New synchronicity indices between real and financial cycles: Is there any link to structural characteristics and recessions in European Union countries?

• Published date: 01 October 2020
• Date: 01 October 2020
• Author: Mariarosaria Comunale
• DOI: http://doi.org/10.1002/ijfe.1770

Received: 10 November 2017 Revised: 20 September 2018 Accepted: 13 September 2019

DOI: 10.1002/ijfe.1770

RESEARCH ARTICLE

New synchronicity indices between real and financial

cycles: Is there any link to structural characteristics and

recessions in European Union countries?

Mariarosaria Comunale

Economics Department, Applied

Macroeconomic Research Division,

Vilnius, Lithuania

Correspondence

Mariarosaria Comunale, Principal

Economist, Economics Department,

Applied Macroeconomic Research

Division, Totori

¸

u g. 4, LT-01121, Vilnius,

Lithuania.

Email:

mariarosaria.comunale@gmail.com

Abstract

In this paper, we examine the comovements and relationships between real,

credit, house, and equity price cycles in European Union countries looking at

their link to structural characteristics and their predictivepower for future reces-

sions. We make use of new within country synchronicity indices, finding that

among the different credit cycles and between the house price cycle and the real

cycle, there is a relatively high level of synchronicity. Credit and gross domes-

tic product (GDP) fluctuations seem to be less synchronised, mostly because

credit volumes tend to lag the real cycle by several quarters. The high rates of

private homeownership tend to be associated with larger cycles in GDP, credit,

and house prices. Higher loan-to-value ratios, seen as a proxy of borrowingcon-

straints, and a higher percentage of flexible-rate mortgages, could also indicate

that a country is more sensitive to shocks and possibly increase procyclicality

and increase cycle volatility. Finally, the procyclicality of the credit and hous-

ing market to the GDP cycle can be linked to the fluctuation in current accounts

and their misalignments with respect to the theoretical equilibrium value. Some

synchronicity measures and, above all, the credit cycles may also be considered

for signalling future recessions.

KEYWORDS

credit cycles, house and equity price cycles, housing market, panel probit, real cycles, synchronicity

JEL CLASSIFICATION

E32; E44; F36

1INTRODUCTION

In this study, we investigate the comovements and rela-

tionships between real, credit, house, and equity price

cycles in European Union (EU) countries looking at their

link to structural characteristics and their predictive power

for future recessions. We make use of newly calculated

within country synchronicity indices for these purposes.

Recently, several papers have focused on the importance

of a deeper analysis of financial cycles (Borio, 2012; Borio,

Disyatat, & Juselius, 2013), on comovement of business

cycles (Gayer, 2007; Cerqueira, 2013; Gächter, Riedl, &

Ritzberger-Grünwald, 2012; Mink, Jacobs, & de Haan,

2012; Belke, Domnick, & Gros, 2017) or of both finan-

cial and credit cycles across countries (Meller and Metiu,

2015, 2017; Samarina, Zhang, & Bezemer, 2015, 2017).

The literature has not considered yet the synchronization

and comovement of real and financial cycles within coun-

tries, especially in the EU looking at different and more

Int J Fin Econ. 2020;1–25. wileyonlinelibrary.com/journal/ijfe ©2020 John Wiley & Sons, Ltd. 1

Int J Fin Econ. 2020;25:617–641. wileyonlinelibrary.com/journal/ijfe © 2019 John Wiley & Sons, Ltd. 617

2COMUNALE

complete data on financial variables. Moreover,other stud-

ies find that housing finance characteristics vary widely

across countries (Cerutti, Dagher, & Dell'Ariccia, 2017,

among others). This latter outcome can be related to the

synchronicity in real and different financial cycles, and

this aspect has not been covered in the past literature.

We believe that a more accurate analysis of these

aspects can be crucial for macroprudential policy mak-

ing and macroeconomic assessment in EU countries.

These are indeed mainly related to country-specific

actions, and a within country analysis (including pos-

sible spillovers, as here reported) seems to be key.

