See discussions, stats, and author profiles for this publication at: https://www.researchgate.

net/publication/323457232

Determinants of Investment in Bulgaria

Article · February 2018

CITATIONS READS

2 263

4 authors, including:

Daniel Kasabov
Bulgarian National Bank
7 PUBLICATIONS   15 CITATIONS   

SEE PROFILE

Some of the authors of this publication are also working on these related projects:

Drivers of Investment Activity View project

All content following this page was uploaded by Daniel Kasabov on 28 February 2018.

The user has requested enhancement of the downloaded file.

 DISCUSSION PAPERS
DP/110/2018

Determinants of Investment
in Bulgaria
Evgeni Ivanov
Neli Ivanova
Daniel Kasabov
Kristina Karagyozova
BULGARIAN NATIONAL BANK

DISCUSSION PAPERS
DP/110/2018

Determinants of Investment
in Bulgaria
Evgeni Ivanov
Neli Ivanova
Daniel Kasabov
Kristina Karagyozova

February 2018
 DISCUSSION PAPERS
Bulgarian National Bank Publications Council
Chairman:
Kalin Hristov, Deputy Governor and Member of the BNB Governing Council
Vice Chairman:
Mariella Nenova, Ass. Prof., Ph. D.
Members:
Andrey Vassilev, Ph. D.
Boryana Pencheva, Member of the BNB Governing Council, Ph. D.
Daniela Minkova, Ph. D.
Ivaylo Nikolov, Ph. D.
Lena Roussenova, Member of the BNB Governing Council, Ass. Prof., Ph. D.
Secretary:
Lyudmila Dimova
Assistant Secretary:
Christo Yanovsky

© Evgeni Ivanov, Neli Ivanova, Daniel Kasabov and Kristina Karagyozova, 2018
© Bulgarian National Bank, series

ISBN 978-619-7409-06-2
ISBN 978-619-7409-07-9 (e-book)

Views expressed in the paper are those of the author and do not necessarily reflect the BNB policy.
Responsibility for the non-conformities, errors and misstatements in this publication lies entirely
D P / 110/2018

with the author.

Send your comments and opinions to:

Publications Council
Bulgarian National Bank
e-mail: [email protected]
Website: www.bnb.bg

2
 Contents

1. Introduction .................................................................................... 5

2. Literature Review ............................................................................ 7

3. Investment Dynamics in Bulgaria ................................................ 10

3.1. General Overview of Investment in Bulgaria .................... 10
3.2. Main Factors behind Investment ....................................... 14

4. Data and Methodology ................................................................. 25

4.1. Methodology ....................................................................... 26
4.2. Data ..................................................................................... 27

5. Results ............................................................................................ 30

6. Conclusion ..................................................................................... 34

Appendices
1. Investment Data .................................................................. 36
2. Survey Results on the Purpose of Investment
and Factors Driving Investment Activity ........................... 39
3. Measures of Uncertainty ..................................................... 41
4. Variables Used ..................................................................... 43
5. EU Funds and Their Role for Government
DISCUSSION PAPERS

and Corporate Investment .................................................. 44

Bibliography ...................................................................................... 47

3
 Abstract: Investment activity in Bulgaria, as in many EU countries, contracted
substantially during the Great Recession and despite the recently witnessed recovery of
business investment remains below pre-crisis levels. The main purpose of this study is
to empirically examine the impact of key drivers of business investment in Bulgaria. We
estimate a Bayesian VAR model to evaluate the importance of cyclical factors, such as
demand, capacity utilisation, economic uncertainty, availability of financing (external
and internal) and cost of financing for investment dynamics over the period 2000Q1 –
2017Q1. We establish external demand as the main driver of business investment in the
country. Investment dynamics in Bulgaria are also strongly dependent on the levels of
economic uncertainty and corporate profits, while foreign financing and lending rates
appear to have lower impact. Our empirical results suggest that the observed slowdown
of business investment in the post-crisis period can be explained by the reduced external
demand and the increased economic uncertainty in the immediate aftermath of the
crisis. After 2013 the negative effects from decreased external demand and heightened
uncertainty fade away and these two factors contribute positively to investment growth.
However, low corporate profits and tightening of financial conditions in the aftermath
of the sovereign debt crisis prevented a more notable recovery. After the second half
of 2015, our results point to a recovery in corporate profits, as well as favourable
financing conditions, which on top of the already observed growth in external demand
and decreasing economic uncertainty support the recovery of investment activity in the
country. A large part of the fluctuations in business investment in most recent years and
especially in 2015 and 2016 that could not be explained by the VAR analysis of demand
factors, uncertainty and financing resources, seems to be driven by the dynamics of EU
co-financed investment.

Keywords: Investment, Structural VAR, Bulgaria

Evgeni Ivanov, Economic Research and Forecasting Directorate, Bulgarian National Bank,
[email protected]
Neli Ivanova, Economic Research and Forecasting Directorate, Bulgarian National Bank,
[email protected]
Daniel Kasabov, Economic Research and Forecasting Directorate, Bulgarian National Bank,
[email protected]
D P / 110/2018

Kristina Karagyozova, Economic Research and Forecasting Directorate, Bulgarian National

Bank, [email protected]
Acknowledgment:
We would like to thank the members of the Publications Council of the Bulgarian National
Bank and particularly Mariella Nenova, Ass. Prof., Ph.D., and Andrey Vassilev, Ph.D., for
their valuable contribution, comments and suggestions.

4
 1. Introduction
Investment plays a key role not only in determining short-run growth prospects
through cyclical swings but also in determining the long-run potential growth
of the country through capital stock accumulation. Without sufficient invest-
ment, the capital stock cannot be renewed regularly, which impedes technologi-
cal progress and prevents capital deepening from increasing labour productivity
(ECB, 2016:1).
Figure 1. Total Investment as Share of GDP
(% of GDP)

Source: IMF WEO April 2017

Investment across countries from the European Union (EU), including Bul-
garia, has been hit hard since the onset of the Great Recession in 2009 and the
DISCUSSION PAPERS

subsequent recovery was slower than in other advanced economies, such as the
US and Japan (see Figure 1). While initially after the Great Recession invest-
ment in the EU started to recover in line with investment in other advanced
economies, the sovereign debt crisis in Europe in 2012 led to a new weakening
of investment in the region (EIB, 2016), including in Bulgaria. The fall in EU’s
total investment witnessed during the Great Recession and the decline in invest-
ment in 2011–2012 were a result not only of lower non-construction invest-
ment1, but also a strong downward correction in overheating housing markets
and persisting budgetary constraints that resulted in lower construction and
public investment in some countries (ECB, 2016:1). As of 2016 total investment

1
Non-construction investment covers investment in (1) machinery, equipment and weapons
systems, (2) intellectual property products and (3) agricultural products.

5
 as a share of GDP remained below its pre-crisis level in the EU even though a
slow recovery has been observed in most Member States since early 2013, driven
mainly by non-construction investment.
Similar dynamics were observed in Bulgaria. Following a period of strong
growth in the period 2002–2008, total investment in Bulgaria decreased sharply
in 2009 with both non-construction and construction investment contributing
to this decrease. In the following years investment in Bulgaria remained weak,
with non-construction investment starting to recover more notably after 2013.
At the same time construction investment remained depressed after the real
2
estate boom and bust and is unlikely to regain its pre-crisis levels soon .
Recent studies on the topic (Barkbu et al, 2015; ECB, 2016:1; Kose et al., 2017)
have identified that the factors that determine the dynamics of non-construc-
tion investment are different from those determining the dynamics of construc-
tion investment. Moreover, these studies suggest focusing on the so-called
“business investment”, which typically encompasses certain components of the
non-construction investment, and is particularly relevant in boosting an econo-
my’s future productive capacity. Relevant studies have identified several factors
which have played a key role for the observed business investment weakness in
the EU after the 2009 crisis. Among them are weak demand, depressed profit
3
margins, impaired financial sectors and economic and political uncertainty .
The aim of this paper is to empirically assess the impact of such cyclical factors
4
on business investment dynamics in Bulgaria and to quantify their contribu-
tions, which will shed more light on the determinants behind the business
investment slowdown observed in the country after the Great Recession. The
estimation is done using a Bayesian Vector Autoregression model (BVAR), based
on quarterly data for Bulgaria for the period 2000Q1 – 2017Q1. The paper also
attempts to examine in more details the role of EU-related investments for the
country’s investment dynamics, given that in recent years public investment,
co-financed by the EU under different operational programs, has played an
important role in Bulgaria, as was the case in most Central and Eastern Euro-
pean countries (EIB, 2016). To the best of our knowledge this paper is the first
D P / 110/2018

attempt to disentangle and quantify the drivers behind business investment

dynamics for the case of Bulgaria.

2
See Section 3 for more details.
3
See Section 2 for a more detailed overview of the main findings in related studies.
4
We use real investment in machinery, equipment and weapons systems as a proxy for business
investment in Bulgaria. For more information on the choice of investment proxy, please refer to
Appendix 1 A.

6
The paper is structured as follows: Section 2 gives an overview of the results
of similar studies on the causes and factors behind the investment slowdown,
focusing mainly on countries from the EU. Section 3 outlines the historical
developments in investment in Bulgaria, offering more descriptive details about
its composition and dynamics, which reinforces the choice of model specifi-
cation and facilitates the interpretation of the empirical results of the paper.
Section 4 presents the technical details on the methodology and data used for
our analysis and the motivation behind choosing them. Section 5 presents and
discusses the results from our analysis. Finally, in Section 6 we conclude by
summarising our results.

