Non-residential real estate and
economic activity: The case of Greece
Economic Analysis & Markets
January 2014
Ilias Lekkos
Lekkosi@piraeusbank.gr
Irini Staggel
Staggelir@piraeusbank.gr
Kostas Kefalas Kefalask@piraeusbank.gr
Paraskevi Vlachou Vlachoupar@piraeusbank.gr
Piraeus Bank
4 Amerikis Street, 105 64, Athens, Greece
Tel: (+30) 210 328 8187, Fax: (+30) 210 328 8605
researchdivision@piraeusbank.gr
Bloomberg Page: <PBGR>
Our aim is to discuss developments in non-residential real estate in Greece. Given the lack of
existing literature, we start by discussing at length the data sources available, and analyzing the
stylized facts of non-residential real estate activity in Greece. Finally, we examine the degree of
covariation (using the Index of Concordance methodology) between non-residential real estate and
the business cycle.
Our results indicate that
i)
the structure of non-residential sector is highly fragmented into various sub-categories and
ii)
at the initial stages of its developments it was strongly affected by the preparations for the 2004
Athens Olympic Games.
iii)
Finally, despite its small share of total GDP, non-residential real estate shows a significant
degree of covariation with the business cycle.
NB: This brief presentation is based on our research paper on “Non-residential real estate and economic activity: the
case of Greece” which was published in the Journal of Property Investment & Finance Vol. 32 No. 1, 2014 pp. 35-55
2
The first step of this analysis is to present the stylized features of non-residential construction activity in
Greece, compare its evolution to the corresponding activity in the Euro Area and finally analyze its relationship
to the Greek business cycle. Given the multi-dimensional nature of non-residential investment activity, we used
a number of different sources in our research:
§
We used quarterly data on Gross Fixed Capital Formation (GFCF) from ELSTAT for Greece and
Eurostat for the Euro Area. Furthermore we decompose GFCF into its main components, namely
“Dwellings”, “Other buildings and structures” – which track closely non-residential investment- and
“Other Investment”.
§
A more detailed picture regarding non-residential activity is provided by the Number of nonresidential building permits, the volume (measured in cubic meters, m3) of non-residential building
permits and the Building permits indicator based on the useful floor area measured in square meters
(m2).
§
An alternative way to analyze the construction sector is through the Production Index in
Construction (PIC), which shows the quarterly activity in the production of building construction and
the production of the civil engineering sector. The index compares the volume of the quarter’s output
with the corresponding figure for the given base period (2005=100) and measures the evolution of
volume trend in value added at factor cost of production in construction for any given reference period.
The main sub-indices of the PIC are the Production Index of Building Construction and the
Production Index of Civil Engineering.
§
Finally, we used the Construction Confidence Index estimated by DG EcFin.
3
Our intention is to determine the general pattern of the data and analyze long-term trends. In order to
examine the relationship between non-residential construction and economic output, we need a
measure of the business cycle for Greece and the Euro Area. The most well established metric for the
evolution of the business cycle is the so-called output gap. In broad terms, the output gap measures
the deviation of GDP from a measure of its long-term trend.
Output Gap Indicators, Greece – Euro Area
2
0
-2
-4
-6
00
01
02
03
04
05
06
Euro Area
Source: European Commission DG ECFIN, EL.STAT., Eurostat, Piraeus Bank Research
07
08
09
10
11
12
Greece
4
According to the system of National Accounts (Eurostat, 1999), total investment in the economy is measured
by Gross Fixed Capital Formation (GFCF). In turn, GFCF can be broken down into residential investment
(“Dwellings”), non-residential investment (“Other buildings and structures”) and “Other investment”. The
evolution of these three components of GFCF is illustrated in the Figure below.
The total GFCF (in real terms) peaked in 2007 and has been declining rapidly ever since. Furthermore, it
becomes obvious that up until 2007, the most significant component of total investment expenditure was
residential investment, with other investment following in second place and non-residential investment
occupying third place.
