Research " FACTORING IN BRAZIL - AN INFORMATION ASYMMETRY APPROACH " potx

2.3 The Contract 2.3.1 Contract Characteristics If a factoring contract is established between the seller and the factor, it will have to specify the time of payment Maturity and whether

FACTORING IN BRAZIL - AN INFORMATION ASYMMETRY APPROACH

BY

CLECIO JOSE BORTONI DIAS

Bach., Universidade de Brasilia, 1993 A.M., University of Illinois at Urbana-Champaign, 1995

UMt Number: 3017068

®

UMI

UMI Microform 3017068

Copyright 2001 by Bell & Howell Information and Learning Company All rights reserved This microform edition is protected against unauthorized copying under Title 17, United States Code

Bell & Howell Information and Learning Company

300 North Zeeb Road P.O Box 1346 Ann Arbor, Mi 48106-1346

UNIVERSITY OF ILLINOIS AT URBANA-CHAMPAIGN

THE GRADUATE COLLEGE

FEBRUARY 1999 (date)

WE HEREBY RECOMMEND THAT THE THESIS BY

CLECIO JOSE BORTONI DIAS

ENTITLED CREDIT CARDS, PRE-DATED CHECKS AND CONSUMER

CREDIT IN BRAZIL - AN INFORMATION ASYMMETRY APPROACH

The Choice between Credit Cards and PDC in Brazil: Empirical Results 37 Buyer’s Credibility and the Choice of Recourse (Sopranzetti’s Paradox) 52

4 9 TP - 4 58 Emmott 60 Curriculum Vitae 4444 63

1 Introduction

Credit cards have never been popular in Brazil Up until June of 1994, inflation rates were very high, variable and, to a great extent, unpredictable It was practically impossible for firms to wait for the 30 days grace period that is normally given on credit card purchases Due to the unpredictability of the inflation rate, adjusting the price was also not a viable option In order to compensate for the lack of a viable short-term consumer credit instrument, stores began to agree to wait a specific number of days before depositing checks that were used on purchases These arrangements were known

as Pre-Dated Checks (PDC)', and gradually became a national institution

The use of PDC presented two important advantages over credit cards, especially

in the context of a nearly hyperinflationary process:

1) PDC allowed for more flexibility with regard to payment schedules This flexibility was especially important in a period in which a one-week grace period could represent a significant discount

2) The number of consumers with checking accounts exceeded by far that of holders of credit cards This ready availability is, in part, a consequence of the PDC flexibility Since there was little acceptance of credit cards, they did not have a great appeal to customers

As inflation rates came down", another explanation had to be found for the persistent popularity of PDC

One natural candidate was the interest rate structure in Brazil After the enactment

of the Real plan, interest rates remained very high, in order to keep inflation in check and

to attract foreign capital There was also a heavy taxation on financial operations, which

kept the spread between savings and borrowing rates very high The use of PDC was a way to avoid payment of financial operation taxes charged on credit card operations Even though credit cards and PDC perform nearly the same function in the economy, they are not perfect substitutes From the seller’s point of view, there is a major difference in the way the risk of default by the customer is borne In credit card sales, the risk is automatically transferred to the card administrator For purchases paid with PDC,

it stays with the seller It is natural to assume then that the seller’s preference of medium

of payment would be a function of her information the customer’s creditworthiness

As the popularity of PDC increased, so did the market for purchases and sales of PDC themselves Instead of waiting for the PDC to mature, storeowners would sell them

to a factor, at a discount, and cash the money immediately

Factoring companies are firms that specialize in the purchase and management of accounts receivable Firms engage in business with factors mainly for two reasons: to make capital advances, and to reduce risk exposure" In Brazil, firms that factor their PDC can do so only with the purpose of advancing capital If the check is not honoured, they are responsible for reimbursing the factor for the loss

The goals of this research are the following:

1) To analyse the role of information asymmetries in the choice between the use of PDC and credit cards

