers who were repeat players, winning only 13.9% of these
cases. They won 32% of the time when they faced one-shot employers.
Employees’ odds were worst when their opponent was a repeat
player-employer who used the same arbitrator more than once. Then,
employees won only 11.3% of the time, compared to a win rate of 21.2%
in cases that did not involve a repeat employer-arbitrator pair.
These findings are consistent with earlier research which has found
that employees arbitrating pursuant to arbitration provisions
contained in personnel manuals or handbooks have relatively low win
rates. In contrast, employees arbitrating as a result of individually
negotiated contracts do quite well. In one study, they won 68.8% of
the time. In another, they won 61.3% of their cases. The employees
arbitrating pursuant to individually negotiated contracts tend to be
highly-paid managers and executives. The employees arbitrating
pursuant to personnel manuals or handbooks are likely to be lower-paid
and lower-ranking employees. . . . 15-25% of all employers have now
adopted employment arbitration. Meanwhile, the rate of unionization in
the United States was only 12% in 2006. . . . “employment arbitration
is likely already a more widespread system for governing employment
relations than collective bargaining and labor arbitration.” . . . the
mean award for employees was $23,233 (including the many cases in
which no damages were awarded to the employee), the mean arbitrator
fee was $10,351 in cases that involving a hearing and award.
This is pretty discouraging.
Is there any way for me to know whether this arbitrator is fair?
Sometimes it is possible to do research on the arbitrator's background, but often information is unavailable. California requires disclosure of arbitration outcomes to a greater extent than many states, and some repeat players keep track of their past experiences. Often even if you know that the arbitrator is biased, you must accept that person as an arbitrator on a take or it leave it basis if you want to do business with a vendor or service provider and other comparable providers also have arbitration clauses naming unfair arbitrators.
In these situations, you are pretty much screwed if there is every a dispute, subject to your ability to influence the vendor or service provider with bad publicity (some contracts also prohibit you from using that tool, although the validity of those contracts is not always clear).
Arbitration involving people of equal bargaining power before a dispute arises (e.g. union-management arbitration or arbitration between large international corporations) is often more fair, but even then, alleged cost and time savings do not materialize.
How should fairness be defined and assessed under the circumstances?
For purposes of setting aside an arbitration award or disqualifying an arbitrator, a personal financial stake in the outcome of the arbitration, bribery, or in certain extraordinary circumstances duress (I'll beat you up if you don't rule in my favor), would be disqualifying, as would some histories of personal relationships between the arbitrator and a party (e.g. the arbitrator is your ex-spouse). This is a very low bar. Lots of unfairness and bias is not judicially reviewable and you are stuck with an outcome even if it is unfair, contrary to the relevant law and gets the facts wrong.
But, assessing fairness is hard even in the best of circumstances because you must know not just how the arbitrator has ruled in various cases, but also what a fair outcome would have been in those cases before a fair arbitrator. Usually, the best you can do is to determine if the person's background suggests an obvious bias, check any word of mouth experiences other people have had with this arbitrator, and review the rules that will be used.
Large corporate arbitration firms like the American Arbitrator's Association are unlikely to have an individualized case by case bias in favor of a particular vendor or service provider, but may have an overall institutional bias in favor of vendors or service providers generally.
Typically arbitration clauses are included in substantial part to discourage litigation at all, and may not function as intended by the party drafting the contract, if many people actually decide to utilize this process.
Arbitration can be tolerable if losing is not a make or break issue for either party, a quick rough justice result is desirable, and having at least a hypothetical unfair remedy will discourage misconduct by the other party more than having no remedy at all, even if the deal is blatantly disregarded. But, it is a very poor forum when a party has very high stakes relative to what they can "survive" in a one time dispute for that party.
For example, Uber's contracts with its drivers have raised that kind of issue with a class action ban in the arbitration process producing about 12,500 individual arbitrations over a key issue (independent contractor v. employee status) resulting in a collapse of the ability of the arbitration process to deal with the cases.