Experience Seamless Full Ebook Downloads for Every Genre at textbookfull.

com

Banking and Financial Markets: How Banks and

Financial Technology Are Reshaping Financial
Markets Andrada Bilan

https://textbookfull.com/product/banking-and-financial-
markets-how-banks-and-financial-technology-are-reshaping-
financial-markets-andrada-bilan/

OR CLICK BUTTON

DOWNLOAD NOW

Explore and download more ebook at https://textbookfull.com

 Recommended digital products (PDF, EPUB, MOBI) that
you can download immediately if you are interested.

The Economics of Money Banking Financial Markets Massimo

Giuliodori

https://textbookfull.com/product/the-economics-of-money-banking-
financial-markets-massimo-giuliodori/

textboxfull.com

Financial markets and institutions Eakins

https://textbookfull.com/product/financial-markets-and-institutions-
eakins/

textboxfull.com

Price and Financial Stability Rethinking Financial Markets

1st Edition David Harrison

https://textbookfull.com/product/price-and-financial-stability-
rethinking-financial-markets-1st-edition-david-harrison/

textboxfull.com

Corporations Financial Markets Law Redmond

https://textbookfull.com/product/corporations-financial-markets-law-
redmond/

textboxfull.com
Emerging Research on Monetary Policy, Banking, and
Financial Markets 1st Edition Cristi Spulbar

https://textbookfull.com/product/emerging-research-on-monetary-policy-
banking-and-financial-markets-1st-edition-cristi-spulbar/

textboxfull.com

The economics of money banking and financial markets

Eleventh Edition, Global Edition Mishkin

https://textbookfull.com/product/the-economics-of-money-banking-and-
financial-markets-eleventh-edition-global-edition-mishkin/

textboxfull.com

Financial Markets and Institutions 5th Edition Coll.

https://textbookfull.com/product/financial-markets-and-
institutions-5th-edition-coll/

textboxfull.com

The Economics of Money, Banking and Financial Markets,

Seventh Canadian Edition Frederic S. Mishkin

https://textbookfull.com/product/the-economics-of-money-banking-and-
financial-markets-seventh-canadian-edition-frederic-s-mishkin/

textboxfull.com

EU securities and financial markets regulation Niamh

Moloney

https://textbookfull.com/product/eu-securities-and-financial-markets-
regulation-niamh-moloney/

textboxfull.com
PALGRAVE MACMILLAN STUDIES IN
BANKING AND FINANCIAL INSTITUTIONS
SERIES EDITOR: PHILIP MOLYNEUX

Banking and
Financial Markets
How Banks and Financial Technology
Are Reshaping Financial Markets

Andrada Bilan · Hans Degryse

Kuchulain O’Flynn · Steven Ongena
Palgrave Macmillan Studies in Banking
and Financial Institutions

Series Editor
Philip Molyneux
University of Sharjah
Sharjah, United Arab Emirates
The Palgrave Macmillan Studies in Banking and Financial Institutions
series is international in orientation and includes studies of banking systems
in particular countries or regions as well as contemporary themes such as
Islamic Banking, Financial Exclusion, Mergers and Acquisitions, Risk
Management, and IT in Banking. The books focus on research and practice
and include up to date and innovative studies that cover issues which impact
banking systems globally.

More information about this series at

http://www.palgrave.com/gp/series/14678
 Andrada Bilan • Hans Degryse •
Kuchulain O’Flynn • Steven Ongena

Banking and Financial

Markets
How Banks and Financial Technology Are
Reshaping Financial Markets
Andrada Bilan Hans Degryse
Department of Banking and Finance Faculty of Economics and Business
University of Zurich Katholieke University of Leuven
Zürich, Switzerland Leuven, Belgium

Kuchulain O’Flynn Steven Ongena

Department of Banking and Finance Department of Banking and Finance
University of Zurich University of Zurich
Zürich, Switzerland Zürich, Switzerland

ISSN 2523-336X ISSN 2523-3378 (electronic)

Palgrave Macmillan Studies in Banking and Financial Institutions
ISBN 978-3-030-26843-5 ISBN 978-3-030-26844-2 (eBook)
https://doi.org/10.1007/978-3-030-26844-2

© The Editor(s) (if applicable) and The Author(s), under exclusive licence to Springer
Nature Switzerland AG 2019
This work is subject to copyright. All rights are solely and exclusively licensed by the Publisher,
whether the whole or part of the material is concerned, specifically the rights of translation,
reprinting, reuse of illustrations, recitation, broadcasting, reproduction on microfilms or in
any other physical way, and transmission or information storage and retrieval, electronic
adaptation, computer software, or by similar or dissimilar methodology now known or
hereafter developed.
The use of general descriptive names, registered names, trademarks, service marks, etc. in this
publication does not imply, even in the absence of a specific statement, that such names are
exempt from the relevant protective laws and regulations and therefore free for general use.
The publisher, the authors, and the editors are safe to assume that the advice and information
in this book are believed to be true and accurate at the date of publication. Neither the
publisher nor the authors or the editors give a warranty, expressed or implied, with respect
to the material contained herein or for any errors or omissions that may have been made.
The publisher remains neutral with regard to jurisdictional claims in published maps and
institutional affiliations.

Cover illustration: Credit Easyturn / E+ / Getty

This Palgrave Macmillan imprint is published by the registered company Springer Nature
Switzerland AG.
The registered company address is: Gewerbestrasse 11, 6330 Cham, Switzerland
 CONTENTS

1 Introduction 1

2 Securitization and Lending 5

3 Interest Rate Risk 31

4 Credit Risk 61

5 Collateral and Lending 105

6 Global Banking 133

7 FinTech and the Future of Banking 179

8 Conclusion 201

References 205

Index 219

v
 LIST OF FIGURES

Fig. 4.1 CDS Contract Payment Diagram

Note: This figure illustrates the payment structure of a CDS
contract. A protection buyer (buyer of the CDS) makes
periodic payments (CDS spread) to the protection seller
(seller of the CDS). The protection seller pays the buyer a
protection value if a “credit event” occurs. The definition of
a “credit event” depends on the standardized contract used,
which is determined based on the domicile country of the
underlying reference entity 64
Fig. 4.2 CDS credit events and bankruptcy law
Note: This figure illustrates the two conditions for the
positive and negative effects of CDS on an underlying firm
to be felt. The first is the ability to restructure a company
under the domicile country’s law. The second condition
only plays a role if the first condition is met. The second
condition is that the CDS contract type should not pay out
when the underlying debt is restructured. If restructuring
is not permitted under law, then the CDS contract type is
irrelevant and the firms have no incentive to strategically
default (as they will be liquidated). However, the incentive
for empty creditors to push a firm into bankruptcy remains. If
restructuring is permitted and defined as a credit event, in this
context, CDS has no effect on the underlying firm. Finally, if
restructuring is permitted and not defined as a credit event,
then CDS acts as a commitment device as well as increases
the probability of default through the effect of empty creditors 83

vii
viii LIST OF FIGURES

Fig. 6.1 Stylized global banking system diagram

Note: This stylized diagram of the global banking system
includes the mechanisms which the global banking literature
seeks to understand. Here, blue represents effects originating
in the home country banks or the foreign affiliates of these
banks. Green represents the same for foreign banks or local
affiliates of foreign banks. The yellow arrows represent the
global interbank market where home banks lend to other
local banks, foreign banks, and local affiliates of foreign banks.
An inward transmission of monetary policy is represented as
follows: the arrows from the foreign central bank, Central
Bank F, to the affiliate of the home bank, Affiliate of Bank
H, and the foreign bank, Bank F, represent the transmission
of monetary policy through these banks. These banks then
transmit this foreign monetary policy to Non-Bank Borrowers
in the home country, in green. This is done directly through
cross-border flows (the curved arrows) or through home
banks, Bank H, or local affiliates of foreign banks, Affiliate of
Bank F. The diagram also shows the outward transmission of
the monetary policy. This follows a similar path as the inward
transmission but where the monetary policy originates from
the home central bank, Central Bank H. The dashed vertical
line depicts the boarded between the home country and the
foreign country. The area around this dashed line represents
the friction of foreign exchange. Finally, the lightning bolts
represent shocks to the foreign bank, blue, or the home bank,
green. Here “shocks” can either be shocks to the assets of
these banks or changes in macroprudential policy 136
Fig. 6.2 Stylized global banking system diagram-Interbank Integration
Note: See Fig. 6.1 for a description of the mechanisms
depicted. The bold parts are the mechanisms considered in
this section; the opaque parts are not considered 144
Fig. 6.3 Stylized global banking system diagram-Banking Affiliates
Note: See Fig. 6.1 for a description of the mechanisms
depicted. The bold parts are the mechanisms considered in
this section; the opaque parts are not considered 150
Fig. 6.4 Stylized Global Banking System Diagram-Global Banks-Local
Funding Shocks
Note: See Fig. 6.1 for a description of the mechanisms
depicted. The bold parts are the mechanisms considered in
this section; the opaque parts are not considered 153
 LIST OF FIGURES ix

Fig. 6.5 Stylized Global Banking System Diagram-Inward

Transmission
Note: See Fig. 6.1 for a description of the mechanisms
depicted. The bold parts are the mechanisms considered in
this section; the opaque parts are not considered 155
Fig. 6.6 Stylized Global Banking System Diagram-Outward
Transmission
Note: See Fig. 6.1 for a description of the mechanisms
depicted. The bold parts are the mechanisms considered in
this section; the opaque parts are not considered 161
Fig. 6.7 Stylized Global Banking System Diagram-Macroprudential
Regulation
Note: See Fig. 6.1 for a description of the mechanisms
depicted. The bold parts are the mechanisms considered in
this section; the opaque parts are not considered 165
Fig. 7.1 The evolution of financial services
Note: Source: IMF (2019) This figure outlines the needs of
consumers for financial services, disaggregated by traditional
and fintech services 180
 LIST OF TABLES

Table 2.1 Securitization and lending 7

Table 3.1 Interest rate risk 56
Table 4.1 Credit risk 95
Table 5.1 Collateral and lending 107
Table 6.1 Global banking 168
Table 7.1 Fintech and lending 182

xi
 CHAPTER 1

Introduction

The traditional role of a bank is to transfer funds from savers to investors,

engaging in maturity transformation, screening for borrower risk, and
monitoring for borrower effort in doing so. Until not so long ago, a
traditional loan contract included as salient dimensions its amount, the
interest rate, its expected credit risk, the required collateral, and the
currency in which the loan was granted, all “wrapped” up in a singular
way of the bank-firm engagement that occurred mainly through the
loan officer. However, the scope of the modern banking industry today
is much broader, offering a range of sophisticated financial products, a
wider geography-including exposure to countries with various currencies,
regulation, and monetary policy regimes-and an increased reliance on
financial innovation and technology.
The new bank business models have had repercussions for the traditional
loan contract and ways of doing business as well. In particular, the main
components and risks of a loan contract can now be altered and hedged on
the market, by means of securitization, interest rate swaps, credit default
swaps, and foreign exchange transactions. Securitized loans can often
be pledged as collateral, thus facilitating new lending. And the lending
technology is evolving from one-to-one dialogues between a loan officer
and a borrower, at a bank branch, towards more informationally neutral
technologies such as peer-to-peer lending or crowd funding.
This book outlines in detail this transition from traditional to modern
banking. In six different chapters, we explore the effects of increased finan-

