Full text of Working Papers (Federal Reserve Bank of Chicago) : The Choice Between Arm's-Length and Relationship Debt : Evidence From e-Loans, Working Paper 2008-10

The full text on this page is automatically extracted from the file linked above and may contain errors and inconsistencies.

Federal Reserve Bank of Chicago

The Choice between Arm’s-Length
and Relationship Debt: Evidence
from eLoans
Sumit Agarwal and Robert Hauswald

WP 2008-10

The Choice between Arm’s-Length and Relationship Debt:
Evidence from eLoans
Sumit Agarwal
Federal Reserve Bank of Chicago

Robert Hauswald
American University

Current Version March 2008

JEL Classi…cation: G21, L11, L14, D44

We thank Hans Degryse, Victoria Ivashina, Robert Marquez, Steven Ongena, Maria-Fabiana Penas, Raghu
Rajan, Phil Strahan, and Greg Udell for stimulating discussions and seminar participants at American University,
the ECB, the ISB 2007 Summer Research Conference in Finance, the 2007 European Finance Association Meetings,
the Conference “Information in Bank Asset Prices: Theory and Empirics,” George Mason University, Mannheim
University, the 3rd New York Fed/NYU Stern Conference on Financial Intermediation, and the 2008 American
Finance Association Meetings for comments. Je¤ Chin provided outstanding research assistance. The views expressed
in this research are those of the authors and do not necessarily represent the policies or positions of the Federal
Reserve Board, or the Federal Reserve Bank of Chicago. Contact information: Sumit Agarwal, Federal Reserve Bank
of Chicago, Chicago, IL 60604-1413, ushakri@yahoo.com, and Robert Hauswald, Kogod School of Business, American
University, Washington, DC 20016, hauswald@american.edu.

The Choice between Arm’s-Length and Relationship Debt:
Evidence from eLoans

Abstract
Using a unique sample of comparable online and in-person loan transactions, we study the
determinants of arm’s-length and inside lending focusing on the di¤erential information content
across debt types. We …nd that soft private information primarily underlies relationship lending
whereas hard public information drives arm’s-length debt. The bank’s relative reliance on public
or private information in lending decisions then determines trade-o¤s between the availability
and pricing of credit across loan types. Consistent with economic theory, relationship debt
leads to informational capture and higher interest rates but is more readily available whereas
the opposite holds true for transactional debt. In their choice of loan type, lender switching,
and default behavior …rms, however, anticipate the inside bank’s strategic use of information
and act accordingly.

1

Introduction

Banks typically o¤er two very di¤erent types of credit to their corporate customers: relationship
loans characterized by inside information and transactional loans for which banks compete on a
much more equal informational footing (see, e.g., Broecker, 1990, Rajan, 1992, Inderst and Müller,
2006, or Hauswald and Marquez, 2006). While the theoretical implications of competition between
informed and uninformed lenders are well understood much of the empirical work has focused on
relationship lending, in part because data on lending relationships is more readily available (see,
e.g., Petersen and Rajan, 1994, Berger and Udell, 1995, or Elsas, 2005). Furthermore, private
transactional debt with the attributes posited by the theoretical literature is hard to identify in
practice. However, recent advances in lending technologies …nally make available new data on
credit-market transactions that closely …t the theoretical de…nition of transactional lending: online
loans. Hence, we propose to …ll this gap in the literature by analyzing the comparative determinants
of online (transactional) and in-person (relationship) credit transactions.
Using a unique sample of all online and in-person loan applications by small businesses to a
large US bank over a 15-months period we investigate a …rm’s choice between transactional (“arm’slength”) and relationship (“inside”) debt and the ensuing bank-borrower interaction to better
understand the economic forces that shape exchange in these two market segments. For each loan
application we collect the bank’s ultimate credit decision and loan terms, its internal credit score,
and the eventual loan performance. Although our bank’s lending standards are identical across
the two modes of origination loan o¢ cers can individually adjust internal credit scores for inperson applications that therefore contain a soft, subjective credit-assessment component supplied
by branch o¢ ces. No such interaction or adjustment takes place for online applications. From
credit-bureau reports we also know each applicant’s Experian Small Business Intelliscore (XSBI)
as a measure of publicly available information and can identify …rms that refuse the o¤ered terms
to switch lenders.
The primary di¤erence between arm’s-length and relationship debt stems from each loan type’s
information content that determines the availability and pricing of credit. Hence, we …rst orthogonalize each applicant’s bank-internal score with the publicly available XSBI score to obtain its
private-information residual (PIR) as a clean measure of the lender’s proprietary intelligence gath-

ered in the screening process. We then follow the typical steps of bank-borrower interaction and
estimate discrete-choice models of the …rm’s choice of lending channel, the bank’s decision to o¤er
credit and the borrower’s to accept the loan terms, and linear-regression models of the o¤ered loan’s
all-in cost. We round o¤ our investigation of the di¤erential information content of arm’s length
and relationship debt by studying the borrower’s decision to switch lenders and the likelihood of
credit delinquency across loan types.
The explanatory variables are proxies for public (Experian score: XBSI), proprietary (lender’s
internal score), and private information (orthogonalization: PIR), and the nature of the lending
relationship or absence thereof. We control for borrower characteristics, loan terms, regional and
business-cycle e¤ects, and the prevailing interest-rate environment. Since the choice between transactional and relationship debt might depend on the local availability of credit we also include the
number of lenders and their branches in each applicant’s zip code to take into account competitiveness e¤ects and, similarly, the …rm’s distance to bank’s branch or online-processing center and to
the nearest full-service competitor as proxies for transaction costs terms of time and e¤ort.
We …nd that public and private information plays very di¤erent roles across lending channels because the bank predominantly relies on one type of intelligence for a particular debt product. Public
information drives transactional credit decisions and pricing whereas private information collected
through prior business interaction and the loan-origination process determines relationship-debt
o¤ers and their terms. We also show that the di¤erential information content across debt types
shapes the predicted trade-o¤ between the availability and pricing of credit for each lending channel
(Broecker, 1990, Rajan 1992, and Hauswald and Marquez, 2006). Arm’s-length debt is less readily
available but at lower rates because symmetrically informed banks, which compete on the basis
of public information, not only drive down its price but also restrict access to credit to minimize
adverse selection ceteris paribus. By contrast, better informed inside lenders strategically use their
information advantage to informationally capture relationship borrowers that pay higher rates but
gain easier access to credit.
Our results also reveal that …rms anticipate the lender’s strategic use of information and rely on
their public credit score as a credit-quality indicator for their own best response. As a consequence,
public information retains some measure of importance even in inside lending and in‡uences …rm
decisions in both arm’s-length and inside transactions whereas the bank’s private information pri2

marily matters to relationship borrowers. Given that …rms take into account the inside bank’s
rent-seeking behavior but nevertheless engage in relationship transactions this …nding strongly suggests that borrowers also bene…t from close ties to their lender, for instance through better access
to credit or intertemporal insurance e¤ects.
The impact and statistical signi…cance of our relationship variables con…rm these e¤ects across
speci…cations and lending channels. Since online applications do not permit banks to generate much
inside knowledge our lender discounts whatever private information might transpire in transactional
lending. By contrast, lending relationships not only o¤er the opportunity to collect such intelligence
but the length and depth of the interaction together with the …rm’s physical proximity are also
good indicators of the information’s quality (see, e.g., Agarwal and Hauswald, 2006). The presence
of established business relationships unsurprisingly enhances the e¤ect of private information on
inside lending but has a much smaller and often insigni…cant e¤ect on arm’s-length transactions.
Our main contribution consists in carefully identifying, measuring, and analyzing the di¤erential
information content of transactional and relationship debt on the basis of a large sample of credit
transactions in a uni…ed framework. Given the chosen mode of bank-borrower interaction we
establish that the extent to which informational considerations shape the choice of debt product
critically depends on the bank’s ability to generate private information and bene…t from it. An
additional contribution consists in showing that borrowers also learn about their bank’s policies
and, in particular, its strategic use of private intelligence that determines the di¤erential response
of …rms to banks’ information-acquisition and lending strategies across debt types. Finally, our
results highlight how technological progress in the form of online banking and credit scoring allows
intermediaries to simultaneously engage in transactional and relationship lending, thereby helping
them to overcome organizational limitations that in the past led to specialization by market segment
or bank size (Berger et al., 2005).
To the best of our knowledge, there is no comparative work on the di¤erential e¤ects of private
and public information by loan type. While Petersen and Rajan (1994), Berger and Udell (1995),
Degryse and Van Cayseele (2000), Elsas (2005), and Schenone (2007) have analyzed the importance
of relationship banking for the collection of inside information they do not consider the respective
use of public and private credit-quality signals across lending modes, which is central to our analysis.
An exception are Bharath et al. (2006) who also …nd that information asymmetries induce borrowers
3

to self-select into lending relationships but who do not consider transactional lending. Focusing on
the bene…ts of relationship lending to borrowers Boot and Thakor (2000) argue that the resulting
close business ties allow banks to fend o¤ competition from other lenders and transactional debt,
which is consistent with our data. Boot (2000) and Boot and Schmeits (2005) o¤er excellent surveys
of recent theoretical and empirical work on relationship banking.
The paper also contributes to the nascent literature on the e¤ect of the internet on …nancial
intermediation. Wilhelm (1999, 2001), who analyzes the impact of the internet on the structure of
banking markets and, especially, relationship banking, argues that technological advances change
the collection and use of (private) information through its codi…cation which is at the heart of
our analysis. Similarly, Petersen (2004) discusses how technology a¤ects the nature of the bankborrower interaction and, hence, the operations of …nancial markets and institutions. Anand and
Galetovic (2006) o¤er empirical predictions on the internet’s e¤ect on …rm-bank relationships in
terms of a shift toward non-relationship modes of interaction, which is only partly borne out by
our results. Bonaccorsi di Patti et al. (2004) investigate demand complementarities between
traditional and online provision of banking services and report that e-banking leads to a reduction
in per-customer pro…tability which mirrors our …ndings on the competitive pricing of transactional
debt. Regarding the importance of online banking Fuentes et al. (2006) study the determinants
of the decision of U.S. banks to create a transactional website for their customers while DeYoung
(2005) investigates the scale economies present in internet banking.
The paper is organized as follows. In the next section we review the theoretical literature on
transactional and relationship debt and distill pertinent empirical predictions. Section 3 describes
our data and estimation strategy. In Sections 4 and 5, we analyze the …rm’s choice of arm’s-length
vs. inside debt and the bank’s decision to o¤er credit and at what price across lending channels.
Section 6 investigates the determinants of the borrower’s decision to reject the banks’ loan o¤er
and obtain credit from a competitor. In Section 7 we report our …ndings on credit default across
loan types. The last section discusses further implications and concludes. We relegate all tables to
the Appendix.

4

2

Transactional and Relationship Lending

The theoretical literature has typically argued that relationship lending o¤ers particular economic
bene…ts to at least one party, if not both, through the closer ties that banks and borrowers forge.
Lending relationships allow intermediaries to gain proprietary information (Rajan, 1992 and Petersen and Rajan, 1994), facilitate renegotiation through the implicit nature of the debt contract
(e.g., Sharpe, 1990), give rise to intertemporal transfers (e.g., Petersen and Rajan, 1995), and allow
borrowers to learn about their bank’s attributes (Iyer and Puri, 2007).1 In fact, the ability to gather
proprietary information (Bhattacharya and Chiesa, 1995) and use it strategically in credit-market
competition has become the de…ning attribute of relationship debt. By contrast, lenders compete
on a more equal informational footing for transactional loans, competing away potential rents but
at the price of less readily available credit (Broecker, 1990 or Hauswald and Marquez, 2003). Hence,
…rms face a trade-o¤ between the availability and pricing of credit across the two lending modes:
informational capture with rent extraction but more ‡exibility in …nancing choices or less readily
available credit at lower rates.
Relationship banking allows lenders to strategically acquire proprietary information and to
create a threat of adverse selection for their rivals, thereby softening price competition. For instance,
Petersen and Rajan (2002) argue that local banks who collect “soft” proprietary information on
small …rms over time have an informational advantage over more remote competitors who might
not enjoy the same degree of access to local information.2 Several empirical predictions follow.
Given a …rm’s credit quality relationship lending facilitates the access to credit and intertemporal
insurance but at the cost of rent extraction. Hence, the more and better proprietary information a
bank has, the more willing it should be to approve loan applications but also the higher the quoted
interest rate will be conditional on the applicant’s credit quality (see, e.g., von Thadden, 2004).
By contrast, symmetrically informed transactional lenders should charge less and be less willing to
grant credit to applicants of comparable credit quality (see, e.g., Broecker, 1990).
By the same token, competition a¤ects each lending channel di¤erently. In purely transactional
credit markets symmetrically informed lenders bid less aggressively because more competition wors1

For a recent survey on relationship banking see Boot (2000).
Agarwal and Hauswald (2006) provide strong evidence for this conjecture. See also Berger, Frame and Miller
(2005) on the role of soft information in lending decisions and the ability of smaller banks that presumably have a
more local focus to collect and process such intelligence.
2

5

ens their inference problem so that credit becomes less available and interest rates rise (Broecker,
1990). By contrast, when relationship and transactional lending directly compete with each other,
e.g., a better informed inside bank against less informed arm’s-length lenders, competition reduces
the incentives for information acquisition so that interest rates should fall in both segments and
credit availability rises because less informed transactional lenders face a diminished threat of adverse selection (Hauswald and Marquez, 2006).
A subtle di¤erence in the adverse-selection problem that lenders face for each loan type is also
behind the respective empirical predictions for borrower switching. In purely transactional credit
markets, banks face symmetric adverse-selection threats so that ceteris paribus they can compete
more aggressively for transactional borrowers who should be more likely to switch. However,
when transactional lenders compete against a better informed inside bank, the greater the latter’s
informational advantage, the greater the threat of adverse selection. As a result, less informed
competitors bid less aggressively (higher interest rates and less frequently) so that relationship
borrowers are less likely to switch providers of credit. Hence, we expect less borrower switching
in relationship lending, the greater the informational advantage of the inside bank is, or the less
competitive a local credit market is. At the same time, better credit risks, which are the primary
targets for rent extraction, should actively respond to such attempts by seeking loans elsewhere so
that publicly observable signals of higher credit quality should induce more lender switching even
by inside borrowers.
Finally, the more private information a lenders has the less likely errors in granting credit
should become. Hence, a bank should experience less credit delinquency in relationship than in
arm’s-length lending. Also, the greater the competition the greater (smaller) adverse-selection
problems become in transactional (relationship) lending so that competition should increase the
incidence of default in transactional loan markets and decrease it in relationship debt.
From an empirical perspective, the de…ning features of transactional and relationship debt then
revolve around the generation and strategic use of proprietary information, di¤erential availability
and pricing of credit, and the resulting competitive reaction as revealed by lender switching across
loan types. While the length and scope of a prior business relationship is thought to reveal the
existence of a lending relationship no such clear-cut identi…er has existed for transactional debt in
the past. However, the advent of online lending to small businesses without any personal interaction
6

between the parties allows us to unambiguously identify purely transactional loans. At the same
time, lenders often engage in extensive information acquisition through their branch o¢ ces so that
in-person applications and the resulting interaction with local loan o¢ cers de…ne relationship debt.