GFCF Components Cumulative Evolution, Greece,
constant prices 2005, mn €
16,000
12,000
8,000
4,000
0
2000
2001
2002
2003
2004
Other investment
2005
2006
Dwellings
2007
2008
2009
2010
Other buildings and structures
2011
2012
Source: ΕLSTAT, Piraeus Bank Research
5
Total GFCF peaked at 26.6% of GDP in 2007, but due to the severe economic recession has declined to 13.1%
in 2012. Nevertheless, non-residential investment follows a completely different pattern. Non-residential
investment was exceptionally high (as a % of GDP) at the beginning of the century due to the preparations for
the 2004 Athens Olympic Games, peaking at 6.8% in 2001. After the end of the Olympic Games, it declined
sharply in 2005 to 3.6% of GDP and remained around this level until 2007. In 2008 and 2009, it enjoyed a brief
renaissance, just when residential investment was plummeting, resulting in an increase of the non-residential
share of total GDP. After 2009, non-residential investment declined sharply and currently contributes only
3.2% of GDP.
Contribution of GFCF main components to its
rate of change, Greece, constant prices 2005
GFCF in Greece, as % of GDP
28.0%
7.5%
26.0%
7.0%
20.0
24.0%
6.5%
15.0
22.0%
6.0%
10.0
20.0%
5.5%
18.0%
5.0%
16.0%
4.5%
14.0%
4.0%
12.0%
3.5%
25.0
%
5.0
0.0
-5.0
-10.0
-15.0
3.0%
10.0%
2000
2001
2002
2003
Total (LHS)
2004
2005
2006
2007
2008
2009
2010
2011
Other buildings and structures (RHS)
2012
-20.0
-25.0
2001
2002
Dwellings
2003
2004
2005
2006
Other buildings and structures
2007
2008
2009
Other investment
2010
2011
2012
Total (YoY change)
Source: ΕLSTAT, Piraeus Bank Research
6
The different behavior in non-residential investment between Greece and the Euro Area is clearly
demonstrated in the Figures below.
At the beginning of the previous decade Greek non-residential investment was clearly driven by the
need for Olympic Games related infrastructure projects, while after 2006, it follows Euro Area trends but
with a 2-year lag. Similarly the construction sector’s confidence at the beginning of the period is much
higher in Greece that in the Euro Area.
Construction Confidence Index,
Greece and Euro Area
GFCF of Other Buildings and Structures, Greece
and Euro Area, constant prices 2005, mn €
40
128,000
2,600
124,000
2,400 20
120,000
2,200
116,000
2,000
112,000
1,800
108,000
1,600
104,000
1,400 -60
0
-20
-40
100,000
1,200
2001
2002
2003
2004
2005
2006
Euro Area (LHS)
2007
2008
2009
Greece (RHS)
2010
2011
2012
Figure 8.
GFCF of Other Buildings and
Structures, Greece and Euro
Area, constant prices 2005,
mn euros
-80
2000 2001 2002 2003 2004 2005 2006 2007 2008 2009 2010 2011 2012
Euro Area
Greece
Source: ΕLSTAT, Eurostat, European Commission DG EcFin, Piraeus Bank Research
7
Two of the best documented features of residential construction activity are the high degree of covariation its
exhibits with overall economic activity within each economy as well as the high level of response to global
economic shocks (Cesa – Bianchi, 2011).
On the contrary, the examination of building permits per m2 both in Greece and the Euro Area (see Figures)
reveals the rather idiosyncratic behavior of non-residential activity. One possible explanation for this
difference in the behavior of residential versus non-residential activity is the rather fragmented nature of nonresidential activity. Whereas residential activity is more or less homogeneous, non-residential construction
consists of many extremely heterogeneous sub-groups.