© The Author(s) 2019 1

A. Bilan et al., Banking and Financial Markets, Palgrave Macmillan
Studies in Banking and Financial Institutions,
https://doi.org/10.1007/978-3-030-26844-2_1
2 A. BILAN ET AL.

cial sophistication on a particular dimension of the loan contract: amount,

interest rate, credit risk, collateral, currency, and lending technology. We
evaluate the economic benefits and costs of this new financial ecosystem,
by relying on recent empirical research in banking and finance.
In Chap. 2, we address the impact on the loan amount that the
phenomenon of loan securitization has had. Securitization implies that
financial intermediaries that grant illiquid loans subsequently pool them
together, diversifying risks and converting them into liquid assets or asset-
backed securities (ABSs). These assets are then sold to outside investors,
in exchange for wholesale funding. While the largest share of ABSs cover
home mortgage loans, other types of loans have also been securitized,
including automobile credit, student loans, or corporate loans.
A liquid market for securitized assets has recognized benefits for the
banking industry, such as improving risk-sharing and reducing banks’ cost
of capital. The volume of credit to both households and firms should
increase as a result. However, the global financial crisis of 2007-2009
started in the securitized subprime mortgage market, which collapsed after
having accumulated unmanageable amounts of risks. The academic litera-
ture has since sought to understand this fact. The explanations advanced
include market imperfections such as increased informational frictions or
banks’ search for yield. Chapter 2 reviews the theories and estimates put
forward in the recent banking literature, to quantify these benefits and costs
of securitization on the volume and the quality of credit.
In Chap. 3, we study how modern finance has affected the interest
rates charged on loans, by enabling banks to hedge interest rates. In an
interest rate swap, banks exchange floating for fixed interest rates, against
a fee. Financial intermediation often exposes banks to interest rate risks
by creating mismatches in the maturity structure and repricing terms of
their assets and liabilities. To reduce these risks, banks use two types of
risk management practices: on-balance sheet and off-balance sheet risk
management. Through on-balance sheet risk management, banks actively
seek to match the maturities of their assets and liabilities. This chapter is
concerned with off-balance sheet risk management, which occurs when
banks trade derivatives in order to hedge interest rate risk.
The ability to use swaps to hedge against interest rate risks improves the
intermediation efficiency of banks, as it allows them to take on more credit
risk. However, the flip side is that hedging activities might also make banks’
lending policies less sensitive to changes in monetary policy. Understanding
to what extent hedging affects the transmission of monetary policy is crucial
 1 INTRODUCTION 3

for both market players and regulators. Chapter 3 reviews the existing
evidence and points towards areas where the interaction between hedging
activities and credit is still to be researched.
Chapter 4 discusses how banks these days can trade away credit risk
(i.e., the risk that the loan amount will not be returned due to borrower
financial distress), by using the Credit Default Swaps (CDS) market. This
is an over-the-counter market, where a limited number of large investment
banks create markets for buyers and sellers of credit risk, such as banks,
insurance companies, and mutual funds. Banks trade in the CDS market
for two reasons: to hedge existing credit exposures and to speculate on
credit risk. We review recent academic contributions that have sought to
explore the interaction between CDS and credit markets, mainly through
banks participation in both.
Empirical evidence shows that when liquidity in the CDS market
improves, banks use more swaps to hedge risky credit exposures. This
enables banks to engage in risk shifting and free up bank capital. Both
effects increase bank credit supply. However, allowing banks to trade in
both markets also generate negative externalities. Because banks become
“empty” creditors when they hedge loan exposures with CDS, accessing
bank credit ex ante might become more difficult or more expensive for
companies that are referenced in the CDS market. Finally, trading in the
CDS derivatives might also suffer if banks trade on private information,
which they acquire by forming relationships with borrowers. Chapter 4
assesses the empirical relevance of these hypotheses.
Chapter 5 reviews the different forms of collateral pledged in bank
loans and their role in easing transactions. The primary reason for the use
of collateral when banks grant loans is to reduce the effects of information
asymmetry between the bank and its borrower. In this case, the collateral is
usually made up of the borrowers’ tangible assets which the bank can take
possession of in case of default. At the core, banks offer different collateral
menus to screen for borrower risk: the lower the probability to default on
a loan, the more substantial the collateral the entrepreneur is willing to
pledge. Further, the collateral attached to a loan provides an incentive for
the bank to more diligently monitor the borrower. This, in turn, increases
the likelihood that the project is successful. Recently, also intangible assets
(e.g., intellectual property rights) have been pledged as collateral and have
improved access to credit markets. In Chap. 5, we assess the empirical
importance of real estate and movable and intangible collateral for loans
to households and firms.
4 A. BILAN ET AL.

Chapter 6 concerns the currency in which the loan amount is denom-

inated. When banks conduct business globally, both their sources of
funding, as well as their volumes of credit, can be expressed in a foreign
currency. Therefore, in addition to supply and demand shocks at home,
global banks are also exposed to economic fluctuations in foreign exchange
markets. Typically, banks hedge this risk ex ante by engaging in cross-
currency FX swaps. These swaps are simultaneous spot purchases and
forward sales of foreign currency, at the prevailing FX forward rates. How-
ever, changes in the forward rates will still affect bank lending decisions,
because they affect the cost of hedging of new loans. Increases in forward
FX rates will make lending in the foreign currency less attractive, with
the reverse being true for domestically headquartered banks. Moreover,
trading frictions and the liquidity of FX markets will also determine the
price and the availability of a hedge, with a direct impact on the relative
cost of lending in foreign markets. In Chap. 6, we review recent empirical
evidence linking the functioning of the FX market to bank lending activity.
In Chap. 7, we address the role of Fintech platforms, which are innova-
tions to the lending technology characterized by dynamiting traditional
customer relationships. Banks gather hard and soft information on their
borrowers, by establishing relationships in which they screen and monitor
their borrowers. This can have both negative effects-when the bank charges
excessive interest rates due to informational rents-and positive effects-by
guaranteeing access to credit to relationship borrowers, even in times of
distress. These two sides of bank-customer interactions are under challenge
by FinTech. A range of innovative platforms, including payment service
providers, aggregators and robo advisors, and peer-to-peer lenders have
the potential to interfere with these traditional links.
On the one hand, the increased competition that FinTech brings might
reduce excessive bank rents. For example, peer-to-peer lending provides a
direct alternative to retail banking, with the potential to increase affordable
credit to small businesses. On the other hand, while FinTech might
decrease transaction costs and borrower choice, it might lead to less
relationship lending in the economy. This could hurt bank profitability,
but, more importantly, it could also hurt access to credit, particularly in
downturns, when stable credit is needed the most. Chapter 7 reviews recent
academic and policy thinking on how FinTech might affect the creation of
value within the bank-borrower relationship.
 CHAPTER 2

Securitization and Lending

While securitization dates back to the seventeenth and the eighteenth

century in Holland (Goetzmann and Rouwenhorst 2008), its massive
usage is a relatively recent development in credit markets. Securitization
implies that financial intermediaries that grant illiquid loans subsequently
pool them together, diversifying risks and converting the loans into liquid
assets or asset-backed securities (ABSs). These assets are then sold to
outside investors, in exchange for wholesale funding. While the largest
share of asset-backed securities cover individual mortgage loans, other
types of loans can also be securitized (including automobile credit, student
loans, or corporate loans).
At the end of 2007, the US-securitized mortgage loan market reached
$6.42 trillion (Loutskina 2011). A liquid market for securitized assets
has recognized benefits for credit markets, such as improving risk-sharing
and reducing banks’ cost of capital (Pennacchi 1988). Credit to both
households and firms should grow as a result. However, the global financial
crisis of 2007-2009 uncovered ways in which the process of securitization
could also hurt lending. It is well understood that the crisis started in the
securitized subprime loan market, which collapsed after having accumu-
lated unmanageable amounts of risks. This prompted both researchers in
finance and policymakers to seek a better understanding of the benefits but
also the vulnerabilities of the markets for asset-backed securities. Below,
we review some of the recent empirical literature addressing this need.
As with the remaining chapters, we first review the main data sources

© The Author(s) 2019 5

A. Bilan et al., Banking and Financial Markets, Palgrave Macmillan
Studies in Banking and Financial Institutions,
https://doi.org/10.1007/978-3-030-26844-2_2
6 A. BILAN ET AL.

used by empirical researchers studying securitization and then we discuss

the most common methodologies employed, before presenting selected
contributions. Table 2.1 offers a compact overview of the papers reviewed.