3

Data Description and Methodology

Our sample consists of all online and in-person applications for new loans over a 15-months span
by small …rms and sole proprietorships to a large US …nancial institution with a particular regional
focus on New England, the Mid-Atlantic, and Florida. During the sample period, this lender ranked
among the top …ve commercial banks and savings institutions according to the FDIC. Since our
bank more or less automatically rolls over prior loans on request unless a signi…cant deterioration in
creditworthiness has occurred very di¤erent considerations drive the decision to grant credit from
the one renewing an existing loan. As a result, most information production takes place around
the origination of a new loan, explaining our sample selection. All loan applications fall under the
de…nition of small- and medium-sized enterprise lending in the Basel I Accord so that the total
obligation of the applying …rm is less than $1 million and its sales are below $10 million.
We focus on small-business lending because borrowers exhibit just the right degree of informational opacity for our purposes and credit products in this market are typically close substitutes.
On the one hand, …rms are su¢ ciently opaque for proprietary information to matter in lending
decisions. On the other hand, small businesses are also quite homogeneous so that bank competition is intense, several lending channels coexist, and third parties provide credit-scoring services
that we can use to measure the contribution of our bank’s own proprietary loan screening to credit
decisions.3

3.1

Operational Policies

The small-business loans originate both from personal visits to branch networks and from websites
without any personal interaction so that we can clearly identify whether credit is granted on an
arm’s-length or relationship basis. In case of an in-person application, the …rm’s representative
3

Since our data provider applies a uniform credit-scoring methodology to all loan requests the internal credit
score is a consistent and meaningful measure of the bank’s proprietary information across applicants, branches, and
distribution channels.

7

(e.g., owner/manager) personally visits one of the 1,408 branch o¢ ces in our sample (out of a
total of 1,552)4 to supply all the relevant information, submit …nancial statements and tax data,
provide a list of assets, etc. The local loan o¢ cer transcribes this information into electronic form
and matches it with credit reports for input into the bank’s proprietary credit-scoring model. The
whole lending process including the credit decision typically takes four hours to a day from the
initial meeting between applicant and loan o¢ cer.
The loan o¢ cer also uses the branch visit to conduct an in-depth interview with the applicant
to gather “soft”information in the sense that it would be hard to verify by a third party. In up to
8% of the cases, the branch will invite the applicant back to follow up on open questions, review
discrepancies in submitted information with credit reports, discuss the prospects of the …rm, etc.
Such information allows the branch manager or account o¢ cer to subjectively adjust the …rm’s
internal score should the applicant deserve credit in their eyes but fail to meet certain commercial,
pro…tability, liquidity, or credit-score requirements. These subjective score revisions represent the
soft-information component of the bank’s internal credit assessment that forms the basis of our
analysis.
Each branch o¢ ce enjoys a considerable amount of autonomy in the assessment, approval, and
pricing of loans but has to justify any deviation from bank-wide practices. As a consequence, credit
decisions ultimately reside with branches because local managers can alter credit scores on the basis
of a standard set of subjective criteria that the …nal score re‡ects. Similarly, they can adapt loan
terms including pricing to the speci…c circumstances of the application. However, branch managers’
career prospects and remuneration depend on the overall success of their credit decisions, and local
“overrides” are closely monitored by the bank’s overall risk management.
In case of online applications, the applicant submits all the requisite information through a
website. The online processing center then requests credit reports to cross-check the information
and computes the …rm’s credit score very much like a branch o¢ ce but does not attempt to resolve
any informational discrepancies. As a matter of operational policy, there is no personal interaction
between the bank and an online applicant so that our lender makes online-credit decision purely
based on its internal credit score, which is not subject to any revisions and computed on the basis of
4

For comparability, the 100 institutions with more than $10 billion in assets in 2002 operated, on average, 364
branch o¢ ces. Their average amount of deposits is about a quarter of our data provider’s deposit base.

8

…rm-supplied information, credit reports, and, possibly, prior interaction. Similarly, any loan terms,
especially interest rates, are solely a function of the …rm’s credit score, its ability to post collateral,
third-party guarantees, etc. As a result, both credit o¤ers and their terms are highly automated in
the online market, closely corresponding to the de…nition of transactional debt because the lender
does not gather additional intelligence beyond publicly available information.
Most monitoring is automated for both loan types and takes place through the daily tracking
of current-account movements or balances5 (whenever available) and prompt debt service. On a
monthly basis, the bank collects new credit reports for the …rm and its owner and updates the account’s risk pro…le. Yearly credit reviews and the treatment of overdue loans, however, di¤erentiate
ongoing information production across lending channels. On each anniversary of the loan’s origination, transactional borrowers submit updated …nancial information online. Relationship borrowers
have to do so in-person at their branch o¢ ce, which uses the visit to discuss the …rm’s prospects,
state of solvency, funding needs, etc. Similarly, if a payment is between 10 and 20 days late on
a relationship loan the account o¢ cer will personally visit the …rm. If the account becomes more
than 20 days overdue, the bank cuts back credit lines to the current balance, i.e., reduces its credit
commitment, but will not take such action on term loans before 60 days past-due.
Although the lending standards are identical across online and in-branch origination the resulting transactions di¤er in their information content because loan o¢ cers and branch managers can
personally revise applicants’credit scores on the basis of subjective impressions. At the same time,
the two lending channels e¤ectively compete within the bank because branches have no incentive
to encourage in-person applicants to also apply online. As a result, the observed loan type allows us to cleanly sort credit applications into transactional or relationship debt with the required
informational attributes.

3.2

Data Description

The sample consists of all applications for new loans to our bank that conform to the Basel I Accord’s
SME lending de…nition between January 2002 to April 2003 (36,723 observations). We match
these records with credit-bureau reports (Experian and Dunn & Bradstreet) on the application
5
Mester et al. (2007) …nd that current-account transactions provide valuable information for loan monitoring in
a setting similar to ours.

9

date to verify the supplied information and delete applications with missing data (e.g., Experian
credit score) or other informational discrepancies such as nonexisting addresses. Our data provider
also engaged in several M&A transactions a¤ecting its branch network so that we also omit all
re-assigned loan records. Overall, we lose 2,868 credit requests leaving a total of 7,945 online
applications and 25,910 in-person ones. Table 1 summarizes our data as a function of the applicant’s
chosen form of interaction with the bank and reports the P -values of t-tests for the each variable’s
mean conditional on the lending channel.6
To analyze informational e¤ects in transactional and relationship lending we rely on the outcome of the bank’s own borrower assessment in terms of the internal credit score calculated for each
loan application. While the methodology is proprietary and subject to con…dentiality restrictions,
the credit-screening procedure is consistent across all branches, lending channels, and applications
because it uses a common set of inputs and the same statistical model. For in-person applications, our bank’s credit scores comprise a subjective element because local branches provide “soft
information” through individual adjustments that can over-ride automated lending decision and
centralized loan pricing. From periodic surveys of loan o¢ cers the data provider estimates that
20% to 30% of the in-person score ultimately consists of subjective (soft) information. We use
the …nal scores whose revisions follow bank-wide guidelines and require detailed justi…cation by
branches. Internal scores for online applications are not subject to revision and therefore comprise
at most hard, i.e., independently veri…able, proprietary information.
Internal scores range from 0 (worst) to 1,850 (best). Their means (medians) are 899 (902) for
online applicants and 930 (949) for in-person ones, and the di¤erence is signi…cant at the 1% level
(P -value of 0.00%). We also collect the applicant’s Experian Small Business Intelliscores (XSBI),
which this leading credit bureau provides together with its report services, as a publicly available
signal for each …rm’s creditworthiness. We reverse the Experian scores, which measure the likelihood
of “serious delinquency”over the next 12 months, and linearly rescale them for comparability with
the better known (retail) FICO scores so that the XSBI variable ranges from 300 (worst) to 850
(best). Contrary to the internal score, the average (median) of online applicants’Experian scores
is statistically signi…cantly higher: 723 (704) against 716 (705) for in-person applicants (P -value
6

For con…dentiality reasons, the data provider did not allow us to report further descriptive statistics because they
could be used to “reverse-engineer” the composition of the loan portfolio.

10

of 0.00%).7 This discrepancy in scores across loan types stems from the subjective revisions to
internal credit assessments for in-person applicants. It highlights not only the informational value
of relationship lending but also shows how banks incorporate subjective information such as personal
impressions of borrower quality into credit decisions.
We assess the nature of the lending relationship, which facilitates the collection of such borrowerspeci…c information, along two dimensions.8 Our …rst variable is the number of months that a
particular …rm has been on the books of the bank, which measures the length of the lending
relationship. We see that in our sample online applicants have, on average, obtained a …rst credit
product 27.7 months prior to the loan application whereas in-person applicants have been borrowers
for 30.8 months. The second variable measures the breadth of the business relationship. To this
end we de…ne a binary variable Scope in terms of the balance of the …rm’s current account (at
least $5,000) together with prior borrowing and the purchase of at least one other banking product
(Scope: about 20% of online against 30% of in-person applications).
To control for the availability of public information and …rm-speci…c attributes we rely on the
months a particular applicant has been in business (64 vs. 103 months for online and in-person
applications, respectively), which is a good proxy for informational transparency, and the …rm’s
monthly net income ($64,734 vs. $101,109 for online and in-person applications, respectively) that
captures size and pro…tability e¤ects. We also use 38 industry dummy variables based on the
applicants’ two-digit SIC codes to account for any industry e¤ects in the data. Table 1 shows
that our sample represents a wide cross-section of industries, albeit with a particular emphasis on
wholesale and retail trade, personal, business and professional services, and construction. Similarly,
we rely on state and quarter dummy variables to account for regional and business-cycle e¤ects.
To measure the competitiveness of local credit markets we collected the number of bank branches
and active lenders in a …rm’s zip code from the FDIC’s Summary of Deposits data base by year.
Concentration measures such as the Her…ndahl-Hirschman Index of deposits or branch shares by
…rm ZIP code are not statistically signi…cant in our speci…cations so that we do not tabulate their
sample statistics or estimation results.
7
The US mean (median) for comparable consumer FICO scores is currently 678 (723). See Experian (2000, 2006)
for further details on the SBI and its ability to forecast credit delinquency.
8
James (1987), Lummer and McConnell (1989), and Elsas (2005) present evidence suggesting that banks gain
access to private information over the course of the lending relationship.

11

In terms of loan characteristics our data contains the requested loan amount (mean of $37,333
and $46,877 for online and in-person applications, respectively, in line with typical small business lending), its maturity (mean: 5.43 and 6.74 years, respectively), and existence of collateral
(about 42% for online against 55% for in-person applications). About 17% (37%) of online (inperson) credit requests were personally guaranteed by guarantors with a monthly income of $23,745
($35,164). 19.6% (28%) of online (in-person) applications are for term loans, the remainder is for
credit lines. As a matter of business policy, our bank only o¤ers term loans at …xed rates and
credit lines at variable rates so that our Term Loan (vs. credit line) binary variable also captures
the nature of the interest rate. Finally, 3.74% of online against 6.41% of in-person applications fall
under the terms of the Small-Business Administration (SBA) guarantee program.
To control for the ease and cost of personally transacting with the bank in terms of time and
e¤ort we use the driving distance in miles between each …rm and their branch o¢ ce for in-person
applications or, for consistency, the processing center for online request, as well as the distance to
the closest full-service branch of a competitor.9 We see that relationship borrowers are on average
located 10.3 (median: 2.8) miles away from their bank branch whereas transactional applicants are
91.7 (median: 31.9) miles away from the bank’s online-loan processing center. By contrast, both
transactional and relationship applicants are about 1 mile on average (median: 0.5 miles) from the
nearest full-service branch of a competing lender.
Since banks and their customers might choose to locate in certain areas based on local economic
conditions, we include the Case-Shiller Home Price Index (CSHPI: see Case and Shiller, 1987,
1989) to account for potential endogeneities in the parties’choice of location and lending channel.
By matching each loan application with the index by zip code and month we also capture loantransaction e¤ects that are due to the local level of economic activity, di¤erences in a- uence across
postal zones, and di¤erential levels of urbanization or road infrastructure as re‡ected in local house
prices.
We see that, contrary to common perceptions, transactional applicants are typically younger and
9

See Degryse and Ongena (2005) on the importance of transportation costs in credit markets. We rely on Yahoo!SmartView and Yahoo!Maps to identify the nearest competitor for all loan applicants and to determine the
driving distances between the …rm, the bank branch for personal applications or the processing center for online ones,
and the competitor’s branch. SmartView has the dual advantage that it does not accept sponsored links and draws on
the combined yellow-page directories of BellSouth and InfoUSA (Mara, 2004) providing objective and comprehensive
bank-branch information.

12

smaller …rms that request smaller loan amounts, o¤er less collateral and personal guarantees, and
are more creditworthy according to publicly available information (XSBI). However, they are less
likely to have a prior business relationship with the bank and, if so, it is shorter than for in-person
applications. As a result, the bank’s internal score as a proprietary measure of credit quality is
higher for relationship borrowers, presumably through subjective revisions that incorporate private
local information into the credit decision.

3.3

Methodology

Our estimation strategy simply retraces the steps of the loan-origination process starting with
a discrete-choice model of the …rm’s choice of loan type as a function of publicly available and
proprietary information, characteristics of the lending relationship, …rm attributes, and our control
variables. We next investigate the bank’s credit decision by estimating a logistic model of its
decision to o¤er credit by lending channel and, if so, at what price. To this end we specify a linear
model of the o¤ered annual-percentage rate (APR: the all-in cost of credit taking into account fees
and commissions) as a function of the same variables once again taking into account the debt type.
Successful loan applicants typically move next by accepting or declining loan o¤ers. Hence, we
explore the di¤erential e¤ect of private and public information across debt type on bank competition
as revealed by an applicant’s decision to switch lenders. Lastly, the respective informational and
competitive dynamics of each lending mode hold di¤erent implications for type II errors in credit
screens and, hence, default across loan types. We therefore estimate the likelihood of borrower
delinquency by lending channel to assess the incidence of debt type on the quality of the bank’s
public and private information in terms of loan performance.
For every decision in the lending process, we specify logistic discrete-choice models with separate
equations for each lending channel so that we can compare informational e¤ects across debt types
and directly test empirical predictions in a uni…ed econometric framework. For instance, we estimate
the likelihood of a loan o¤er Yi = 1 as

E [Yi jxi ] = E [(1
where

(x0i ) =

expfx0i

1+expfx0i

1eloan ) Yi + 1eloan Yi jxi ] = Pr fYi = 1 jxi g =

x0i +1eloan x0i

(1)

g
is the logistic distribution function. The binary variable 1eloan ; which
g
13

takes the value 1 for online applications and 0 otherwise, allows us to report results by debt type
because we have
h

i
E Y^i jxi =

x0i ^ +1eloan

x0i ^

=

8
>
<
>
:

x0i ^ + ^

for transactional debt (1eloan = 1)

x0i ^

for relationship debt (1eloan = 0)

Similarly, we specify the following linear-regression model of the o¤ered loan’s all-in cost (APR) ri :
ri = x0i +1eloan x0i + "i

(2)

We focus on the following key variables in our investigation of the di¤erential information production in transactional and relationship lending: each …rm’s Experian Small Business Intelliscore
(XSBI) as a measure of publicly available information, its internal credit score as a measure of the
lender’s proprietary information, the scope and months-on-book variables measuring the depth of
the lending relationship, and a measure of soft private information. To extract this purely private
component of credit screens we orthogonalize the internal and Experian scores because the former
relies on a mix of public and private intelligence as inputs into the proprietary scoring model.
Speci…cally, we estimate the bank’s private credit assessment as the residual u
^i of the regression

ln (IntScorei ) =

0

+

1

XSBIi + 1eloan (

0

+

1

XSBIi ) + ui

(3)

which we label the Private-Information Residual (PIR). Incidentally, the R2 of the above regression
are 0.67 and 0.71 for the online and in-person equations, respectively, which con…rms our data
provider’s contention that up to 30% of the internal score is based on soft, subjective information.10
The Private-Information Residual u
^i represents a clean measure of our data provider’s soft
private information whenever it exists. Given its construction, the online PIR captures hard private
intelligence only to the degree that it exists for eLoans through repeat business, veri…cation of
self-reported information with credit reports, and the lender’s proprietary scoring methodology.
In addition to such hard private information, the in-person PIR also comprises a soft subjective
component stemming from the loan o¢ cer’s personal impressions of borrower quality incorporated
10

For con…dentiality reasons we cannot provide further details on the orthogonalization nor report any results. The
log-linear speci…cation best agrees with the nonlinear nature of Experian’s Small Business Intelliscore.