Building Permits Index, Greece,
m2 of useful floor area (2010=100)
500
Building Permits Index, Euro Area,
m2 of useful floor area (2010=100)
320
280
400
240
300
200
200
160
100
120
0
2004
2005
2006
2007
T otal
2008
2009
Non residential
2010
Residential
2011
2012
80
2004
2005
2006
2007
T ot al
2008
2009
Non-residential
Source: Εurostat, Piraeus Bank Research
2010
2011
2012
Residential
8
According to the breakdown of the number of non-residential building permits, 32% of the total is for retail spaces,
i.e. shops and shopping malls, 10% for industrial structures and 9% for offices. Indicative of the heterogeneity of
non-residential real estate activity is the fact that unclassified structures comprise 29% of all permits while the
“Others” category, which consists of classified yet too small to report structures, accounts for 12% of the total.
Furthermore, the decomposition of non-residential activity based on volume – reveals an even more nuanced
picture. When examined from this perspective, industrial structures dominate making up 38% of total, with retail
second at 19% and offices at a distant 5%. Again unclassified and Others account for 14% and 10% respectively.
Greece, Number of building permits
(% of total, average 1997 -2011)
Collective
0%
Livestock
2%
Hotels
3%
Greece, Building permits volume in m3
(% of total, average 1997 -2011)
Medical care
0%
Collective
0%
Medical care
1%
Hotels
6%
Others
10%
Others
12%
Non Specified
29%
Livestock
3%
Retail Spaces
19%
Retail Spaces
32%
Industrial
38%
Industrial
10%
Offices
Educational 9%
Institutes
1%
Agricultural
2%
Non Specified
14%
Educational
Institutes Offices
1%
5%
Agricultural
3%
Source: ΕLSTAT, Piraeus Bank Research
9
A final, but equally important question relates to the degree of comovement between the variables
that track non-residential investment and the business cycle given by the output gap.
The most common method of measuring the degree of association or comovement uses the
standard linear correlation, “ρ” coefficient. Although correlation is the most well known and
intuitive measure of association it suffers from a well documented shortcoming. It is affected by both
the duration of the comovement of two variables as well as the size of the comovement. In other
words, a high correlation value can be estimated when either both variables move for a long time in
the same direction, or when the two variables move briefly but sharply in the same direction
(McDermott and Scott 2000).
One statistic that is immune to this shortcoming is the concordance index proposed by Harding
and Pagan (2001). According to Harding et al (2001) the concordance index measures the
fraction of time two variables spend in the same phase of the cycle. In what follows we first
estimate the turning point of the cycle for all variables under examination and then proceed
to make an estimation of the correlation and concordance indices.
10
The turning points are derived according to a Bry - Broschan algorithm (BBQ). According to the
BBQ algorithm (Bry and Boschan, 1971 & Harding, and Pagan, 2002), in the BBQ algorithm, turning
points are identified when a local peak (trough) is observed around two quarters. Each phase of a
cycle should have a minimum duration of 2 quarters, and a whole cycle should have a minimum
duration of five quarters.
In the case of output gap, a contraction (expansion) period is defined as the period between a peak
(trough) and a trough (peak) indicating declining (increasing) GDP growth rates. The turning points of
the business cycle in Greece and the Euro Area are reported below, with the shaded regions indicating
periods of recession.
Euro Area
Greece
2.0
16.0
12
2.0
14.0
10
1.0
12.0
1.0
8
10.0
0.0
8.0
-1.0
6
0.0
6.0
4
4.0
-2.0
2.0
-3.0
2
-1.0
0.0
-2.0
-4.0
0
-2.0
-2
-4.0
-5.0
-6.0
-4
-3.0
-8.0
-6
-6.0
-10.0
2000
2001
2002
2003
2004
GDP growth rate (RHS)
2005
2006
2007
2008
Output Gap - Greece
2009
2010
2011
2012
Turning Points
Note: shaded areas represent recession periods.