2.1 DATA
Without aiming to be exhaustive, we go over some of the main sources
of data that are used in this literature. Most of the data sets have a
panel structure, with a time and a cross-section dimension, typically across
banks, firms, or loans. To study the banking market, researchers usually
employ bank financial indicators containing information extracted from
balance sheets and income statements. Such financial data is available
either through regulatory reports (e.g., the Federal Reserve’s Report of
Condition and Income or “call reports”) or from private data providers
(such as SNL Financial). Indicators for the intensity of securitization can
be obtained from private data provider Dealogic, which reports banks’
involvement in different types of securitized assets (mortgages, corporate
loans, covered bonds) or from rating agencies such as Moody’s. Trading
data involving asset-backed commercial paper or bonds can be found
either in trade repositories, such as the Depository Trust and Clearing
Corporation (DTCC), or by approaching the dealer banks that make
markets for these products.
For loan-level analyses, the source of data depends on the type of
credit product concerned. Most of the empirical research on mortgage
lending has focused on the US market, where data is available either
from the Home Mortgage Disclosure Act (HMDA) or the First American
CoreLogic LoanPerformance databases. Detailed information is available,
including mortgage rates, loan size, maturity, year of origination, contract
types (hybrid, fixed-rate, adjustable rate), default information, as well as
borrower characteristics (credit scores, debt-to-income ratios) and loan
application status (denied, approved, originated). Data aggregated at ZIP
code level is also available from Equifax, another data provider that
maintains consumer credit history and related indicators on US house-
holds. Equifax provides annual aggregated data for outstanding credit and
amounts in default, broken down by type of loan: mortgages, home equity
lines, credit card debt, auto loans, student loans, and consumer loans.
Loan-level data on corporate credit has come from a variety of sources.
In the US, the most widely used data source is the syndicated loan market
Table 2.1 Securitization and lending

Paper Research question Data Methodology Main findings

Loutskina (JFE, Does securitization Bank financial statements Correlations: between From 1976 to 2007, as
2011) change bank liquidity extracted from the Federal the intensity of banks’ ability to securitize
management? Reserve call reports in a securitization and bank loans has increased, the
long-term panel (1976 to on-balance sheet liquidity percentage of total assets
2007) indicators. Enhances with held as liquid securities
difference-in-differences decreased on average by
specifications around 7.33 percentage points
regulatory interventions
Maddaloni and Does securitization European data from the Exogenous Variation in Low monetary policy
Peydró (RFS, impact the Bank Lending Survey monetary policy driven by short-term interest rates
2011) transmission of (BLS) on bank the fact that the soften lending standards
monetary policy or of self-reported lending Euro-wide target rate is for both firms and
banking regulation? volumes across countries relatively orthogonal to households, which is
the individual countries’ amplified by high
macroeconomic securitization activity
conditions
Acharya et al. Moody’s rating reports on Correlations and Increasing conduit
(JFE, 2013) conduits with information cross-sectional variation exposure from 0% to
on sponsors, type, assets, across banks with 100% of bank equity
guarantees, and ratings; different levels of capital reduced the bank’s stock
ABCP transactions and different exposures to return during a three-day
collected by the losses on conduits event window at the
Depository Trust and beginning of the financial
Clearing Corporation crisis by 1.1%
(DTCC); bank financial
data from Bankscope
2 SECURITIZATION AND LENDING

(continued)
7
 8

Table 2.1 (continued)

Paper Research question Data Methodology Main findings

Demyanyk and Was lax screening by Loan-level data on half of Proportional odds Adjusting for observed
van Hemert banks a reason for the all US subprime mortgage duration models, characteristics and
(RFS, 2011) distress in subprime contracts and estimating the macroeconomic conditions,
A. BILAN ET AL.

mortgages that led to characteristics from probability of default on loan quality deteriorated
the 2007-2009 LoanPerformance, a loan controlling for progressively between 2001
financial crisis? originated over 2001 to observables and 2007, but this was
2007 masked by a simultaneous
boom in house prices
Mian and Sufi Data on mortgage Correlations, across Over 2002 to 2005,
(QJE, 2009) lending activity from priors derived from subprime US ZIP codes
Equifax with information theoretical predictions experience an increase in
on volumes and defaults and using microdata mortgage credit twice as
at the level of US ZIP large as prime ZIP codes,
codes; new mortgage while their income growth
lending collected through is negative. Simultaneously,
the Home Mortgage subprime securitization
Disclosure Act (HMDA) increases. From 2006,
and aggregated at ZIP defaults are three times as
code large in subprime ZIP codes
Keys et al. Individual data on US Regression Doubling the likelihood of
(QJE, Keys mortgage lending from discontinuity around securitization on otherwise
et al. (2010)) LoanPerformance, over the credit score which similar loans is associated
2001 to 2006 serves as securitization with a 10%-25% increase in
threshold for mortgage the likelihood of default.
loans This can be attributed to lax
bank screening at loan
origination
Purnanandam Loan-level mortgages Difference-in- Banks with higher rates of
(RFS, 2010) collected through the differences, exploiting securitization prior to the
Home Mortgage the intensity of 2007 liquidity shock have
Disclosure Act (HMDA) securitization across significantly higher rates
and bank financial different banks subject to of defaults on entire loan
characteristics from the a liquidity shock on portfolios afterwards.
Federal Reserve call securitized assets Therefore, lax screening
reports incentives were prevalent,
and not limited to narrow
credit score segments
Benmelech et al. Does credit-risk Loan-level data on the US Difference-in- No evidence that
(JFE, 2012) transfer lead to lax syndicated loan market differences, comparing securitized corporate
credit standards in and collateralized loan the performance of loans were riskier than
corporate lending? obligations (CLOs) loan securitized syndicated similar loans that were not
portfolios from Creditflux loans with non-securitized securitized between 1997
syndicated loans and 2007
Albertazzi et al. Loan-level data from the Joint estimation of the The tests confirm the
(WP, 2016) Bank of Italy Credit probability of presence of adverse
Register and Supervisory securitization and default selection, reject the
Records on the entire on the sample of loans presence of moral hazard,
population of firms and show that the
borrowing from Italian negative effect of adverse
banks over the years selection is more than
2002-2007 accounted for by positive
selection on observables
at the time of loan
origination

(continued)
2 SECURITIZATION AND LENDING
9
 10

Table 2.1 (continued)

Paper Research question Data Methodology Main findings

Carbo-Valverde Are other modalities Firm and bank financials Simultaneous estimation Different types of
et al. (JFS, of credit-risk transfer from Amadeus as well as of credit supply and securitization generate
2015) of mortgage loans less securitization information demand from firm different effects: one
vulnerable to lax from Dealogic financials; correlations standard deviation in
A. BILAN ET AL.

screening? between the estimate of securitization activity

credit constraints and the during a crisis implies a
degree of securitization at 1.2% positive impact on
bank level loan supply when
securitization is achieved
through covered bonds,
compared to a negative
impact of ?2.4% in the
case of asset-backed
securitization
Gorton and Are securitized Securitized bonds from Correlational evidence, The Libor-OIS index is
Metrick (JFE, markets vulnerable to the dealer banks, covering horseracing the Libor-OIS significant when
2012) runs? non-subprime credit spread versus the ABX explaining spreads, repo
products such as credit index to explain contagion rates, and haircuts on the
cards, student loans, auto from the subprime to the four classes of
loans, and commercial non-subprime market non-subprime bonds,
mortgage-based securities, segments which suggests that the
enhanced with indicators perceptions of bank
of subprime distress (the counterparty risk
ABX index) and bank increased consistently
counterparty risk (the with a run on banks
Libor-OIS spread)
 2 SECURITIZATION AND LENDING 11

available from Thomson Reuters’ Dealscan database, although information

on securitized loans is also available from other providers such as Creditex.
In Europe, loan-level data is held by the credit registers of national central
banks and they include all bank-firm lending above a certain threshold. As
in the case of mortgages, these data sets offer an array of characteristics on
the loan itself (maturity, interest rate, performance) or on the borrower.
Finally, information on bank lending behaviour is sometimes available also
from surveys (an example being the Bank Lending Survey carried out by
the Eurosystem).
In addition, in many empirical analyses, credit data is supplemented
with firm financial information, generally available either from Compustat
(worldwide) or from Bureau van Dijk and Moody’s Analytics (European),
and financial market indicators.

2.2 METHODOLOGY
The empirical literature investigating the relationships between securitiza-
tion and bank lending uses extensively panel data econometrics. In general,
the research design seeks to establish causality between securitization and
the functioning of credit markets.
One way to take a stab at establishing causality is to exploit cross-
sectional heterogeneity between subjects along theoretical priors. At a
country level, for example, Maddaloni and Peydró (2011) investigate
whether monetary policy has a larger impact on lending standards in
countries with a higher intensity of securitization. Or, at a bank level,
Loutskina (2011) studies whether banks with a higher share of loans
that can be securitized have lower holdings of liquid assets. Acharya
et al. (2013) investigate whether banks with lower levels of regulated
capital engage more intensely in securitization. Gorton and Metrick (2012)
exploit cross-sectional heterogeneity among several prime and subprime
products. Then they study the time dimension to establish that distress
in the subprime market preceded the prime market in the run-up to the
financial crisis. Finally, at loan level, Benmelech et al. (2012) and Albertazzi
et al. (2016) argue that securitization does not negatively impact all credit
markets, by showing that corporate loans that were securitized did not
underperform similar loans which had not been securitized. All these
studies exploit correlations in the variables of interest at an individual-time
dimension, allowing researchers to saturate the models with controls at
12 A. BILAN ET AL.

the same level of detail, but also with fixed effects along each of the two
dimensions.
However, while many times a useful tool, correlational evidence is vul-
nerable to endogeneity. In the cross-sectional dimension, banks able to use
securitization might be different from banks that are not. Or, securitized
loans might be different from non-securitized loans, along dimensions that
are unobservable to the econometrician. In the time series dimension, there
might be macroeconomic factors and policies with a heterogeneous impact
on lending, which might correlate with the intensity of securitization.
For example, increasing house prices or extending government guarantees
might lead to the securitization of riskier loans. Finally, observed loan
volumes and spreads are the equilibrium result of transactions between
banks and borrowers, and, as such, they are subject to the simultaneity
between supply and demand.
The literature has addressed these issues in several ways. The recent
availability of banking microdata allowed researchers to construct increas-
ingly homogeneous samples, thereby comparing individuals with similar
characteristics, within the same product or geographical market segment
(Albertazzi et al. 2016; Benmelech et al. 2012; Mian and Sufi 2009;
Demyanyk and Van Hemert 2009). Other studies relied on established
empirical techniques that specifically tackle endogeneity, such as difference-
in-differences or regression discontinuity designs (Purnanandam 2010;
Keys et al. 2010). And, even when detailed microdata has not always
been available, simultaneous estimation has helped researchers disentangle
supply and demand (Carbo-Valverde et al. 2015).
Below, we review these contributions in detail.