14

into the internal score through the interview, follow-up, and revision process. Since we compare
the PIR across two equations in the same speci…cation the transactional eLoans become the de
facto benchmark which we use to measure the additional and, hence, soft information content of
in-person credit applications. Note, however, that we can also interpret the residual u
^i as a proxy
for the bank’s informational advantage over publicly available information regardless of debt type.
To control for systematic e¤ects in self-selection and approval practices across branches and
lending channels we estimate all our speci…cations including the internal-score orthogonalization
with branch …xed e¤ects and rely on clustered standard errors that are adjusted for heteroskedasticity across bank branches and autocorrelation within o¢ ces including the online-loan processing
center. The estimation of all discrete-choice models proceeds by full-information maximum likelihood; we report their pseudo R2 which is simply McFadden’s likelihood ratio index whenever
appropriate.
It is worthwhile to point out that the unique nature of our data set allows us to sidestep pervasive
endogeneity problems that arise in the study of the credit terms when the sample only consists of
booked loans (see, e.g., Berger et al., 2005). Since our data comprise all applications and loan o¤ers
potential borrowers have not chosen yet whether to accept or to refuse the lender’s terms. The
omission of declined loan o¤ers could give rise to the joint endogeneity of borrower characteristics,
bank attributes, and loan terms, which we avoid through sample selection by including the 1,335
ultimately declined o¤ers in this part of the analysis. Since several of the variables …t better in
logarithms than levels we use the former whenever appropriate.

4

The Choice between Arm’s-Length and Relationship Debt

Speci…cation 1 in Table 2 reveals that public credit-quality perception is by far the most important
criterion in a …rm’s choice of loan type. Applicants, who presumably have a good sense of their own
creditworthiness, are the more likely to choose arm’s-length debt the higher their public credit score
is: a 10% increase in the …rm’s Experian score raises the likelihood of applying online by 2.15%.
The second important determinant is Months on Books. The longer a …rm has been a borrower
at our lender the more likely it is to apply in-person for a relationship loan. We also see that,
contrary to widespread perceptions, the …rm’s size, pro…tability, age, and ability to post collateral

15

do not seem to enter into the applicant’s choice of loan type: Net Income, Months in Business, and
Collateral are all statistically insigni…cant.
However, lending relationships allow information to ‡ow in both directions so that borrowers
typically learn about their bank’s operational policies, too (e.g., Iyer and Puri, 2007). Furthermore,
verbal communication between the loan o¢ cer and …rm representative during the origination interview often reveals bank-internal information to applicants who use such knowledge in their own
decision making. Hence, we would expect the lender’s proprietary and private information as measured by the Internal Score and PIR to be correlated with borrower perceptions of bank-internal
credit assessments. To capture this facet of lending relationships we successively add these two
variables to the speci…cation. We see that the inclusion of the Internal Score dramatically reduces
the marginal e¤ect of the public score lending support to our contention that the former can also
serve as a proxy for borrower impressions of their bank’s credit-quality signal, which increases their
likelihood of applying online (Speci…cation 2, Table 2). However, in terms of economic signi…cance
the marginal e¤ect of public information, i.e., the XSBI score, is almost four times that of the
Internal Score.
To the extent that loan o¢ cers communicate their subjective impressions not only to their own
institution (through score revisions) but also to customers (during the interview), applicants might
also become aware of the bank’s private information. Hence, we next replace the Internal Score
with its orthogonalization in terms of the XSBI, the Private-Information Residual (PIR). Comparing Speci…cations 2 and 3 in Table 2 we see that the distinction between proprietary (Internal
Score) and private (PIR) information is crucial. Only when we properly measure the latter as the
former’s orthogonal complement to public information do we …nd the predicted sign pattern so
that public signals of high credit quality are associated with transactional debt and private signals
with relationship lending. To preclude any possibility of spurious correlation between the PIR and
dependent variable arising from our two-equation estimation (3) we reestimate the speci…cation
with the residuals from a pooled orthogonalization but do not report the results because they are
virtually identical.
The two overriding factors for the …rm’s choice of debt type are now the public credit-quality
signal, whose marginal e¤ect is almost unchanged from the previous estimation, and our privateinformation measure PIR ( Speci…cation 3, Table 2). Not only are their marginal e¤ects of com16

parable magnitude but their opposite signs also conform to perceived notions of the di¤erential information content present in transactional and relationship lending. A better public credit-quality
signal makes the …rm more likely to apply online for a transactional loan because applicants that
are presumably aware of their own credit risk know that a higher public score improves their access
to (cheaper) arm’s-length debt and act accordingly.
Conversely, a …rm with a longstanding banking relationship might be able to infer its lenders’s
credit-quality assessment if only because of the signalling value of repeated loan o¤ers. It can
count on being well regarded and, hence, on preferential treatment by its bankers, who, in turn,
gain better access to inside information. As a result, we would expect the …rm’s application
loan-type decision and the bank’s private credit-quality signal to be correlated. The PIR’s large
negative marginal e¤ect in Speci…cation 3 of Table 2 bears out this conjecture. The better the
private credit-quality assessment, the less likely the …rm will request a transactional loan and
instead apply for relationship credit in-person at a branch o¢ ce. Since the PIR also measures the
inside bank’s informational advantage vis-à-vis competitors this …nding suggest that despite the
danger of informational capture better private information actually increases a …rm’s likelihood of
choosing relationship debt through the promise of future bene…ts such preferential access to credit
or intertemporal transfers.
To further investigate this hypothesis we next add interaction terms between the PIR and
relationship variables to capture the potential for collecting private information and the borrower’s
awareness of such e¤orts (Speci…cation 4, Table 2). Both the PIR-Months-on-Books and PIRScope e¤ects further support our interpretation that despite the danger of informational capture
borrowers well known to their bank seek relationship debt precisely because loan o¢ cers can better
communicate their (high) opinion of good credit risks to those customers during negotiations. The
longer (Months on Books) or broader (Scope) the parties’interaction the more likely the …rm will
choose relationship debt and the more important the existence of private information becomes for
this choice of loan type.
The fact that both the lender’s informational advantage and prior borrowing strongly increase
the probability of a relationship-loan request provides additional support for our conjecture that
…rms not only are aware of their lender’s information but also bene…t from special ties to their bank.
Firms know that longstanding business relationships facilitate the access to credit precisely because
17

loan o¢ cers tend to have a better picture of their prospects. Exposed to the danger of informational
capture by their bank, applicants of high perceived credit quality might as well bene…t from more
readily available credit that inside debt typically o¤ers in such circumstances, a topic that we turn
next to.

5

Credit Decision by Lending Channel

In this section, we analyze the availability and pricing of credit by origination mode to determine
the di¤erential information content of arm’s-length and relationship debt. Table 3 reports summary
statistics for the key variables by credit decision and lending channel, in particular loan terms and
pricing. Two facts consistent with the theoretical predictions on debt type stand out: rejection rates
are much higher for online applications (about 61% as compared to 49% for in-person requests), and
credit spreads are on average much lower for transactional than for relationship loans (279 and 453
basis points, respectively). Credit appears to be much less readily available through transactional
channels but, when it is, loan rates are much more favorable.

5.1

Credit Availability

The results for the bank’s decision to grant credit show that transactional debt is much harder to
obtain than relationship debt ceteris paribus. Both speci…cations in Table 4 reveal that applying
online lowers the probability of a loan o¤er by up to 11.2%. Transactional lenders know that they
compete on a much more level informational playing …eld in this segment, if not at an outright
disadvantage should the …rm also be seeking inside credit elsewhere. To avoid potential adverseselection problems they have to be much more circumspect in their arm’s-length lending and refrain
from o¤ering credit more often, thereby lowering the probability of an online loan o¤er (see, e.g.,
Broecker, 1990 or von Thadden, 2004).
Speci…cation 1 in Table 4 shows that the likelihood of obtaining transactional credit increases in
both the public and proprietary credit-quality signal (XSBI and Internal Score, respectively): the
better the outcome of the credit screen, be it public or bank-internal, the easier access to online loans
becomes. However, an increase in the Internal Score has only a small, albeit statistically highly
signi…cant, impact on the likelihood of obtaining transactional credit. By contrast, the Experian

18

score (XBSI) is not statistically signi…cant in the relationship-loan equation. Instead, positive
proprietary credit assessments containing a mixture of soft private and hard public information
primarily decide the access to inside credit. This …nding suggests that not only the origin of the
bank’s information but also how it processes and interprets its intelligence matters for relationship
lending.
To carefully distinguish private from public information we again replace the Internal Score
with its Private-Information Residual (PIR) and add the relationship-PIR interaction terms to the
model (Speci…cation 2 in Table 4). Our results con…rm that di¤erent types of information shape
each credit-market segment. Although both the PIR and Experian score are statistically signi…cant
in each equation, the relative magnitudes of the variable’s marginal e¤ects are reversed across loan
types. Transactional-credit decisions primarily rely on public information (XSBI score) whereas
private information (PIR) only has a small impact; in fact, the marginal e¤ect of a positive public
credit signal is almost 8 times larger than that of a positive private credit-assessment. By contrast,
private information is the overriding factor in the decision to o¤er relationship credit because its
marginal e¤ect is almost …ve times larger than the small positive impact of public information.
Comparing the relative impact of public and private information on credit availability across
loan types we see that the marginal e¤ect of positive private information is 15 times greater for
relationship than for transactional lending. Interestingly, the importance of a high public credit
score does not di¤er as much across the two lending modes (only 5.5 times lower) and retains its
statistical signi…cance at 5% in the relationship-loan equation (Speci…cation 2). In light of the fact
that lenders and loan types compete with each other this …nding is less surprising than it might
otherwise be. The theoretical literature has long argued that good credit risks are the primary
targets for informational capture in relationship lending (e.g., von Thadden, 2004) and, therefore,
more likely to switch providers of credit. Hence, banks know that public perceptions of credit
quality matter in the competitive response of other lenders that try to poach borrowers. As a
result, the Experian score not only captures credit-quality e¤ects but also acts as a proxy for the
expected intensity of competition for the borrower.
We conclude from both speci…cations in Table 4 that, consistent with theoretical predictions,
private information primarily determines access to inside debt whereas public information drives
arm’s-length lending. Banks speci…cally gather more costly private information for borrowers that
19

through their chosen mode of interaction with the lender facilitate its collection and signal their
willingness to be informationally captured. The di¤erential impact of the length and scope of the
banking relationship across loan types con…rms this interpretation. Scope and Months on Books are
statistically insigni…cant in the decision to o¤er arm’s-length credit but highly signi…cant both in
statistical and marginal terms for relationship-loan o¤ers. Taken together these e¤ects suggest that
a prior lending relationship enhances the likelihood of obtaining inside credit precisely because they
facilitate the collection and interpretation of (private) information. By contrast, prior interaction is
less relevant for the decision to grant transactional loans because there is no opportunity to revise
online applicants’scores in light of additional information.
Similarly, we see that the …rm-bank distance is only statistically signi…cant (at around 5%) in
the in-person-loan equation. The closer a potential relationship borrower is to a branch o¢ ce the
higher the likelihood of obtaining credit becomes. In addition to capturing physical transaction
costs, the bank-borrower distance is an excellent proxy for the quality of the lender’s private information and, hence, informational advantage (see Agarwal and Hauswald, 2006). Petersen and
Rajan (2002) argue that soft subjective information, whose collection borrower proximity and prior
lending relationships facilitate, is crucial for lending decision. No such opportunity to collect soft
information and incorporate it into credit decisions exists in the case of transactional loans, which
might explain the statistical insigni…cance of the relevant variables in the eLoan equation.
A comparison of the two speci…cations in Table 4 shows that all the other e¤ects remain virtually
una¤ected by the inclusion of the Private-Information Residual. The …rm’s size or pro…tability (Net
Income) and its ability to post collateral or to guarantee the loan raises the likelihood of a loan o¤er
for each lending channel and the marginal e¤ects are very comparable. The local-competitiveness
e¤ects closely correspond to theoretical predictions. More competition, i.e., a higher number of
competing lenders or branches in the …rm’s zip code, decreases the likelihood of obtaining a loan
of either type because competition decreases the average quality of the applicant pool (see, e.g.,
Broecker, 1990) so that banks refrain more often from o¤ering credit.
Our …ndings suggest that the use and, hence, quality of proprietary intelligence radically differs across lending channels. The limited ability to gather inside information or, equivalently, its
high cost in transactional lending forces banks to discount any private knowledge and instead to
rely on publicly available signals of credit quality. As a result, banks compete on a much more
20

equal informational footing, which borrowers recognize and incorporate into their choice of loan
product. By contrast, banks heavily rely on private information gathered through inside lending
in relationship-credit decisions. Although lenders can use their informational advantage to soften
competition through the threat of adverse selection and to extract information rents it also facilitates relationship borrowers’ access to credit (see, e.g., Hauswald and Marquez, 2006). By the
same token, our credit-decision results validate the …rm’s perception of the importance of personal
interaction and private information for obtaining relationship loans discussed in Section 4.

