Financial integration and the Great Leveraging

• DOI: http://doi.org/10.1002/ijfe.1649
• Published date: 01 January 2019
• Date: 01 January 2019

Received: 15 May 2016 Revised: 26 February 2018 Accepted: 26 June 2018

DOI: 10.1002/ijfe.1649

RESEARCH ARTICLE

Financial integration and the Great Leveraging

Daniel Carvalho

Banco de Portugal, Lisbon, Portugal

Correspondence

Daniel Carvalho, Banco de Portugal, Rua

Francisco Ribeiro 2 1150-165 Lisbon,

Portugal.

Email:

daniel.sousa.carvalho@bportugal.pt

JEL Classification: E51, F30, G15

Handling Editor: Taylor Mark

Abstract

This paper studies the joint dynamics of international capital flows and

domestic credit and money growth. It finds a strong statistical relationship

between cross-border flows and the decoupling of these two variables—the

Great Leveraging—a stylized fact documented for several economies in the past

decades, due to faster credit than money growth, and associated with the expan-

sion of banks nonmonetary liabilities. Results indicate that, in addition to the

equity/debt breakdown, it is important to split capital flows according to the

domestic recipient sectors: banking sector flows display, in general, a stronger

correlation with the growth of credit and with its excess growth over money.

Furthermore,thecountry'sFXregimealsoplaysarole,asmoneyholdingstend

to be less responsive to cross-border capital flows in fixed FX regimes. In broad

terms, this approach sheds light on the mechanisms through which the inter-

national banking activity might have consequences for the composition of the

domestic aggregate bank balance sheet.

KEYWORDS

capital flows, credit and money, financial integration, financial stability

1INTRODUCTION

Cross-border capital flows and, more generally, financial

integration, interact with domestic credit growth via dif-

ferent channels. On the one hand, capital inflows provide

the resident banking sector with more resources which

can be channeled to domestic credit provision. On the

other hand, capital flows exert an upwardpressure on asset

prices, generating wealth effects, which can be translated

into higher consumption and demand for credit—see, for

instance, Aizenman and Jinjarak (2009), Sá and Wieladek

(2010), and Vásquez-Ruíz (2012). Furthermore, higher

asset values improve household and corporate balance

sheets, boosting collateral values and, in that fashion, may

facilitate the access to credit.

The outburst of the global financial crisis has high-

lighted the need to better understand the connection

between credit growth and financial stability.To the extent

that surges in international capital flows have commonly

been associated with periods of rapid credit growth, they

too have been linked to the likelihood and magnitude of

crises. Examples of studies in this area include Mendoza

and Terrones(2008) and Calderon and Kubota (2012), who

argue that credit booms tend to follow periods of high cap-

ital inflows. In contrast, Sá (2006) does not find evidence

of a causal relationship between both variables. Reinhart

and Reinhart (2008) examine the links between capital

flows and financial crises and draw unconditional proba-

bilities of a crisis given episodes of capital flow bonanzas.

Against the backdrop of the global financial crisis, Lane

and Milesi-Ferretti (2012) study the process of external

adjustment and conclude that, in the pre-crisis period,

countries with higher current account deficits than would

be explained by underlying economic fundamentals expe-

rienced sharper corrections once the crisis erupted.

Recent studies have explored in more detail the relation

between private credit growth and different types of inter-

national capital flows. Lane and McQuade (2014) show

54 © 2018 John Wiley & Sons, Ltd. wileyonlinelibrary.com/journal/ijfe Int J Fin Econ. 2019;24:54–79.

CARVALHO 55

that the instrument composition is relevant in this context:

although debt flows exhibit a strong covariation pattern

with private domestic credit growth, the same is not the

case regarding equity flows. More recently, Igan and Tan

(2015) investigate a deeper flow decomposition and find

that foreign direct investment (FDI), portfolio, and other

investment flows have distinct implications for corporate

and household credit.

Despite the body of research on the relationship between

cross-border finance and private domestic credit, to my

knowledge, the literature is yet to explore how financial

integration affects money holdings. The main contribu-

tion of this paper is to extend the previous literature by

assessing how not only the different types of capital flows

but also their domestic recipient sectors interact with both

domestic credit and money holdings.