2.3 SECURITIZATION AND BANK BUSINESS MODELS

The possibility to engage in loan securitization has changed the way banks
do business. The impact has been either direct, by altering bank manage-
ment decisions, or indirect, as securitization affected the transmission of
monetary policy or of bank regulations. Below, we review some empirical
contributions that have documented these changes.
The conventional benefits of securitization are put to the test in Lout-
skina (2011). The author uses bank balance sheets and income statements
extracted from the Federal Reserve call reports to study how securitization
changes bank liquidity and funding management. Relying on panel data
 2 SECURITIZATION AND LENDING 13

analysis conducted over 30 years (1976 to 2007), this study documents

that securitization indeed reduces banks’ needs to hold liquid assets. In
turn, this increases banks’ ability to extend credit. The reason is clear: loans
that are long term and illiquid can now be sold in the market in exchange
for new liquidity. Banks, thus, adapt their business from holding loans on
balance sheet up to maturity to selling them to outside investors.
To capture this effect empirically, Loutskina (2011) constructs an index
measuring the potential liquidity of each bank’s portfolio. More precisely,
this is a weighted average of a bank’s potential to securitize its own loans,
based on the shares of loans of the same type already securitized on the
market. She then finds that the index is inversely correlated with the banks’
holdings of liquid assets, measured as the share of marketable assets and
federal funds sold to total assets. Further tests conducted around regulatory
changes affecting liquidity for securitized assets confirm that this market
offers a substitute to bank on-balance sheet liquidity.
But securitization can also have indirect effects on the banking market.
These stem from its interaction with monetary policy or bank regulation.
In this sense, Maddaloni and Peydró (2011) show that while low monetary
policy interest rates lead to a softening in bank lending standards, the effect
can be magnified by securitization.
That an expansionary monetary policy might spur risk-taking in lending
by banks has been recognized in the banking literature as the “risk-taking
channel of monetary policy”. Agency problems in banking-due to bailouts
and liquidity assistance-mean that low interest rates may induce banks to
soften their lending standards by improving banks’ liquidity (Allen and
Gale 2009) and net worth (Adrian and Shin 2010). Moreover, low interest
rates make riskless assets less attractive and may lead to a search-for-yield
by financial intermediaries (Rajan 2006).
Above, we discussed how the possibility to securitize loans improves
bank liquidity. But this behaviour could prove excessive during a monetary
expansion. In addition, loan securitization is one other way of creating
assets that yield attractive returns for investors. As a result, while expan-
sionary monetary policy alone could lead to softer lending standards,
securitization may further amplify this phenomenon. To investigate this
channel, Maddaloni and Peydró (2011) use European data from the
14 A. BILAN ET AL.

Bank Lending Survey (BLS), covering Euro-area countries.1 The authors

assemble survey data from 12 countries which were in the European
Monetary Union between 2002:Q4 to 2008:Q3.2
Directly assessing the effects of monetary policy on bank lending
standards can lead to biased estimates, because both tend to be endoge-
nously determined, together with local economic conditions. However,
the authors argue, this is less a concern in the Euro-area. Here, monetary
policy rates are set by the Governing Council of the European Central
Bank (ECB) and are identical across countries, while there are significant
differences in terms of GDP and inflation. Based on this observation,
the authors devise an identification strategy which exploits cross-country
variation in monetary policy conditions across Euro-area countries.
A first empirical specification investigates the effect of a low monetary
policy rate on lending standards:

LSt,i = αi + νt + βST ratet −1,i + γ LT ratet −1,i + δControlst −1,i + t,i

(2.1)

where LSt,i is the net percentage of banks that report having tightened
credit standards in quarter t and country i. ST ratet −1,i captures the local
effect of monetary policy conditions, proxied by Taylor rule residuals. The
residuals are country-specific and are obtained by regressing the EONIA
rate on local GDP growth and inflation.3 A positive residual indicates
contractionary monetary policy (or high short-term monetary policy rates),
while negative residuals proxy for expansive monetary conditions (low
rates). The main macroeconomic controls include the ten-year government
bond interest rates, as well as GDP and inflation rates for each Euro-
area country. Country and time fixed effects control for all unobservable,
time-invariant local conditions and, respectively, all unobservable but time-
varying shocks that affect monetary policy and lending standards.
A second specification measures whether the effect of an expansionary
monetary policy is amplified when securitization activities intensify, by

1 Jiménez et al. (2014) and Ioannidou et al. (2014) already provide empirical evidence for
the existence of the risk-taking channel of monetary policy.
2 The 12 countries are Austria, Belgium, France, Finland, Germany, Greece, Ireland, Italy,
Luxembourg, the Netherlands, Portugal, and Spain.
3 The EONIA rate is the Euro OverNight Index Average, the short-term rate on the
interbank market targeted by European monetary policy.
 2 SECURITIZATION AND LENDING 15

interacting the short-term rates with indicators for securitization. Here, as

well, the authors rely on cross-country variation in securitization practices
and regulation. To measure securitization activity, they use the ratio
between the volume of all the issuances of asset-backed and mortgage-
backed securities in each quarter and country (as reported by Dealogic),
normalized by GDP.

LSt,i = αi + νt + βST ratet−1,i + ζ Securitizationt−1,i

+ θ(ST rate ∗ Securitization)t−1,i + γ LT ratet−1,i + δControlst−1,i + t,i

(2.2)

The results confirm that low monetary policy short-term interest rates
soften lending standards for both firms and households. And, crucially, this
softening is amplified by high securitization activity, especially in the case
of mortgages (the θ coefficient on the interaction between securitization
and short-term rates is positive and statistically significant).
Aside from its interaction with monetary policy, securitization may also
dilute banking regulation if, for example, it affects how banks respond to
capital requirements. Acharya et al. (2013) argue that, prior to the financial
crisis, commercial banks relied on securitization in order to circumvent
regulatory capital constraints.
Because securitization transfers loan exposures from the banking sector
to outside investors in exchange for cash, banks can use it to reduce their
capital requirements (i.e., the level of capital that they have to hold against
risky assets). Focusing on the asset-backed commercial paper (ABCP)
market prior to 2007, Acharya et al. (2013) show that some banks use
heavily this product, but without adequately transferring the underlying
risk. As a result, these banks are in a vulnerable position when the crisis
starts. Below, we first describe the most relevant institutional features of
the ABCP market and then we review the empirical methods employed by
the authors.
In the run-up to the financial crisis, banks set up new institutions, called
“conduits”, which were meant to purchase medium to long-term assets
from the balance sheet of the originating bank. The conduits securitized
the assets, and financed them by issuing short-term ABCP to investors.
These conduits were risky: if the value of the assets in their portfolio
deteriorated, the conduits might not raise enough cash to refinance matur-
ing commercial paper (hence, they were vulnerable to “rollover risk”).
16 A. BILAN ET AL.

However, many of the investors in ABCP at the time were money market
funds, who preferred safe assets. To attract these investors, banks insured
the assets with explicit guarantees. The guarantees were structured in a
way that the investors got paid off even if the conduit’s cash flow was
insufficient to satisfy all claims. Crucially, in spite of the guarantees, banks
relying on this type of securitization could reduce significantly regulatory
capital requirements. In some cases, this reduction was down to at most a
tenth of the capital required to back on-balance sheets assets.
As a result, banks retained most of the credit risk even after securitizing
the loans. At the beginning of the financial crisis, when the value of the
assets of the conduits suffers a negative shock,4 losses to outside investors
are actually very small. But the impact on banks is negative and large,
because they held the financial responsibility of the conduits and had
previously committed to repaying maturing ABCPs at par to investors
because of the guarantees.
To document these events empirically, Acharya et al. (2013) use com-
prehensive panel data. The main data source is rating reports for all
938 conduits rated by Moody’s Investors Service from January 2001
to December 2009, containing data on conduit sponsor, type, assets,
and guarantees. This is enhanced with Moody’s Weekly Announcement
Reports of rating downgrades on the conduits, from January 2007 to
December 2008. In addition, the authors access a proprietary data set
on all US ABCP transactions between January 2007 and February 2008,
collected by the Depository Trust and Clearing Corporation (DTCC).
Finally, they include bank financial data from Bankscope and bank share
prices, which are publicly available.
The empirical analysis follows three steps. An initial descriptive analysis
shows that the majority of conduits were sponsored by commercial banks
and were structured with liquidity guarantees which permitted capital
reductions. Within this set of banks, the paper shows, capital-constrained
commercial banks were more likely to sponsor conduits. The authors use
panel regressions in order to assess the relation between bank exposure to
ABCP (measured as the ratio of ABCP relative to bank equity) and bank

4 In July 2007, ABCP outstanding amounted to $ 1.3 trillion. But, on 9 August 2007,
the French bank BNP Paribas halts withdrawals from three funds invested in mortgage-
backed securities due to losses from the underlying assets. This has a profound impact on the
ABCP market, which experiences the equivalent of a banking run with outstanding volumes
dropping to $ 833 billion in December 2007.
 2 SECURITIZATION AND LENDING 17

capital:

Exposureit = αi + δt + βCapitalRatioit + γ Xit + it (2.3)

where Exposureit measures the ABCP exposure of bank i at time t,

CapitalRatioit is the capital ratio of bank i at time t, αi are bank fixed
effects, δt are time fixed effects, and Xit are controls for time-varying bank
characteristics (size, return on assets and short-term debt, deposits, and
loans, each as a share of assets). Bank capital is measured both as economic
capital (book equity to assets) and regulatory capital (Tier 1 regulatory
capital to risk-weighted assets). Economic capital is included because
regulatory capital is vulnerable to reverse causality. If banks had already
engaged in securitization in order to arrive at their existing capital ratios,
then, in equilibrium, the econometrician would only observe relatively
larger levels of regulated capital. In this case, economic capital might
provide a better measure of capital constraints than regulatory capital, as
the former is not monitored by bank supervisors.
A second specification measures whether, ex post, when asset qual-
ity deteriorates, conduits experience a run from their short-term credit
providers, leading to lower ABCP issuance and higher spreads. To do this,
the paper exploits cross-sectional variation in the strength of the guarantee:
while some guarantees cover completely credit and liquidity risks, others
are weaker. The analysis focuses on the period of four months before and
after the start of the financial crisis on August 9, 2007, as follows:

yit = α + βj Guaranteeij + γj Af tert Guaranteeij + δAf tert + it (2.4)

where yit is either the natural logarithm of the face value of ABCP
outstanding of conduit i in week t or the overnight ABCP spread over
the federal funds rate on new issues by conduit i on day t. Guaranteeij
is an indicator variable for guarantee of type j of conduit i and Af tert
is an indicator variable that equals one after the start of the crisis. Addi-
tional specifications include time fixed effects, conduit fixed effects and
sponsor-time fixed effects. If the financial crisis made investors increasingly
concerned about conduit risks, then conduits with weaker guarantees
should have a lower performance and less ability to roll over maturing assets
(γj should be stronger for relatively weaker guarantees).
Thus, for conduits that are fully guaranteed, no realized losses should
be passed to investors. In exchange, banks with greater exposure to these
18 A. BILAN ET AL.

conduits should experience a negative shock, as they would have to absorb

the losses on conduit assets. To verify whether this was indeed the case, a
final specification tests whether these banks experience lower stock returns,
in an event study around August 9, 2007. The baseline specification is:

Ri = α + βConduitExposurei + γ Xi + i (2.5)

where Ri is the cumulative equity return of bank i, computed over the

three-day period from August 8 to 10, 2007. ConduitExposurei is bank
i’s conduit exposure relative to equity in January 2007, and Xi are bank
characteristics. Consistent with the hypothesis, we would expect a negative
β.
The estimations show that, first, banks that are more capital constrained
have larger ABCP exposures. Second, when asset quality deteriorates at the
beginning of the financial crisis, ABCPs with weaker guarantees experience
larger investor outflows. And, third, banks highly exposed to conduits
appear to be the main losers from the underperformance of the conduits:
an increase in conduit exposure from 0% to 100% of bank equity reduces
the stock return during the three-day event window by 1.1%.