5.2

Loan Pricing

To investigate di¤erential credit pricing across lending channels we next estimate linear models of
the loan’s o¤ered all-in cost (APR) as a function of our previously described explanatory variables.
Like the internal score of in-person applicants, branches can adjust both the loan terms and pricing
in light of local conditions and information. No such adjustment opportunity exists for eLoans whose
price is a simple function of the internal score, the ability to post collateral or personally guarantee
the loan, etc. Table 3 provides descriptive statistics for the o¤ered loan terms by credit channel.
To control for the interest-rate environment, we rely on the maturity-matched (interpolated) US
Treasury yield on the loan date and the di¤erence between the 5-year and 3-months US Treasury
yield (Term Spread: yield-curve shape). We estimate the model with the Heckman correction for
sample-selection bias (Lambda) to take into account the lender’s prior credit decision.
Table 5 shows that arm’s-length debt is up to 138 basis points less expensive than inside debt
ceteris paribus. Speci…cation 1 summarizes the e¤ects of relationship variables, …rm attributes, loan
terms, and various controls on o¤ered loan rates. Adding the informational variables (Speci…cations
2 and 3), we observe the same relative importance of public, proprietary, and private information
in the determination of o¤ered loan rates across lending channels that we found for the prior credit
decision. Even with our measure of proprietary information Speci…cation 2 in Table 5 shows that
the impact of the public (XSBI) and internal score on the quoted all-in cost symmetrically varies
across lending channels. An increase in the Experian score (XSBI) greatly reduces transactional
loan rates whereas bank perceptions of higher credit quality (Internal Score) lead to a much more
modest reduction in rates. The exact opposite is true for relationship loans whose price is much
more a¤ected by a rise in the Internal Score than in the XSBI one. These e¤ects are all the more
21

pronounced that the Experian score is highly nonlinear in implied credit quality.
Replacing the bank’s credit score with the Private-Information Residual reinforces this conclusion (Speci…cation 3, Table 5). Our measure of private credit assessments now becomes statistically insigni…cant in the eLoan equation but retains its high statistical signi…cance in the in-person
equation. The same is true for the relationship-PIR interaction variables that increase the privateinformation e¤ect for relationship loans but are statistically insigni…cant in the transactional-loan
equation. Any pure private information the bank can gather is mostly valuable in inside lending to
limit competition and informationally capture relationship borrowers. Its poorer quality for online
borrowers does not o¤er any signi…cant bene…ts over publicly available creditworthiness signals.
Hence, our bank disregards the purely private component of its credit assessments in the pricing of
transactional debt which primarily results from symmetrically informed competition on the basis
of public credit-quality signals. We also note that competition e¤ects do not seem to signi…cantly
…gure in the pricing of transactional or relationship loans.
Interestingly, the relationship variables Scope and Months on Books (statistically) signi…cantly
reduce not only the o¤ered APR of relationship debt but also the cost of transactional debt. Contrary to the credit decision, the prior purchase of other products from the bank and the length
of a lending relationship enters into the pricing of transactional loans. One possible explanation
might revolve around rewarding customer loyalty in the presence of very low switching costs in
online lending (see also Schenone, 2007). As a result, prior lending could be a signi…cant factor in
banks’pricing policy but less for informational considerations, which the bank addresses through
the decision to grant credit, than to retain a customer of proven pro…tability. Adding the interaction terms in Speci…cation 3 lends further credence to this interpretation. In the eLoan equation,
the interaction terms are statistically insigni…cant whereas the relationship variables retain their
signi…cance. In the in–person equation the interaction terms are highly signi…cant so that the relationship variables enhance the bene…cial e¤ect of a higher private credit-quality signal. Hence, prior
business interaction a¤ects inside-loan rates more by improving the quality of credit assessments
so that banks place greater weight on their private information in the pricing of relationship debt.
It is also worthwhile to point out that a …rm’s age matters for the pricing of transactional but
not relationship debt. Older, more established …rms pay less for loans but the e¤ect is statistically
signi…cant only for online o¤ers. The opposite is true for …rm pro…tability (Net Income) that
22

only matters for the pricing of relationship debt. Again, informational e¤ects might be at work.
The longer a …rm has been in existence the more publicly available information exists which is
particularly valuable in the pricing of transactional debt. By contrast, …nancial data such as net
income are self-reported in online-loan applications and, therefore, susceptible to manipulation. It
is very costly to follow up on …nancial information for online applications so that our data provider
seems to disregard it in this case. By contrast, loan o¢ cers can easily verify such information
during the branch visit by in-person applicants (from, e.g., tax …lings) and, hence, place more trust
in …nancial statements.
The other explanatory variables have very similar e¤ects across the two loan types. In particular,
we note that the ability to post collateral or to personally guarantee a loan reduces loan rates by
210 to 239 and 30 to 81 basis points, respectively, depending on the lending channel. This …nding
contrasts with previous work such as Berger and Udell (1995) or Carey, Post, and Sharpe (1998)
who report that collateral is associated with higher spreads. However, their results are probably
due to the fact that collateral acts as a proxy for nonmeasured risk characteristics. Our …nding
that, once we explicitly control for borrower risk through the inclusion of various credit-quality
measures (XSBI, internal score, PIR), collateral and guarantees reduce loan rates and, given our
speci…cation, credit spreads bears out this conjecture (see also Inderst and Müller, 2006). In fact,
Booth and Booth (2006) also …nd that, controlling for the interdependence between the decision to
pledge collateral and borrowing costs, secured loans typically carry lower spreads.
Taken together our results provide very strong empirical evidence for the predicted trade-o¤
between the availability and pricing of credit across lending modes. In their choice of loan type,
…rms face a choice between easier access to relationship debt and lower priced transactional debt.
Furthermore, we establish that di¤erent types of information lead to this trade o¤. The limited
ability to gather proprietary intelligence in transactional lending forces banks to rely more on public
information that further levels the playing …eld. Hence, online borrowing combines lower interest
rates with a lower probability of receiving credit ceteris paribus. By contrast, the bank’s ability
to collect private information and to strategically use it enhances the likelihood that an in-person
applicant receives credit albeit at the price of higher rates and informational capture, a topic we
turn to next.

23

6

Lending Competition and Borrower Choice

By comparing credit o¤ers to actually booked loans and matching the observations with creditbureau information on competing loan o¤ers we identify 420 transactional and 915 relationship
borrowers that decline the bank’s terms and seek credit from a competitor around the same time.11
Table 6 provides summary statistics by debt type in function of the borrower’s decision to accept
or to decline the o¤er. We see that, on average, the declined loan o¤ers are very similar to accepted
ones for each lending channel.
When the degree of information asymmetry varies by borrower credit transactions become more
contested as the informational advantage of the better informed lender falls. Less precise credit
assessments decrease the threat of adverse selection so that less informed competitors can bid more
aggressively by o¤ering credit more often and at lower rates, thereby eroding the more informed
bank’s ability to earn information rents (see, e.g., Hauswald and Marquez, 2006). Hence, the
smaller our bank’s informational advantage becomes the more frequently borrowers should switch
lenders. In the limit, when all banks are symmetrically informed, price competition dissipates any
information rents and transactional borrowers frequently switch lenders. The implied switching
rates in Table 6 bear out this prediction: transactional borrowers are almost twice as likely as
relationship ones to decline a loan o¤er and seek credit elsewhere (13.39% against 6.98%).
To investigate di¤erences in competitiveness across debt type we next estimate a logistic discretechoice model of the successful loan applicant’s decision to switch lenders. Speci…cation 1 in Table 7
shows that, in line with theoretical predictions, transactional borrowers are almost 5% more likely
to decline loan o¤ers and seek credit elsewhere. As we conjectured earlier, the public credit-quality
signal (XSBI) is by far the most important factor in inducing applicants to decline loan o¤ers. The
higher a …rm’s public score, the easier it becomes to switch lenders explaining the variable’s high
marginal e¤ect even for inside borrowers whose decisions otherwise are more strongly correlated
with private or proprietary information revealed to them over the course of the business relationship
or origination process.
By contrast, private credit-quality signals (PIR), which act as proxies for the borrower’s perception of the lender’s private credit assessment, have a large marginal e¤ect only in the in-person
11

This decision is very di¤erent from borrower’s choice of single vs. multiple banking relationships; see Detragiache
et al. (2000) and Farinha and Santos (2002).

24

equation of Speci…cation 2 in Table 7. Firms rationally anticipate that banks attempt to informationally capture inside borrowers and act accordingly so that the amount and quality of private
information predicts switching behavior. The better the bank’s own credit assessment of a relationship borrower, who can infer the existence of a positive private credit-quality signal from the o¤er
alone, the more likely the latter is to switch lenders. As before, the …rm’s decision and the bank’s
private information should therefore be correlated, which explains why the Private-Information
Residual has such a large impact on the switching behavior of relationship borrowers.
Curiously, the relationship variables (Scope, Months on Books) reduce the likelihood of declining
a loan o¤er for both transactional and relationship borrowers. The large marginal e¤ects and high
statistical signi…cance of the relationship-PIR interaction terms in the in–person equation suggest
that informational e¤ects are at work. The bank’s desire to retain prior customers might explain a
similar e¤ect for transactional borrowers. Unsurprisingly, the higher the quoted loan rate the more
likely are …rms to decline the o¤er and seek credit elsewhere irrespective of the chosen loan type.
Not only is it easier for better credit risks to obtain competing loan o¤ers, they are also the primary
targets for rent extraction through loan pricing and, hence, have a larger incentive to switch lenders.
Consistent with theoretical predictions the e¤ect is more pronounced for relationship borrowers that
face a greater threat of informational capture.
We also see that competition matters but that the e¤ects display an interesting pattern across
loan products. For transactional borrowers, the likelihood of switching lenders increases in the
number of locally competing institutions presumably because of name recognition, geographically
targeted promotional campaigns, etc. By contrast, the switching behavior of relationship borrowers
rises in the number of locally competing branch o¢ ces so that branch-proliferation e¤ects and the
easier access to personal interaction they o¤er matter.
Our results are broadly consistent with strategic lending by inside banks that use private information to informationally capture high-quality relationship borrowers.12 The better the bank’s
information, i.e., the higher the quality of its credit screen or the closer a borrower is located
to a branch, the easier it becomes to extract rents because our lender has a larger informational
advantage over its competitors. Such attempts, however, fail in the transactional-loan segment
12

See also Sharpe (1990), Rajan (1992), or von Thadden (2004) on this point. For evidence on the resulting winner’s
curse in banking see Sha¤er (1998).

25

where symmetrically informed competitors can compete more aggressively for online borrowers. As
a result, the public perception of credit quality drives a …rm’s decision to switch lenders all the
more that borrowers are more likely than not aware of their own Experian scores, which is a good
indicator for the likelihood of receiving a competing loan o¤er.

7

Information Production and Credit Delinquency

Our credit-bureau data also allow us to trace type I (denying a loan to a good credit risk) and type
II (o¤ering a loan to a bad credit risk) errors in credit decisions across loan types. Regarding the
former, out of the 4,810 unsuccessful online applicants 3,321 …rms (69%) managed to obtain a loan
from another source within a month of their loan-application’s rejection. By contrast, less than half
(6,347 out of 12,808) in-person applicants were able to do so. Although transactional borrowers
have a lower ex ante probability of obtaining a loan (see Tables 3 and 4) their cost of seeking credit
online is lower, too so that they typically …le more loan applications than relationship borrowers
and, therefore, have a higher success probability ex post.
In terms of type II errors in screening, our sample contains 91 transactional loans (about 3.5%)
and 319 relationship ones (2.7%) that have fallen 60 days past-due, which corresponds to our data
provider’s internal de…nition of a non-performing loan, within 18 months of origination.13 We …rst
note that the incidence of credit delinquency is signi…cantly higher in the transactional subsample.
To put the respective default rates into perspective, we also trace the credit delinquency of successful
applicants that switched lenders. Their default rates are 4.5% and 5.8% for arm’s-length and
relationship loans, respectively, which is much worse than our bank’s own loan performance. Default
rates for unsuccessful applicants that were able to obtain a loan elsewhere are very high but do not
vary much by lending channels: 24.79% and 24.63% for online and in-person (denied) applications,
respectively. We interpret these default frequencies as evidence that our lender minimizes type II
error in credit decisions by trying to avoid lending to bad credit risks. In doing so, the bank is
more successful for relationship loans than transactional ones, for which intermediaries generally
13

We choose this window so that the likelihood of a loan becoming overdue is still related to the initial credit
assessment and not to subsequent economic events beyond the bank’s control. Although the technical de…nition of
default is 180 days past-due lenders typically take action after at most 60 days past-due either writing o¤ the loan,
selling it o¤, or assigning it for collection. As a result we do not know which of the delinquent loans ultimately
experience default although over 90% of loans that are 60 days overdue eventually do according to our data provider.

26

su¤er higher adverse-selection problems.
To investigate the di¤erences in loan performance across transactional and relationship lending
we estimate a logistic model of credit delinquency in terms of our usual information, relationship,
and control variables by lending channel. Table 8 shows that transactional borrowers are up to
2.9% more likely to default than relationship ones ceteris paribus. The results also exhibit the
usual pattern in information e¤ects across equations. Public information (XSBI score) has by far
the largest impact on the likelihood of default for both loan types. Positive private information
(internal score, PIR) only a¤ects the performance of relationship loans in an economically signi…cant
manner. Again, proprietary intelligence is primarily useful for mitigating credit risk in relationship
lending but adds less to the bank’s ability to predict the performance of transactional loans.
The marginal e¤ects of the relationship variables that are much larger for in-person than online
loans and, especially, the PIR-relationship interaction terms con…rm this e¤ect. Banks bene…t
from lending relationships through better private information that allows them to decrease their
borrower-speci…c credit exposure. Similarly, the Months in Business variable has quite a large and
statistically signi…cant marginal e¤ect on decreasing the risk of default across both lending modes
presumably because there is more information - private or public - available for older …rms. The
loan amount has a large negative e¤ect on the likelihood of default that is more or less constant
across lending channels and speci…cations. In contrast to DeYoung et al. (2004) who report that
the probability of default on small-business loans increases in the distance between borrower and
lender we do not …nd any signi…cant distance e¤ects for either loan type.
The small but highly signi…cant positive marginal e¤ects of the competitiveness measures are
consistent with theoretical predictions that more competition implies more adverse selection and,
hence, more default. The informational e¤ects, however, suggest that di¤erent forces are responsible
for each lending channel. In transactional lending, more competition decreases the average quality
of the borrower pool so that each lender su¤ers more adverse selection (Broecker, 1990). When
competition increases for relationship borrowers, the informed lender has less of an incentive to
acquire private information and the overall quality of its loan portfolio falls (see, e.g., Gehrig,
1998).

27

8

Conclusion

This paper presents an in-depth comparative analysis of the respective roles of private and public
information in arm’s-length and inside-debt transactions. The advent of online lending and banks’
distinct operational practices across lending channels o¤er the opportunity to unambiguously identify transactional loans that match in all other respects traditional relationship debt. Using an
exhaustive sample of online and in-person loan requests by small businesses we are able to determine the relative importance of private and public information for each debt type. At the same
time, our data also allows us to investigate how the chosen form of bank-borrower interaction a¤ects
the lender’s acquisition of private and proprietary information, its strategic use in credit decisions,
and the borrowers’response for each form of debt.
Our results reveal that banks rely on di¤erent types of information for each lending mode.
Public information primarily drives credit availability and pricing in transactional lending whereas
private information determines credit decisions for relationship loans. Since banks have less opportunity to generate borrower-speci…c information from arm’s-length debt they compete on a more
symmetrically informed basis and rely more heavily on public information in their transactional
credit decisions. The opposite is true for relationship loans. We …nd strong evidence that banks
disregard publicly available information when they have access to better “soft”private information
through inside lending that becomes the foundation of their relationship-credit decision and pricing.
By the same token, borrowers base their choice of debt type mainly on public credit-quality
information that is readily available to them and provides them with a sense of their success chances
in each credit-market segment. Furthermore, we …nd evidence that inside borrowers anticipate on
the existence and consequences of private information. Longstanding business relationships imply
more inside information together with preferential treatment so that the likelihood that a …rm
will seek a relationship loan increases in the lender’s private credit-worthiness signal. Similarly, a
…rm’s decision to decline relationship debt or to default on it depends more on the bank’s private
information than transactional debt although public information retains some importance for these
choices, too. These …ndings are consistent with the notion that borrowers recognize the value of
lending relationships for banks’ability to acquire proprietary information and to strategically use
it.