Source: ELSTAT, Eurostat, European Commission DG EcFin, Piraeus Bank calculations
-8
-4.0
Q1
Q2
Q3
Q4
Q1
Q2
Q3
Q4
Q1
Q2
Q3
Q4
Q1
Q2
Q3
Q4
Q1
Q2
Q3
Q4
Q1
Q2
Q3
Q4
Q1
Q2
Q3
Q4
Q1
Q2
Q3
Q4
Q1
Q2
Q3
Q4
Q1
Q2
Q3
Q4
Q1
Q2
Q3
Q4
Q1
Q2
Q3
Q4
Q1
Q2
Q3
Q4
-12.0
Q1
Q2
Q3
Q4
Q1
Q2
Q3
Q4
Q1
Q2
Q3
Q4
Q1
Q2
Q3
Q4
Q1
Q2
Q3
Q4
Q1
Q2
Q3
Q4
Q1
Q2
Q3
Q4
Q1
Q2
Q3
Q4
Q1
Q2
Q3
Q4
Q1
Q2
Q3
Q4
Q1
Q2
Q3
Q4
Q1
Q2
Q3
Q4
Q1
Q2
Q3
Q4
-7.0
2000
2001
2002
2003
2004
GDP growth rate (RHS)
2005
2006
2007
2008
Output Gap - Euro area
2009
2010
2011
2012
Turning Points
11
Having identified the periods of expansion and contraction of all relevant variables we turn our
attention to the analysis of the relationship between the cyclical behavior of non-residential
activity and economic activity.
In order to review the similarity of the economic cycles and to measure how closely the cycles
of the indicators are synchronized to the output gap index we use the index of concordance.
The index of concordance (IC) measures the portion of time that two business cycles
spend in the same phase and is defined as:
1
ICir
T
å {S S + (1 - S )(1 - S )}
T
i r
t t
i
t
r
t
i =1
where, Si/t and Sr/t denote the state variable of the business cycle of the indicator i and the
reference variable (i.e the output gap). St: takes the value 1 in expansions and zero in
contractions.
An initial analysis was conducted with the use of the correlation coefficient. The results of this
analysis showed that the correlation index is in rather low levels in a range of 0.23 to 0.51. On
the contrary, the concordance index indicates a relative strong synchronicity between the
business cycles of the variables and the reference variable (i.e output gap).
12
Based on the concordance index we see that in Greece, the GFCF is on average roughly three quarters
of the time in the same cyclical phase as the output gap and in most cases follows the cyclical
fluctuations of the reference variable with an average lag of 2 quarters.
An interesting finding is that, even though all three indicators have entered the recession phase of the
economic cycle, the GFCF followed a downward trend with a 4 - period lag, showing some kind of
resistance to the overall decline of the economic activity.
At the same time, according to GFCF, only one cycle is noted in the Euro Area.
Euro Area
Greece
3.0
3.0
2.0
2.0
1.0
1.0
0.0
-1.0
0.0
-2.0
-1.0
-3.0
-4.0
-2.0
-5.0
-3.0
-6.0
Q1
Q2
Q3
Q4
Q1
Q2
Q3
Q4
Q1
Q2
Q3
Q4
Q1
Q2
Q3
Q4
Q1
Q2
Q3
Q4
Q1
Q2
Q3
Q4
Q1
Q2
Q3
Q4
Q1
Q2
Q3
Q4
Q1
Q2
Q3
Q4
Q1
Q2
Q3
Q4
Q1
Q2
Q3
Q4
Q1
Q2
Q3
Q4
Q1
Q2
Q3
Q4
-4.0
Q1
Q2
Q3
Q4
Q1
Q2
Q3
Q4
Q1
Q2
Q3
Q4
Q1
Q2
Q3
Q4
Q1
Q2
Q3
Q4
Q1
Q2
Q3
Q4
Q1
Q2
Q3
Q4
Q1
Q2
Q3
Q4
Q1
Q2
Q3
Q4
Q1
Q2
Q3
Q4
Q1
Q2
Q3
Q4
Q1
Q2
Q3
Q4
Q1
Q2
Q3
Q4
-7.0
2000
2001
2002
2003
2004
2005
GFCF of other buildings & structures - GR
Output Gap - Greece
2006
2007
2008
2009
2010
GFCF - Turning Points
Output Gap - Turning Points
2011
2000
2001
2002
2003
2004
2005
2006
2007
2008
2009
2010
2011
2012
2012
GFCF of other buildings and structures - EA
Output Gap - Euro Area
GFCF - Turning Points
Output Gap - Turning Points
Source: ΕLSTAT, Eurostat, European Commission DG EcFin, Piraeus Bank Research
13
The production index both in Greece and the Euro Area, seems to be in the same cyclical phase with
the output gap throughout most of the observed period. However, in the case of Greece, we note that in
the output gap there is a cycle fluctuation in 2001, which is not reflected in the production index.