The relevance of putting together the asset and liability

sides of the banking sector balance sheet in an integrated

manner is further strengthened by the evidence by Schu-

larik and Taylor (2012) of two distinct periods concerning

the dynamics of credit and money. Using a groundbreak-

ing historical dataset, they describe how money and credit

were growing at roughly the same pace since the end of

the Second World Waruntil the early 1970s but, from that

period on, credit grew faster than money. This decoupling

between the two variables—labeled the Great Leverag-

ing by Taylor (2012)—was achieved by the fast expansion

of banks' nonmonetary liabilities, such as long-term debt

securities, which enabled them to grant credit beyond

their deposit base, and can be seen as a measure of

banking sector leverage. Furthermore, based on the cri-

sis classifications in Bordo, Eichengreen, Klingebiel, and

Martinez-Peria (2001), they providethe link between faster

credit growth and crises, by noting that the decoupling

between credit and money went hand in hand with a resur-

gence of these episodes since the 70s—in contrast, there

were barely any crisis episodes before that. Moreover,they

show that credit is a predictor of crises, whereas money is

not. Their intuition for this result is that credit is a more

encompassing measure of bank balance sheets as it cap-

tures features such as leverage and nonmonetary liabilities

which money does not.1

I rely on the monetary presentation of the balance

of payments to track how cross-border flows affect

domestic monetary holdings. In a nutshell, it estab-

lishes a statistical link between balance of payments

flows of the different resident sectors—specifically, bank-

ing or money-issuing and nonbanking or money-holding

sectors—and monetary aggregates. Furthermore, to the

extent that cross-border banking activity is generally seen

as a key driver in credit funding, focusing on the sec-

toral decomposition of flows is also relevant from that

perspective—on the role of global banks, how they oper-

ate and provide liquidity worldwide, see Cetorelli and

Goldberg (2011; 2012), Bruno and Shin (2013), McCauley

(2012) and Niepmann (2013).

To formally assess the relationship between, on the one

hand, credit, money, and the ratio between credit and

money and, on the other hand, banking and nonbanking

sector cross-border capital flows, a cross-sectional econo-

metric specification is carried out, for a group of countries

encompassing OECD members plus other mostly Asian

and Latin American countries. Importantly, the goal is to

determine the covariation patterns of these variables and,

therefore, inferring causality is, in this case, not possible

and left for future research.

More specifically, I ask two main questions: (a) What

is the comovement of capital flows to money-issuing and

money-holding sectors and the growth of domestic credit,

broad money, and the ratio between these two variables

and (b) could this comovement be dependent on the FX

regime of the recipient country? The second question fol-

lows from the evidence in Magud, Reinhart, and Vesper-

oni (2014) that countries with fixed FX regimes tend to

attract more capital from abroad and experience larger

credit expansions than countries with floating FX regimes.

I also ask a third question that naturally emerges from

the previous two: Is it the case that both the sectoral and

instrument composition of flows are relevantto fully assess

the relationship between cross-border flows and credit and

money?

Results indicate that, indeed, there is a differenti-

ated comovement of cross-border flows according to not

only the domestic recipient sectors but also the instru-

ment composition and the growth of credit and money:

While banking sector debt flows significantly and strongly

comove with the decoupling of credit and money—they

display a strong positive covariation with credit growth

but not with money growth—nonbanking sector debt

flows only weakly comove with growth in the credit to

money ratio, likely because there is a coincidence of strong

comovement of these flows with both credit and money

growth. Regarding equity flows, although both the flows

of the banking and nonbanking sectors comove with the

growth of the ratio of credit to money, those of the former

do so on account of a negative comovement with money

growth—a denominator effect—whereas those of the lat-

ter on account of positive comovement with credit—a

numerator effect. Furthermore, the country's FX regime

also matters: There is no evidence of statistically signif-

icant positive correlation of capital flows with money

growth in fixed FX as opposed to floating FX regimes. In

a medium to longer term perspective, using a multiyear

period approach, only banking sector flows comove with

credit and the credit to money ratio and only those of

the nonbanking sectors comove with money. Finally, the