2) To investigate the existence of reinforcing mechanisms that might explain the prevalence of PDC or credit cards in different economies or even in different sectors

of the economy

3) To correct and extend Sopranzetti’s (1998) factoring model, upon which this dissertation is based

This dissertation is organized as follows:

In chapter 2, we create a simple economic model with asymmetric information in order to analyze the seller’s preference for different types of factoring options (and, thus, for the medium of payment)

In chapter 3, we focus on the consumer’s choice of medium of payment and formulate a model combining the seller and customer’s preferences As a consequence,

we observe the existence of reinforcing mechanisms that might explain the prevalence of PDC or credit cards in specific sectors of the economy

In chapter 4, we test some of the hypotheses raised in the two previous chapters against data collected in Brazil We observe strong evidence of information-based explanations for the sellers’ preference of medium of payment and some qualitative evidence of the presence or reinforcing mechanisms determining the prevalence of PDC

or credit cards in specific sectors of the economy

The model presented in chapter 2 is an extension of Sopranzetti (1998) In chapter

5, we correct some errors in Sopranzetti’s analysis, and offer an alternative explanation for his empirical findings

Chapter 6 presents our conclusions

2 The Model

In order to analyze the consumer credit market in Brazil, we present an asymmetric information economy with three types of agents: consumers, sellers and factors

The model presented here is based on Sopranzetti’s (1998) analysis of the factoring market in the US That model, in turn, is an extension of Gorton and Pennacchi (1995), originally created for the analysis of loan sales In Sopranzetti’s model, the seller’s monitoring leve! was considered private information Here, we extend that idea

by also making the factor’s level of monitoring private information As a consequence, the design of the factoring contract is the result of efforts to overcome moral hazard problems of both parties Another departure from the original model is that sellers are distinguished by different levels of internal capital costs

Those extensions will allow us to correct some errors in Sopranzetti’s analysis, while keeping all of its major results Those results were derived from empirical observations of the American factoring market Therefore, keeping them in our model is

an important step towards generality

As shown by Smith and Schnucker (1994), credit card sales can also be viewed as factoring operations The reasoning goes as follows: in order to compensate for the lack

of information that the seller might have of the buyer’s creditworthiness, the credit card company steps in and buys the accounts receivable generated by the sale at a discount Since the credit card company specializes in information administration and credit collection, it can offer the seller an attractive enough discount rate

The chapter is divided as follows: Sections 2.1-2.3 present the model characteristics, the game and the factoring contract Sections 2.4-2.6 present the main facts Section 2.7 discusses the effects not having totally creditworthy sellers in the model Section 2.8 presents the full information solution, and the comparison between that and the one of the original model Section 2.9 discusses the consequences of having factors with different capital costs Section 2.10 examines the meaning of recourse factoring

2.1 Model Description

2.1.1 Agents, Endowments and Preferences

There are three types of agents in the economy: Buyers, sellers and factors

e Buyers want to purchase one unit of a good in period 0 In exchange, they promise to pay the seller L monetary units in period ' This payment, however, is not certain

We will assume that the probability of payment will be a function of the buyer’s creditworthiness c, and of how much monitoring effort the seller and the factor spend

on the buyer Different levels of creditworthiness distinguish buyers

e Sellers can produce the good in period 0 at a cost of one monetary unit They have enough rotating capital to extend consumer credit to the seller They also have the option to negotiate the accounts receivable generated by the sale with a factor If the seller so desires, she can monitor the buyer at a cost A level c of monitoring will cost the seller uc Sellers are divided into two categories, h and 1, according to their discount rates We define jas the inverse of the discount rate and assume that 4,<7

e Factors specialize in the purchase and management of accounts receivable Their discount rates fall in between that of the two types of sellers (y,<¥<y) Like sellers, factors can also monitor buyers and the expected value of payment will increase with the amount of monitoring m exerted by the factor Monitoring the buyer at a level m will cost the factor wen All agents are assumed to be risk neutral and to maximize expected incomes

2.1.2 Information Asymmetries

The level c and m of monitoring are assumed to be private information for the seller and the factor, respectively