28

However, the bene…ts of a lending relationship must ultimately outweigh the cost of informational capture for …rms that otherwise would not selfselect into inside debt. Hence, our …ndings
also provide support for the contention that relationship borrowers bene…t from the closer ties with
their banks. The fact that in-person loan applicants have, on average, a much longer and deeper
relationship with their bank than online applicants lends additional credence to this interpretation. Such bene…ts typically revolve around intertemporal transfers between the parties, i.e., the
notion that banks are more willing to …nance borrowers that would otherwise not be able to …nd
funding if they can recover the initial costs through future rent extraction or better loan performance. To directly investigate the existence of such bene…ts, however, one would need panel data
on bank-borrower interaction over a longer time period. We leave this question for future research.

29

Table 1: Descriptive Statistics for All Loan Applications
Lending Channel
Variable
Loan Amount
Maturity (years)
Term Loan (vs. Credit-Line)
Collateral
Primary Guarantor
Primary Guarantor’s Monthly Salary
SBA Guarantee
Internal Credit Score
Public (XSBI) Credit Score
Private-Information Residual
Scope of Banking Relationship
Months on Books
Monthly Deposit Account Balance
Months in Business
Firm’s Monthly Net Income
Case-Shiller House Price Index
Firm-Bank Distance (miles by car)
Firm-Comp Distance (miles by car)
State CT
State MA
State ME
State NH
State NJ
State NY
State PA
State RI
Other States
Q1 2002
Q2 2002
Q3 2002
Q4 2002
Q1 2003
SIC 0: Agriculture, Forestry, Fishing
SIC 1: Mining, Construction
SIC 2: Manufacturing (Consumer)
SIC 3: Manufacturing (Industrials)
SIC 4: Transport., Comm., Gas, Elect.
SIC 5: Wholesale and Retail Trade
SIC 6: Finance, Insurance, Real Estate
SIC 7: Personal & Business Services
SIC 8: Professional Services
SIC 9: Administration
Number of Branches
Number of Institutions
Number of Observations

Online Application
Mean
Median Std Dev
$37,333 $34,320 $124,921
5.43
5.18
2.05
19.57%
38.01%
41.54%
41.75%
17.09%
39.83%
$23,745 $20,821 $107,148
3.74%
14.59%
899.22
902.05
73.40
723.57
704.67
55.87
0.0059
0.0003
0.4975
19.92%
35.03%
27.68
23.32
48.52
$12,649 $10,835
$16,040
63.88
54.30
41.62
$64,734 $58,521
$77,808
168.53
152.10
36.08
91.74
31.90
80.91
0.89
0.54
1.16
8.32%
10.32%
23.53%
41.26%
2.32%
14.36%
2.88%
16.34%
16.32%
34.85%
35.53%
45.73%
0.27%
5.08%
4.86%
21.29%
2.00%
1.76%
17.02%
34.62%
15.19%
36.00%
17.45%
36.03%
20.62%
37.80%
23.99%
35.06%
2.20%
14.49%
9.93%
27.70%
2.80%
15.49%
3.38%
17.09%
4.26%
19.33%
25.94%
42.21%
4.48%
19.94%
19.68%
37.61%
13.62%
31.33%
0.30%
5.46%
4.51
2.77
4.53
3.58
2.57
4.15
7,945

In-Person Application
Mean
Median Std Dev
$46,877 $39,749
$42,693
6.74
6.20
5.34
28.06%
46.86%
54.91%
48.64%
36.59%
47.59%
$35,164 $32,012
$88,582
6.41%
15.90%
930.51
949.38
133.29
716.74
703.78
57.55
0.0003
0.0005
0.6316
30.40%
43.66%
30.79
22.65
43.20
$14,363 $11,014
$41,669
103.56
89.24
103.08
$101,109 $90,108 $315,463
166.40
154.79
31.14
10.29
2.82
25.12
1.02
0.52
1.53
12.89%
35.04%
15.18%
35.73%
3.14%
17.28%
2.58%
15.67%
24.53%
42.62%
35.77%
47.61%
3.08%
17.07%
3.20%
17.46%
0.17%
4.00%
18.34%
38.43%
18.61%
38.99%
17.48%
37.70%
19.01%
38.66%
27.13%
32.96%
3.02%
16.99%
13.29%
33.62%
2.41%
15.20%
3.05%
17.02%
4.94%
21.66%
31.01%
45.78%
3.35%
17.64%
19.19%
39.10%
13.21%
33.28%
0.13%
3.51%
4.78
3.00
5.36
3.55
2.99
3.38
25,910

t-Test
P -val
0.0000
0.0000
0.0000
0.0000
0.0000
0.0000
0.0000
0.0000
0.0000
0.4676
0.0000
0.0000
0.0003
0.0000
0.0000
0.0000
0.0000
0.0000
0.0000
0.0000
0.0001
0.1500
0.0000
0.6947
0.0000
0.0000
0.0000
0.0062
0.0000
0.9577
0.0011
0.0000
0.0001
0.0000
0.0486
0.1341
0.0124
0.0000
0.0000
0.3235
0.3264
0.0006
0.0000
0.4629
33,855

This table presents summary statistics for the variables described in Section 3 for our full sample of 33,855 data points
in function of the …rm’s choice of lending channel. The last column indicates the P -values of a two-sided t-test for the
equality of the variables’mean conditional on the loan’s type (wherever appropriate).

30

Table 2: The Choice of Lending Channel and Loan Type
Speci…cation
Variable
Constant
ln(1+XSBI)
ln(1+Internal Score)
Private-Info. Res.
Scope
ln(1+M. on Books)
Scope PIR
ln(1+MOB) PIR
ln(1+M. in Business)
ln(1+Net Income)
ln(1+CSHPI)
ln(1+F-B Dist)
ln(1+F-C Dist)
Collateral
Primary Guarantor
SBA Guarantee
Term Loan
ln(1+# Branches)
ln(1+# Competitors)
4 Quarterly Dum.
8 State Dummies
38 SIC Dummies
Number of Obs
Pseudo R2

Coe¤
-2.1933
0.5347

1
P -val
0.0001
0.0001

Coe¤
-2.0659
0.4646
0.3582

2
P -val
0.0001
0.0001
0.0001

21.46%

14.15%
3.80%

-0.1345
-0.6809

0.7590
0.0001

-0.03%
-6.88%

-0.1202
-0.6896

0.5331
0.0001

0.1193
-0.0738
-0.1020
0.9874
-0.2644
-0.2148
-0.0444
-0.0791
-0.0769
0.1929
0.1001

0.9820
0.4932
0.9223
0.0001
0.0001
0.6349
0.9320
0.0158
0.9202
0.7745
0.6338
Yes
Yes
Yes
33,855
4.84%

0.15%
-0.98%
-0.78%
1.87%
-1.04%
-0.71%
-0.01%
-0.32%
-0.08%
0.03%
0.02%

0.1048
-0.0756
-0.1007
1.0053
-0.2663
-0.2162
-0.0480
-0.0766
-0.0818
0.2135
0.1013

0.8862
0.4871
0.9210
0.0001
0.0001
0.6400
0.7368
0.0001
0.9447
0.6971
0.6329
Yes
Yes
Yes
33,855
5.30%

Marg

Coe¤
-2.0708
0.4502

3
P -val
0.0001
0.0001

14.18%

-0.02%
-6.29%

-0.8958
-0.1201
-0.6959

0.0001
0.7879
0.0001

-9.83%
-0.03%
-7.14%

0.10%
-1.01%
-0.71%
1.93%
-0.95%
-0.75%
-0.01%
-0.37%
-0.09%
0.03%
0.02%

0.1104
-0.0789
-0.1049
1.0121
-0.2912
-0.2356
-0.0480
-0.0842
-0.0823
0.2060
0.1094

0.9080
0.5000
0.9270
0.0001
0.0001
0.6893
0.9110
0.0130
0.9848
0.7760
0.6377
Yes
Yes
Yes
33,855
5.21%

0.17%
-1.00%
-0.81%
1.92%
-1.05%
-0.73%
-0.01%
-0.34%
-0.09%
0.03%
0.02%

Marg

Marg

Coe¤
-2.0612
0.4502
-0.9009
-0.1194
-0.6796
-0.3522
-0.1014
0.1096
-0.0776
-0.1035
1.0183
-0.2863
-0.2311
-0.0481
-0.0834
-0.0821
0.2071
0.1093

4
P -val
0.0001
0.0001
0.0001
0.7920
0.0001
0.0346
0.1179
0.9011
0.4987
0.9860
0.0001
0.0001
0.6780
0.9170
0.0136
0.9740
0.7760
0.6351
Yes
Yes
Yes
33,855
5.23%

This table reports the results from estimating a logistic discrete-choice model of the …rm’s choice of loan type by fullinformation maximum likelihood for our full sample (33,855 observations). The dependent variable is the …rm’s decision
to apply online for a transactional loan (Y = 1: 7,945 observations) or in-person for a relationship loan (Y = 0: 25,910
observations). We estimate the speci…cation Pr fYi = 1 jxi g = (x0i +1eloan x0i ), where 1eloan = 1 for online applicaexpfx0
kg
tions and 0 otherwise and is the logistic distribution function (x0ik k ) = P exp ik
, with branch …xed e¤ects and
fx0in n g
n
compute clustered standard errors that are adjusted for heteroskedasticity across branch o¢ ces and correlation within.
The explanatory variables are our proxies for public (Experian’s Small Business Intelliscore XSBI), proprietary (Internal Score) and private (Private-Information Residual) information, bank-borrower relationship characteristics (Scope,
Months on Books abbreviated “MOB” in the interaction terms), …rm attributes, the competitiveness of local credit markets (number of competing lenders and competing branches), proxies for the ease of transacting with lenders (…rm-bank
and …rm-competitor distances abbreviated F-B and F-C Dist, respectively), and control variables for local economic
conditions (Case-Shiller house-price index abbreviate CSHPI), the business cycle (quarterly dummies), state, and …rm’s
industry (see Section 3 for a description of the variables).
The Private-Information Residual (abbreviated “PIR" in the interaction terms) measures the bank’s pure private information that we obtain from orthogonalizing the internal and Experian scores. Speci…cally, the PIR for each observation
is the residual u
^i of the branch …xed-e¤ects regression ln (IntScorei ) = p + p XSBIi + 1eloan ( e + e XSBIi ) + ui .
We report the coe¢ cients ( “Coe¤”), their P -values (“P -val”), and marginal e¤ects (“Marg”) for the decision to apply
online (Y = 1) but suppress the results for the business-cycle, state, and industry control variables in the interest of readability. Since the probabilities of applying online or in person sum to 1 the marginal e¤ects for the choice of a relationship
loan are simply the opposite of the reported ones. We obtain the marginal e¤ects by simply evaluating @@xPrj = 0 (x0i ) j
at the regressors’sample means and coe¢ cient estimates ^ . The pseudo-R2 is McFadden’s likelihood ratio index 1 log L .
log L0

31

Marg
14.30%
-9.77%
-0.03%
-7.03%
-2.58%
-1.80%
0.17%
-0.98%
-0.81%
1.90%
-1.05%
-0.72%
-0.01%
-0.34%
-0.09%
0.03%
0.02%

Table 3: Descriptive Statistics for the Credit Decision by Lending Channel
Panel A: Online Loan Applications
Loan-Application Outcome
Variable
Loan Rate (APR: all-in cost of loan)
Loan Amount
Maturity (years)
Term Loan (vs. Credit-Line)
Collateral
Primary Guarantor
SBA Guarantee
Internal Credit Score
Public (XSBI) Credit Score
Private-Information Residual
Scope of Banking Relationship
Months on Books
Monthly Deposit Account Balance
Months in Business
Firm’s Monthly Net Income
Firm-Bank Distance (miles by car)
Firm-Comp Distance (miles by car)
Maturity-Matched UST Yield
5Y - 3M UST Yield Spread (bpts)
Number of Observations

Mean
6.91%
$37,102
5.42
14%
51%
26.87%
0.80%
1040.85
729.96
0.0289
22.12%
38.53
$13,902
73.57
$80,948
82.40
0.95
4.12%
202.40

Accept
Median Std Dev
6.86%
1.93%
$34,270 $124,787
5.16
2.03
34%
32%
34.30%
2.37%
1022.32
80.70
717.40
47.95
0.0180
0.4764
30.33%
30.35
54.04
$12,075
$15,499
60.22
43.57
$75,409 $102,184
31.18
82.45
0.51
1.25
3.74%
2.35%
196.32
55.87
3,135

Mean
N/A
N/A
N/A
23.33%
35.92%
10.74%
5.64%
804.46
717.99
-0.0183
18.42%
20.94
$11,925
57.93
$54,033
98.32
0.86
N/A
N/A

Reject
Median
N/A
N/A
N/A

822.01
702.93
-0.0098
18.77
$9,939
50.30
$47,538
32.49
0.56
N/A
N/A
4,810

Std Dev
N/A
N/A
N/A
40.36%
48.15%
43.37%
22.53%
68.97
60.48
0.5813
37.83%
45.01
$16,318
39.87
$61,859
79.76
1.12
N/A
N/A

t-Test
P -val
N/A
N/A
N/A
0.0000
0.0000
0.0000
0.0000
0.0000
0.0000
0.0002
0.0000
0.0000
0.0000
0.0000
0.0000
0.0000
0.0010
N/A
N/A
7,945

Panel B: In-Person Loan Applications
Loan-Application Outcome
Variable
Loan Rate (APR: all-in cost of loan)
Loan Amount
Maturity (years)
Term Loan (vs. Credit-Line)
Collateral
Primary Guarantor
SBA Guarantee
Internal Credit Score
Public (XSBI) Credit Score
Private-Information Residual
Scope of Banking Relationship
Months on Books
Monthly Deposit Account Balance
Months in Business
Firm’s Monthly Net Income
Firm-Bank Distance (miles by car)
Firm-Comp Distance (miles by car)
Maturity-Matched UST Yield
5Y - 3M UST Yield Spread (bpts)
Number of Observations

Mean
8.46%
$46,648
6.71
22.49%
60.12%
34.14%
0.56%
1039.64
716.89
0.0379
35.37%
43.26
$17,127
116.30
$110,525
9.97
1.11
3.93%
220.78

Accept
Median
8.13%
$39,881
6.19

1045.41
709.96
0.0112
30.65
$11,845
96.86
$95,203
2.64
0.55
3.86%
210.80
13,102

Std Dev
2.72%
$42,479
5.38
46.72%
48.23%
46.80%
4.68%
138.47
57.62
0.7157
43.87%
56.28
$62,579
107.20
$255,740
21.42
1.59
1.94%
57.11

Mean
N/A
N/A
N/A
33.79%
49.98%
39.23%
12.32%
817.65
715.19
-0.0350
25.55%
17.69
$11,652
91.79
$92,210
10.68
0.94
N/A
N/A

Reject
Median
N/A
N/A
N/A

855.09
698.66
-0.0106
14.50
$10,108
81.80
$85,187
2.99
0.48
N/A
N/A
12,808

Std Dev
N/A
N/A
N/A
47.11%
48.60%
48.48%
27.23%
128.33
57.67
0.5806
43.32%
29.60
$20,964
98.83
$373,508
28.68
1.46
N/A
N/A

t-Test
P -val
N/A
N/A
N/A
0.0000
0.0000
0.0000
0.0000
0.0000
0.0173
0.0000
0.0000
0.0000
0.0000
0.0000
0.0000
0.0228
0.0000
N/A
N/A
25,910

This table reports descriptive statistics for the key variables described in Section 3 in terms of the lending channel (7,945
online applications in Panel A and 25,910 in-person ones in Panel B) and the bank’s decision to o¤er or to deny credit.
The last column indicates the P -values of a two-sided t-test for the equality of the variables’ mean conditional on the
bank’s decision (wherever appropriate). For summary statistics of the control variables by lending channel see Table 1.