2. Literature Review
Given the importance of investment for the overall economic growth, the topic
of what the drivers of investment are and why the recovery in investment has
been weak in most advanced economies in the years after the Great Recession
has become one of the central research topics in many international institutions,
among which are the IMF (2014), the ECB (2016:1), the World Bank (Vashak-
madze et al., 2017) and the EIB (2015). The preferred method used in empirical
studies on the determinants of investment typically is structural VAR models,
which is also the approach that we have adopted in this paper. Evidence suggests
that financial crises have long-lasting effects on investment with a decline of 3 to
3.5 percentage points in the investment-to-GDP ratio three years after the crisis
(Barkbu et al., 2015). The recovery in the aftermath of the 2008–2009 financial
crisis was particularly slow across most EU countries, including Bulgaria, and
the factors behind it have been thoroughly analysed. Some of the identified fac-
tors in the literature include weak demand, economic uncertainty, financing
conditions, profit expectations and capacity utilisation.
DISCUSSION PAPERS

Weak demand and the prospects for sluggish future economic growth have
been established as one of the core reasons for the weak post-crisis recovery
of investment across Europe (EIB, 2015). The European Commission (2017)
has repeatedly identified insufficient demand as one of the most important fac-
tors limiting firms’ production in its industrial surveys, which is likely to have
a negative impact on firm’s investment decisions. Furthermore, according to
analysis conducted by the IMF (2014), output changes can explain trends in
investments activity in most euro area (EA) member states. This was further
supported by an empirical study on the case of Spain which empirically proves,
using an accelerator model, that output almost entirely explains the dynamics of
investment in the country (Barkbu et al, 2015). Busetti et al. (2015) examine the
causes of exceptionally low investment in Italy since 2007 and find that demand

7
 conditions, along with cost of capital, uncertainty and credit supply have a
statistically significant impact on firms’ capital spending. External demand, in
particular, has been found to play a dominant role in explaining investment
dynamics in small open economies (CNB, 2017). Kose et al. (2017) explore the
evolution of investment growth in emerging market and developing economies
(EMDE), using a structural VAR model. They conclude that the investment
slowdown in EMDE could be attributed to a large extent to the weak economic
activity in major economies.
Another important factor for investment dynamics that has been established in
the existing literature is economic uncertainty. In general increased uncertainty
is considered to lead to a postponement of investment (Busetti et al., 2015).
As Bloom et al. (2007) present, heightened uncertainty, apart from affecting
directly investment, also reduces the responsiveness of investment to a demand
increase, which reduces the effects of policy stimulus. There are many other
studies that examine the potentially adverse impact of uncertainty on invest-
ment. A recent study by Meinen and Rohe (2016) develops five types of uncer-
tainty measures to investigate the role of uncertainty for investment decisions
with a focus on Germany, France, Italy and Spain. Based on both descriptive
analysis and VAR models, the authors conclude that investment responds nega-
tively to uncertainty shocks and that uncertainty can account for a significant
portion of the decrease in investment in the course of the Great Recession.
A report published by the EIB (2013) proves that uncertainty, together with
demand, have been the main drivers of the investment slowdown in the euro
area since 2010. A study by the European Commission (2013) concludes that the
increase in uncertainty has a significant negative effect both on investment and
on consumption. Busetti et al. (2015) confirm that uncertainty had a negative
impact on investment during the Great Recession and establish it as one of the
factors explaining the delayed recovery of the economy.
Credit constraints can also have a sizeable negative impact on investment since
they limit the availability of resources that firms could use to expand their
activities. However, studies for the euro area find that financing conditions
D P / 110/2018

have been a constraining factor only in certain countries. Barkbu et al. (2015)
establish a negative correlation between financial constraints and investment for
Italy, Portugal and Spain. In addition, applying an extended version of an accel-
erator model, Busetti et al. (2015) estimate that for the case of Italy constraints
imposed by tight credit conditions accounted for about one-third of the fall in
investment in the periods 2008–2009 and 2012–2013. The ECB (2016:1) uses a
VAR model and finds that credit constraints have been one of the main factors
supressing the euro area’s business investment in the years after the sovereign
debt crisis.

8
Earned profit is another factor that has been found to play a role in firms’
decisions to invest. Sustaining profits means that firms can invest without
having to borrow funds from credit markets and, thus, could easily and more
quickly expand their production if they assess that there is enough demand for
their products. Profitability is also a positive signal which may lead to further
investments in case returns are high enough to compensate for the risk and the
cost of capital (EIB, 2016). The ECB (2016:1) finds that the current business
investment recovery in the euro area is being driven mostly by profit growth,
improving demand and falling interest rates. Mizen and Vermeulen (2005) con-
firm that investment in Germany and the UK is sensitive to sales growth and
operating profits.
Capacity utilisation could also be considered as a factor that can determine invest-
ment activity (ECB, 2016:1, Busetti et al., 2015), since it reflects firms’ choices on
how to meet current and expected demand. Capacity utilisation can give an indi-
cation of the cyclical position of the economy. For example, in the aftermath of the
Great Recession demand was greatly reduced and as a result capacity utilisation
decreased, which in turn made firms’ new investments in production extension
unjustified. On the other hand, in times of economic expansion when demand is
booming and capacity utilisation is rising, firms are under pressure to invest more
in order to extend or improve their production capabilities.
In addition to traditional borrowed and own financial resources, transfers from
the EU have also played an important role in determining investment dynamics,
especially in new member state countries (EC, 2016:2). EU co-financed invest-
ment is unevenly spread across the years and as such affects significantly not
only the level, but also the growth rate of economy wide investment (Eesti Pank,
2016). In some member states, EU transfers have played a countercyclical role
in the aftermath of the crisis, offsetting declines in other investment (EC, 2016:2
DISCUSSION PAPERS

and Eesti Pank, 2016). For other member states, positive contributions from
EU co-financed investment were concentrated mostly in 2014–2015 and apart
from their impact on government investment they have also played a prominent
role in explaining the dynamics of investment by non-financial corporations
(CNB, 2016). In 2016 EU funds absorption declined at the switchover between
the programing periods, having an adverse impact on investment growth in all
Central and Eastern European countries (NBP, 2017).
Having already identified the main cyclical factors proposed in the economic
literature as key drivers of investment dynamics, it is worth discussing in more
details how investment in Bulgaria has evolved over the years and which of the
outlined factors have been in play.

9
 3. Investment Dynamics in Bulgaria
This section analyses the investment dynamics in Bulgaria over the period
2000–2016 and aims at identifying the main drivers of investment in Bulgaria
before, during and after the global financial crisis, outlining the possible factors
behind the slow recovery of investment after 2009. The main findings are used
as an input for section 5 where we quantify the impact of the identified factors
on investment activity through the lens of a Bayesian VAR model.

3.1. General Overview of Investment in Bulgaria

Gross fixed capital formation in Bulgaria accounts for more than 20% of GDP.
As such it plays a key role not only in determining short-run growth prospects
through cyclical swings but also in determining the long-run potential growth
and the productive capacity of the economy.
Following the introduction of the currency board arrangement in July 1997 and
the achieved macroeconomic stabilisation, Bulgaria had significant investment
needs. This reflected the low initial level of the capital stock (relative to the
5
size of the economy) compared with other new member state countries (NMS)
and the high rate of depreciation of the existing capital stock (see Table 1). In
addition to this, a mix of cyclical factors that we will discuss here has also con-
tributed to developments of investment in the country.

Table 1. Indicators Related to Investment Activity in Bulgaria

Indicators 2000–2003 2004–2008 2009–2011 2012–2016
Investment (% of GDP) 18.8% 27.2% 23.7% 20.7%
Nomianl interest rates on new business, non-fnancial corporations 10.9% 8.5% 8.7% 7.0%
Real (ex-post) interest rates on new business, non-fnancial corporations 6.4% 1.7% 5.6% 7.8%
Capital stock-to-GDP ratio 1.9 1.8 2.1 2.2
Capital stock-to-GDP ratio (new member states*) 2.0 2.0 2.2 2.3
Capital intensity (net capital stock at 2010 prices per person employed) 30.04 32.98 42.78 51.19
Consumption of fixed capital (% of previous period net capital stock) 8.1% 9.1% 7.5% 6.7%
D P / 110/2018

Consumption of fixed capital (% of GDP) 14.4% 15.1% 14.8% 14.4%

Marginal efficiency of capital (change in GDP at constant market prices of
year T per unit of gross fixed capital formation at constant prices of year T-5) 0.28 0.26 0.01 0.09
Net returns on net capital stock (2010 = 100) 104.68 127.59 103.57 84.87
*New member state countries (NMS) are Czech Republic, Estonia, Croatia, Latvia, Lithuania, Hungary,
Poland, Romania, Slovenia and Slovakia.
Source: European Commission (AMECO database), BNB

5
Here we include the countries that joined the EU after 2004 and that have similar economic and
historical development to Bulgaria. These countries are Czech Republic, Estonia, Croatia, Latvia,
Lithuania, Hungary, Poland, Romania, Slovenia and Slovakia.

10
Based on its dynamics, investment activity in Bulgaria could be divided into
three main periods. The first period spans from 2000 to 2008 and is charac-
terised by an investment boom. The share of investment during that period
increased substantially from a low of 16.8% of GDP in 2000 to a peak of 33.0%
in 2008. This period was characterised by favourable cyclical developments
in Bulgaria and in our main trading partners from the EU, the construction
boom in Bulgaria over the period 2004–2008 and a decline in investors’ risk
perceptions. The second period covers the post-crisis years (2009–2011) and is
characterised by a decline in investment. The cumulative decline in real gross
fixed capital formation for that period was around 35%. The observed weak
recovery of investment activity in 2012–2016 can be distinguished as the third
period. During this period the growth rate of real gross fixed capital formation
in Bulgaria averaged 0.8% per year for a total increase of 4.1%, with government
investment playing an important role. The EU funds programming cycle has
had a notable effect on investment in this third period, with EU funds support-
ing investment activity in the period 2012–2015 and leading to a temporary
decline in 2016.
We can generally distinguish between two broad categories of investment – con-
struction investment and non-construction investment. Examining the break-
down of gross fixed capital formation by asset type could shed more light on
how the identified factors have affected investment activity over the post-crisis
period and which components have driven the observed investment recovery
in recent years. Data suggests that the decline in investment after the global
financial crisis was the consequence both of a strong downward adjustment in
the construction sector and of lower non-construction investment activity (see
Figures 2 and 4). Eight years after the crisis gross fixed capital formation in
construction, both residential and non-residential, is still around 40% below its
DISCUSSION PAPERS
pre-crisis peak level. Although gross fixed capital formation in assets other than
construction has recovered somewhat recently, it is still more than 20% below
the peak reached in 2008 (see Figure 4).