2.1.3 The Characteristics of the Promised Payment

Let x be a random variable representing the expected value of the payment promised by the buyer to the seller We assume that x is an increasing function of a, c and m We also assume that «x is strictly concave with regard to those three variables;

The model presented above yields no closed-form general results In line with Sopranzetti (1998) and Gorton and Penacchi (1995), we choose the following exponential functional form for x:

® x increases at a decreasing rate with @ Since x is bounded between 0 and L, we restrict Z such that œ€ [I.=} In the limit, as @ goes to infinity (the buyer is perfectly creditworthy), x=L and monitoring is irrelevant

e The relative efficiency of c and m — monitoring by the seller and the factor - is regulated by # and 6, the rates at which x converges to L as c and m, respectively,

increase

2.2 The Game

The game is structured as follows:

1) The buyer applies for consumer credit with a seller

2) The seller observes &, and approaches a factor in order to offer him a factoring

contract for the accounts receivable to be generated by the sale The contract specifies:

® Apnce

® A time provision, i e., whether payment will be made as soon as the contract is signed or upon maturity of the receivable

e A recourse provision

3) The factor observes o and the contract characteristics, and decides whether or not

to accept the offer If he does so, he also decides how much monitoring m he will perform

4) If the factor accepts the contract, the seller extends credit to the buyer and decides

how much monitoring c she will undertake

2.3 The Contract

2.3.1 Contract Characteristics

If a factoring contract is established between the seller and the factor, it will have

to specify the time of payment (Maturity) and whether the factor will be able to call upon the seller for payment in case the buyer defaults on his promised payment (Recourse)

With regard to time, factoring payment can be made as soon as the contract is signed and the receivable factored (advance factoring) Alternatively, it can be made once the receivable matures (maturity factoring) We shall observe that the sellers’ internal capital costs play a major role in determining the choice between two types of contract

The recourse provision determines whether it is the seller or the factor who ultimately bears the defaults In no-recourse contracts, the factor bears the loss One can interpret that type of contract as if the factor bought the risk together with the receivable

In recourse contracts, the seller bears the loss The price of the receivable will, of course, vary according to the recourse provision

The same line of thought can be applied to the time provision in the factoring business In advance factoring, the factor buys the necessary wait for the maturity time together with the receivable Thus, those are the two variables that are traded in a factoring contract: Time and Risk.‘

2.3.2 The Optimal Factoring Contract

Given the game structure, all the expected profits are captured by the seller; the factor makes 0 expected profits We can reformulate the game as a constrained maximization problem on the part of the seller The optimal contract will be individually

rational and incentive compatible for both the seller and the factor In other words, profits must be non-negative and the agents must have the proper incentives to act according to their private information

Let R be the seller’s expected profit, F the factor’s expected profit and P the price offered by for the receivable The optimal contract can be described by the following equations:

(5) Mm = Arg max F (Factor’s IC) (6) C = Arg max R (Seller’s IC)

where / represents an index for the time provision (/=1 for advance factoring, /=0 for maturity factoring) and ¢, an index for the recourse provision (C=0 for no recourse and C=1 for recourse contracts)

By varying / and ¢ we can observe the following types of contracts:

e I=1, C=0 — Advance factoring without recourse;

e I=l, É=l — Advance factoring with recourse;

e [=0, C=0— Maturity factoring without recourse and

e I[=0, C=1 — Maturity factoring with recourse, equivalent, equivalent to no factoring at all

Notice that P is not a function of c or m Since the level of monitoring by the factor and the seller are private information, the price paid by the receivable cannot depend on those variables

Before we proceed with the analysis using the chosen functional form for x, we can establish some facts regarding the choice of advance or maturity contracts by buyers according to their discount rates We also establish the relation between the choice of recourse provision in the contract and the monitoring efforts exerted by the seller and the factor It is important to notice that many of the facts are fairly general and independent

of the functional form chosen for the expected value of the receivable As a general rule,

if a fact is dependent on the functional form, it will be apparent in the proof

Fact 1 - Low discount rate sellers will always opt for maturity contracts High discount rate sellers will always opt for advance contracts