The properties of the various cycles vary widely across

EU countries and are only sometimes synchronous

across types and within the countries themselves. These

differences may partly reflect the links between pri-

vate homeownership, the share of mortgage financ-

ing held by middle-income households, and the role

played by collateral constraints in driving medium-term

cycles. Cross-country differences, here reported, sug-

gest that potential benefits could arise from implement-

ing country-specific macroprudential policies (as shown

previously in dynamic stochastic general equilibrium

modelling [DSGE] analyses) and also macroeconomic

assessment should take into account the relationship

between cycles. Macroprudential policies, in this regard,

limiting leverage cycles could also contribute to con-

taining macroeconomic imbalances (European Central

Bank, 2018).

In this paper, indeed we try to fill a gap in the literature

as well as to contribute to a more accurate analysis useful

for macroprudential and macroeconomic policy making.

Therefore, we examine the relationships between the main

properties of real, credit, and house price cycles within

17 EU countries and their structural characteristics.1The

contribution of this paper is threefold. First, it looks at

within countries cyclical comovements, especially linked

to real and financial cycles, by using a novel index of syn-

chronicity.2Moreover, it analyses the relationshipbetween

cyclical properties and this within-country synchronicity

with macrofinancial structural characteristics. Then the

computed cycles and synchronicity indices are assessed as

their early warning signalling power to recessions.

In this paper, we make use of cycles computed by

the band-pass filter alaChristiano and Fitzgerald (2003),

with 8-80 quarters as lower–upper bounds to capture

both short-term and medium-term fluctuations. The data

refer to a data set for 17 EU countries (from 1999Q2 to

max 2016Q2) and includes Greece and the new mem-

ber states, if not otherwise specified.3Among housing

markets characteristics, we especially consider homeown-

ership rates, maximum loan-to-value (LTV) ratios4and

the share of flexible-rate mortgages. Moreover, we look

at some macroeconomic measures, that is, the current

account and a measure of current account misalignments,

as in Comunale (2016, 2017b).

Weconsider two properties of the computed cycles: their

standard deviations and, as a novel feature, the pairwise

synchronicities among them. We make use of the syn-

chronicity index based on a within country version of Mink

et al. (2012) and Samarina et al. (2015), and we then com-

pute the “synthetic” synchronicity index for each pair of

cycles as the share of instances in which the cycles are

synchronised over the total number of observations.5

Ultimately, we provide a probability model to check if

the synchronicity indices and the cycles can be seen also

as a signal for a forthcoming recession up to 1 year ahead.

In the spirit of Schularick and Taylor (2012), this is applied

for the countries with longer time series data sets with

the recession indicator for each of the 10 countries from

FRED-Organization of Economic Development (OECD).6

We test for their predictive ability by using the area under

the receiver operating characteristic curve (AUROC). Fol-

lowing Gourinchas and Obstfeld (2012), we also provide

some extensions by using data from the mid-90s with

macroeconomic control variables.

We find that the cycles between credit variables seem

relatively well synchronised. Credit and GDP cycles seem

to be less synchronised, mostly because credit volumes

tend to lag the real cycle by several quarters. Other cycles

which experience, by contrast, a high synchronicity level,

are related to GDP and property or equity prices. The

high rates of private homeownership tend to show larger

cycles in GDP, credit, and house prices. The three cycles

also turn out to be more synchronised in this case. Higher

LTV ratios, seen as a proxy of borrowing constraints and

a higher percentage of flexible-rate mortgages, could also

make a country more sensitive to shocks and possibly

increase procyclicality. Countries with higher LTV ratios

and a higher percentage of flexible-rate mortgages on

average indeed experience more synchronised GDP cycle

with loans to nonfinancial corporations (NFCs) and higher

volatility in loans to households. Finally,the procyclicality

of the credit and housing market to the GDP cycle seems to

be linked to the fluctuation in current accounts and their

misalignments with respect to the theoretical equilibrium

value.

Lastly, the cycles and synchronicity of GDP and credit

may also be considered for signalling recessions. This is

especially true with respect to loans to NFCs and real GDP.

Moreover, the cycle of long-term interest rates also has

a substantial predictive ability. The synchronicity of real

and credit cycles loses predictive value in the presence of

macroeconomic controls and the application of a shorter

time span. Under such conditions, the results for the cycles

are instead robust.

618 COMUNALE