11
 Figure 2. Breakdown of Real Total Investment
(annual percentage changes, percentage points)

Figure 3. Non-construction Investment Growth Rate and Contribution

by Major Asset Type
(annual percentage changes, percentage points)

Source: Eurostat

The decline in construction investment after the global financial crisis reflected
D P / 110/2018

mainly a prolonged period of housing market adjustment to the changed eco-

nomic environment in Bulgaria. The heightened economic uncertainty, negative
labour market developments, increased saving rate, negligible FDI inflow to
the residential sector and tightened financial conditions suppressed household
demand for dwellings, resulting in a prolonged period of decline in house prices.
Starting from the second half of 2008, house prices tended to decline strongly
and by the end of 2011 they were already more than 30% lower than the pre-
crisis levels. This dampened investment in the construction sector (BNB, 2017).

12
Apart from housing, the depressed economic activity reduced the need for new
non-residential buildings and also forced the public sector to cut back on infra-
structure expenditures, especially in the period 2010–2011.

Figure 4. Real Investment Compared with the Pre-crisis Peak Levels

(index, 2008 = 100)

Figure 5. Ratio of Real Non-construction Investment to Value Added

(index, 2008 = 100)

DISCUSSION PAPERS

Note: Long-term average is calculated for the period 1998–2016.

Source: Eurostat

Non-construction investment declined markedly in period 2009–2010 and

started to recover afterwards, though at much slower growth rates compared
with the pre-crisis period. Currently, the level of non-construction investment is
significantly below the pre-crisis peak observed in 2008 (see Figure 4). Figure 5
indicates that in the run-up to the global financial crisis there was some “over-
investment” in Bulgaria, measured by the ratio of real non-construction invest-

13
 ment to value added6, which possibly reflected favourable cyclical conditions,
high marginal return on capital and positive firms’ expectations about future
demand, lower cost of finance and large FDI flows. In the aftermath of the global
financial crisis both the level and growth rate of non-construction investment
fell by much more than the value added. Following the cyclical recovery of the
economy, however, the ratio of real non-construction investment to value added
started to rise and now it is close to its long-run average7. In the years after the
2008/09 crisis, non-construction investment, which in Bulgaria consists almost
entirely of investment in “machinery, equipment and weapon systems”, has
become the main driver of gross fixed capital formation (see Figures 3 and 4).
Having briefly discussed how investment in Bulgaria has evolved over the years
we now turn to a more detailed discussion on the likely significance of key fac-
tors that drive investment decisions in Bulgaria.

3.2. Main Factors behind Investment

As already discussed in Section 2, according to the economic literature the level
of firms’ desired capital stock is determined mostly by demand conditions,
expectations of return on capital, the cost of financing, availability of funding
and economic uncertainty. In the case of Bulgaria, EU funds are also likely to
play a role for investment dynamics. The detailed elaboration of the potential
effects of the identified factors on investment dynamics in Bulgaria would be
useful in order to interpret the results from our VAR model in Section 5.

Demand Conditions
Demand for investment in Bulgaria was strong in the post-1997 period taking
into account the low initial level of the capital stock and the high rate of depre-
ciation of the existing capital stock (see Table 1). Since mid-1997 the Bulgarian
economy began to open up, with the exports of goods and services-to-GDP ratio
reaching 64% in 2016, up from 36% in 2000. The EU became Bulgaria’s most
important trading partner and the trade and financial linkages between them
deepened further as a result of Bulgaria’s accession in the EU. Apart from trade,
D P / 110/2018

this process was also supported by the observed at the time large FDI inflows
from EU countries in the financial sector and in the export-oriented manufac-

6
The ratio of real non-construction investment to value added gives a measure of the size of
investment in the overall economy. As this ratio tends to be pro-cyclical over time, comparing it with
the long-term value gives an indication of the cyclical behaviour of investment and the presence of
over/under-investment (ECB, 2016:1).
7
Long-run average is calculated for the period 1998–2016. In addition to the business cycle, the ratio
of real business investment to gross value added reflects factors such as the capital intensity of the
economy (see Table 1) and structural changes over time.

14
turing sector, which helped facilitate the integration of Bulgaria in Global Value
Chains. As a result, investment activity in Bulgaria became more and more
closely related to the cyclical developments in the EU, and, respectively, to the
external demand for Bulgarian goods and services (see Figure 6).

Figure 6. Non-construction Investment and External Demand

for Bulgarian Exports (at constant prices)
(%, percentage change)

Figure 7. Change in Capacity Utilisation in the Manufacturing Sector

and Growth Rate of Real Non-construction Investment
(%) (percentage points)

DISCUSSION PAPERS

Note: Non-construction investment in Figures 6 and 7 is defined as overall investment excluding construction
and is used as a proxy for investment in manufacturing.
Source: ECB, NSI

Over the period 2000–2008 demand for investment was stimulated by favour-
able cyclical conditions in both Bulgaria and our main trading partners from
the EU. On the one hand, firms were faced with strong external and domestic

15
 demand at a time of high capacity utilisation, which pushed up non-construc-
tion investment (see Figure 7). On the other hand, the rising demand for new
dwellings and non-residential buildings, stimulated by the positive labour
markets developments, the easing financial conditions and FDI inflows in real
estate activities, supported investment in the construction sector (Kotseva and
Yanchev, 2017).
Survey data reveals that during the pre-crisis period more than 40% of firms in
the manufacturing sector tended to report demand as the most important posi-
tive driver of investment8. As a result, the share of investment serving extension
purpose increased from 32% in 2002 to 50% in 2008. The increase in capacity
utilisation observed at that time in the manufacturing sector was another factor
which necessitated further expansions of firms’ capital stock (see Figure 7).
During the post-crisis period (2009–2011) investment in Bulgaria was pushed
down by cyclical conditions as the economy (especially the construction sector)
underwent large downward economic adjustments, accompanied by a gradual
deleveraging of the private sector. Due to a prolonged depressed domestic
demand, investments were aimed mostly at rationalisation rather than expan-
sion of productive capacity. At the same time, given the importance of exports
to the EU for the total growth of the Bulgarian economy, weak external demand
over the period 2009–2011 amid the sovereign debt crisis in Europe amplified
negative domestic developments. The deterioration of demand conditions cre-
ated significant production capacity under-utilisation during the post-crisis
period which limited the need for further increases in gross fixed capital for-
mation. With the strengthening of economic activity in Bulgaria and the EU
in recent years, investment activity has started to recover and since 2014 the
share of firms that undertake investment in order to expand productive capacity
has been increasing. Data on the capacity utilisation rate in the manufacturing
sector also suggests increasing needs for further investment.

Economic Uncertainty
Economic theory suggests that uncertainty has a detrimental effect on economic
D P / 110/2018

activity by giving economic agents the incentive to postpone investment, con-

sumption and employment decisions until uncertainty is resolved, and by push-
ing up the cost of capital through increased risk premia (EC, 2013).
In this paper we use five uncertainty indicators, constructed by Ivanov (2018),
following the methodology of Girardi and Reuter (2017). These indicators are
based on monthly qualitative survey data, taken from the European Commis-

8
See Appendix 2 “Survey results on the purpose of investment and factors, driving investment
activity” for more information.

16
sion’s Business and Consumer Survey. The core underlying assumption is that the
more economic agents disagree in their assessment and expectations about cur-
rent and future economic activity, the higher is the uncertainty in the economy9.
In this section we present only the composite uncertainty indicator for Bulgaria,
which consolidates the information of the different uncertainty indicators (see
Figure 8) but for the purposes of the VAR model we use all five uncertainty indi-
cators before selecting the most appropriate one for the empirical estimation.

Figure 8. Composite Uncertainty Indicator, Bulgaria

(standard deviation from mean)

Note: Since the index is standardised a value of zero would mean that the level of uncertainty in the economy
is “average”, while values above zero indicate periods with heightened uncertainty and values below zero
are associated with periods with “below-average” uncertainty. Movements along the Y-axis are measured in
standard deviations due to the standardisation transformation.
Source: EU Commission BCS Survey, own calculations

Based on this indicator we can see that the global financial crisis and sovereign
debt crisis triggered an increase in uncertainty in Bulgaria (indicated by the
shaded area in Figure 8), likely depressing economic activity and investment. In
DISCUSSION PAPERS
recent years (2013–2017) uncertainty has declined in line with the strengthen-
ing of economic activity in Bulgaria and our trading partners, which is likely to
have had a positive effect on investment dynamics.

9
For more information on the construction and specificities of these indicators please refer to
Appendix 3.

17
 Figure 9. Uncertainty and the Business Cycle in Bulgaria
(%) (standard deviation from mean)

Notes: “Output gap estimates” refers to our assessment of the output gap in Bulgaria, estimated using a
production function approach on quarterly data. The composite uncertainty indicator is also transformed to
have quarterly frequency in order to be comparable to the output gap estimates.
Source: NSI, EU Commission BCS Survey, own calculations

As expected, the uncertainty index in Bulgaria moves in the opposite direction

of the business cycle in the country (see Figure 9). One reason is that negative
news shocks (such as terrorist attacks or wars) that can trigger recessions also
cause higher uncertainty. Another reason for heightened uncertainty during
recessions is that recessions might themselves increase uncertainty (ECB,
2016:2).