Proof - We start by observing that the choice of I does not affect the seller’s IC constraint (5) Since the structure of the game guarantees that the sellers IR will be met with equality, we can solve it for P and substitute it into the maximization problem:

By substituting /=1 and 1=0 into (7), we have respectively:

(8) R=maxy,lx-wm)+(y,-1,)$(L-x)-yuc i=hl P.c.m.C

(9) R= max/.[x-wm-uc] i=hl

Prem

It is easy to observe that, for 7,<¥, (8)>(9), and that the inverse is true for 4> 7

2.4 The Recourse Provision and the Monitoring Efforts by

the Seller and Factor

From the general maximization problem, we can also establish the relationship between the recourse provisions in the factoring contract and the monitoring efforts exerted by the seller and the factor

Fact 2 - In no-recourse contracts (C=0), c will also be equal to zero, i.e., the seller will exert no monitoring effort

Proof -By substituting C=0 into (6), the seller’s IC constraint, we have:

Fact 3- In recourse contracts (C=1), m will be equal to zero, i.e., the factor will exert no monitoring effort

Proof - By substituting C=1 into the factor’s [C constraint (5), we have:

(1U max Fn) =y,[L-wm]-yP 20

The partial derivative of (11) w.r.t m is equal to — y,w <0 Thus, in order to maximize F, the factor will choose m=0

Again, we can observe intuitively that, in recourse contracts, the factor has no incentive to monitor In case of default, he can recur to the seller for payment He will therefore maximize expected earnings by not exerting any monitoring efforts

We can now solve the maximization problem for the advance and maturity factoring cases, and to perform some comparative static analysis

2.5 Advance Factoring

Fact one has established that only high discount rate sellers will opt for advance factoring Facing a consumer credit application from the buyer, the seller will compare the expected profits for recourse and no-recourse contracts

In order to analyse the considerations that will influence the seller’s decision, we can now

- Be-dm

e

substitute the functional form (1) x= uf — a into the maximization problem, and perform the comparison

2.5.1 Advance, no-Recourse Factoring

By substituting I=1 and C=0 into the maximization problem, we have:

2.5.2 Advance, Recourse Factoring

For /=1 and C=1, we have the following maximization problem:

Since the factors IR will be met with equality, we can solve it for P, substitute it into R The first order condition for c then yields:

The solution for the maximization problem (Raa) will be given by:

w w wa@ H u ua

Proof: We obtain (14) by simply subtracting (13) from (12)

From (14), we can perform some comparative static analysis We observe that the choice of the recourse provision will depend on the relative costs (u/8)and (w/d) of monitoring by the seller and the factor, respectively For instance, the first derivative of (14) with respect to (u/A) is equal to:

Besides weighting relative monitoring costs, the high discount rate seller also has

to take into account the difference between her and the factor’s internal capital cost Recourse factoring allows her to borrow more from the factor In fact, given our assumptions, she can borrow the entire promised payment L Therefore, the greater the difference between y, and y the more likely it will be that the factoring contract will have

a recourse provision

2.6 Maturity Factoring

Low discount rate sellers will only sign maturity factoring contracts We now compare the expected returns for the seller for recourse and no-recourse factoring, in order to determine in which way the variables of the model will affect her decision between one choice and the other

2.6.1 Maturity, no-Recourse Factoring

By substituting [=0 and z=0 into the maximization problem, we have:

a From the factor’s IC we have the following first order condition:

wa

—e"=w or m=—lIn

From the factor’s IR, P will be given by: P= L— : +ăn SE] , and the solution for R,

in this situation:

w w wa

as) R NRM 71 Ũ “5 Bin 2)

2.6.2 Advance, Maturity Factoring (No Factoring)

If the low discount rate seller decides to keep her accounts receivable, then her expected income will be the solution to the following problem (/=0,¢=1):

Fact 5 — Given the chosen functional form, low discount rate sellers will choose recourse over no-recourse factoring if:

