4 February 2020
We
all know that the UK SFO (Serious Fraud Office) has implicated AirAsia
directors/staff for involvement in bribery with Airbus.
I will not go through all the details but here are the key points. You can also refer to the links below for further reading.
- SFO enters into €991m DPA (Deferred Prosecution Agreement) with Airbus.
- Improper payment consists of $50m (RM240m) to AirAsia directors/staff as sponsorship for the Caterham F1 team in exchange for securing Airbus fleet orders.
- AirAsia, Tony Fernandes, Kamarudin Meranun denies allegation.
- MAVCOM to probe AirAsia if they broke aviation law.
- MACC (Malaysia Anti-Corruption Commission) launched an investigation against Tony Fernandes and Kamarudin Meranun.
***
Further readings
***
If
you’re an AirAsia long-term shareholder, this bribery scandal is
something worth the time to ponder because it can significantly impact
the business long-term prospect. Unlike coronavirus that will only
dampen earnings for the next few quarters, this scandal can
significantly impact the business’ medium to long-term cash flow.
How should you deal with this uncertainty so you can make the best investment decision?
We can use Bayes’ theorem. I have written a bit about Bayes’ theorem here
(point 4) so I’m not going to repeat it here. But in short, it is a
simple framework to update and quantify your belief every time you
receive more information.
Before
we can apply Bayes’ theorem in this case, we need to understand what we
are trying to find out. Generally, what everyone wants to know is
whether Tony Fernandes and/or Kamarudin Meranun did accept the bribe
from Airbus. But that is not very useful. Because bribery is subjective.
What is considered as a bribe for UK SFO might not be considered as
bribe for MACC, for example. Therefore, whether Tony and/or Kamarudin
are found guilty is not as important as what the consequences are if
they’re found guilty. If authorities found Tony guilty but only give him
a warning, that won’t affect AirAsia’s long-term prospect or alter your
original investment thesis.
From
a long-term shareholder’s point of view, there are only three things
that will dramatically change AirAsia’s valuation negatively:
- Any outcome that would force Tony Fernandes to relinquish his control of AirAsia. Which includes selling all of his stakes, unable to take charge of the direction of AirAsia and so on.
- Any outcome that would stop Tony Fernandes from running AirAsia i.e >3 years jail time, >3 years ban from managing AirAsia and so on.
- Any outcome that would significantly impact AirAsia’s future cash flow i.e significant fines or any restriction that inhibit their revenue or increase the cost of doing business.
There
is little dispute that most of AirAsia’s success can be attributed to
Tony Fernandes. So for a long-term shareholder, the two black swans from
the bribery allegation are:
- AirAsia receive fines (big ones)
- Tony Fernandes (and/or Kamarudin) lose control of AirAsia
These
are quite different from forecasting the probability of whether Tony is
guilty of bribery. Because as mentioned earlier, the verdict is not
important. The consequences of the verdict are.
To make this easier, we will apply Bayes’ theorem separately to both hypotheses. We need 3 things to apply Bayes’ theorem:
- Prior probability
- Probability as a condition of hypothesis the being true
- Probability as a condition of hypothesis the being false
Hypothesis: AirAsia would receive a fine
Prior probability - What is the initial estimate of how likely AirAsia will receive a fine for bribery before the bribery allegation appears?
It
is not that uncommon for AirAsia to receive fines; $2 mil fines from
MAVCOM in 2019; $200K by Australian Court in 2012 etc. But bribery
fines? I would consider that as rare if not impossible considering
AirAsia’s reputation. Or we can use base rate to think about how often
do we see bribery fines? Airbus $4 bil fines, Ericsson’s $1 bil fines
bribing government etc but they’re far in between. I’ll put it at 13%.
Keep in mind, this is subjective. There’s no right or wrong, but this is
what Bayes’ theorem is all about. The starting point is not important
as long as you continue to move towards the direction of less wrong.
Condition of the hypothesis being true - What is the probability of bribery allegation if AirAsia receives a fine for bribery?
Next,
we need to estimate the probability that if AirAsia receives a fine for
bribery, how likely will we see this allegation? The probability will
be 100% because bribery investigation always precedes punishment.
Condition of the hypothesis being false - What is the probability of bribery allegation if AirAsia doesn't receive a fine for bribery?
Then
we consider the opposite: How likely is it for us to see a bribery
allegation if AirAsia doesn't receive a fine at all? There are many
reasons for not receiving a fine despite the allegation. AirAsia could
get acquitted by MACC and MAVCOM, or receive a warning without any fines
and so on. I’ll put it at 50%.
As
shown above, my belief of whether AirAsia will be fined has increased
from 13% to 23% following the bribery allegation. This 23% will become
the prior probability when we receive more new evidence. Which will be
the continuation of this post.
Next, we are going to update our belief on whether Tony will lose control of AirAsia.
Hypothesis: Tony Fernandes to lose control of AirAsia
Prior probability - What is the initial estimate of how likely for Tony to lose control of AirAsia before the bribery allegation appears?
There
are several ways for someone to ‘lose control’ on a company they own.
Voluntary selling down their stakes, like what Steve Jobs did when he
got ousted by Apple's board in 1985. Forced selling, although rare in
corporate history, but not entirely impossible. Forced selling can come
from government authorities banning a person from the aviation industry
for several years. Most often being pressured to do so. The last one
would be a resignation, effectively relinquishing power to have any say
in the future direction of AirAsia.
How
often do we see any of the above scenarios happen to a majority
shareholder? I’ll put it at 3% prior probability given Tony is the
largest shareholder of AirAsia and in the corporate history of Malaysia,
there have been little cases where a majority shareholder is forced to
relinquish control of his company.
Condition of the hypothesis being true - What is the probability of bribery allegation if Tony loses control of AirAsia?
If
Tony loses control of Airasia, how likely would we see the evidence of
bribery allegation? In one way or many, Tony can lose control of AirAsia
for many reasons other than bribery allegations, but definitely bribery
is a big reason. I’ll give this 60%.
Condition of the hypothesis being false - What is the probability of bribery allegation if Tony does not lose control of AirAsia?
If
Tony doesn’t lose control of AirAsia, how likely is it to see the
evidence of bribery allegation? Again, Tony can get acquitted by MACC,
or get fined, which I believe is a more likely scenario. So a bribery
allegation doesn't always mean Tony will lose control of AirAsia. Far
from it. So I’ll put 50% on this being false
My initial belief that there’s a 3% probability of Tony losing control of AirAsia has gone up to 4%, a very small change.
Of course, these are my subjective guess, and an initial estimate as to whether these two hypotheses are going to happen. As more new evidences show up, I'll revise my belief as necessary. Your belief will not be the same as mine, could be higher or lower. Someone might think the possibility of getting fined is 90% rather than 23%, and that's fine. The beauty of Bayes's theorem is that over time as more evidences show up, it will allow our belief to become less wrong and 'converge' towards a smaller range of possibility.
Conclusion
The
advantage of using Bayes’ theorem is by quantifying your belief, you
have a better idea of what your belief entails. And as you practice
Bayesian thinking, you get better at making decisions because it forces
you to consider alternative outcomes, which allow you to have a more
balanced judgment. In the scenarios above, we consider the possibility
of something else that could happen as much as what is likely to happen.
What can explain the allegation aside from the obvious? What do we
expect to see if this evidence shows up? You start thinking in
decision-tree. You become more comfortable dealing with uncertainty and
cultivate an open-minded attitude rather than succumb to confirmation
bias.
https://klse.i3investor.com/blogs/JTYeo/2020-02-09-story-h1483734824-AirAsia_How_to_think_about_the_uncertainty.jsp