Cost of Financing and Availability of Funding

The investment boom in Bulgaria over the period 2000–2008 was sustained by
large FDI inflows (see Figure 10). FDI inflows followed an upward trend from
5.6% of GDP in 2002 to a peak of 27.9% in 2007 and 18.1% in 2008. Most of the
FDI inflows were absorbed by the construction, real estate, financial, manufac-
turing, trade and transport sectors. On average, around half of the FDI inflows
in Bulgaria were used to finance corporate investment (mostly manufacturing
D P / 110/2018

and development of domestic trade infrastructure), one fifth were allocated to

the banking sector (stimulating domestic credit), while the remainder went to
the corporate real estate sector – largely financing the building of vacation real
estate aimed at foreigners (Mitra, 2011). Rapid credit growth in the pre-crisis
period and declining interest rates further supported accumulation of capital
(see Table 1 and Figure 11). This reflected sound macroeconomic fundamentals
(successful macroeconomic stabilisation, restoration of confidence in the bank-

18
ing sector and strong GDP growth), capital inflows and increased competition
in the banking sector following the privatisation of state banks.

Figure 10. FDI Inflows by Sector of Economic Activity

(percent of GDP)

Figure 11. Investment and Cost of Financing

(%)

DISCUSSION PAPERS

Source: European Commission (AMECO database), BNB

The global financial crisis and sovereign debt crisis increased economic uncer-
tainty across Europe and FDI inflows decelerated markedly due to investors’
increased risk-aversion. As a result one of the financing channels that Bulgarian
firms had at their disposal in the pre-crisis period was substantially reduced. At
the same time, real financing costs for firms (proxied by real (ex-post) inter-
est rates) increased in the post-crisis period (see Table 1) as spreads rose and
banks experienced problems with deteriorating portfolios. Thus, bank lending

19
 conditions deteriorated and credit-supply shocks limited investment activity.
According to the Bulgarian National Bank’s quarterly bank lending survey,
banks tightened credit standards markedly over the period from the third quar-
ter of 2008 to the first quarter of 2010 (Karamisheva, 2016). Since the beginning
of 2010 an easing of credit standards has been observed. In recent years nominal
and real costs of external financing have trended downwards and have become
increasingly supportive of business investment due mostly to the expansionary
monetary measures in the euro area. With the recovery of bank lending, more
firms have been able to take advantage of low interest rates to invest more.

Corporate Profits and Expectations of Returns on Capital

Other factors that determine investment decisions are the dynamics of cor-
porate profits and expectations of returns on invested capital. The levels of
corporate profits and retained earnings influence firms’ capacity to finance
investment with internal means. During the pre-crisis period gross investment
rate tended to increase in line with the upward trend of gross profit share10 (see
Figure 12 or net returns on capital stock in Table 1). This profitability-type
indicator corresponds to the share of the value added created during the pro-
duction process remunerating capital and could be viewed as an approximation
of operating profit before depreciation. The increase in gross profit share for
the overall economy reflected the low initial level of the capital stock and thus
the high marginal efficiency of capital (see Table 1). Declining risk perceptions
and expectations of higher future returns based on the EU membership and
prospects of euro adoption (IMF, 2007) further stimulated investment activity.
After the global financial crisis firms’ profitability, measured by gross profit
share, deteriorated, likely having a negative effect on gross fixed capital for-
mation. Gross profit share dynamics was a consequence of successive years of
lacklustre GDP growth rate and a prolonged period of housing market adjust-
ment. In addition to the business cycle, structural factors also contributed to
the weakening of investment activity. The post-crisis period was marked by a
substantial slowdown in TFP growth and potential output growth, which were
D P / 110/2018

not expected prior to the crisis (see Figure 13). As a result economic agents
revised corporate profit growth expectations downwards compared with the
pre-crisis period, thereby depressing domestic and foreign investment.

10
Gross profit share is a measure of profitability and is calculated as the ratio of gross operating
surplus to value added.

20
 Figure 12. Gross Investment Rate and Profit Share (Total Economy)
(%) (%)

Figure 13. Breakdown of Total Factor Productivity (TFP) Growth Rate

(%, percentage points)

DISCUSSION PAPERS

Notes: (i) Non-construction investment in Figure 12 is defined as overall investment excluding construction;
(ii) TFP decomposition in Figure 13 is based on estimates of the European Commission. According to the
methodology used, potential output is estimated by using a Cobb-Douglas production function, and unlike
standard approaches it explicitly accounts for the level of efficiency with which inputs (capital and labour)
are being used in the production process. As a result, TFP is explicitly modelled as a function of the level of
efficiency of factor inputs (Havik, K. et al., 2014).
Source: NSI, European Commission (AMECO database)

21
 Impact of EU Funds on Public and Private Investment
The programming cycle of EU funds is another factor that influences investment
activity since the country joined the EU in 200711. Similar to other NMS coun-
tries, the absorption of EU funds in Bulgaria has probably played an important
role in explaining recent developments in business investment that could not be
explained by demand factors, uncertainty or financing resources.
The largest part of the EU structural and investment funds in Bulgaria under
both 2007–2013 and 2014–2020 financial frameworks have been allocated in
12
support of government investment in transport and environmental projects .
Some of the infrastructure projects have also been undertaken by state owned
companies that are classified outside the general government sector (a notable
example is the expansion of the underground railway in Sofia with financial
support from the EU and the national budget amounting to BGN 1 billion or
8% of the 2007–2013 ERDF and Cohesion Fund allocations)13. EU transfers
to private companies have remained more limited, even though their role has
increased in the 2014–2020 programming period as compared to the 2007–2013
financial framework. As an alternative to grant funding, the EU-wide priorities
for the current programming period have focussed more on various financial
instruments, which mobilise banks’ financing and private capital to support
investment projects. The role of these instruments for stimulating productive
private investment has been strengthened by the additional mechanisms, avail-
able under the Investment plan for Europe, launched in 201414.
Given the financing priorities of the EU cohesion policy, one could expect three
main channels through which EU co-financed investment could stimulate busi-
ness investment: 1) providing direct financial support for individual investment
projects in non-financial corporations, 2) creating favourable conditions for
business investment through better infrastructure, technologies and enhanced
human capital as a result of higher government investment and 3) lowering the
cost of bank lending for business investment projects as a result of either explicit
guarantees through various financial instruments in place or lower risk profile
D P / 110/2018

of EU co-financed investment projects.

There are, however, concerns that the stimulating impact of EU funds might
remain limited if EU co-financed investment does not add to but rather replaces

11
Before joining the EU in 2007 Bulgaria had access to pre-accession EU funds. For further details of
the pre-accession EU funds please refer to Paliova and Lybek (2014).
12
Details on the allocations and financing priorities in the 2007–2013 and 2014–2020 programming
periods can be found in MF (2016).
13
Source: Unified management system for the EU structural instruments in Bulgaria, www.eufunds.bg.
14
See Appendix 5 on detailed information for EU funds allocations and financial priorities.

22
market-based investment. Moreover, if national government spending on co-
financing EU projects increases considerably and this translates into tax or
government debt increases, real interest rates may also increase, resulting in
private investment crowding-out. Last but not least, if EU grants are not spent
efficiently, their impact on business investment might be rather limited and
short-lived.
Given the large scope of EU co-financed extensions and improvements in
15
existing infrastructure in Bulgaria , there is a strong argument in support of a
positive overall impact of EU funds on business investment. While substitution
effects may not be completely ruled out, the access to cheaper and more certain
EU-related financing is likely to have supported the implementation of projects
that may not have been undertaken otherwise, especially in the aftermath of
the global financing crisis, when investment activity was impeded by growing
uncertainty and scarce external and domestic sources of financing.
The available data on EU funds absorption rates, EU co-financed government
investment and capital transfers from EU budget to non-financial corporations
(see Appendix 1B on available data sources) indicates that EU related flows to
both the government and non-financial corporations have played an important
role in explaining the dynamics of business investment, especially in most recent
years. The impact of EU funds-related flows was relatively weak until the end
of 2011 as the absorption rate (as defined by the rate of payments relative to the
16
overall funding available ) was less than a quarter of the allocated Cohesion
policy budget (23.8%). The absorption rate strongly increased in 2013–2015
and as a result EU co-financed government investment had a notable positive
contribution for economy-wide investment in 2012–2014 and was the main
factor behind investment growth in 2015 (see Figure 15). Capital transfers from
government to non-financial corporations, related to the implementation of
DISCUSSION PAPERS

EU co-financed projects, have also increased substantially after 2012 and have
peaked in 2015.

15
The implementation of infrastructural investment projects under the 2007–2013 programming
period resulted in the completion of 175 km. of new roads and the reconstruction of respectively
1040 km. and 234 km. of roads and railroads (data as of 2014, EC, 2016:1)
16
MoF data on Cohesion policy payments (both EU and nationally co-financed). The total budget
has been augmented in the course of the implementation of the programming periods to reflect shift
in financing and deductions.

