Summary Incentives & Control Lectures
Week 1: Introduction to Economic models
There are two basic assumptions in all economic models:
1. Greed; people like more over less. Think of salary negotiation: do you want 70k or 80k? most
people would say 80k. or do you want 100k, everybody would go for the 100K and not for
50K. This can be illustrated by a concave of utility function.
And Risk-aversion;
The concave function increases at a certain
point at a decreasing rate. The figure shows
both options. It shows ‘more over less’ and
the ‘Risk-aversion’. Each additional unit of
income makes you more happy, but the
increase in happiness does not decrease with
larger amounts of money. The figure has to do
with risk-aversion. Compared to the 200k you
will getting the 300k, but not so much as you
are afraid to lose the additional 100K. you
want to avoid losing 100K more than you are
happy getting 100k more. That is basically
Risk-Aversion.
2. Asymmetry; Some people are always
better informed than other people. The problem with asymmetric information is, it does not
help because people want to have transactions. And asymmetric information does not help
with completing transactions and has to be resolved. Two parties can take the initiative: the
uniformed party or informed moves first. If you combine these two together, thus what is
the private information about; hidden in actions, or hidden in characteristics? And the
second one is who takes the initiative in resolving the asymmetric information problem?
Then you can come up with three types of asymmetry information models;
- Moral hazard; About what do we have information? This is about hidden action/effort or
about hidden characteristics. Example: ‘the boss would like me to work hard, but I like to
work normally and Not work at the weekends, not work in the evenings if I am not
compensated for this’. This is a misalignment between principal and agent. The worker
(agent) prefers leisure activities over productive (costly) effort. The problem has to be
resolved, how does this happen? By the uniformed party moves first (the boss, the
principle, superior) he/she offers a contract, and the contract has a performance
measure in it. The performance measure picks up how hard the agent works. The harder
the agents works, the higher the performance measure will be, the more compensation
the agent will take home.
- Schreening/Adverse Selection; has to do with hidden characteristics. ‘you (agent) know
something about yourself about a characteristic that another party doesn’t know, which
is in assentation having these transaction’. How is this problem resolved? The
uninformed party moves first. And they offer you a contract which is only beneficial for
the high type, and not for the low type.
- Singnalling; This is about hidden characteristics, but the difference is that the informed
party moves first. You know something about yourself, private info, and you want to
, communicate this to somebody who doesn’t know this info. The problem is: ‘How are
you going to communicate this in an credible manner?’. It is all about credible
communication. The requirement for credible communication is has to be separating
equilibrium instead of pooling equilibrium. The last one means I can copy paste the same
strategy as you do. Separating equilibrium means that it is less costly for you to come up
with credible communication than it is for me.
What is the scope of “Incentives & Control”?
First, there are always people (employees/managers) that know something about the market they
are operating in, that headquarters do not know.
Second, some people (managers) have personal objectives, that are not perfectly aligned with
corporate objectives. There is some mis alignment of interest. Many organizations have both
characteristics.
The scope of this course: How can we apply methods and techniques to align the interest of
employees throughout the organization with the corporate objectives in a setting that is generally
characterized by asymmetric information.
Why is this relevant to us?
Because people commit fraud, see wells fargo example.
From an assurance perspective it is important to know how are the incentives and controls within the
firm organized? Because it might also indicate a risk factor when setting up the audit engagement ->
Some firms have a questionnaire to assist engagement teams with the documentation of risk
assessments on fraud and non-compliance by law regulation.
Why do incentives & control problems emerge?
First thing relevant to know is: what do we need to know to make good business decisions?
Knowledge and (authority to make) Decision rights. You need both, we have to co-locate decision
right with knowledge. However, in reality knowledge is often dispersed throughout the organization.
To co-locate decision right with knowledge, we can do two things: Moving the knowledge to those
with decision rights (KTC) or move decision right to those with knowledge (CC). We can distinguish
between two kinds of knowledge: specific knowledge and general knowledge. General knowledge is
easy and inexpensive to transmit (who are your main competitors, what is your market share?, etc.)
Specific knowledge is more difficult and costly to transfer/transmit. Thus the only option is:
decentralizing the decision rights. This arises the next questions: Who gets the decision right? and
how do you make sure that the people who get the decision rights not only take actions who are in
their own best interest, but also in the interest of the whole firm?
You can look at outside markets (Meghan 1992 paper), but we can’t learn much from it because
outside markets have a system of private alienable rights (rights you can buy and/or sell on the
market). How much you want to pay for these right depends on how much value you expect to
extract from this license. The rights are assigned by the one who value (for example) this license the
most, and they are motivated to make good use of this right because all the sales that they do are all
to their own benefit. That is why they are motivated to make the best out of it. However, this does
not work for firms systems, because e.g. employees do not receive rights to alienate decision rights
under their own control.
To what extend should firm decentralize the decision rights? You can look at two things: look at firms
who are completely centralized (left) or completely decentralized (right). Left you see two costs; ‘cost
due to poor information’ and ‘cost due to inconsistent objectives’. When a firm is completely
centralized, the cost due to poor information is very high. (Which makes sense; headquarters who
make all the decisions), but a lot of knowledge is out there in the organization. The cost due to
, inconsistent objectives are very
low. This makes sense; there is
no other chance that other
people will make decisions
which are not for the best
interest of the firm. Look at the
right, (decentralized) the cost
due to poor information is very
low, because most people have
knowledge and decision rights
to act on this knowledge. But
the cost due to inconstant
objectives is very high, because
all (product) managers in the
different regions have decision
rights, but they sometimes
make decisions that are best for them but not for the firm as a whole.
The total cost are the sum of cost due to poor information and cost due to inconsistent objectives. So
on average they can conclude it is kind of halfway.
The optimal level of delegation is a trade-off between ‘the cost of poor info’ versus ‘cost of
inconsistent objectives’. People are not assigned totally to all decisions rights to make any major
decision, it will give a breakdown of decision rights. (different types of decision; initiation,
ratification, implementation, monitoring decision rights). If we delegate decision rights to managers,
we are going to evaluate; subsequently we provide rewards (successful) and punishments (not
successful) tied to performance measure outcomes.
Bottom-line: Why do incentives & control problems emerge?
- Not all decision rights are maintained in headquarter, a lot of decision rights are delegated to lower
levels in the organization.
- We need to create an incentive system that provide measures of performance and specifies
relationship between reward and punishment and performance measure outcomes.
Framework of Merchant & Van Der Stede (1992 paper)
Created a control system framework, that is composed on four different types of controls:
1. Result controls
Result controls focus on the results of the actions. By holding people accountable for the results
delivered through their actions. Good results is a reward, poor results is a “punishment”. Especially
important in settings of asymmetric information. Advantage of result controls is Empowerment:
Head Quarter asks for certain profit target, managers are free to decide how to achieve this number
(autonomy). HQ will evaluate the profits at end of period.
To influence profit, manager needs to be able to influence both sales and costs (controllability
principle). When using result controls you can do two things:
- use result controls in ex-post fashion, this an evaluation afterwards. If the result are good, you can
conclude that the manager worked hard and complemented the right actions. Problem is: external
factors (e.g., pandemic/ economic recession) can influence the profit results of the manager.
- You can also use performance measure in an ex-ante fashion, this means the managers knows the
performance measure used. Advantage: manager knows from the beginning where to focus on. But
performance measures are affected by external factors, thus risk is imposed on the manager. The