2.4 TYPES OF SECURITIZATION AND ASYMMETRIC

INFORMATION
The previous section discussed how, prior to 2007, the market for securi-
tized assets increased significantly, with positive effects on lending volumes
but also negative effects on risk-taking. In this section, we review the
literature exploring the reasons for the increased risk-taking. While this
effect could simply be the direct consequence of a shift in the distribution of
new borrowers, when supply increases sufficiently to satisfy a credit demand
of lower quality, this section shows how securitization could also increase
informational frictions in the credit market.
In their traditional role as financial intermediaries, banks should reduce
information asymmetries between lenders and borrowers. But securitiza-
tion changes this traditional lending market from an “originate-to-hold”
to an “originate-to-distribute” model (Purnanandam 2010). When the
originator of a loan holds it to maturity, they have incentives to select
the loans that perform best. This is done by both ex ante screening the
borrowers and by monitoring them ex post (Pennacchi 1988; Gorton and
 2 SECURITIZATION AND LENDING 19

Pennacchi 1995; Petersen and Rajan 1994; Parlour and Plantin 2008;
Holmstrom and Tirole 1997; Diamond and Rajan 2005). But if banks
offload the loans shortly after origination by securitizing them, then their
screening efforts might be reduced. In fact, if screening is costly and loans
can be easily securitized, banks are likely to only screen borrowers for hard
information. This is information that can easily be observed by outside
investors and contracted upon. Any soft information that might predict the
future performance of a loan, but it is costly to process and, unobservable to
outsiders (such as indicators of the future income capacity of the borrowers,
or the diligence with which they prepared their application and its quality),
could be overlooked.
The empirical evidence suggests that securitization did increase infor-
mational asymmetries, but this effect was limited to the subprime lending
market. Researchers have sought to explain the increased risk-taking in this
market segment by looking for causal evidence of, first, an increase in bank
credit supply, and second, of a simultaneous reduction in screening efforts.
Below, we review this evidence. Then, we go through studies involving
other types of lending (credit to large corporations and to small businesses)
or other forms of securitization (covered bonds), and show that these
markets appear less vulnerable to informational frictions.

2.4.1 Subprime Mortgage Lending

Among the first studies exploring the reasons for the boom and bust
behaviour of the securitized mortgage market in the run-up to the financial
crisis are Demyanyk and Van Hemert (2009) and Mian and Sufi (2009).
Demyanyk and Van Hemert (2009) use loan-level data on subprime
mortgages from LoanPerformance and document a decrease in loan quality
in this market segment between 2001 to 2007. They employ proportional
odds duration models, estimating the probability of first-time delinquency
on a subprime mortgage loan as a function of loan characteristics, borrower
characteristics, macroeconomic conditions, and origination year effects.
The evidence is consistent with a sustained increase in risk-taking in this
market. Conditional on observables, loan quality deteriorates monoton-
ically between 2001 and 2007, but it is masked by the simultaneous
appreciation in house prices. As house prices start to decline in 2006, poor
loan quality materializes into realized defaults in 2006 and 2007.
Discovering Diverse Content Through
Random Scribd Documents
 212.3 Thureau-Dangin, Les cylindres de Goudéa, p. 57: Les héros
morts leur bouche auprès d’une fontaine il plaça.
212.4 Winckler, op. cit., p. 41.
212.5 Jeremias, op. cit., p. 15.
213.1 E.g. Peiser, Sketch of Babylonian Society, in the Smithsonian

Institute, 1898, p. 586, speaks as if it was ancestor-worship that held

the Babylonian family together.
213.2 Vide my article on “Hero-worship” in Hibbert Journal, 1909, p.
417.
214.1 V. Landau, Phönizische Inschr., p. 15.
214.2 Jeremias, Hölle u. Paradies, p. 37.
215.1 It would be idle for my purpose to distinguish between the so-
called “Achaean” and “Pelasgian” elements in the Homeric Νέκυια;
even if the latter ethnic term was of any present value for Greek
religion.
215.2 Hesiod, Ἔργ. 110-170 (the men of the golden and the silver
ages and the heroes).
216.1 Vide Zimmern in K.A.T.3, pp. 636-639; Jeremias, Hölle u.
Paradies, p. 25; cf. his Die Babyl. Assyr. Vorstellungen rom. Leben
nach dem Tode.
216.2 Vide supra, p. 160.
216.3 Zimmern, op. cit., p. 520; King, op. cit., p. 188.
217.1 King, op. cit., p. 138.
217.2 Lagrange, Études sur les religions sémitiques, p. 493.
218.1 Cf. Keil. Bibl., ii. 109; Jeremias, Hölle u. Paradies, pp. 13-14.
219.1 Jastrow, op. cit., pp. 472-473.
219.2 Ib., p. 473.
219.3 Ib., p. 472.
219.4 Zimmern in Sitzungsber. d. Kön. Sächs. Gesell. Wiss. 1907,

“Sumerisch-Babylonische Tanzlieder,” p. 220.

219.5 Vide Jeremias in his article on “Nergal” in Roscher’s Lexikon,
iii. p. 251.
219.6 It is doubtful if any argument can be based on the name
Ningzu, occasionally found as the name of the consort of Ereshkigal
(Zimmern, K.A.T.3, p. 637) and said to mean “Lord of Healing,” in
reference, probably, to the waters of life.
219.7 Only in the story of Adapa he appears as one of the warders

of the gates of heaven (Zimmern, K.A.T.3, p. 521).

220.1 The story of Aphrodite descending into Hades to seek Adonis
is much later than the period with which we are dealing. Nergal’s
descent to satisfy the wrath of Allatu and his subsequent marriage
with her (Jeremias, Hölle und Paradies, p. 22) is a story of entirely
different motive to the Rape of Kore.
 CHAPTER XIII NOTES
223.1 Cook, The Religion of Ancient Palestine, p. 17.
223.2 Researches in Sinai, p. 72, etc., 186: he would carry back the
foundation to the fourth millennium B.C.
223.3 Vide Arch. Anzeig., 1909, p. 498.
223.4 Vide Cults, iii. p. 299.
224.1 Vide Hogarth’s evidence for the date of the earliest

Artemision, Excavations at Ephesus, p. 244.

224.2 Il., i. 38.
224.3 Ib., vi. 269, 299-300.
224.4 Ib., ii. 550.
224.5 Ib., ix. 405.
224.6 Vide Stengel, Griechische Sacral-Altertümer, p. 17.
224.7 Vide Athen. Mittheil., 1911, pp. 27, 192.
225.1 Vide Jeremias in Roscher, Lexikon, ii. p. 2347, s.v. “Marduk.”
225.2 Something near to it would be found in the cult-phrase Ζεὺς
Νᾶος of Dodona, which is a form commoner in the inscriptions than
Ζεὺς Νάϊος, if, with M. Reinach (Rev. Archéol., 1905, p. 97), we
regarded this as the original title and interpreted it as “Zeus-Temple.”
But the interpretation is hazardous.
225.3 A disk on the top of a pole, vide Jastrow, Rel. Bab. Assyr.,
vol. i. p. 203.
226.1 Cook, op. cit., p. 28.
226.2 Religion of the Semites, pp. 185-195; “Mycenaean Tree and
Pillar Cult,” Hell. Journ., 1901. It is interesting to note that Baitylos, a
name derived from the Semitic description of the sacred stone as the
“House of God,” is given as the name of a divine king in the
cosmogony of Philo Byblius, Müller, Frag. Hist. Graec., iii. p. 567; cf.
the baitylos with human head found at Tegea inscribed Διὸς
Στορπάω (fifth century B.C.), “Zeus of the lightning” (Eph. Arch.,
1906, p. 64).
 227.1 Vide Evans, op. cit., and Annual of British School, 1908,
1909.
227.2 Vide my Cults, i. pp. 13-18, 102; ii. pp. 520, 670; iv. pp. 4,
149, 307; v. pp. 7, 240, 444.
227.3 For the evidence of a pillar-cult of Apollo Agyieus and
Karneios coming from the north, vide Cults, vol. iv. pp. 307-308.
227.4 The pillars known as “Kudurru,” with emblems of the various
divinities upon them, served merely as boundary-stones (vide
Jastrow, op. cit., i. p. 191; Hilprecht in Babylonian Expedition of
University of Pennsylvania, vol. iv.).
228.1 6, 269.
228.2 Cults, ii. 445.
228.3 Op. cit., vol. v. p. 8.
229.1 Arnob. Adv. Gent., 5, 19 (in the mysteries of the Cyprian
Venus), “referunt phallos propitii numinis signa donatos.”
229.2 Cook, Religion of Ancient Palestine, p. 28; cf. Corp. Inscr.
Sem., i. 11. 6, inscription found in cave, dedicated perhaps by the
hierodulai, “pudenda muliebria” carved on the wall.
229.3 Rel. of Sem., pp. 437-438.
229.4 De Dea Syria, c. 16 and c. 28.
229.5 Histoire de l’Art, iv. pl. viii, D.
230.1 Jeremias, in his articles on “Izdubar” and “Nebo” in Roscher’s
Lexikon, ii. p. 792 and iii. p. 65, concludes that a phallic emblem was
employed in the ritual of Ishtar; but he bases his view on the
translation of the word ibattu in the Gilgamesh Epic, which is
differently rendered by King, Babylonian Religion, p. 163, and
Zimmern, K.A.T.3, p. 572.
230.2 Thureau-Dangin, Les Cylindres de Goudéa, p. 69.
231.1 This may explain the double phrase, used concerning the
institution and endowment of temple-rites in an inscription of the time
of Tiglath-Pileser III., which Zimmern translates by “Opfer-
Mahlzeiten,” Keil. Bibl., iv. p. 103; cf. especially K.B., iii. p. 179 (inscr.
of ninth century); Zimmern, Beiträge zur Kenntniss der Babyl. Relig.,
ii. p. 99 (sacred loaves offered before consultation of divinity).
231.2 Vide Robertson Smith, op. cit., p. 200.
 231.3 Vide Cults, i. p. 88; v. p. 199.
232.1 Judges ix. 13; cf. Robertson Smith, op. cit., p. 203.
232.2 Lagrange, Études sur les religions sémitiques, p. 506. This