23
 Figure 14. Capital Spending on EU Programmes
(% of GDP)

Note: Four quarters rolling data, includes capital spending in government units and transfers to non-
government units on programmes financed by the pre-accession funds, Structural funds, the Cohesion fund,
European Agriculture Fund for Rural Development and European Fisheries Fund, as well as other projects
and programs, financed with EU funds. Includes co-financing from the national budget.
Source: MoF monthly reports on the implementation of the Consolidated fiscal programmes

Figure 15. Contribution of EU Co-financed Government Investment to the

Growth of Economy Wide Investment
(%, percentage points)
D P / 110/2018

Note: EU co-financed government investment is proxied by quarterly data on capital transfers received by
the government and an additional adjustment is made to account for co-financing from the national budget.
Source: Eurostat, own estimates

The switchover of the 2007–2013 and the 2014–2020 financial frameworks

resulted in delays in the implementation of projects under the new program-

24
ming period17. As a result there was a sharp decline in EU co-financed invest-
ment in 2016 that caused a decline of economy wide investment, despite
favourable non-government investment and national government investment
dynamics (see Figure 15). EU flows to direct SME support and other capital
transfers to non-financial corporations were also lower as compared to the peak,
recorded in 2015.
In summary of this section, it can be argued that mostly cyclical factors sup-
ported the slow investment recovery in recent years. The investment recovery in
Bulgaria reflected mainly increased non-construction investment, particularly
business investment, and is likely to have been driven by improving demand,
lower economic uncertainty and easing financing conditions, with additional
support from EU-fund related investments.

4. Data and Methodology

Having reviewed the related literature on the topic and having identified the
most relevant drivers behind investment dynamics we can now focus on build-
ing the econometric framework which will be used to analyse business invest-
ment in Bulgaria. The descriptive analysis above reveals that investment activity
in Bulgaria over the period 2000Q1–2017Q1 was driven mostly by the cyclical
developments in factors, such as external demand, uncertainty and financing
conditions. The quantification of the relative importance of each of these factors
could be estimated through the lens of a Bayesian VAR model.

4.1. Methodology
From a methodological perspective, we use Bayesian VAR models to estimate
the impact of the identified factors on business investment activity. Vector
DISCUSSION PAPERS

autoregression models (VARs) are very flexible time series models that can
capture complex dynamic interrelationships between macroeconomic variables.
The flexibility and ability to fit the data comes from the rich parameterization
of VAR models. However, that comes at the risk of overfitting the data, of having
imprecise inference and large uncertainty about the future paths projected by
the model. The relatively short time span of most macroeconomic time series for
countries like Bulgaria further aggravates the problem of dense parametrization.
A solution to this issue is to employ Bayesian techniques to estimate the VAR
model. In Bayesian estimation we use informative priors in order to shrink the

17
The delay was common across all NMS and was explained largely by common factors: complex
legislative and administrative procedures, which require: 1) adoption of relevant EU regulations at
the EU level, and 2) negotiations between member states and the EU for finalising the operational
programmes and partnership agreements.

25
 richly parameterised unrestricted model towards a parsimonious naïve bench-
mark model, hence reducing estimation uncertainty (Giannone et al., 2012).
A general VAR model with n endogenous variables, p lags and m exogenous
variables can be written as:
yt = A1yt-1 + A2yt–2 + … + Apyt-p + Cxt + εt, where t = 1, 2 …, T (1)
E(εt) = 0 (2)
E(εt ε's) = Σ if t = s (3)
E(εt ε's) = 0 if t ≠ s (4)
yt = (y1,t , y1,t , ..., yn,t) is a n×1 vector of endogenous variables, A1, A2, …, Ap are
p matrices of dimension n×n, C is a n×m matrix, and xt is a m×1 vector of
exogenous regressors which can be constant terms, time trends, or exogenous
data series. εt = (ε1,t , ε1,t , ..., εn,t) is a vector of residuals following a multivariate
normal distribution, i.e. et ~ N(0, Σ)
In more compact notation, the VAR model can be written as:
Y = XB + ε (5)

Where Y is a T×N matrix, where T denotes the total number of observations

and N stands for the number of variables (endogenous and exogenous), while B
is a matrix of suitable dimension containing the model’s unknown parameters.
In the VAR framework the parameters of interest are the coefficients of
the model, gathered in a matrix B in (5), along with the residual covariance
matrix Σ. Unlike the traditional frequentist approach, in Bayesian econometrics
every parameter of interest is treated as a random variable, characterised by
some underlying probability distribution. The ultimate goal in Bayesian econo-
metrics is to identify these distributions in order to produce estimates and carry
inference on the model. This is done by combining the prior information we
may have about the distribution of these parameters (the prior distribution)
with the information contained in the data (the likelihood function) to obtain
D P / 110/2018

an updated distribution accounting for both these sources of information. This

updated distribution is commonly known as the posterior distribution.
For the estimation of the VAR model we use a Matlab package, called Bayesian
Estimation, Analysis and Regression (BEAR) Toolbox, developed by Dieppe et al.
(2016). The model parameters are estimated using an Independent Normal-

26
Inverse-Wishart prior distribution18, similar to Meinen and Rohe (2016). This
prior has an advantage over the standard Minnesota and the Normal-Inverse-
Wishart priors because it allows for more estimation flexibility. Unlike the
Normal-Inverse-Wishart, the Independent Normal-Inverse-Wishart prior
does not impose dependence between the variance of the residual term and the
variance of the VAR coefficients (Dieppe et al., 2016). As the name suggests,
it involves setting the prior for the VAR coefficients and the error covariance
independently. Under this prior, analytical expressions for the marginal poste-
rior distributions are not available, but numerical methods, such as the Gibbs
sampling algorithm, could be used for posterior inferences (Blake and Mumtaz,
2012).
Since the number of parameters in the VAR model is large and presenting all
of them is cumbersome, in Section 5 we adhere to the commonly accepted
practice to report functions of the VAR coefficients, which summarise informa-
tion better. In particular, we report impulse response functions and historical
decompositions. Impulse response functions describe the response of variable
yi following a shock in variable yj, everything else being constant. In turn, his-
torical decomposition allows us to decompose the variance of each endogenous
variable into its constituent components, whereby each component is due to one
structural shock in the model. Structural shocks are identified using a recursive
ordering (Cholesky decomposition).

4.2. Data
In this section we describe the choice of variables that we have used for the
empirical estimation in the analysis. For the purpose of our VAR specifications
we use quarterly data on a number of macroeconomic series, most of which
cover the period from 2000Q1 up to 2017Q1. For a detailed list of the sources
DISCUSSION PAPERS

and individual series coverage please refer to Appendix 4.

The series that we use as a proxy for business investment in Bulgaria is real
investment in machinery, equipment and weapons systems published by EURO-
STAT. For more information on the choice of a measure of investment and the
issues related to this choice please refer to Appendix 1 A.
External demand in the model is proxied by the weighted real import of goods
and services of Bulgaria’s main trading partners – a series compiled by the
ECB. As indicated in the literature review section, weak demand and sluggish
growth outlook have been empirically identified as some of the core reasons
18
Specifically, we set hyperparameter l 1 = 0.8, l 2 = 0.5, l 3 = 1, and l 4 = 100 which are the
recommended values in the BEAR Toolbox for this particular prior. We also set the number of
iterations to 10 000 with the number of burn-in iterations equal to 9 000.

27
 for the investment slowdown, especially for small open economies. This is
why we expect higher external demand to lead to higher investment and in
the case of Bulgaria this effect is likely to be quite significant as the country is
export-oriented.
To account for economic uncertainty in Bulgaria we use 5 different indicators
that have been calculated from the EU Commission’s Business and Consumer
Survey (BCS), following a methodology by Giradi and Reuter (2017). For more
information on the calculation and the interpretation of these uncertainty
indicators please refer to Appendix 3. Although each of the five uncertainty
measures has been separately used in our estimations, in the final model we
present the results of the uncertainty indicator that gives us the best fit and
economic interpretation. We expect that heightened uncertainty will have a
negative impact on investment in the country as suggested by both theoretical
and empirical work.
Capacity utilisation in the manufacturing sector is included in the specification
in order to account for the economy’s supply side motives for investment. We
expect that higher levels of capacity utilisation will result in higher investments.
The effect of profits on investment activity in our model is represented by
including in the VAR the share of gross operating surplus in gross value added
for the manufacturing sector. Data is taken from the National Statistical Insti-
tute. In line with relevant literature we expect that an increase in profits would
lead to higher investment.
To account for availability and importance of external funding for Bulgaria we
have included FDI flows, the data for which we have obtained from the BNB sta-
tistics. An increase in FDI flows in the economy is expected to increase invest-
ment activity through the external financing of investment channel.
To account for possible credit constraints and cost of financing we include the
weighted interest rate on new loans to the corporate sector. This data is taken
from the BNB’s interest rate statistics. We expect that the lower the interest rates
to the corporate sector are, the easier the access to loan financing is, and, thus,
D P / 110/2018

the higher will be the increase in investment.

As suggested by recent studies certain variables enter the VAR model in (log)
levels (CNB, 2017; Meinen and Rohe, 2016; Manteu and Serra, 2017). Invest-
ment and external demand series are transformed by taking their natural loga-
rithm and FDI flows are taken as a share of GDP.

28
Using the outlined variables, we have estimated a VAR model with six lags of
each endogenous variable for the case of Bulgaria over the 2000:Q1 – 2017:Q1
period. We have chosen the following recursive ordering of variables, which
would allow us to obtain the structural shocks of our VAR model:
External demand
Uncertainty
Capacity Utilisation
Gross Profit Share
FDI Flow
Real NFC Lending Rates
Business Investment
This Choleski ordering is similar to imposing short-run restrictions, such
that we impose a particular causal chain on the data. In this setup the vari-
able ordered first (in our case external demand) does not contemporaneously
react to variables ordered below it, but will affect all the ones ordered below
it contemporaneously. The variable ordered second (uncertainty in Bulgaria)
responds contemporaneously only to the variable above it in the ordering.
Finally, the variable ordered last (business investment) responds contemporane-
ously to shocks in all variables ordered above it but shocks in investment do not
affect the variables above it in the current period.
A possible shortcoming of our specification is that we do not explicitly include a
proxy for future demand expectations, such as the Economic Sentiment Indica-
tor (ESI), published by the EC. However, expectations indicators such as the
ESI focus mainly on the near-term future, typically 3 months for firms and 12
months for consumers. As investment decisions tend to have long-term conse-
quences for firms, they are planned carefully and medium to long-term demand
expectations are likely to be more important. In the short-run firms can face
DISCUSSION PAPERS

changes in demand by altering their capacity utilisation. Similar to Barkbu et al.