32

Table 4: The Credit Decision by Loan Type
Speci…cation
Loan Type
Variable
Constant
eLoan (1eloan = 1)
ln(1+XSBI)
ln(1+Internal Score)
Private-Info. Res.
Scope
ln(1+M. on Books)
Scope PIR
ln(1+MOB) PIR
ln(1+M. in Business)
ln(1+Net Income)
ln(1+CSHPI)
ln(1+F-B Dist)
ln(1+F-C Dist)
Collateral
Primary Guarantor
SBA Guarantee
Term Loan
ln(1+# Branches)
ln(1+# Competitors)
4 Quarterly Dum.
8 State Dummies
38 SIC Dummies
Number of Obs
Pseudo R2

1

2

Coe¤

eLoans
P -val

In-Person Loans
Coe¤
P -val
Marg
-2.0817 0.0001

Marg

-1.6314
0.4259
0.1419

0.0001
0.0001
0.0046

-9.33%
19.99%
2.63%

0.2556
0.1683

0.1460
0.0001

0.25%
11.40%

0.2714
0.3797

0.2228
0.7968

0.27%
0.12%

0.9190
0.9313

0.0001
0.0001

2.53%
1.65%

0.9189
0.6910
0.0884
-0.4299
0.0894
0.5451
0.0510
-0.3711
-0.0263
-1.2680
-1.0295

0.0001
0.0001
0.1314
0.8420
0.4177
0.0001
0.0147
0.9280
0.0791
0.0001
0.0001
Yes
Yes
Yes

0.67%
1.38%
0.23%
-0.02%
0.02%
2.40%
0.19%
-0.34%
-0.07%
-1.14%
-1.07%

0.3772
0.8964
1.0297
-0.8729
0.6474
0.6006
0.6481
-0.1260
-0.5006
-0.5528
-0.0701

0.0001
0.0001
0.0390
0.0515
0.6820
0.0001
0.0001
0.4330
0.0001
0.0342
0.0085
Yes
Yes
Yes

2.71%
1.17%
0.52%
-1.15%
0.22%
2.01%
4.12%
-0.41%
-0.66%
-1.60%
-2.11%

33,855
12.06%

Coe¤

eLoans
P -val

Marg

In-Person Loans
Coe¤
P -val
Marg
-2.1270 0.0001

-1.6037
0.4019

0.0001
0.0001

-11.19%
18.63%

0.2629

0.0397

3.43%

0.1928
0.2560
0.3559
0.0636
0.3019
0.8907
0.6577
0.0908
-0.4238
0.0899
0.5473
0.0513
-0.3419
-0.0263
-1.2836
-1.0031

0.0425
0.3032
0.7968
0.4529
0.3960
0.0001
0.0001
0.0900
0.8220
0.3832
0.0001
0.0133
0.9148
0.0839
0.0001
0.0001
Yes
Yes
Yes

1.03%
0.33%
0.12%
0.42%
0.24%
0.67%
1.67%
0.23%
-0.02%
0.02%
2.81%
0.25%
-0.36%
-0.07%
-1.21%
-1.18%

0.6425
0.8737
0.8594
0.1488
0.0415
0.3630
0.8767
0.9659
-0.8956
0.6012
0.5879
0.5646
-0.1197
-0.4655
-0.4720
-0.0640

0.0001
0.0001
0.0001
0.0657
0.0001
0.0001
0.0001
0.0145
0.0440
0.6380
0.0001
0.0001
0.5343
0.0001
0.0390
0.0073
Yes
Yes
Yes

15.79%
2.29%
1.79%
1.72%
1.63%
2.74%
1.00%
0.18%
-0.99%
0.23%
1.87%
4.02%
-0.32%
-0.64%
-1.77%
-1.96%

33,855
12.04%

This table reports the results from estimating a logistic discrete-choice model of the bank’s credit decision by loan type
for our full sample (33,855 observations) using maximum likelihood. We estimate the speci…cation Pr fYi = 1 jxi g =
(x0i +1eloan x0i ), where 1eloan = 1 for online applications and 0 otherwise and is the logistic distribution function,
with branch …xed e¤ects and compute clustered standard errors that are adjusted for heteroskedasticity across branch
o¢ ces and correlation within. The dependent variable is the bank’s decision to o¤er (Y = 1: 3,135 and 13,102 observations
for online and in-person loans, respectively) or to deny (Y = 0: 4,810 and 12,808 observations for online and in-person
loans, respectively) credit. The explanatory variables are our proxies for public, proprietary, and private information, bankborrower relationship characteristics, …rm attributes, measures of the local credit market’s competitiveness and various
control variables. See Section 3 for a description of the variables and the notes to Table 2 for further methodological
details.

33

Table 5: Determinants of the O¤ered Loan Rate
Speci…cation
Loan Type
Variable
Constant
eLoan (1eloan = 1)
ln(1+XSBI)
ln(1+Internal Score)
Private-Info. Res.
Scope
ln(1+M. on Books)
Scope PIR
ln(1+MOB) PIR
ln(1+M. in Business)
ln(1+Net Income)
ln(1+CSHPI)
ln(1+F-B Dist)
ln(1+F-C Dist)
Collateral
Primary Guarantor
SBA Guarantee
Term Loan
ln(1+Maturity)
ln(1+# Branches)
ln(1+# Competitors)
UST Yield
Term Spread
Lambda
4 Quarterly Dum.
8 State Dummies
38 SIC Dummies
Number of Obs
Adjusted R2

1
eLoans
Coe¤
P -val
-1.3454

2
In-Person Loans
Coe¤
P -val
7.8736 0.0001

3
In-Person Loans
Coe¤
P -val
7.4440 0.0001

-1.2883
-1.2408
-0.2695

0.0001
0.0001
0.0001

-0.6463
-1.6531

0.0001
0.0001

0.0001
0.0001

-0.4614
-0.7186

0.0001
0.0466

-0.2988
-0.3573

0.0001
0.0001

-0.9037 0.0877 -0.1458 0.2950
-0.3180 0.2907 -0.7772 0.0001
-0.5259 0.0590 -0.6326 0.0001
-1.7713 0.0010 -1.9030 0.0012
0.7113 0.0001
0.9647 0.0052
-2.3422 0.0001 -2.3782 0.0001
-0.8130 0.0290 -0.2991 0.0009
0.4457 0.3134
0.3260 0.0247
1.3000 0.0370
0.3586 0.0001
-0.3996 0.0001 -0.7342 0.0001
-0.1847 0.8730 -0.0534 0.7580
-0.3311 0.9314 -0.3531 0.3885
0.2622 0.0001
0.2925 0.0001
0.2788 0.0001
0.4356 0.0041
0.6642 0.0463 -0.3814 0.0062
Yes
Yes
Yes
Yes
Yes
Yes
16,237
14.06%

-0.8338
-0.3075
-0.5180
-1.1433
0.1970
-2.3203
-0.7462
0.4257
1.2480
-0.3752
-0.1752
-0.3197
0.2472
0.2739
0.5893

0.1816 -0.1374 0.3741
0.3167 -0.7621 0.0001
0.0953 -0.5737 0.0001
0.5154 -1.0574 0.4736
0.9350
0.5463 0.2450
0.0001 -2.2421 0.0001
0.0459 -0.2815 0.0012
0.3878
0.3057 0.0260
0.0408
0.3508 0.0001
0.0001 -0.6924 0.0001
0.8977 -0.0515 0.8930
0.8997 -0.3244 0.5000
0.0001
0.2723 0.0001
0.0001
0.4184 0.0081
0.2420 -0.2920 0.4678
Yes
Yes
Yes
Yes
Yes
Yes
16,237
17.28%

-0.4919
-0.7862

0.0001

eLoans
Coe¤
P -val

0.0001
0.0457

-0.3281
-0.3787

eLoans
Coe¤
P -val
-1.3791
-1.2904

0.0001
0.0001

In-Person Loans
Coe¤
P -val
7.7973 0.0001
-0.6844

-0.1488 0.2720 -0.4764 0.0001
-0.4309 0.0012 -0.3091 0.0001
-0.7367 0.0330 -0.3807 0.0001
-0.0311 0.7882 -0.1985 0.0001
-0.0522 0.7008 -0.1289 0.0188
-0.8329 0.0560 -0.1455 0.3932
-0.2994 0.3842 -0.7600 0.0001
-0.5487 0.1345 -0.5847 0.0001
-1.1096 0.5960 -1.0688 0.5194
0.1880 0.9161
0.6134 0.4235
-2.4330 0.0001 -2.1043 0.0001
-0.7275 0.0248 -0.2788 0.0002
0.4032 0.3830
0.3231 0.0283
1.2241 0.0493
0.3171 0.0001
-0.3019 0.0001 -0.5951 0.0001
-0.1701 0.8911 -0.0544 0.9015
-0.2989 0.9616 -0.3113 0.4671
0.2691 0.0001
0.3016 0.0001
0.2704 0.0001
0.4516 0.0003
0.5861 0.2489 -0.2960 0.4689
Yes
Yes
Yes
Yes
Yes
Yes
16,237
17.15%

This table reports the results from estimating linear models of the o¤ered loan rate (APR: all-in cost of the loan) of
the form ri = x0i +1eloan x0i + "i , where 1eloan = 1 for online applications and 0 otherwise, by OLS with branch
…xed e¤ects and clustered standard errors that are adjusted for heteroskedasticity across branch o¢ ces and correlation
within. The explanatory variables are our proxies for public, proprietary, and private information, bank-borrower
relationship characteristics, …rm attributes, and various control variables. Lambda is the inverse Mills ratio (hazard
rate) for the logistic distribution required by the Heckman procedure for sample-selection bias. See Section 3 for a
description of the variables.

34

0.0001

Table 6: Descriptive Statistics for Accepted and Declined Credit O¤ers
Panel A: Online (Transactional) Loan O¤ers
Loan-O¤er Decision
Variable
Loan Rate (APR: all-in cost of loan)
Loan Amount
Maturity (years)
Term Loan (vs. Credit-Line)
Collateral
Primary Guarantor
SBA Guarantee
Internal Credit Score
Public (XSBI) Credit Score
Private-Information Residual
Scope of Banking Relationship
Months on Books
Monthly Deposit Account Balance
Months in Business
Firm’s Monthly Net Income
Firm-Bank Distance (miles by car)
Firm-Comp Distance (miles by car)
Maturity-Matched UST Yield
5Y - 3M UST Yield Spread (bpts)
Number of Observations

Mean
6.81%
$37,789
5.34
14%
54%
28.55%
0.76%
1062.51
728.31
0.03
23.29%
37.88
$14,050
76.15
$81,123
84.22
0.97
4.33%
207.35

Accept
Median Std Dev
6.63%
1.85%
$34,987 $124,485
5.20
2.02
33%
31%
33.88%
2.33%
1027.93
79.75
716.08
45.24
0.02
0.46
30.03%
30.79
53.57
$12,056
$15,067
60.66
43.27
$75,927 $101,117
31.31
81.20
0.51
1.21
3.67%
2.31%
196.46
55.14
2,715

Mean
7.65%
$32,833
5.87
13%
34%
18%
1.08%
987.04
734.82
0.03
16%
44.59
$14,071
56.60
$85,128
75.50
0.88
2.94%
184.49

Decline
Median Std Dev
8.16%
2.58%
$32,737 $138,135
5.36
2.51
41%
39%
42%
2.85%
1070.79
97.66
726.94
49.24
0.02
0.56
36%
30.62
65.75
$12,756
$18,660
60.43
53.23
$75,440 $124,760
31.86
100.00
0.53
1.52
4.28
2.87%
216.48
66.99
420

t-Test
P -val
0.0000
0.4546
0.0000
0.6450
0.0000
0.0000
0.0115
0.0000
0.0067
0.8598
0.0000
0.0208
0.9797
0.0000
0.4652
0.0476
0.1743
0.0000
0.0000
3,135

Panel B: In-Person (Relationship) Loan O¤ers
Loan-O¤er Decision
Variable
Loan Rate (APR: all-in cost of loan)
Loan Amount
Maturity (years)
Term Loan (vs. Credit-Line)
Collateral
Primary Guarantor
SBA Guarantee
Internal Credit Score
Public (XSBI) Credit Score
Private-Information Residual
Scope of Banking Relationship
Months on Books
Monthly Deposit Account Balance
Months in Business
Firm’s Monthly Net Income
Firm-Bank Distance (miles by car)
Firm-Comp Distance (miles by car)
Maturity-Matched UST Yield
5Y - 3M UST Yield Spread (bpts)
Number of Observations

Mean
8.58%
$46,833
6.25
21.97%
62.07%
35.45%
0.52%
1046.79
715.94
0.04
35.66%
43.81
$18,013
118.47
$114,131
10.11
1.12
3.39%
216.19

Accept
Median
8.16%
$39,970
6.15

1046.92
708.63
0.01
30.48
$11,954
97.83
$94,861
2.64
0.55
3.37%
208.40
12,187

Std Dev
2.59%
$42,806
5.41
47.48%
48.51%
47.68%
4.56%
141.72
54.03
0.68
43.67%
58.79
$65,413
111.38
$271,336
21.87
1.62
1.16%
55.65

Mean
8.52%
$49,300
6.46
29.66%
60.97%
33.21%
1.39%
1047.86
719.03
0.04
35.29%
46.78
$19,238
103.14
$114,889
10.02
1.12
3.87%
240.21

Decline
Median
8.28%
$41,089
6.29

1066.65
713.03
0.01
31.31
$11,942
97.91
$96,580
2.17
0.39
3.70%
211.66
915

Std Dev
2.76%
$57,307
5.36
37.45%
45.84%
45.35%
3.46%
85.15
58.94
0.75
41.79%
48.51
$48,615
93.35
$175,969
23.75
1.72
1.04%
61.23

t-Test
P -val
0.5266
0.1017
0.2462
0.0000
0.5093
0.1689
0.0000
0.8216
0.0968
0.8812
0.8023
0.1358
0.5788
0.0000
0.9337
0.9107
0.9879
0.0000
0.0000
13,102

This table provides summary statistics for key variables described in Section 3 as a function of the borrower’s decision
to accept (online and in-person applications: 2,715 and 12,187 observations, respectively) or to decline (420 and 915
observations, respectively) the bank’s loan o¤er by lending channel. The last column indicates the P -values of a two-sided
t-test for the equality of the variables’mean conditional on the applicant’s decision.