seems to agree with the statement in Diodorus (19, 94) that the
Nabataeans tabooed wine; yet Dusares, the Arabian counterpart of
Dionysos, was a Nabataean god.
232.3 Gray, Shamash Religious Texts, p. 21.
232.4 Dhorme, Choix, etc., p. 41, l. 136.
232.5 Vide Cults, iii. p. 390, R. 57h.
232.6 Ib., ii. p. 646.
234.1 Robertson Smith, op. cit., pp. 272-273.
234.2 Athenae. 376a (Cults, i. p. 141).
234.3 Cults, ii. pp. 646-647.
234.4 O. Weber, Dämonenbeschwörung, p. 29; his note on the
passage “that the unclean beast is offered as a substitute for an
unclean man” is not supported by any evidence.
234.5 Zimmern, K.A.T.3, pp. 409-410.
235.1 Robertson Smith’s theory that the gift-sacrifice was a later
degeneracy from the communion-type is unconvincing; vide specially
an article by Ada Thomsen, “Der Trug von Prometheus,” Arch. Relig.
Wissensch., 1909, p. 460.
236.1 “Sacrificial Communion in Greek Religion,” in Hibbert Journal,
1904.
236.2 E.g. Il., 1, 457-474; Od., 3, 1-41; 14, 426.
236.3 Cf. Schol. Od., 3, 441 (who defines οὐλοχύται as barley and

salt mixed with water or wine… καὶ ἔθυον αὐτὰ πρὸ τοῦ ἱερείου…
κριθὰς δὲ ἐνέβαλον τοῖς θύμασι χάριν εὐφορίας); Schol. Arist. Equ.,
1167, τοῖς θύμασιν ἐπιβαλλόμεναι [κριφαί]. Vide Fritz. Hermes, 32,
235; for another theory, vide Stoll, “Alte Taufgebraüche,” in Arch.
Relig. Wissensch., 1905, Beiheft, p. 33.
237.1 Vide Evans, “Mycenaean Tree and Pillar Cult,” Hell. Journ.,
1901, pp. 114-115.
237.2 Od., 14, 426; cf. the custom reported from Arabia of mingling
hair from the head of a worshipper with the paste from which an idol
is made.
 237.3 Aristoph. Pax., 956.
237.4 Athenae, p. 419, B.
237.5 Vide Arch. Rel. Wiss., 1909, p. 467; Thomsen there explains
it wholly from the idea of tabu.
237.6 The common meal of the thiasotaï is often represented on
later reliefs, vide Perdriyet, “Reliefs Mysiens,” Bull. Corr. Hell., 1899,
p. 592.
238.1 Vide Cults, i. pp. 56-58, 88-92.
239.1 In my article on “Sacrificial Communion in Greek Religion,”
Hibbert Journal, 1904, p. 320, I have been myself guilty of this, in
quoting the story told by Polynaenus (Strategem. 8, 43), about the
devouring of the mad bull with golden horns by the Erythraean host,
as containing an example of a true sacrament.
239.2 Vide Cults, vol. i. p. 145.
239.3 See Crusius’ article in Roscher’s Lexikon, s.v. “Harpalyke.”
240.1 Vide Cults, v. pp. 161-172.
240.2 Ib., v. p. 165.
241.1 K.A.T.3, p. 596.
241.2 Jeremias, Die Cultus-Tafel von Sippar, p. 26.
241.3 Zimmern, Beiträge zur Kennt. Bab. Rel., p. 15.
242.1 Vide Frazer, Adonis-Attis-Osiris, p. 189; cf. “Communion in
Greek Religion,” Hibbert Journ., 1904, p. 317.
242.2 Jeremias, Die Cultus-Tafel von Sippar, p. 28.
243.1 Weber, Dämonenbeschwörung, etc., p. 29.
243.2 iv. R2, pl. 26, No. 6; this is the inscription quoted by Prof.

Sayce (vide infra, p. 182, n.) as a document proving human sacrifice.

I owe the above translation to the kindness of Dr. Langdon; it differs
very slightly from Zimmern’s in K.A.T.3, p. 597.
243.3 Jeremias, op. cit., p. 29.
243.4 Renan’s thesis (C. I. Sem., i. p. 229) that the idea of sin, so
dominant in the Hebrew and Phoenician sacrifice, was entirely
lacking in the Hellenic, cannot be maintained; he quotes Porph. De
Abstin., 1, 2, 24, a passage which contains an incomplete theory of
Greek sacrifice. The sin-offering is indicated by Homer, and is
recognised frequently in Greek literature and legend; only no
technical term was invented to distinguish it from the ordinary
cheerful sacrifice.
244.1 Cults, ii. p. 441.
244.2 Vide K.A.T.3, pp. 434, 599, where Zimmern refers to the
monuments published by Ménant, Pierres gravées, i. figs. 94, 95, 97,
as possibly showing a scene of human sacrifice. But Ménant’s
interpretation of them is wrong; vide Langdon, Babyloniaca, Tome iii.
p. 236, “two Babylonian seals”; the kneeling figure is the owner of
the seal; the personage behind him is no executioner, but Ramman
or Teschub holding, not a knife, but his usual club. The inscriptions
published by Prof. Sayce (Trans. Soc. Bibl. Arch., iv. pp. 25-29) are
translated differently by Dr. Langdon, so that the first one (iv. R2, pl.
26, No. 6) refers to the sacrifice of a kid, not of an infant. The
misinterpretation of the inscription has misled Trumbull (Blood
Covenant, p. 166). The statement in 2 Kings xvii. 31 about the
Sepharvites in Samaria does not necessarily point to a genuine
Babylonian ritual, even if we are sure that the Sepharvites were
Babylonians.
245.1 Babylonian and Assyrian Laws, p. 95.
245.2 The excavations at Gezer have revealed almost certain
evidence of the early practice of human sacrifice; a number of
skeletons, one of a girl sawn in half, were found buried under the
foundation of houses (vide Cook, op. cit., pp. 38-39).
246.1 Stengel, Die griechischen Kultusaltertümer, p. 89.
246.2 K.A.T.3, p. 599.
246.3 Jastrow, op. cit., i. p. 500.
246.4 Might this be the meaning of a line in a hymn translated by
Jastrow, op. cit., p. 549, “I turn myself to thee (O Goddess Gula), I
have grasped thy cord as the cord of my god and goddess” (vide
King, Babyl. Magic, No. 6, No. 71-94); or of the phrase in the
Apocrypha (Epist. Jerem., 43), “The women also with cords about
them sit in the ways”?
246.5 Zimmern’s Beiträge, etc., p. 99.
247.1 On the famous bronze plaque of the Louvre (Jeremias, Hölle
und Paradies, p. 28, Abb. 6) we see two representatives of Ea in the
fish-skin of the god; and on a frieze of Assur-nasir-pal in the British
Museum (Hell. Journ., 1894, p. 115, fig. 10; Layard, Monuments of
Nineveh, 1, pl. 30), two men in lions’ skins; but these are not skins of
animals of sacrifice.
247.2 Vide my Evolution of Religion, pp. 118-120.
248.1 K.A.T.3, p. 49.
248.2 3, 300; 19, 265-267.
248.3 Polybius, 3, 25, ἐγὼ μόνος ἐκπέσοιμι οὕτως ὡς ὅδε λίθος
νῦν.
248.4 Op. cit., ii. p. 217.
250.1 According to Dr. Langdon (op. cit., p. xvi.), the wailing for
Tammuz was developed in the early Sumerian period of the fourth
millennium.
251.1 Langdon, op. cit., 300-341; cf. Zimmern, “Sumerisch-
Babylonische Tamuzlieder,” in Sitzungsber. König. Sächs. Gesell.
Wissen., 1907, pp. 201-252, and his discussion, “Der Babylonische
Gott Tamuz,” in Abhandl. König. Sächs. Gesell. Wissen., 1909.
251.2 Vide supra, p. 105.
251.3 Vide Langdon, op. cit., p. 501.
251.4 Antiqu., 8, 5, 3; cf. Clem. Recogn., 10, 24; Baudissin in his
Eschmun-Asklepios (Oriental. Stud. zu Nöldeke gewidmet, p. 752)
thinks that the Healer-god, Marduk Asclepios Eschmun, is himself
one who died and rose again in Assyrian and Phoenician theology.
For Asklepios of Berytos we have the almost useless story of
Damascius in Phot. Bibl., 573 H.; the uncritical legend in Ktesias (c.
21) and Ael. Var. Hist., 13, 3, about the grave of Belitana at Babylon
(to which Strabo also alludes, p. 740), does not justify the view that
the death of Marduk was ever a Babylonian dogma.
252.1 Perrot-Chipiez, Histoire de l’Art, iv. pl. viii.
253.1 Rev. de Philol., 1893, p. 195.
253.2 Vide Frazer, op. cit., pp. 98-99.
253.3 K. O. Müller, Kleine Schriften, vol. ii. pp. 102-103.
253.4 Journ. Roy. Asiat. Soc., 1909, pp. 966, 971; the information
about the true meaning of the ideogram I owe to Dr. Langdon.
254.1 Vide supra, p. 91; cf. Cults, ii. pp. 644-649; iii. pp. 300-305.
 254.2 The Babylonian myths of Etana and Adapa, and their ascent
to heaven, may have given the cue to the Phrygian stories of
Ganymede and Tantalos.
256.1 Dr. Frazer, in Magic Art and the Evolution of Kings (G. B., vol.
ii. p. 45), quotes from N. Tsackni (La Russie Sectaire, p. 74) an
example of a fanatic Christian sect in modern Russia practising
castration. I have not been able to find this treatise.
257.1 Vide Cults, iii. pp. 300-301. Dr. Frazer’s theory is that the act
of castration was performed in order to maintain the fruitfulness of
the earth (op. cit., pp. 224-237). But this is against the countless
examples which he himself has adduced of the character and
function of the priest or priest-king as one whose virile strength
maintains the strength of the earth; the sexual act performed in the
field by the owner increases the fruitfulness of the field (Frazer, GB2,
ii. p. 205). Why should the priest make himself impotent so as to
improve the crops? The only grounds of his belief appear to be that
the priest’s testicles were committed to the earth or to an
underground shrine of Kybele (Arnob. Adv. Gent., v. 14, and Schol.
Nikand. Alexipharm., 7; vide Cults, 3; Kybele Ref. 54a); but such
consecration of them to Kybele would be natural on any hypothesis,
and Arnobius’ words do not prove that they were buried in the bare
earth.
259.1 Vide Cults, i. pp. 36-38.
259.2 Vide Evolution of Religion, p. 62.
260.1 Porph. Vit. Pyth., 17; cf. Callim. H. ad. Jov., 8; Diod. Sic., 3,
61; vide Cults, i. pp. 36-37.
260.2 Vide A. Evans in Hell. Journ., xvii. 350.
261.1 Vide Cults, vol. ii. p. 651; cf. Clem. Recogn., 10, 24,

“sepulcrum Cypriae Veneris apud Cyprum.”