(2015), ECB (2016:1) and Meinen and Rohe (2016) the setup of our model does
not include an expectations proxy, which would suggest that economic agents
are backward-looking and future demand prospects are assessed primarily based
on current demand developments. However, short-term future expectations of
economic agents are partly accounted for by the uncertainty indicator, which is
based on economic agents’ disagreement on future production dynamics.

29
 5. Results
In order to determine the direction in which each of the factors above affects
business investment and to establish their significance we look at the impulse
response functions (IRFs) from the estimated VAR Model (see Figure 16).

Figure 16. Impulse Response Functions

D P / 110/2018

Notes:
The line denotes the median response of investment and the blue area around it denotes the 90% confidence
bands.
On the X-axis are plotted the 20 time periods (which in our case are quarters) for which we examine the
effect of shocks in our endogenous variables. On the Y-axis we have the response of business investment (in
percent from baseline path) to a positive one standard-deviation structural shock in each variable of interest.

30
The results from the impulse response functions suggest that in Bulgaria one
of the factors that has the biggest and most persistent effect on business invest-
ment is external demand. This is consistent with findings in the literature on
investment drivers in small open economies (CNB, 2017) and supports the
initial hypothesis based on the descriptive analysis in Section 3. The maximum
impact on business investment from one standard deviation increase in external
demand is observed in the second year after the shock.
The effects of uncertainty shocks to investment in Bulgaria are provided in the
top right panel. Here, we present the results of the model when the U3 uncer-
tainty indicator is used, as it appears to be the most significant and the one that
fits best in the model. This is intuitive given that U3 is an indicator that was
specifically designed to account for the economic uncertainty in the industrial
sector, which is the main investor in gross fixed assets. When we substitute
the U3 uncertainty indicator in our model with another of our uncertainty
measures we obtain impulse responses that are correct in terms of the expected
direction of the effect but there is a certain loss of statistical significance. The
statistically significant response of investment to uncertainty shocks proves that
uncertainty has a crucial and long-lasting negative effect on economic agents’
investment decisions in Bulgaria. As expected, higher uncertainty constrains
investment and leads to firms postponing their investment decisions. Under
higher uncertainty agents in Bulgaria tend to adopt a “wait-and-see” approach,
depressing investment activity for up to two years.
Judging from the impulse response functions, increases in capacity utilisation
do not have a significant impact on business investment in Bulgaria. One pos-
sible explanation is that high capacity utilisation on its own does not neces-
sarily transfer into an immediate investment increase. If firms are uncertain
about how sustainable demand prospects are, they may choose to respond to
DISCUSSION PAPERS

the increased demand by pushing up further their production capacity, thereby

deferring new investments. This would imply that positive effects on investment
from increased capacity utilisation would tend to materialise in the medium
term conditional on continuous improvement in demand conditions and firms’
sentiments.
Increases in profit (proxied by the share of gross operating surplus in gross value
added for the manufacturing sector) have a positive impact on business invest-
ment, reflecting firms’ increased capacity to finance investment with internal
means. The positive reaction of business investment is very persistent and is
maximised in the end of the first year after the initial shock in profit. This result
is consistent with empirical research that establishes a positive link between
increasing profits and investment decisions (ECB, 2016:1).

31
 Moving to external sources of financing, the reaction of business investment
to increases in foreign direct investment is positive and significant (after the
second quarter), confirming our findings in Section 3 that FDI has been an
important investment financing channel in Bulgaria.
The results from the applied VAR suggest that the reaction of business invest-
ment to changes in the cost of financing is relatively less significant. The nega-
tive response of investment activity to an increase in the non-financial corpora-
tions’ lending rates materialises from the first quarter after the initial shock in
interest rates and the maximum (negative) impact from an increase in lending
rates is observed in the sixth quarter after the shock.
The Choleski ordering in our VAR model also allows us to construct a historical
decomposition of business investment dynamics and in this way to quantify
the effects of different factors that have driven business investment in different
periods of time. Figure 17 shows that results from the VAR broadly confirm
our initial expectations about the main drivers of business investment that were
stated in Section 3.

Figure 17. Business Investment Growth – Historical Decomposition

(%, percentage points)
D P / 110/2018

Note: “Other” here includes the combined effects from the lagged investment and from the exogenous shocks.

32
Based on the historical decompositions of the VAR model it can be noted that
business investment in Bulgaria in the years right before the crisis was largely
driven by the observed increases of external demand associated with the cyclical
upturn in our main trading partners and optimism, related to Bulgaria’s acces-
sion in the EU. Increasing corporate profits and low levels of uncertainty further
supported the dynamics of business investment in the years prior to the crisis.
Furthermore, we can conclude that financing conditions also had quite a nota-
ble contribution to the investment boom in the run-up to the crisis. In 2005 and
2006 we observe a positive contribution from favourable domestic lending rates,
which in 2007 and 2008 is overtaken by the positive contribution of foreign
financing in the form of FDI inflows. Firms’ capacity to finance investment with
internal means, proxied by the profit share, was also a particularly important
factor for business investment after 2005.
In the 2009 – 2010 period lower external demand and the negative impact it had
on the overall economy appear to have been the biggest drag on business invest-
ment. The negative effect of the lower external demand was further amplified
by the heightened uncertainty triggered by the global financial crisis. Both the
internal and the external channels of financing of investment were impaired,
which exacerbated the decline in investment. Firms’ profits and ability to
finance investment using internal sources were negatively affected, which is
suggested by the observed negative contribution of the share of gross operating
surplus in GVA to business investment dynamics. Moreover, the higher risk-
aversion of investors as a result of the crisis was reflected in lower FDI flows
which also limited business investment in the period. The historical decompo-
sition of the VAR model also indicates that the increase in real non-financial
corporations’ lending rates up until the end of 2009 had a negative impact on
business investment.
DISCUSSION PAPERS

With the recovery of external demand and the exhaustion of the negative impact
of uncertainty, business investment in Bulgaria also started to recover in the
beginning of 2011. However, the 2012 sovereign debt crisis in Europe lead to a
new increase in uncertainty, followed by a supressed external demand, which
resulted in another weakening of business investment in Bulgaria that lasted up
to the second half of 2013.
As evidenced by the historical decomposition, the decrease in uncertainty sup-
ported the recovery of business investment in Bulgaria after the second half of
2013. Higher external demand also contributed to investment growth especially
after the second half of 2014, offsetting a temporary increase in uncertainty
around the end of 2014 and the first half of 2015 triggered by domestic and
international developments (see Appendix 3 for more information). In 2014 and

33
 2015 the increasing external demand, the recovery of FDI flows and the increas-
ing absorption of EU funds (assumed to be reflected in “other factors”) were the
main factors that helped investment return to positive growth rates despite the
negative contribution from low corporate profits and tight financial conditions.
The sharp decline of EU co-financed investment observed throughout 2016
could explain the decreasing business investment dynamics in spite of the posi-
tive impact of demand factors, lower uncertainty and recovering profitability.
In the beginning of 2017 business investment recorded positive growth rates on
account of continued recovery of external demand.
Overall, the results from the VAR model suggest that external demand has been
the most significant determinant of business investment. In addition, the effect
of uncertainty on investment activity has also been quite substantial. Availabil-
ity of financing (both domestic and external) has had a more pronounced effect
on investment in the pre-crisis years as compared to the post-crisis period. Con-
straints regarding availability of funding have been mainly a drag on investment
in the aftermath of the crisis. The absorption of EU funds could also be identi-
fied as an important factor for business investment in Bulgaria, judging by the
contribution of the “other factors” component throughout the years. Together
with demand factors it could explain most of the observed investment growth
in the beginning of 2017.

6. Conclusion
Gross fixed capital formation is a key driver of the business cycle and plays a
prominent role in determining the potential growth of a country. As such, the
observed slowdown in investment activity across advanced economies, and
particularly countries in the EU, including Bulgaria, observed after the Great
Recession, has raised questions on the drivers behind this slowdown.
This paper looks into the drivers behind business investment in Bulgaria and
quantifies their effect in order to explain the observed slowdown of investment
in the aftermath of the Great Recession and the followed weak recovery in more
D P / 110/2018

recent years. This is done using a Bayesian VAR estimation with which we have
evaluated the relative importance of factors, such as demand, capacity utilisa-
tion, economic uncertainty, firms’ profitability, foreign direct investments and
lending rates for business investment dynamics during the period 2000Q1 –
2017Q1. Separately from the VAR analysis, we have also examined the role of
EU-related funds for explaining business investment dynamics in Bulgaria,
given the considerable amount that Bulgaria managed to absorb in recent years
under EU programmes.

34
The results from our VAR model suggest that external demand is the main
driver of business investment in the country, which is consistent with the find-
ings of similar studies on small open economies. Apart from external demand,
we find that the level of economic uncertainty and corporate profits also play a
crucial role in determining the level of business investment in the economy. At
the same time the importance of foreign direct investment and lending rates for
investment activity appears to be less significant.
From the historical decomposition, obtained from our VAR model, we can con-
clude that almost all of the factors considered in the analysis contributed posi-
tively for the business investment growth in the run-up to the Great Recession.
External demand increases, associated with the cyclical upturn in our main
trading partners, low levels of uncertainty and increasing corporate profits were
the main factors that stimulated investment growth in Bulgaria in that period.
Lending rates (in 2005–2006) and FDI inflows (in 2007–2008) also contributed
to the business investment growth. Our estimations suggest that the observed
slowdown of business investment in the post-crisis period can be explained
by the reduced external demand and the increased economic uncertainty in
the immediate aftermath of the crisis, which lasted up to mid–2013. After the
second half of 2015 our results point to a recovery in corporate profits, as well as
favourable financing conditions, which on top of the already observed growth in
external demand and decreasing economic uncertainty support the recovery of
investment activity in the country. Additional evidence suggest that the dynam-
ics of EU co-financed investment has also been a notable factor in explaining
business investment dynamics, especially in 2015 and 2016.