35

Table 7: The Decision to Decline a Loan O¤er
Speci…cation
Loan Type
Variable
Constant
eLoan (1eloan = 1)
ln(1+XSBI)
ln(1+Internal Score)
Private-Info. Res.
Scope
ln(1+M. on Books)
Scope PIR
ln(1+MOB) PIR
ln(1+M. in Business)
ln(1+Net Income)
ln(1+CSHPI)
ln(1+F-B Dist)
ln(1+F-C Dist)
Collateral
Primary Guarantor
SBA Guarantee
Term Loan
APR
ln(1+Loan Amount)
ln(1+Maturity)
ln(1+# Branches)
ln(1+# Competitors)
UST Yield
Term Spread
4 Quarterly Dum.
8 State Dummies
38 SIC Dummies
Number of Obs
Pseudo R2

1

2

Coe¤

eLoans
P -val

In-Person Loans
Coe¤
P -val
Marg
-4.4031 0.0001

Marg

1.0927
1.6612
0.2839

0.0001
0.0001
0.0001

4.88%
22.89%
2.75%

0.7772
0.5800

0.0001
0.0001

25.52%
8.72%

-2.5589
-1.5607

0.0001
0.0001

-4.96%
-3.52%

-1.0172
-1.8461

0.0242
0.0001

-3.98%
-4.17%

-0.2590
2.3368
0.9326
2.1461
-1.0768
0.0422
2.0609
1.2450
-0.7127
0.2651
-2.0620
-0.1770
0.4441
0.6081
0.8489
-1.0548

0.3837
0.0001
0.0088
0.0001
0.0001
0.9850
0.0001
0.0001
0.0001
0.0001
0.0001
0.0001
0.4853
0.0222
0.0001
0.0290
Yes
Yes
Yes

-0.30%
1.72%
0.66%
1.82%
-0.31%
0.15%
3.92%
0.03%
-0.38%
9.48%
-4.12%
-0.95%
0.16%
0.29%
2.36%
-1.70%

-0.3124
1.9730
0.0262
2.0528
-1.0918
0.1780
2.1411
0.2710
-0.0122
0.3955
-2.0135
-0.2959
0.4857
0.0237
1.1566
-1.0086

0.0207
0.0001
0.95
0.0001
0.0001
0.5530
0.0001
0.0248
0.9890
0.0001
0.0001
0.0001
0.0440
0.9307
0.0001
0.0001
Yes
Yes
Yes

-0.20%
2.48%
0.31%
0.97%
-0.27%
0.20%
4.94%
0.11%
-0.03%
12.82%
-2.27%
-1.37%
0.35%
0.13%
3.50%
-2.65%

16,237
6.72%

Coe¤

eLoans
P -val

Marg

In-Person Loans
Coe¤
P -val
Marg
-4.8724 0.0001

1.1135
1.6532

0.0001
0.0001

4.28%
27.10%

0.7697

0.0001

30.35%

0.3753
-2.4749
-1.6741
0.1303
0.0627
-0.2741
2.3335
1.0016
2.0690
-1.0525
0.0431
2.0262
1.2261
-0.6795
0.2587
-2.1222
-0.1793
0.4024
0.6429
0.8587
-1.0109

0.0200
0.0001
0.0001
0.6991
0.9680
0.3005
0.0001
0.0004
0.0001
0.0001
0.9240
0.0001
0.0001
0.0001
0.0001
0.0001
0.0001
0.7751
0.0229
0.0001
0.0479
Yes
Yes
Yes

3.65%
-4.91%
-3.69%
0.44%
0.23%
-0.34%
2.62%
0.65%
1.66%
-0.35%
0.15%
4.00%
0.08%
-0.34%
9.62%
-4.59%
-1.01%
0.17%
0.27%
2.69%
-1.66%

0.5379
-1.0038
-1.7965
0.6093
0.7223
-0.3325
1.9760
0.0276
2.0314
-1.0580
0.1812
2.2082
0.2630
-0.0116
0.3752
-1.9767
-0.2752
0.4835
0.0251
1.1876
-0.9967

0.0001
0.0390
0.0001
0.0001
0.0001
0.0289
0.0001
0.9315
0.0001
0.0001
0.7690
0.0001
0.0289
0.9276
0.0001
0.0001
0.0001
0.0204
0.9586
0.0001
0.0001
Yes
Yes
Yes

10.70%
-3.20%
-4.09%
3.63%
3.32%
-0.22%
2.43%
0.34%
0.85%
-0.25%
0.28%
4.98%
0.11%
-0.07%
11.80%
-2.63%
-1.66%
0.34%
0.18%
3.80%
-2.78%

16,237
6.83%

This table reports the results from estimating a logistic discrete-choice model of the borrower’s decision to refuse the
bank’s loan o¤er and to seek credit elsewhere by full-information maximum likelihood for the subsample of successful
loan applications (15,897 observations). As before, we use branch …xed e¤ects and clustered standard errors that are
adjusted for heteroskedasticity across branch o¢ ces and correlation within. The dependent variable is the applicant’s
decision to decline (Y = 1: 420 online and 915 in-person observations) or to accept (Y = 0: 2,715 online and 12,187
in-person observations) the bank’s o¤er; the explanatory variables are our usual proxies for public, proprietary, and private
information, bank-borrower relationship characteristics, …rm attributes, and various control variables. See Section 3 for a
description of the variables and the notes to Table 2 for further details.

36

Table 8: The Likelihood of Credit Delinquency
Speci…cation
Loan Type
Variable
Constant
eLoan (1eloan = 1)
ln(1+XSBI)
ln(1+Internal Score)
Private-Info. Res.
Scope
ln(1+M. on Books)
Scope PIR
ln(1+MOB) PIR
ln(1+M. in Business)
ln(1+Net Income)
ln(1+CSHPI)
ln(1+F-B Dist)
ln(1+F-C Dist)
Collateral
Primary Guarantor
SBA Guarantee
Term Loan
APR
ln(1+Loan Amount)
ln(1+Maturity)
ln(1+# Branches)
ln(1+# Competitors)
UST Yield
Term Spread
4 Quarterly Dum.
8 State Dummies
38 SIC Dummies
Number of Obs
Pseudo R2

1

2

Coe¤

eLoans
P -val

In-Person Loans
Coe¤
P -val
Marg
-1.1865 0.0001

Marg

1.1476
-1.8501
-0.2720

0.0001
0.0001
0.0001

2.72%
-19.97%
-4.49%

-0.9446
-0.4469

0.0001
0.0001

-21.35%
-9.82%

-0.4961
-0.9643

0.0001
0.0289

-1.06%
-0.96%

-0.7620
-0.3010

0.0001
0.0001

-2.78%
-3.50%

-0.8820
-0.5422
-0.8409
0.2819
-0.7299
-0.5385
-0.3894
2.9317
0.2625
2.0545
-0.9771
-0.4464
2.7084
3.7498
0.4750
1.1993

0.0582
0.0001
0.0001
0.5480
0.3930
0.0001
0.0001
0.0001
0.0001
0.0001
0.0001
0.0001
0.0001
0.0001
0.2840
0.0001
Yes
Yes
Yes

-2.87%
-2.18%
-0.63%
0.11%
-0.04%
-1.43%
-2.80%
0.33%
0.42%
4.99%
-8.57%
-1.09%
0.20%
0.51%
0.47%
1.47%

-0.0165
-0.0842
-0.0706
0.2339
-0.2175
-0.1912
-0.5462
0.5753
0.6510
1.1254
-1.5777
-0.8171
0.1071
0.1644
0.4661
1.8662

0.8617
0.0001
0.0001
0.3728
0.5770
0.0001
0.0001
0.0001
0.0001
0.0191
0.0001
0.0001
0.0001
0.0001
0.0001
0.0001
Yes
Yes
Yes

-3.24%
-1.97%
-0.49%
0.12%
-0.04%
-1.91%
-1.31%
2.95%
0.26%
7.06%
-9.79%
-1.44%
0.37%
0.12%
0.71%
1.43%

14,902
12.44%

Coe¤

eLoans
P -val

Marg

In-Person Loans
Coe¤
P -val
Marg
-1.1872 0.0001

1.1523
-1.7680

0.0001
0.0001

2.95%
-22.79%

-0.9028

0.0001

-20.31%

-0.2656
-0.5133
-1.0539
-0.5721
-0.4899
-0.8561
-0.5806
-0.9012
0.2674
-0.6994
-0.5871
-0.3952
3.0405
0.2733
1.9412
-1.0305
-0.4987
2.6964
3.8523
0.4515
1.2402

0.0001
0.0001
0.0001
0.0001
0.3507
0.0770
0.0001
0.0070
0.6844
0.6510
0.0001
0.0001
0.0001
0.0001
0.0001
0.0001
0.0001
0.0001
0.0001
0.3355
0.0001
Yes
Yes
Yes

-4.12%
-1.20%
-0.91%
-1.84%
-0.32%
-3.50%
-2.59%
-0.70%
0.12%
-0.04%
-1.82%
-2.48%
0.23%
0.59%
4.97%
-10.88%
-1.29%
0.25%
0.47%
0.42%
1.71%

-0.1027
-0.7598
-0.3225
-0.3474
-0.2096
-0.0172
-0.0919
-0.0736
0.2529
-0.2308
-0.1967
-0.5377
0.5554
0.6747
1.1269
-1.5018
-0.8007
0.1146
0.1691
0.4427
1.8641

0.0001
0.0001
0.0001
0.0001
0.0140
0.8720
0.0001
0.0001
0.2015
0.4900
0.0001
0.0001
0.0001
0.0001
0.0121
0.0001
0.0001
0.0001
0.0001
0.0001
0.0001
Yes
Yes
Yes

-12.88%
-2.94%
-3.18%
-3.38%
-1.62%
-3.83%
-1.83%
-0.47%
0.19%
-0.08%
-2.35%
-1.63%
3.20%
0.34%
7.02%
-9.37%
-1.59%
0.41%
0.15%
0.75%
1.09%

14,902
12.11%

This table reports the results from estimating a logistic model of the likelihood that a loan becomes 60 days overdue
within 18 months of origination by full-information maximum likelihood for the subsample of actual loans booked by
the bank (14,613 observations). Again, we use branch …xed e¤ects and clustered standard errors that are adjusted for
heteroskedasticity across branch o¢ ces and correlation within. The dependent variable is the performance status of the
loan during its …rst 18 months: at most 60 days overdue (corresponding to our bank’s internal de…nition of a delinquent
loan Y = 1: 91 and 319 online and in-person observations, respectively), or current (Y = 0: 2,624 and 11,868 online
and in-person observations, respectively). The explanatory variables are our proxies for public, proprietary, and private
information, bank-borrower relationship characteristics, …rm attributes, and various control variables; see Section 3 for a
description of the variables and the notes to Table 2 for further details.

37

References
[1] Agarwal, S. and R. Hauswald (2006), “Distance and Information Asymmetries in Lending,”
mimeo, FRB of Chicago and American University.
[2] Anand, B. and A. Galetovic (2006), “Relationships, Competition and the Structure of Investment Banking Markets,” Journal of Industrial Economics 54: 151-199.
[3] Berger, A., W. Frame and N.Miller (2005), “Credit Scoring and the Availability, Price, and
Risk of Small Business Credit,” Journal of Money, Credit and Banking 37: 191-222.
[4] Berger, A., N. Miller, M. Petersen, R. Rajan and J. Stein (2005), “Does Function Follow Organizational Form? Evidence from the Lending Practices of Large and Small Banks,”Journal
of Financial Economics 76: 237-269.
[5] Berger, A. and G. Udell (1995), “Relationship Lending and Lines of Credit in Small Firm
Finance,” Journal of Business 68: 351-382.
[6] Bharath, S., S. Dahiya, A. Saunders and A. Srinivasan (2006), “So What Do I Get? The
Bank’s View of Lending Relationships,” forthcoming Journal of Financial Economics.
[7] Bhattacharya, S. and G. Chiesa (1995), “Proprietary Information, Financial Intermediation
and Research Incentives,” Journal of Financial Intermediation 4: 328-357.
[8] Bonaccorsi di Patti, E., G. Gobbi and P.E. Mistrulli (2004), “Testing for Complementarity
between Stores and E-Commerce: The Case of Banking Services,” mimeo, Banca d’Italia.
[9] Boot, A. (2000), “Relationship Banking: What Do We Know?” Journal of Financial Intermediation 9: 7–25.
[10] Boot, A. and A. Thakor (2000), “Can Relationship Banking Survive Competition?" Journal
of Finance 55: 679-713.
[11] Boot, A. and A. Schmeits (2005), “The Competitive Challenge in Banking,”Amsterdam Center
for Law & Economics Working Paper No. 2005-08.
[12] Booth, J. and L.C. Booth (2006), “Loan Collateral Decisions and Corporate Borrowing
Costs,"Journal of Money Credit and Banking 38: 67-90.
[13] Broecker, T. (1990), “Credit-Worthiness Tests and Interbank Competition,”Econometrica 58:
429-452.
[14] Carey, M., M. Post, and S. Sharpe (1998), “Does Corporate Lending by Banks and Finance
Companies Di¤er? Evidence on Specialization in Private Debt Contracting,” Journal of Finance 53: 845-878.
[15] Case, K.E., and R.J. Shiller (1987), “Prices of Single-Family Homes since 1970: New Indexes
for Four Cities,” New England Economic Review September /October.
[16] Case, K.E. and R.J. Shiller (1989), “The E¢ ciency of the Market for Single-Family Homes,"
American Economic Review 79: 125-137.
[17] Degryse, H. and S. Ongena (2005), “Distance, Lending Relationships, and Competition," Journal of Finance 60: 231-266.
38

