Lecture 1: Theoretical foundations of financial regulation
3 dismal truths about financial regulation:
- it is complicated: a lot of rules, a lot of details and the coverage of a niche
- it involves choosing winners and losers between two groups: whenever we introduce
a rule, we reduce someone’s utility
- everyone who understands anything about it works for one group or another (either a
regulator, or works in a financial institution -> has some skin in the game). No expert
is a neutral expert
Why do we have financial intermediaries?
- banking as a monitoring device
- banking as an asset transformation device
Banking as a monitoring device
General investors are uniformed and are small enough to create free riding problems. Small
investors do not have the access, skills, time or resources to analyse and check what firms
are doing with the money you give them as investors. You do not have the incentive as a
small investor to invest this time/energy. -> freeriding problems: everyone holds that
someone else is checking what the firm is doing. As a result, no one actually checks the firm.
This can lead to problems if firms have incentives to do something suboptimal for
shareholders.
Investors may not be skilled, but they are rational. If they realize that firms are doing
something bad with the capital they are providing, they will not give capital in the first place.
Hence, we may miss good projects in society because firms cannot commit to doing good
things with the capital. -> need for banks
Banks have the necessary knowledge and incentives to monitor borrower’s behaviour.
Hence, a project that would not be financed by the market, is financed by the banks.
Bank = an institution whose current operations consist in granting loans and receiving
deposits from the public.
Do we still need banks?
Corporate finance is becoming more and more market based. In the US, corporate bonds’
face value is around 10 times larger than bank loans to firms. There exist more and more
alternatives for raising money from a traditional bank loan. Tech companies try to disrupt the
banking sector and they remove financial intermediation.
What do banks do?
- monitoring and information processing tool
- asset transformation
- denomination (transformation from small deposits to large loans)
- quality transformation (loan to bank is risky but deposit is safe)
- maturity transformation (deposit is short term and mortgage is long term)
- risk management = evaluation of assets risk
1
,Our model:
Firms need debt capital to invest, two possibilities:
- direct lending from small investors
- banks loans
- banks are different from small investors as they can monitor the firms
Setup of our model:
- 1 firm. Initial investment cost: I0 = 1
- Risk neutral competing investors. Rf = 0%
- Two possible investment projects
- Good. Low risk, positive net present value -> higher payoff than cost
- Bad. high risk, negative net present value
- what is attractive about a bad investment project? -> upward potential
= it can produce a lot of good value if it goes well, but there is a low
probability of it going very well. But if it goes well, a lot of value is
created.
Expected payoff good project: πG * G
Expected payoff bad project: πB * B
Expected NPV good project: πG * G - 1 > 0
Expected NPV bad project: πB * B - 1 < 0
We assume: πB < πG (probability of success)
You do direct lending to gain interest rate. Return offered to creditors: R = 1 + r. 2 scenarios:
- If the project goes well, the creditor gets R. They get back the money which they
lended and get the interest.
- If the project does not go well, the firm defaults. Both the firm and the creditors get 0.
The firm opts for the good project when: expected return with G > expected return with B
Expected return with G:
Expected return with B = πB * (B - R)
Firm opt for G (good project) if: πG * (G - R) > πB * (B - R)
So: expected return with G > expected return with B
2
,Hence:
𝜋(𝐺) ∗ 𝐺 − 𝜋(𝐵) ∗ 𝐵
R< 𝜋(𝐺) − 𝜋(𝐵)
𝜋(𝐺) ∗ 𝐺 − 𝜋(𝐵) ∗ 𝐵
Where 𝜋(𝐺) − 𝜋(𝐵)
is called Rc -> the threshold value of R
R = total return promised to creditor
If R > Rc, the firm does the bad project because the interest rate asked by the creditors is too
high
If there is too much debt/cost of debt is too high, firm chooses B
Creditors influence the firm’s choice by choosing R by demanding a lower or higher return.
By asking a higher return (interest rate), they incentivize the firm to do a bad project (high
potential, but low probability -> gamble). This is called an asset substitution problem.
Asset substitution problem = when a firm’s management willingly deceives another by
replacing higher quality projects/assets with lower quality projects/assets after a credit
analysis has already been performed. E.g. they could sell a project as low-risk to get
favourable terms from creditors. After loan funding, they could use the proceeds for risky
endeavors, thus passing the unforeseen risk to creditors.
If the creditors ask a low enough return, they will do a good project. Creditor know that:
π = f(R) -> probability of success is the probability of seeing their money back
If R < Rc, they will do a good project (πG). Financing through loans is only possible with R <
Rc.
If R > Rc, they will do a bad project (πB). With project B, creditors can expect to lose money:
πB * R < 1. Hence, creditors will never finance a bad project as they cannot expect to get the
full money back.
Creditors cannot ask for more than a project can create, so: R < B. Maximum value it can
create is B. The bad project has a negative NPV of: πB * B < 1
Creditor will not finance a bad project. Equilibrium is only possible with: π(R) = πG. Hence, it
can only be a good project, because if it is a bad project, they will not give money in the first
place.
For risk-neutral, perfect competition where the risk free rate is 0:
π(R) * R = 1 -> creditors expected 0 return in equilibrium.
Hence, the equilibrium return is:
πG * R = 1
Hence: R = (1 / πG)
This equilibrium is only possible when R < Rc, otherwise the firm chooses a bad project
instead of a good project. Hence:
(1 / πG) < Rc
This condition is more likely to be satisfied (so we choose for a positive NPV instead of
gambling) if the difference between B and G is small. Direct lending only works when moral
hazard is not a big issue.
Summary:
3
, If (1 / πG) < Rc -> direct lending is possible at R = (1 / πG)
If (1 / πG) > Rc -> direct lending (= equity or, market funding) is not possible because:
- creditors know the firm would do the bad project
- creditor must then in principle ask R = (1 / πG) > B
- they ask for a very high interest rate for the low expected return of project B
- however, no project can produce more value than B. Hence, firms do not
accept the terms of this loan
Hence, good projects (NPV > 0) do not get financed as firms cannot credibility commit to
them, which brings a social cost. Hence, we need banks who can monitor firms. They pay C
(monitoring costs) to make sure that the firm chooses the good project.
The equilibrium interest banks require is the recovery of lending and the money spend on
monitoring:
πG * Rm = 1 + C
Expected return for the monitoring banks = costs for the monitoring bank (lending + monitor
costs)
-> Rm = (1 + C) / πG where Rm is the return of monitoring banks
Due to the cost of monitoring, banks loans are in principle more costly (Rm > R) than equity.
Hence, firms ask for a bank loan only when direct lending is not possible, when (1 / πG) > Rc.
The bank cannot ask for more than G (Rm < G) -> Rm < (1 + C) / πG
Where π = probability of success
Banks allow for good projects to be financed which do not get financed by the market
(directly lending), as banks can monitor. Small firms which get denied a loan by banks look
more and more for funding through P2P or crowdfunding.
How can we reconcile these two twins?
- the assumptions of the model are violated:
- investors in P2P platforms may not be not rational/not well informed about
possible projects (while this was an assumption). They do not understand
what they are doing and the risk of these firms.
- they ask R = 1 / πG even if 1 / πG > RC
- and if R > RC, they will do a bad project and direct lending is
not possible
4