Regarding the Euro Area, the production index of civil engineering implied more cycles than all the other
indices.
Greece
Euro Area
2.0
2.0
1.0
1.0
0.0
-1.0
0.0
-2.0
-1.0
-3.0
-4.0
-2.0
-5.0
-3.0
-6.0
2000
2001
2002
2003
2004
2005
2006
2007
2008
2009
2010
2011
2012
-4.0
Q1
Q2
Q3
Q4
Q1
Q2
Q3
Q4
Q1
Q2
Q3
Q4
Q1
Q2
Q3
Q4
Q1
Q2
Q3
Q4
Q1
Q2
Q3
Q4
Q1
Q2
Q3
Q4
Q1
Q2
Q3
Q4
Q1
Q2
Q3
Q4
Q1
Q2
Q3
Q4
Q1
Q2
Q3
Q4
Q1
Q2
Q3
Q4
Q1
Q2
Q3
Q4
Q1
Q2
Q3
Q4
Q1
Q2
Q3
Q4
Q1
Q2
Q3
Q4
Q1
Q2
Q3
Q4
Q1
Q2
Q3
Q4
Q1
Q2
Q3
Q4
Q1
Q2
Q3
Q4
Q1
Q2
Q3
Q4
Q1
Q2
Q3
Q4
Q1
Q2
Q3
Q4
Q1
Q2
Q3
Q4
Q1
Q2
Q3
Q4
Q1
Q2
Q3
Q4
-7.0
2000
2001
2002
2003
2004
2005
2006
2007
2008
2009
2010
Production index of civil engineering
PIC - Turning Points
Production index of civil engineering - EA
PIC - Turning Points
Output Gap - Greece
Output Gap - Turning Points
Output Ga p - Euro Area
Output Gap - Turning Points
2011
2012
Source: ΕLSTAT, Eurostat, European Commission DG EcFin, Piraeus Bank Research
14
In Greece, the building permits indicator shows a significant degree of concordance with the
business cycle.
In the case of the Euro Area, the concordance index between the time-series is strongly positive,
but still in levels below the respective in Greece.
Greece
Euro Area
3.0
3.0
2.0
2.0
1.0
1.0
0.0
-1.0
0.0
-2.0
-1.0
-3.0
-4.0
-2.0
-5.0
-3.0
-6.0
2001
2002
2003
2004
2005
2006
2007
2008
2009
2010
Non residential building permits (M2) index – GR
M2 - Turning Points
Output Gap - Greece
Output Gap - Turning Points
2011
2012
-4.0
Q1
Q2
Q3
Q4
Q1
Q2
Q3
Q4
Q1
Q2
Q3
Q4
Q1
Q2
Q3
Q4
Q1
Q2
Q3
Q4
Q1
Q2
Q3
Q4
Q1
Q2
Q3
Q4
Q1
Q2
Q3
Q4
Q1
Q2
Q3
Q4
Q1
Q2
Q3
Q4
Q1
Q2
Q3
Q4
Q1
Q2
Q3
Q4
Q1
Q2
Q3
Q4
Q1
Q2
Q3
Q4
Q1
Q2
Q3
Q4
Q1
Q2
Q3
Q4
Q1
Q2
Q3
Q4
Q1
Q2
Q3
Q4
Q1
Q2
Q3
Q4
Q1
Q2
Q3
Q4
Q1
Q2
Q3
Q4
Q1
Q2
Q3
Q4
Q1
Q2
Q3
Q4
Q1
Q2
Q3
Q4
Q1
Q2
Q3
Q4
-7.0
2000
2001
2002
2003
2004
2005
2006
2007
2008
2009
2010
Non residential building permits (M2) index – EA
M2 - Turning Points
Output Gap - Euro Area
Output Gap - Turning Points
2011
2012
Source: ΕLSTAT, Eurostat, European Commission DG EcFin, Piraeus Bank Research
15
We presented one of the first comprehensive discussions regarding the evolution of non-residential real estate activity in
Greece.