261.2 Ib., pp. 651-652.
261.3 Vide Cults, vol. ii. pp. 447, n. c., 478, 638, n. a.
261.4 Aristot. Rhet., 2, 23.
262.1 Athenae, p. 620 A (ζητεῖν αὐτὸν τοὺς ἀπὸ τῆς χώρας μετά

τινος μεμελῳδημένου θρήνου καὶ ἀνακλήσεως); Pollux., 4, 54.

262.2 Frazer, GB2, vol. ii. p. 106.
 263.1 Vide Thureau-Dangin, Vorderasiatische Bibliothek, i. p. 77.
263.2 Weber, Arabien vor dem Islam, p. 19.
264.1 Vide Evans in Hell. Journ., 1901, p. 176.
264.2 Cults, i. pp. 184-191.
264.3 Ib., iii. pp. 123-124.
264.4 Ib., iii. p. 176; cf. vol. iv. p. 34 n. b.
264.5 Ib., i. pp. 189-190.
265.1 1, 181.
265.2 Vide, for instance, Dr. Langdon in the Expositor, 1909, p. 143.
265.3 Winckler, Die Gesetze Hammurabi, p. 182.
266.1 Vide Dieterich, Mithras-Liturgie, pp. 126-127; Reizenstein,

Die hellenistischen Mysterien-religionen.

266.2 Vide Herzog’s Real-Encyclop., s.v. “Montanismus.”
266.3 Jourdanet et Siméon transl. of Sahagun, pp. 147-148.
266.4 Golther, Handbuch der Germanischen Mythologie, p. 229; cf.
Mannhardt, Baumkultus, p. 589.
267.1 Pausan., 2, 33, 3; 9, 27, 6; cf. my article in Archiv. für
Religionswiss., 1904, p. 74; E. Fehrle, Die Kultische Keuschheit im
Alterthum, p. 223, gives other examples which appear to me more
doubtful.
267.2 Paus., 3, 16, 1.
267.3 Cults, v. pp. 217-219.
268.1 Vide Cults, v. p. 109.
268.2 Winckler, op. cit., p. 110; Johns, op. cit., p. 54.
269.1 Code, § 182.
269.2 Jastrow, op. cit., ii. 157.
269.3 Vide Winckler’s interpretation of §§ 178, 180, 181; cf. also

Zimmern in K.A.T.3, 423.

269.4 1, 199.
270.1 E.g. Zimmern in K.A.T.3, p. 423.
270.2 Verse 43.
271.1 The first to insist emphatically on the necessity of their
distinction was Mr. Hartland, in Anthropological Essays presented to
E. B. Tylor, pp. 190-191; but he has there, I think, wrongly classified
—through a misunderstanding of a phrase in Aelian—the Lydian
custom that Herodotus (1, 93) and Aelian (Var. Hist., iv. 1) refer to;
both these writers mention the custom of the women of Lydia
practising prostitution before marriage. Aelian does not mention the
motive that Herodotus assigns, the collection of a dowry; neither
associates it with religion. Aelian merely adds that when once
married the Lydian women were virtuous; this need have nothing to
do with the Mylitta-rite.
272.1 E.g. Hosea iv. 13; Deut. xxiii. 18; 1 Kings xiv. 24.
272.2 Weber, Arabien vor dem Islam, p. 18.
272.3 C. I. Sem., 1, 263.
272.4 Strab., 272.
272.5 Strab., 559.
272.6 Pind. Frag., 87; Strab., 378; (Cults, ii. p. 746, R. 99g).
273.1 Cities and Bishoprics, i. 94. In his comment he rightly points

out that the woman is Lydian, as her name is not genuine Roman;
but he is wrong in speaking of her service as performed to a god
(Frazer, Adonis, etc., p. 34, follows him). This would be a unique
fact, for the service in Asia Minor is always to a goddess; but the
inscription neither mentions nor implies a god. The bride of Zeus at
Egyptian Thebes was also a temple-harlot, if we could believe
Strabo, p. 816; but on this point he contradicts Herodotus, 1, 182.
273.2 Et. Mag., s.v. Ἱκόνιον.
274.1 De Dea Syr., 6; cf. Aug. De Civ. Dei, 4, 10: “cui (Veneri) etiam
Phoenices donum dabant de prostitutione filiarum, antequam eas
jungerent viris”: religious prostitution before marriage prevailed
among the Carthaginians in the worship of Astarte (Valer. Max., 2,
ch. 1, sub. fin.: these vague statements may refer either to
defloration of virgins or prolonged service in the temple).
274.2 See Frazer, op. cit., p. 33, n. 1, quoting Sozomen. Hist.
Eccles., 5, 10, 7; Sokrates, Hist. Eccles., 1, 18, 7-9; Euseb. Vita
Constantin., 3, 58. Eusebius only vaguely alludes to it. Sokrates
merely says that the wives were in common, and that the people had
the habit of giving over the virgins to strangers to violate.
Sozomenos is the only voucher for the religious aspect of the
practice; from Sokrates we gather that the rule about strangers was
observed in the rite.
274.3 18, 5.
274.4 This is confirmed by the legend given by Apollodoros (Bibl., 3,
14, 3) that the daughters of Kinyras, owing to the wrath of Aphrodite,
had sexual intercourse with strangers.
275.1 Justin, 21, 3; Athenaeus, 516 A, speaks vaguely, as if the
women of the Lokri Epizephyrii were promiscuous prostitutes.
275.2 Pp. 532-533.
275.3 The lovers, Melanippos and Komaitho, sin in the temple of
Artemis Triklaria of the Ionians in Achaia; the whole community is
visited with the divine wrath, and the sinners are offered up as a
piacular sacrifice (Paus., 7, 19, 3); according to Euphorion,
Laokoon’s fate was due to a similar trespass committed with his wife
before the statue of Apollo (Serv. Aen., 2, 201). It may be that such
legends faintly reflect a very early ἱερὸς γάμος once performed in
temples by the priest and priestess: if so, they also express the
repugnance of the later Hellene to the idea of it; and in any case this
is not the institution that is being discussed.
276.1 Antike Wald u. Feld Kulte, p. 285, etc.
277.1 Why should not the priestess rather play the part of the
goddess, and why, if we trust Plutarch (Vit. Artaxerx., 27), was the
priestess of Anaitis at Ekbatana, to whose temple harlots were
attached, obliged to observe chastity after election?
277.2 Vol. i. pp. 94-96.
277.3 Op. cit., p. 35, etc.
277.4 Op. cit., p. 44.
278.1 I pointed out this objection in an article in the Archiv. f. Relig.
Wissensch., 1904, p. 81; Mr. S. Hartland has also, independently,
developed it (op. cit., p. 191).
278.2 Vol. ii. p. 446.
278.3 Origin of Civilisation, pp. 535-537.
279.1 Vide Westermarck, History of Human Marriage, p. 76.
279.2 Mr. Hartland objects (loc. cit., p. 200) to this explanation on
the ground that the stranger would dislike the danger as much as
any one else; but the rite may have arisen among a Semitic tribe
who were peculiarly sensitive to that feeling of peril, while they found
that the usual stranger was sceptical and more venturesome: when
once the rule was established, it could become a stereotyped
convention. His own suggestion (p. 201) that a stranger was alone
privileged, lest the solemn act should become a mere love-affair with
a native lover, does not seem to me so reasonable; to prevent that,
the act might as well have been performed by a priest. Dr. Frazer in
his new edition of Adonis, etc. (pp. 50-54), criticises my explanation,
which I first put forth—but with insufficient clearness—in the Archiv.
für Religionswissenschaft (1904, p. 88), mainly on the ground that it
does not naturally apply to general temple-prostitution nor to the
prostitution of married women. But it was never meant to apply to
these, but only to the defloration of virgins before marriage. Dr.
Frazer also argues that the account of Herodotus does not show that
the Babylonian rite was limited to virgins. Explicitly it does not, but
implicitly it does; for Herodotus declares that it was an isolated act,
and therefore to be distinguished from temple-prostitution of
indefinite duration; and he adds that the same rite was performed in
Cyprus, which, as the other record clearly attests, was the
defloration of virgins by strangers. Sozomenos and Sokrates attest
the same of the Baalbec rite, and Eusebius’s vague words are not
sufficient to contradict them. One rite might easily pass into the
other; but our theories as to the original meaning of different rites
should observe the difference.
280.1 But vide Gennep, Les Rites de passage, p. 100.
280.2 Cf. Arnob. Adv. Gent., 5, 19, with Firmic. Matern. De Error.,
10, and Clemens, Protrept., c. 2, p. 12, Pott.
281.1 1, 199.
281.2 The lady who there boasts of her prostitute-ancestresses
describes them also as “of unwashed feet”; and this is a point of
asceticism and holiness.
282.1 Op. cit., p. 199.
282.2 K.A.T.3, p. 423.
283.1 Vide supra, p. 163. The writer of the late apocryphal
document, “The Epistle of Jeremy,” makes it a reproach to the
Babylonian cult that “women set meat before the gods” (v. 30), and
“the menstruous woman and the woman in child-bed touch their
sacrifices” (v. 29), meaning, perhaps, that there was nothing to
prevent the Babylonian priestess being in that condition. But we
cannot trust him for exact knowledge of these matters. Being a Jew,
he objects to the ministration of women. The Babylonian and Hellene
were wiser, and admitted them to the higher functions of religion.
283.2 Vide Cults, iv. p. 301.
283.3 Vide Inscription of Sippar in British Museum, concerning the
re-establishment of cult of Shamash by King Nabupaladdin, 884-860
B.C. (Jeremias, Die Cultus-Tafel von Sippar).
284.1 Sumerian and Babylonian Psalms, p. 75.
284.2 Vide Langdon in Transactions of Congress for the History of
Religions (1908), vol. i. p. 250.
284.3 Vide Zeitung für Assyriologie, 1910, p. 157.
284.4 Formula for driving out the demon of sickness, “Bread at his

head place, rain-water at his feet place” (Langdon, ib. p. 252).