Proof: Similar to Fact 4

Just as in the advance factoring case, relative efficiency in monitoring between the seller and the factor determines the choice between recourse and no-recourse factoring However, in this case the seller is not borrowing money from the factor As a consequence, discount rates have no influence on the decision taken by the seller

2.7 On the Implicit Assumption of Limited Borrowing by the

Seller

After observing the results of our model, one could pose the following question: Since high discount rate sellers are assumed to repay their debts with certainty, why don’t they simply borrow enough money from factors such as to bring their discount rate to the same level of factors, and then behave like low discount rate sellers?

In fact, that option has only been prevented by an implicit assumption limiting the amount that could be borrowed by the seller to the face value of the receivable L

Since there is no intrinsic justification for such assumption, we have the following options:

1) Drop it altogether — High discount rate sellers would behave like low discount

rate ones, and only the relative efficiencies in monitoring (u/f) and (w/d) would influence the decision of factoring receivables with or without recourse

Instead of implicitly limiting the amount that could be borrowed by the seller to

L, we could explicitly assume that in case of default, she would only honor the recourse clause of the contract up until the expected value that the receivable would have for her

In order to adapt the model to this new assumption, we change the timing of the game as follows:

1) The good is sold, the receivable is factored in period 0 at a price P and a recourse

provision are established”

In period 1, the factor tries to receive payment without any monitoring If there is any default, the factor will face to options:

2.1) He can spend monitoring efforts in order to maximize the expected value

of the receivable (equivalent to no-recourse factoring)

nN tv ~ He can resell the receivable to the seller for a price Q The receivable’s price will be given by the expected value that it has to the seller (Equivalent to recourse factoring with limited liability for the seller.) Obviously, the original price P paid by the factor for the receivable will be a function of the expected value that the factor expects to get at time 1 We should also observe that the competitive structure of the game guarantees that all expected profits will still be transferred to the seller

2.7.1 The Factor’s Choice of Monitoring or Reselling the Receivable

If the Factor keeps the receivable and decides to monitor, his expected profits are given by:

a lS oa) a aM ee) la)

If the opposite is true, the seller will factor its receivable without recourse

From Fact 6, we can observe the following: The partial derivative of (20) with respect to

Yn is always negative That is, if the seller’s discount rate increases (¥, decreases), she will

be more inclined to opt for recourse factoring The partial derivatives of (20) with respect

to (⁄) and (w/d) are, respectively, negative and positive In other words, even though the final equation is slightly more complicated, all major results from Fact 4 are preserved For simplicity, in the remainder of the text, we shall assume that facts 1 — 4 are still valid

2.8 The Full Information Solution

In order to better evaluate the loss in efficiency caused by information asymmetries in the model, we find the solution for the factoring contract when c and m are public information

We start by observing that Fact 1 is still valid under this new hypothesis Comparing with the original maximization problem, the full information one will not have the incentive compatibility constraints, since now each party in the contract can observe the amount of monitoring exerted by the other one For the purpose of Fact 1, all other equations remain the same

R Frac O) = max 7," P(c,m)— y,{¢(L - x(đ,c,m)) + uc]

cml S.T

For Recourse contracts (¢=1), we have:

(22) Rar ago = max yr(L— wm)— 7,[Ù — x(Œ,c,m) + wc]

For no-Recourse contracts (¢=0),

(23) Ret anr (c,m) = max Y,[x(Œ,c,m) — wm]— „cu

By subtracting (23) from (22), we have (7; —ƑY,„)(L~ x(Œ,c,m)) >0, i.e., the recourse contract dominates the no-recourse one

The intuition behind this result is that high discount rate sellers want to borrow as much as possible from factors They can do so if they promise to honor the buyer’s commitment when he defaults Since there are no incentive compatibility constraints in the full information world, there are nc reasons for them not to make that promise

The F.O.C for (23) will be:

Fact 7- If monitoring is public information, then high discount rate sellers will only sign recourse factoring contracts Moreover, if

re oO Fae}

then only factors will monitor If the opposite is true, only sellers will monitor