§
The first of our key findings is the documentation of the highly fragmented nature of the non-residential real
estate sector.
§
Furthermore, we present evidence that, at least at the beginning of the last decade, non-residential real estate
development in Greece was highly affected by the preparations for the 2004 Olympic Games. That fact in
combination with differences in the rates of economic growth between Greece and the Euro Area meant that
Greek non-residential real estate displays little connection with non-residential activity in the rest of the Euro Area.
§
Regarding the degree of covariation (using the Index of Concordance methodology) between non-residential real
estate and the business cycle, our findings suggest that, despite its small share of total GDP, the various
indicators of non-residential real estate activity show a significant degree of covariation (or concordance) with
the business cycle.
Based on the concordance index we see that in Greece:
§
GFCF is for, on average, roughly three quarters of the time in the same cyclical phase as the output gap and in
most cases follows the cyclical fluctuations of the reference variable with an average lag of 2 quarters.
§
The production index and the building permits indicator seem to be in the same cyclical phase as the output
gap throughout most of the observed period.
§
Even though all three indicators have entered the recession phase of the economic cycle, the GFCF followed a
downward trend with a 4-period lag, showing some kind of resistance to the overall decline in economic activity.
On the contrary, non-residential permits of useful floor usage index (in m2) reacted with a one-period lead.
16
Disclaimer: This note constitute an investment advertisement, is intended solely for information purposes and it cannot in any way be considered investment
advice, offer or recommendation to enter into any transaction. The information included in this note may not be construed as suitable investment for the holder,
nor may it be considered as an instrument to accomplish specific investment goals or relevant financial needs of the holder and may neither be reckoned as a
substitute to relevant contractual agreements between the Bank and the holder. Before entering into any transaction each individual investor should evaluate the
information contained in this note and not base his/her decision solely on the information provided. This note cannot be considered investment research and
consequently it was not compiled by Piraeus Bank according to the requirements of the law that are intended to ensure independence in the sector of investment
research. Information comprised in this note is based on publically available sources that are considered to be reliable. Piraeus Bank cannot be held accountable
for the accuracy or completeness of the information contained in this note. Views and estimates brought forward in this note represent domestic and international
market trends on the date indicated in the note and they are subject to alteration without previous warning. Piraeus Bank may also include in this note investment
research done by third parties. This information is not modified in any way, consequently the Bank cannot be held accountable for the content. The Piraeus Bank
Group is and organization with a considerable domestic and international presence, and provides a great variety of investment services. In cases where conflicts
of interest issues should arise while Piraeus Bank or the rest of the companies of the group provide investment services in relation to the information provided in
this note, Piraeus Bank and the companies of the Group should be underlined that (the list is not exhaustive): a) No restrictions apply in dealing for own account,
or with regards to trading in relation to portfolios managed by Piraeus Bank or companies of its group before the publication of this note, or with regards to trading
before an initial public offer. b) It is possible that investment or additional services are provided to the issuers included in this note against a fee. c) It is possible
that Bank or any of its subsidiaries participate in the share capital of any of the issuers included in this note or may attract other interests financial or not from them.
d) The Bank or any of its subsidiaries may act a market maker or an underwriter for any of the issuers included in the note. e) Piraeus Bank may have issued
similar notes with different or incompatible content with the content of this note. It should be explicitly noted that: a) Figures refer to past performances and past
performances do not constitute a safe indication for future performances. b) Figures constitute simulation of past performances and they are not a safe indication
of future performances. c) Any projections or other estimates are not safe indications for future performances. d) Taxation treatment of information provided in this
note may differ according to the rules that govern each individual investor. Therefore the holder should seek independent advice in relation to taxation rules that
may affect him/her. e) Piraeus Bank is not under any obligation to keep data and information provided herein updated.
17