284.5 Delitsch, Wörterbuch, i. 79-80.
284.6 Zeit. für Assyr., 1910, p. 157.
284.7 Vide Hippocrates (Littré), vi. 362; Stengel, Griechischer
Kultusaltertümer (Iwan Müller’s Handbuch, p. 110).
285.1 Referred to in the comedy of Eupolis called the “Baptai.”
285.2 Jastrow, op. cit., p. 500.
285.3 Op. cit., p. 297, 487; the priest-exorciser, the Ashipu, uses a
brazier in the expulsion of demons.
285.4 Vide Golther, Handbuch der Germanischen Mythologie, p.
580; cf. my Cults, v. p. 196.
285.5 Cults, vol. v. pp. 383-384; cf. iv. p. 301.
286.1 Cults, v. p. 356; cf. p. 363 (the purifying animal carried round
the hearth).
286.2 Eur. Herc. Fur., 928.
286.3 Dio Chrys. Or., 48 (Dind., vol. ii. p. 144), περικαθήραντες τὴν
πόλιν μὴ σκίλλῃ μηδὲ δαδί, πολὺ δὲ καθαρωτέρῳ χρήματι τῷ λόγῳ
(cf. Lucian, Menipp., c. 7, use of squills and torches in “katharsis,” (?)
Babylonian or Hellenic); Serv. ad Aen., 6, 741, “in sacris omnibus
tres sunt istae purgationes, nam aut taeda purgant aut sulphure aut
aqua abluunt aut aere ventilant.”
286.4 “To take fire and swear by God” is a formula that occurs in the
third tablet of Surpu; vide Zimmern, Beiträge zur Kenntniss Babyl.
Relig., p. 13; cf. Soph. Antig., 264.
286.5 Salt used as a means of exorcism in Babylonia as early as
the third millennium (vide Langdon, Transactions of Congress Hist.
Relig., 1908, vol. i. p. 251); the fell “of the great ox” used to purify the
palace of the king (vide Zimmern, Beiträge, p. 123; compare the Διὸς
κῴδιον in Greek ritual).
287.1 Vide Thureau-Dangin, Cylindres de Goudéa, pp. 29, 93.
287.2 Vide Evolution of Religion, pp. 113, 114, 117; Cults, v. p. 322
(Schol. Demosth., 22, p. 68).
287.3 5, 13, 6.
287.4 Vide Cults, iii. pp. 303-304; Evolution of Religion, p. 121.
288.1 Vide supra, p. 146.
288.2 Vide Cults, iii. p. 167.
288.3 Published in Zimmern’s Beiträge, p. 123; cf. Weber,
Dämonenbeschwörung, pp. 17-19.
289.1 Il., xvi. 228.
289.2 Od., ii. 261.
289.3 Il., i. 313.
290.1 Od., xxii. 481: In the passage referred to above, Achilles uses
sulphur to purify the cups.
290.2 Od., xiii. 256-281: This is rightly pointed out by Stengel in his
Griechische Kultusaltertümer, p. 107.
290.3 Evolution of Religion, pp. 139-152; Cults, iv. pp. 295-306.
291.1 Vide Cults, iv. pp. 144-147, 300: To suppose that Hellas learnt
its cathartic rites from Lydia, because Herodotus (i. 35) tells us that
in his time the Lydians had the Hellenic system of purification from
homicide, is less natural. Lydia may well have learnt it from Delphi in
the time of Alyattes or Croesus. Or it may have survived in Lydia as
a tradition of the early “Minoan” period; and, similarly, it may have
survived in Crete.
291.2 Vide supra, pp. 176-178.
 292.1 Vide Cults, iv. pp. 268-284.
292.2 For similar practices, vide Cults, pp. 415-417.
292.3 Clem. Alex. Strom., p. 755, Pott.
293.1 Paus., 9, 33, 4.
293.2 For the facts vide Zimmern, K.A.T.3, p. 592.
294.1 Works and Days, l. 824.
294.2 Ib., l. 804.
294.3 Expositor, 1909, p. 156.
294.4 Vide Photius and Hesych., s.v. Μιαραὶ ἡμέραι.
295.1 Hell., 1, 4, 12.
295.2 Vide Cults, v. pp. 215-216.
295.3 Cults, iv. p. 259.
295.4 Vide supra, pp. 176-177.
296.1 Sumerian and Babylonian Psalms, p. 196.
296.2 King, Babylonian Religion, p. 196.
296.3 Vide Fossey, La Magie Assyrienne, p. 96.
297.1 Knudtzon, Assyrische Gebete an den Sonnengott, p. 78
(texts belonging to period of Asarhaddon, circ. 681).
297.2 Zimmern, Beiträge, etc., p. 161.
298.1 Zimmern, Beiträge, etc., p. 163.
298.2 Fossey, op. cit., p. 399.
298.3 iv. R. 56, 12; Fossey, op. cit., p. 401.
298.4 Expositor, 1909, p. 150, giving text from iv. R. 40.
299.1 Fossey, op. cit., p. 209.
299.2 Zimmern, Beiträge, p. 173.
299.3 Supra, p. 176.
299.4 Zimmern, op. cit., p. 169.
300.1 Zimmern, Beiträge, pp. 30-31; he mentions also the similar
practice of tying up a sheepskin or a fillet of wool and throwing it into
the fire.
300.2 Zimmern, op. cit., p. 33: note magic use of knots in general,

vide Frazer, G.B.2, vol. i. pp. 392-403; Archiv. für Religionsw., 1908,
pp. 128, 383, 405. The superstition may have prevailed in Minoan
Crete (see A. Evans, Annual British School, 1902-1903, pp. 7-9) and
was in vogue in ancient Greece.
300.3 W. Warde Fowler, The Religious Experiences of the Roman
People, Gifford Lectures, p. 49.
301.1 Vide supra, pp. 248-249; Cults, iv. p. 191.
301.2 For the main facts relating to the Babylonian system and the

“baru”-priests, vide Zimmern, Beiträge, etc., pp. 82-92; for the

Hellenic, vide Cults, iv. 190-192, 224-231; also vol. iii. 9-12.
301.3 The documentary evidence, from a very early period, is given
by Zimmern, Beiträge, etc., pp. 85-97.
301.4 L. 322: Clytemnestra speaks of pouring oil and vinegar into
the same vessel and reproaching them for their unsociable
behaviour.
302.1 We have also one example of an oracle of Ishtar (in plain
prose), Keil. Bibl., ii. p. 179.
303.1 Zimmern, op. cit., p. 89.
303.2 Cults, iii. p. 297.
303.3 Lucian, De Dea Syr., 43.
303.4 Cults, iii. p. 297.
303.5 Vide Cults, iv. pp. 191-192; iii. p. 11.
 TRANSCRIBER’S NOTES.
Page numbers are given in {curly} brackets.
Plain text version only: endnote markers are given in [square]
brackets.
Minor spelling inconsistencies (e.g. coexist/co-exist, temple-
ritual/temple ritual, etc.) have been preserved.
Add title, subtitle, and author’s name to cover image.
Alterations to the text:
Convert footnotes to endnotes, relabel note markers (append the
original note number to the page number), and add a corresponding
entry to the TOC.
[Title page]
Add commas to author’s bibliography.
[Chapter I]
Change “from the tyranny of a morbid ascetism” to asceticism.
[Chapter III]
“In his Historie des anciennes Religions, Tiele classifies” to
Histoire.
“and their aboriginal god was Possidon” to Poseidon.
[Chapter IV]
“and expecially the powers of the lower world” to especially.
“Even Allat, the goddess of Hell, she who” to Allatu.
“the great Assyrian god Ahshur is quaintly expressed” to Asshur.
“the idea that Istar is the compeer in power” to Ishtar.
“between the Hittites and the Assyrian Babylonian kingdom” to
Assyrian-Babylonian.
“no clear trace of theriomophism either in the” to theriomorphism.
“how far the Minaon religion was purely anthropomorphic” to
Minoan.
[Chapter V]
 “I formerly developed in the second volume of my cults” capitalize
and italicize cults.
[Chapter VI]
“Still less is Allalu, the monstrous and grim Queen” to Allatu.
[Chapter VII]
(Alalkomenai, “the places of Athena Alalkomene; Nemea, “the…)
add right double quotation mark after Alalkomene.
[Chapter VII]
“about whom he is particulurly thoughtful” to particularly.
[Chapter IX]
“and regards this Hititte goddess as the ancestress” to Hittite.
[Chapter XIII]
“modern savagery and the history of ascetism” to asceticism.
(and bewail her”: “If you regard her as a deity, do) delete right
double quotation mark.
[Index]
“Hell, Babylonian conception of, 205-206” add period at end of
line.
[Endnotes]
(Page 17, note 1) “Archiv fur Religionswissenschaft, 1904.” to für.
(Page 42, note 1) “that the idiogram of Enlil, the god of” to
ideogram.
(Page 84, note 3) “last of the Babylonian kings, Nabuna ’id, who
prays” to Nabuna’id.
(Page 124, note 1) “Die Phoenizischen Imschriften,” to
Phönizischen Inschriften.
(Page 148, note 1) “Weber, Dämonenbeschworung bei den
Babyloniern…” to Dämonenbeschwörung.
(Page 183, note 3) “pp. 502 503, n. 2” add comma after 502.
(Page 232, note 2) “Lagranges, Études sur les religions
sémitiques” to Lagrange.
(Page 246, note 1) “Stengel, Die griechischen Kultusalterthümer,
p. 89” to Kultusaltertümer.
(Page 286, note 5) “vide Zimmern, Beitrage, p. 123;” to Beiträge.

[End of text]
*** END OF THE PROJECT GUTENBERG EBOOK GREECE AND
BABYLON ***

Updated editions will replace the previous one—the old editions

will be renamed.

Creating the works from print editions not protected by U.S.

copyright law means that no one owns a United States copyright
in these works, so the Foundation (and you!) can copy and
distribute it in the United States without permission and without
paying copyright royalties. Special rules, set forth in the General
Terms of Use part of this license, apply to copying and
distributing Project Gutenberg™ electronic works to protect the
PROJECT GUTENBERG™ concept and trademark. Project
Gutenberg is a registered trademark, and may not be used if
you charge for an eBook, except by following the terms of the
trademark license, including paying royalties for use of the
Project Gutenberg trademark. If you do not charge anything for
copies of this eBook, complying with the trademark license is
very easy. You may use this eBook for nearly any purpose such
as creation of derivative works, reports, performances and
research. Project Gutenberg eBooks may be modified and
printed and given away—you may do practically ANYTHING in
the United States with eBooks not protected by U.S. copyright
law. Redistribution is subject to the trademark license, especially
commercial redistribution.

START: FULL LICENSE

Welcome to our website – the ideal destination for book lovers and
knowledge seekers. With a mission to inspire endlessly, we offer a
vast collection of books, ranging from classic literary works to
specialized publications, self-development books, and children's
literature. Each book is a new journey of discovery, expanding
knowledge and enriching the soul of the reade

Our website is not just a platform for buying books, but a bridge
connecting readers to the timeless values of culture and wisdom. With
an elegant, user-friendly interface and an intelligent search system,
we are committed to providing a quick and convenient shopping
experience. Additionally, our special promotions and home delivery
services ensure that you save time and fully enjoy the joy of reading.

Let us accompany you on the journey of exploring knowledge and

personal growth!

textbookfull.com