DISCUSSION PAPERS

35
 Appendices

Appendix 1. Investment Data

A. Business Investment Data
In this paper we follow the example of recent studies on the topic (Barkbu et al,
2015; ECB, 2016:1; Kose et al., 2017) and for the purpose of our VAR analysis
we separate construction investment from total investment in order to extract
the part of investment that is particularly relevant in boosting an economy’s
future productive capacity. In order to obtain our investment proxy we use
the breakdown of Gross Fixed Capital Formation by main asset type, which is
available from Eurostat for all EU countries. As can be seen in Figure 18, non-
construction investment in Bulgaria is driven almost entirely by investment in
“machinery, equipment and weapons systems” (encompassing transport invest-
ment, information and communication technology (ICT) equipment, other
machinery and equipment, and weapons systems).

Figure 18. Non-construction Investment Growth Rate and Contributions

by Major Asset Type
(annual percentage changes, percentage points)
D P / 110/2018

Source: Eurostat

As such we have decided to use investment in machinery, equipment and weap-

ons systems as our proxy for “business” investment, similar to Meinen and Rohe
(2016). Our choice is further supported by the fact that machinery and equip-
ment investment arises mainly in the corporate sector and, thus, its dynamics
should reflect firms’ investment decisions (ECB, 2016:1). We should note that

36
in the way that we have defined business investment it is not only private sector
investment but also contains some part of public investment. Unfortunately,
due to data limitations, we cannot separate out this public investment compo-
nent which for Bulgaria and other NMS could be relevant in particular periods,
especially in the years with high absorption rates of EU-related funds, related
to the programming cycle of EU funds. For more advanced countries such as
the “core” euro area economies this public investment component is likely to
be relatively smaller, with the ECB (2016:1) estimating it at around 10% of euro
area investment, with some variation across countries.

B. EU Funds-related Data
Regarding the impact of EU funds on government investment we can use two
sources of information: 1) annual data on EU transfers to government and non-
government units (source: ESCB Government Finance Statistics, EU budget
transactions) and 2) quarterly data on capital transfers received by the govern-
ment (source: Eurostat, quarterly non-financial accounts of the general govern-
ment sector). The latter is preferred in this analysis due to its higher frequency
and the inclusion of capital transfers from pre-accession funds, which had a
significant share in government GFCF in the 2007–2009 period19.
Assessing the impact of EU funds on non-government investment is not
straightforward. Data on capital transfers from the EU budget to non-govern-
ment units from the ESCB GFS on EU budget transactions is not available for
20
Bulgaria, while data from the same source for NMS indicates that the share of
capital transfers to non-government in all capital transfers from the EU budget
has varied a lot across years and across countries. On average for these coun-
tries, capital transfers to non-government units represented between 34-42% of
all capital transfers in the period 2007–2015, but have increased notably in 2016
DISCUSSION PAPERS

to 48.1% of all capital transfers.

As alternative sources of data that provide indications for the relative impor-
tance of EU funds in financing private investment we use 1) Consolidated
Fiscal Programme national methodology cash-based data on capital transfers
to non-financial corporations, related to the implementation of EU co-financed
projects from all five EU funds and 2) payments on projects financed by Opera-
tional Programme “Competitiveness” excluding payments related to financial

19
While Table GFS Table 1b also includes information on transfers from pre-accession funds, these
flows are not split between capital and current transfers and their impact on government investment
cannot be assessed.
20
Data is available for Czech Republic, Estonia, Hungary, Latvia, Lithuania, Poland, Slovakia and
Slovenia.

37
 instruments. Both sources provide information only on projects’ costs, financed
by the EU budget and the national budget, while the amount provided by final
beneficiaries could also be quite substantial (up to 50% of the total budget).
These sources are clearly second best alternatives as they are both cash-based
and their coverage does not correspond to the sectoral coverage of the national
accounts. In particular, the first source includes EU financed and national
budget financing to non-financial corporations that have been classified in
the general government sector (notably the National Railway Infrastructure
Company). The second source has a more limited coverage and includes data
on the one operational programme aimed to support projects in SME. It does
not include support for investment by agricultural producers from the EAFRD.
The beneficiaries of OP “Competitiveness” are typically private companies and
both financial and no-financial corporations are eligible for support from this
programme.
D P / 110/2018

38
 Appendix 2. Survey Results on the Purpose of Investment
and Factors Driving Investment Activity
Two times per year, in spring and autumn, the European Commission conducts
a survey among the companies in the manufacturing sector, collecting infor-
mation on realised and planned investment, purpose of investment and main
factors influencing investment. This appendix summarises the results of this
survey for Bulgaria over the period 2002–2016.
Survey results suggest that over the period 2002–2003 around 60% of overall
investment was dedicated to the replacement of worn-out equipment and/
or rationalisation. In the following years, however, this pattern changed and
firms tended to report more investments as serving extension purposes (see
Figure 19). The gradual shift in the focus of investment projects could be attrib-
uted to the significant FDI inflows over the period 2004–2008, the integration
of domestic firms in global value chains as well as to the business cycle (Ivanova
and Ivanov, 2017). During periods of economic upswings it could be expected
that more investments are dedicated to the expansion of production capacity,
compared with downturns when investments are likely to aim mainly at replac-
ing worn-out plant or equipment. Similar to findings for other EU countries,
the relative share of investments dedicated to the extension of production sights
in Bulgaria is positively correlated with the growth rate in gross fixed capital
formation (EC, 2017). Contrary to the pre-crisis period, in the aftermath of the
global financial crisis (2010–2014) investment in Bulgaria has mainly served
rationalisation purposes as firms aimed at improving their efficiency and
competitiveness. In 2013 the share of investment for rationalisation stood at its
highest level since the start of the survey 2002. With the cyclical recovery of the
economy in 2015–2017, firms started again to report investments that are more
aimed at the extension of production capacity.
DISCUSSION PAPERS

In terms of the factors that stimulate or limit investment activity, the survey dis-
tinguishes between demand, financial, technical (e.g. technological factors and
the availability of labour) and other factors (e.g. taxation). Demand and finan-
cial conditions are by far the most important factors behind investment deci-
sions in Bulgaria, regardless of the phase of the business cycle (see Figure 20).
In the pre-crisis period demand and financial factors exerted a stimulating
effect on investment. This reflected the robust domestic demand growth over
the period 2000–2008 combined with favourable external environment. At the
same time, the downward trend of interest rates and the availability of credits
contributed to easing in financial conditions. In 2009 all factors declined while
investment activity contracted severely following the global financial crisis. In
2011–2013 firms in the manufacturing sector that participated in the survey

39
 reported an increase in realised investment, with demand and financial factors
being the two main contributors. This was in line with the increase in capac-
ity utilisation over the same period and the gradual recovery of the economy.
In addition to demand and financial conditions, increasingly more managers
started to report technical factors as relevant for investment decisions, with the
share of those managers outpacing pre-crisis levels. According to survey results,
investment has declined over the last two years but this possibly could be attrib-
uted to the end of the EU programming period.

Figure 19. Purpose of Investing

(%, share)

Figure 20. Factors Driving Investment and Realised Investment in the

Manufacturing Sector
(%, percenrage change) (%, share of firms)
D P / 110/2018

Source: European Commission (AMECO database)

40
 Appendix 3. Measures of Uncertainty
There are a number of methods (EC, 2017) that are used to indirectly measure
uncertainty, such as financial market volatility indicators, disagreement among
professional forecasters, survey-derived indicators and Economic Policy Uncer-
tainty indicators. All of these approaches have their pros and cons.
For our analysis we have chosen to use measures for uncertainty constructed
from survey data, since economic uncertainty is measured “at source”, i.e.
directly at the level of the economic agents who make investment and consump-
tion spending decisions (EC, 2013). To that aim we use data from the European
21
Commission Business and Consumer Survey (BCS) and by following the
methodology of Girardi and Reuter (2017) we obtain 4 different uncertainty
indicators and a composite uncertainty indicator for Bulgaria, similar to
Ivanov (2018)22. The derived indicators measure the divergence of business
and consumer expectations about the economy and about their finances. The
core underlying assumption is that the more economic agents disagree in their
assessment or expectations about economic activity, the higher is the uncer-
tainty in the economy. Uncertainty can, thus, be illustrated as a mean preserv-
ing increase in the “tails” of the probability distribution of an event (EC, 2013).
This distinguishes it from risk, where economic agents are able to form views
about the probability distribution of possible future states, based on logic or on
past experience (ECB, 2016:2).
The first indicator (U1) measures the disagreement for future activity across
all sectors (industry, services, retail trade, consumers, construction) and, thus,
should capture uncertainty at the level of the whole economy. This is an expec-
tations-based uncertainty indicator.
The second indicator (U2) does not only focus on disagreement for future activ-
DISCUSSION PAPERS

ity but also takes into account respondents’ backward-looking assessments for
the same variables in subsequent survey rounds across all sectors (Girardi and
Reuter, 2017). This indicator is a type of ex-post uncertainty measure, based on
indirect evaluation of respondents’ forecast errors.
The third indicator (U3) is similar to the first indicator (U1) but it measures the
disagreement only related to the future production in the industry sector (single
question in the survey), rather than the uncertainty in the whole economy
21
For more information see European Commission Business and Consumer Surveys: https://
ec.europa.eu/info/business-economy-euro/indicators-statistics/economic-databases/
business-and-consumer-surveys_en
22
In this section we give a short overview of the methodology and finding of Ivanov (2018). For
more information and a detailed discussion on how the uncertainty measures for Bulgaria were
derived, please refer to that paper.