The conclusions to be drawn from the comparison between the full and the private information cases are, of course, dependent on the functional form chosen for the expected value of payment x However, some fairly general conclusions can be inferred:

e If, in the full information world, the seller is expected to monitor (or to be the main monitor), then there will be little or no loss in efficiency due to information asymmetries

e If however the system relies mainly on the factor for monitoring, then there will be considerable loss in efficiency Normally, moral hazard situations should not give rise

to inefficiencies in a risk neutral economy The natural conclusion is that inefficiency arises as a consequence of the implicit assumption regarding financing restrictions by the sellers We can observe that the factor will tend to over monitor the buyer

[nlf < 0 , and expected earnings will drop accordingly

f

2.8.2 Maturity Factoring

As in the asymmetric information case, low discount rate sellers will opt for

Rep y oS) = max lr Plevm)— 11S (L— x(@,c,m)) + ue]

Notice that Gis no longer present in the maximization problem In words, if there

is full information concerning the monitoring efforts by the seller and the factor, then the choice of recourse provision becomes irrelevant

The F.O.C for (28) are:

By comparing those results with the expected profits in the private information solution, we can observe that they are equivalent to (15) and (16)

2.9 Different Discount Rates for Different Types of Factors

Up until this point in the model, we have assumed the existence of one single type

of factor, that signs both recourse and no-recourse contracts In the Brazilian consumer credit market, this is not the case Credit card companies perform no-recourse operations Recourse factoring is performed by so called factors of PDC

In this section we analyze the consequences of having different parameters for different types of factors Since it has been established that factors that sign recourse contracts do not monitor, we shall ignore changes in monitoring parameters 6 and w Instead, we should concentrate the consequences of having different discount rates for different types of factors

Let ¥ and y be the inverse of the discount rates for credit card companies and factors, respectively with (¥-<yg) Possible justifications for having higher discount rates for credit card companies, when compared to PDC factors, would be differences in their tax structures and lower operating costs

2.9.1 Expected Profits for Different Factoring Options

By doing the necessary modifications to (12), (13), (15) and (16), we can observe and compare expected profits for the four factoring possibilities:

2.9.2 Low Discount Rate Sellers

By substituting i=/ into (29) - (32) we observe that, so long as Yn<Vie<Ver<yi, there are no major changes in the factoring habits of low discount rate contracts Depending on relative monitoring efficiency, they will either sign maturity factoring without recourse or keep their receivables

If instead we have yn<Vre<Vi<¥r we will observe a change in pattem Now (31)>(29) and (30)>(32) In words, the low discount rate seller will, depending on relative efficiency in monitoring, either sign maturity contracts without recourse or advance contracts with recourse

2.9.3 High Discount Rate Sellers

By substituting y, into (29) - (32), we can draw similar conclusions for high discount rate sellers Again, if Yn<Vie<Ya<yi there are no changes in the seller’s factoring choices However, if Yic<Yn<Ya<yi, then (30)>(32) and (31)>(29) Depending on relative

monitoring efficiency, high discount rate sellers will either sign advance recourse or maturity no-recourse factoring contracts

2.10 On the Meaning of Recourse Factoring — Why doesn’t Everybody Factor?

Fact 1 establishes that high discount rate sellers will always advance factor their accounts receivable, either with or without recourse Since a considerable proportion of firms in any economy would fall into that category, the proposition deserves some consideration

Mian and Smith (1992) describe the following necessary actions for the extension

of consumer credit to a buyer:

19) “the credit risk of the potential account debtor must be assessed,

2) the credit granting decision (including setting credit terms) must be made,

3) the receivable must be financed until maturity,

4) the receivable must be collected and

5) the default risk must be borne.”

In order to maximize its expected profits, a firm may decide to assign part or all

of those tasks to an third party

Table I, reproduced from Mian & Smith, establishes the way those tasks are divided according to the type of outside agency that the firm chooses to manage its accounts receivable