[18] Degryse, H. and P. Van Cayseele (2000). “Relationship Lending within a Bank-Based System:
Evidence from European Small Business Data,”Journal of Financial Intermediation 9: 90-109.
[19] Detragiache, E., P. Garella and L. Guiso (2000), “Multiple versus Single Banking Relationships:
Theory and Evidence," Journal of Finance 55: 1133-1161.
[20] DeYoung, R., D. Glennon and P. Nigro, (2004), “Borrower-Lender Distance, Credit Scoring,
and the Performance of Small Business Loans,” mimeo, Federal Reserve Bank of Chicago.
[21] DeYoung, R. (2005), “The Performance of Internet-based Business Models: Evidence from the
Banking Industry,” forthcoming Journal of Business.
[22] Elsas, R. (2005), “Empirical Determinants of Relationship Lending,” Journal of Financial
Intermediation 14: 32–57.
[23] Experian (2000), Small Business Intelliscore, Experian Information Solutions, Inc. September
2000; available at www.experian.com.
[24] Experian (2006), Predicting Risk: The Relationship between Business and Consumer Scores,
Experian Information Solutions, Inc. June 2006; available at www.experian.com.
[25] Farinha , M. and J. Santos, (2002), “Switching from Single to Multiple Bank Lending Relationships: Determinants and Implications,” Journal of Financial Intermediation 11: 124-151.
[26] Fuentes, R., R. Hernandez-Murillo and G. Llobet (2006), “Strategic Online-Banking Adoption,” Federal Reserve Bank of St. Louis WP 2006-058A.
[27] Gehrig, T., (1998), “Screening, Cross-Border Banking, and the Allocation of Credit," Research
in Economics 52: 387-407.
[28] Hauswald, R. and R. Marquez (2003), “Information Technology and Financial Services Competition,” Review of Financial Studies, 16: 921-948.
[29] Hauswald, R. and R. Marquez (2006), “Competition and Strategic Information Acquisition in
Credit Markets,” Review of Financial Studies 19: 967-1000.
[30] Inderst, R. and H. Müller (2006), “A Lender-Based Theory of Collateral,”forthcoming Journal
of Financial Economics.
[31] Iyer, R. and M. Puri (2007), “Who Runs? The Importance of Relationships in Bank Panics,”
mimeo, Duke University.
[32] James, C. (1987), “Some Evidence on the Uniqueness of Bank Loans,” Journal of Financial
Economics 19: 217-235.
[33] Lummer, S. and J. McConnell, (1989), “Further Evidence on the Bank Lending Process and
the Capital Market Response to Bank Loan Agreements,”Journal of Financial Economics 25:
99-122.
[34] Mara, J. (2004), “SmartView Launches on Yahoo! Maps,” ClickZ Internet Advertising News
March 9, 2004; available at www.clickz.com/news/print.php/3323441.
[35] Mester, L., L. Nakamura and M. Renault (2007), “Transactions Accounts and Loan Monitoring,” Review of Financial Studies 20: 529-556.
39

[36] Petersen, M. (2004), “Information: Hard and Soft,” mimeo, Northwestern University.
[37] Petersen, M. and Rajan, R., (1994), “The Bene…ts of Lending Relationships: Evidence from
Small Business Data,” Journal of Finance 49: 3-37.
[38] Petersen, M. and Rajan, R. (1995), “The E¤ect of Credit Market Competition on Lending
Relationships," Quarterly Journal of Economics 110: 407-443.
[39] Petersen, M. and R. Rajan (2002), “Does Distance Still Matter? The Information Revolution
in Small Business Lending,” Journal of Finance 57: 2533-2570.
[40] Rajan, R., (1992), “Insiders and Outsiders: The Choice between Informed and Arm’s-Length
Debt," Journal of Finance 47: 1367-1400.
[41] Schenone, C. (2007), “Lending Relationships and Information Rents: Do Banks Exploit Their
Information Advantages?” mimeo, University of Virginia.
[42] Sha¤er, S., (1998), “The Winner’s Curse in Banking," Journal of Financial Intermediation 7:
359-392.
[43] Sharpe, S., (1990), “Asymmetric Information, Bank Lending and Implicit Contracts: A Stylized Model of Customer Relationships," Journal of Finance 45: 1069-1087.
[44] von Thadden, E.-L., (2004), “Asymmetric Information, Bank Lending and Implicit Contracts:
The Winner’s Curse,” Finance Research Letters 1: 11-23.
[45] Wilhelm, W. (1999), “Internet Investment Banking: The Impact of Information Technology
on Relationship Banking,” Journal of Applied Corporate Finance 12, Spring 1999.
[46] Wilhelm, W. (2001), “The Internet and Financial Market Structure," Oxford Review of Economic Policy 17: 235-247.

40

Working Paper Series
A series of research studies on regional economic issues relating to the Seventh Federal
Reserve District, and on financial and economic topics.
Firm-Specific Capital, Nominal Rigidities and the Business Cycle
David Altig, Lawrence J. Christiano, Martin Eichenbaum and Jesper Linde

WP-05-01

Do Returns to Schooling Differ by Race and Ethnicity?
Lisa Barrow and Cecilia Elena Rouse

WP-05-02

Derivatives and Systemic Risk: Netting, Collateral, and Closeout
Robert R. Bliss and George G. Kaufman

WP-05-03

Risk Overhang and Loan Portfolio Decisions
Robert DeYoung, Anne Gron and Andrew Winton

WP-05-04

Characterizations in a random record model with a non-identically distributed initial record
Gadi Barlevy and H. N. Nagaraja

WP-05-05

Price discovery in a market under stress: the U.S. Treasury market in fall 1998
Craig H. Furfine and Eli M. Remolona

WP-05-06

Politics and Efficiency of Separating Capital and Ordinary Government Budgets
Marco Bassetto with Thomas J. Sargent

WP-05-07

Rigid Prices: Evidence from U.S. Scanner Data
Jeffrey R. Campbell and Benjamin Eden

WP-05-08

Entrepreneurship, Frictions, and Wealth
Marco Cagetti and Mariacristina De Nardi

WP-05-09

Wealth inequality: data and models
Marco Cagetti and Mariacristina De Nardi

WP-05-10

What Determines Bilateral Trade Flows?
Marianne Baxter and Michael A. Kouparitsas

WP-05-11

Intergenerational Economic Mobility in the U.S., 1940 to 2000
Daniel Aaronson and Bhashkar Mazumder

WP-05-12

Differential Mortality, Uncertain Medical Expenses, and the Saving of Elderly Singles
Mariacristina De Nardi, Eric French, and John Bailey Jones

WP-05-13

Fixed Term Employment Contracts in an Equilibrium Search Model
Fernando Alvarez and Marcelo Veracierto

WP-05-14

1

Working Paper Series (continued)
Causality, Causality, Causality: The View of Education Inputs and Outputs from Economics
Lisa Barrow and Cecilia Elena Rouse

WP-05-15

Competition in Large Markets
Jeffrey R. Campbell

WP-05-16

Why Do Firms Go Public? Evidence from the Banking Industry
Richard J. Rosen, Scott B. Smart and Chad J. Zutter

WP-05-17

Clustering of Auto Supplier Plants in the U.S.: GMM Spatial Logit for Large Samples
Thomas Klier and Daniel P. McMillen

WP-05-18

Why are Immigrants’ Incarceration Rates So Low?
Evidence on Selective Immigration, Deterrence, and Deportation
Kristin F. Butcher and Anne Morrison Piehl

WP-05-19

Constructing the Chicago Fed Income Based Economic Index – Consumer Price Index:
Inflation Experiences by Demographic Group: 1983-2005
Leslie McGranahan and Anna Paulson

WP-05-20

Universal Access, Cost Recovery, and Payment Services
Sujit Chakravorti, Jeffery W. Gunther, and Robert R. Moore

WP-05-21

Supplier Switching and Outsourcing
Yukako Ono and Victor Stango

WP-05-22

Do Enclaves Matter in Immigrants’ Self-Employment Decision?
Maude Toussaint-Comeau

WP-05-23

The Changing Pattern of Wage Growth for Low Skilled Workers
Eric French, Bhashkar Mazumder and Christopher Taber

WP-05-24

U.S. Corporate and Bank Insolvency Regimes: An Economic Comparison and Evaluation
Robert R. Bliss and George G. Kaufman

WP-06-01

Redistribution, Taxes, and the Median Voter
Marco Bassetto and Jess Benhabib

WP-06-02

Identification of Search Models with Initial Condition Problems
Gadi Barlevy and H. N. Nagaraja

WP-06-03

Tax Riots
Marco Bassetto and Christopher Phelan

WP-06-04

The Tradeoff between Mortgage Prepayments and Tax-Deferred Retirement Savings
Gene Amromin, Jennifer Huang,and Clemens Sialm

WP-06-05

2

Working Paper Series (continued)
Why are safeguards needed in a trade agreement?
Meredith A. Crowley

WP-06-06

Taxation, Entrepreneurship, and Wealth
Marco Cagetti and Mariacristina De Nardi

WP-06-07

A New Social Compact: How University Engagement Can Fuel Innovation
Laura Melle, Larry Isaak, and Richard Mattoon

WP-06-08

Mergers and Risk
Craig H. Furfine and Richard J. Rosen

WP-06-09

Two Flaws in Business Cycle Accounting
Lawrence J. Christiano and Joshua M. Davis

WP-06-10

Do Consumers Choose the Right Credit Contracts?
Sumit Agarwal, Souphala Chomsisengphet, Chunlin Liu, and Nicholas S. Souleles

WP-06-11

Chronicles of a Deflation Unforetold
François R. Velde

WP-06-12

Female Offenders Use of Social Welfare Programs Before and After Jail and Prison:
Does Prison Cause Welfare Dependency?
Kristin F. Butcher and Robert J. LaLonde
Eat or Be Eaten: A Theory of Mergers and Firm Size
Gary Gorton, Matthias Kahl, and Richard Rosen
Do Bonds Span Volatility Risk in the U.S. Treasury Market?
A Specification Test for Affine Term Structure Models
Torben G. Andersen and Luca Benzoni

WP-06-13

WP-06-14

WP-06-15

Transforming Payment Choices by Doubling Fees on the Illinois Tollway
Gene Amromin, Carrie Jankowski, and Richard D. Porter

WP-06-16

How Did the 2003 Dividend Tax Cut Affect Stock Prices?
Gene Amromin, Paul Harrison, and Steven Sharpe

WP-06-17

Will Writing and Bequest Motives: Early 20th Century Irish Evidence
Leslie McGranahan

WP-06-18

How Professional Forecasters View Shocks to GDP
Spencer D. Krane

WP-06-19

Evolving Agglomeration in the U.S. auto supplier industry
Thomas Klier and Daniel P. McMillen

WP-06-20

3

Working Paper Series (continued)
Mortality, Mass-Layoffs, and Career Outcomes: An Analysis using Administrative Data
Daniel Sullivan and Till von Wachter
The Agreement on Subsidies and Countervailing Measures:
Tying One’s Hand through the WTO.
Meredith A. Crowley

WP-06-21

WP-06-22

How Did Schooling Laws Improve Long-Term Health and Lower Mortality?
Bhashkar Mazumder

WP-06-23

Manufacturing Plants’ Use of Temporary Workers: An Analysis Using Census Micro Data
Yukako Ono and Daniel Sullivan

WP-06-24

What Can We Learn about Financial Access from U.S. Immigrants?
Una Okonkwo Osili and Anna Paulson

WP-06-25

Bank Imputed Interest Rates: Unbiased Estimates of Offered Rates?
Evren Ors and Tara Rice

WP-06-26

Welfare Implications of the Transition to High Household Debt
Jeffrey R. Campbell and Zvi Hercowitz

WP-06-27

Last-In First-Out Oligopoly Dynamics
Jaap H. Abbring and Jeffrey R. Campbell

WP-06-28

Oligopoly Dynamics with Barriers to Entry
Jaap H. Abbring and Jeffrey R. Campbell

WP-06-29

Risk Taking and the Quality of Informal Insurance: Gambling and Remittances in Thailand
Douglas L. Miller and Anna L. Paulson

WP-07-01

Fast Micro and Slow Macro: Can Aggregation Explain the Persistence of Inflation?
Filippo Altissimo, Benoît Mojon, and Paolo Zaffaroni

WP-07-02

Assessing a Decade of Interstate Bank Branching
Christian Johnson and Tara Rice

WP-07-03

Debit Card and Cash Usage: A Cross-Country Analysis
Gene Amromin and Sujit Chakravorti

WP-07-04

The Age of Reason: Financial Decisions Over the Lifecycle
Sumit Agarwal, John C. Driscoll, Xavier Gabaix, and David Laibson

WP-07-05

Information Acquisition in Financial Markets: a Correction
Gadi Barlevy and Pietro Veronesi

WP-07-06

Monetary Policy, Output Composition and the Great Moderation
Benoît Mojon

WP-07-07

4

Working Paper Series (continued)
Estate Taxation, Entrepreneurship, and Wealth
Marco Cagetti and Mariacristina De Nardi

WP-07-08

Conflict of Interest and Certification in the U.S. IPO Market
Luca Benzoni and Carola Schenone

WP-07-09

The Reaction of Consumer Spending and Debt to Tax Rebates –
Evidence from Consumer Credit Data
Sumit Agarwal, Chunlin Liu, and Nicholas S. Souleles

WP-07-10

Portfolio Choice over the Life-Cycle when the Stock and Labor Markets are Cointegrated
Luca Benzoni, Pierre Collin-Dufresne, and Robert S. Goldstein

WP-07-11

Nonparametric Analysis of Intergenerational Income Mobility
with Application to the United States
Debopam Bhattacharya and Bhashkar Mazumder

WP-07-12

How the Credit Channel Works: Differentiating the Bank Lending Channel
and the Balance Sheet Channel
Lamont K. Black and Richard J. Rosen

WP-07-13

Labor Market Transitions and Self-Employment
Ellen R. Rissman

WP-07-14

First-Time Home Buyers and Residential Investment Volatility
Jonas D.M. Fisher and Martin Gervais

WP-07-15

Establishments Dynamics and Matching Frictions in Classical Competitive Equilibrium
Marcelo Veracierto

WP-07-16

Technology’s Edge: The Educational Benefits of Computer-Aided Instruction
Lisa Barrow, Lisa Markman, and Cecilia Elena Rouse

WP-07-17

The Widow’s Offering: Inheritance, Family Structure, and the Charitable Gifts of Women
Leslie McGranahan

WP-07-18

Demand Volatility and the Lag between the Growth of Temporary
and Permanent Employment
Sainan Jin, Yukako Ono, and Qinghua Zhang

WP-07-19

A Conversation with 590 Nascent Entrepreneurs
Jeffrey R. Campbell and Mariacristina De Nardi

WP-07-20

Cyclical Dumping and US Antidumping Protection: 1980-2001
Meredith A. Crowley

WP-07-21

The Effects of Maternal Fasting During Ramadan on Birth and Adult Outcomes
Douglas Almond and Bhashkar Mazumder

WP-07-22

5

Working Paper Series (continued)
The Consumption Response to Minimum Wage Increases
Daniel Aaronson, Sumit Agarwal, and Eric French

WP-07-23

The Impact of Mexican Immigrants on U.S. Wage Structure
Maude Toussaint-Comeau

WP-07-24

A Leverage-based Model of Speculative Bubbles
Gadi Barlevy

WP-08-01

Displacement, Asymmetric Information and Heterogeneous Human Capital
Luojia Hu and Christopher Taber

WP-08-02

BankCaR (Bank Capital-at-Risk): A credit risk model for US commercial bank charge-offs
Jon Frye and Eduard Pelz

WP-08-03

Bank Lending, Financing Constraints and SME Investment
Santiago Carbó-Valverde, Francisco Rodríguez-Fernández, and Gregory F. Udell

WP-08-04

Global Inflation
Matteo Ciccarelli and Benoît Mojon

WP-08-05

Scale and the Origins of Structural Change
Francisco J. Buera and Joseph P. Kaboski

WP-08-06

Inventories, Lumpy Trade, and Large Devaluations
George Alessandria, Joseph P. Kaboski, and Virgiliu Midrigan

WP-08-07

School Vouchers and Student Achievement: Recent Evidence, Remaining Questions
Cecilia Elena Rouse and Lisa Barrow

WP-08-08

Does It Pay to Read Your Junk Mail? Evidence of the Effect of Advertising on
Home Equity Credit Choices
Sumit Agarwal and Brent W. Ambrose
The Choice between Arm’s-Length and Relationship Debt: Evidence from eLoans
Sumit Agarwal and Robert Hauswald

WP-08-09

WP-08-10

6