41
 (which was the case for U1). This indicator is particularly relevant in explaining
investment dynamics in Bulgaria, since the industrial sector is the main investor
in gross fixed assets in the country.
The fourth indicator (U4) is constructed in a slightly different way, compared
to the previous three indicators. Rather than using the detailed breakdown of
“increase” and “decrease” responses from the BCS, this indicator uses dispersion
of changes in the overall balance of replies between consecutive rounds of the
survey (Girardi and Reuter, 2017). The idea is that a high degree of uncertainty
may not be only reflected in disagreement between respondents to a given ques-
tion but may also be reflected in the resulting balance scores diverging across
different questions, i.e. increased dispersion across questions rather than within
questions (Manteu and Serra, 2017).
Given that no indicator is a perfect proxy for uncertainty, it might be preferable
to compile a composite measure of uncertainty which captures and consolidates
the information content of the different uncertainty proxies (ECB, 2016:2).
We have chosen to use the median aggregation approach, similar to the ECB
(2016:2), in order to obtain a composite uncertainty indicator. The four uncer-
tainty indicators and the composite indicator are presented in Figure 21.

Figure 21. Uncertainty Indicators, Bulgaria

(standard deviationfrom mean)
D P / 110/2018

Source: EU Commission BCS Survey, own calculations

42
 Appendix 4. Variables Used
The table in this section contains detailed information on the variables used in
the empirical part of our analysis.

Name Variable Source Coverage Transformation

Business Investment – real
Eurostat, sea-
investment in machinery, 2000Q1- Natural logarithm
INV_M sonally adjusted
equipment and weapon 2017Q2 transformation
systems
External Demand for Bul-
garian Goods and Services –
ECB, seasonally 2000Q1- Natural logarithm
WTV weighted index of the real
adjusted 2017Q2 transformation
imports of Bulgaria’s main
trading partners
Uncertainty Indicator 1 –
BCS, own 2000Q1- Standardised
U1 related to the future expec-
calculations 2017Q1 index
tations in the economy
Uncertainty Indicator 2 –
BCS, own 2000Q1- Standardised
U2 related to backward-looking
calculations 2017Q1 index
assessments
Uncertainty Indicator 3 –
BCS, own 2000Q1- Standardised
U3 related to the future produc-
calculations 2017Q1 index
tion in the industry sector
Uncertainty Indicator 4 –
BCS, own 2000Q1- Standardised
U4 obtained from balance
calculations 2017Q1 index
scores
Composite Uncertainty BCS, own 2000Q1- Standardised
U_C
Indicator calculations 2017Q1 index
National Ac-
Share of Gross Operating 2000Q1–
counts, four Natural logarithm
GOS_P Surplus in GVA for the 2017Q1
DISCUSSION PAPERS
quarters rolling transformation
manufacturing sector
data
Business Sur-
Capacity Utilisation in the veys, 2000Q1 –
CU N/A
Manufacturing Sector BNB seasonally 2017 Q1
adjusted
Real Non-Financial Cor- 2000Q1–
BNB Interest
IR porations (NFC) Lending 2017Q1 N/A
Rates Statistics
Rates (GVA deflated)
BNB Balance of
FDI Inflow as a Share of Payment Statis- 2000Q1–
FDIFLOW_GDP N/A
GDP tics (BPM5 and 2017Q1
BPM6)

43
 Appendix 5. EU Funds and Their Role for Government
and Corporate Investment
As described in Appendix 1B two alternative sets of data are used to quantify the
impact of EU funds on government investment and investment of non-financial
corporations.
As already noted, EU funds have been one of the major factor for the dynamics
of government investment in recent years. The share of EU co-financed projects
in total government investment has increased from just 11.7% in 2008 to the
substantial 65% in 2015, when EU co-financed investment peaked at 4.3% of
GDP (see Figure 22).
Capital transfers from government to non-financial corporations, related to the
implementation of EU co-financed projects, have also increased substantially
after 2012 and have peaked in 2015, when they accounted for 1.2% of GDP, or
8.8% of total investment by non-financial corporations (see Figure 23). It has
to be notes that these transfers also included co-financing on projects, imple-
mented by state owned companies, while direct support to small and medium
enterprises was more limited. Under the 2007–2013 framework, the total budget
of the Operational programme “Competitiveness of the Bulgarian economy”
amounted to EUR 1.2 billion or 1.9% of private investment (0.4% of GDP) and
under the 2014–2020 framework, there are two operational programmes (OPs
“Innovations and Competitiveness” and “SME Initiative”) that finance projects
related to enterprise support and innovation with a total budget of EUR 1.4
billion, or 2.4% of private investment (0.4% of GDP).
In addition to grants, EU funded financial instruments are also meant to play an
important role in supporting private investment. Around 30% of the 2007–2013
Operational programme “Competitiveness of the Bulgarian economy” budget
was allocated to financial instruments for enterprises (JEREMIE initiative). The
funds allocated to financial instruments for enterprises under the 2014–2020
financial framework amount to EUR 337 million and include 100% of the
budget of the “SME Initiative” (EUR 102 million) and the contracted funds of
D P / 110/2018

Fund Manager of Financial Instruments in Bulgaria EAD on OP “Innovations

23
and Competitiveness” (EUR 235 million) .

23
As of April 2017 the Fund Manager of Financial Instruments in Bulgaria EAD has contracted
financial resources from four operational programmes totalling BGN 1.2 billion. In the previous
programing period, the majority of the funds allocated to financial instruments were the funds
under the JEREMIE programme (finances as part of OP “Competitiveness), while financial
engineering instruments for urban development, energy efficiency and renewable energies (Jessica
initiative) were of marginal important (EUR 33 million).

44
Figure 22. EU Co-financed Investment as a Share of Total Government Investment
(% of GDP) (%)

Source: Eurostat, own estimates

Figure 23. Capital Transfers from Government to Non-financial Corporations,

Related to the Implementation of EU Co-financed Projects
(% of GDP)

DISCUSSION PAPERS

Note: Payments to direct SME support represent yearly payments under OP “Competitiveness” excluding
payments related to financial instruments for enterprises.
Source: MoF, Unified management system for the EU structural instruments in Bulgaria, own
estimates

45
 With the launch of the Investment plan for Europe (“Juncker Plan”) in 2014,
the scope for interventions to stimulate investment in EU countries has been
broadened further. The Juncker Plan’s European Fund for Strategic Investments
(EFSI) aims to support strategic investments in key areas such as infrastructure,
education, research and innovation, as well as risk finance for small businesses.
Its main goal is to overcome bottlenecks in current investment environment and
to mobilise private investment in times when public resources are scarce. As of
May 2017 total financing in Bulgaria under the Juncker Plan’s for SME support
and support of infrastructure and innovation projects amounts to of EUR 206.9
million, with expected overall impact on investment of nearly EUR 1.124 bil-
lion (after financial intermediaries’ and private companies’ capital has been
mobilised).24
D P / 110/2018

24
According to latest information published on MoF website as of July 2017.

46
 Bibliography
Barkbu, B., S.P. Berkmen, P. Lukyantsau, S. Saksonovs and H. Schoelermann (2015).
“Investment in the Euro Area: Why Has It Been Weak?”, IMF Working Paper.
Blake, A., H. Mumtaz (2012). “Technical Handbook – No. 4 Applied Bayesian econometrics
for central bankers”, Centre for Central Banking Studies, Bank of England, p. 38.
Bloom, N., S. Bond and J. Van Reenen (2007). “Uncertainty and Investment Dynamics”,
Review of Economic Studies, 74(2), pp. 391-415.
BNB (2017). “Dynamics of House Prices in Bulgaria between 2000 and 2016”, Economic
Review Issue 1/2017.
Busetti, F., C. Giordano, G. Zevi (2015). “Main Drivers of the Recent Decline in Italy’s
non-construction investment” Banca D’Italia. Occasional Papers 276.
Czech National Bank (2016). “The impact of drawdown of EU funds on investment by
non-financial corporations”, Box in the CNB Inflation Report, Issue IV/2016.
Czech National Bank (2017). “Private investment – an analysis of individual data on non-
financial corporations”, Box in the CNB Inflation Report, Issue I/2017.
Dieppe, A., R. Legrand and B. Roye (2016). “The BEAR toolbox”, ECB Working Paper
Series № 1934/July 2016.
EC (2013). “Assessing the Impact of Uncertainty on Consumption and Investment”, Special
focus in “Quarterly Report on the Euro Area”, Volume 12, N.2 (2013).
EC (2016:1). “Ex post evaluation of Cohesion Policy programmes 2007–2013, focusing on
the European Regional Development Fund (ERDF) and the Cohesion Fund (CF): Task 3
Country Report Bulgaria”, Directorate-General for Regional and Urban Policy European
Commission B-1049 Brussels.
EC (2016:2). “Public finances in EMU–2016”, institutional paper 045/2016
EC (2017). “European Business Cycle Indicators: 4th Quarter 2016”, Technical Paper 013. DISCUSSION PAPERS
ECB (2016:1). “Business investment developments in the euro area since the crisis”, Article
in ECB’s Economic Bulletin, Issue 7/2016.
ECB (2016:2). “The impact of uncertainty on activity in the euro area”, Article in ECB
Economic Bulletin, Issue 8/2016.
Eesti Pank (2016). “Transfers from the European Union”, Box in the Estonian Economy
and Monetary Policy quarterly bulletin, Issue 1/2016.
EIB (2013). “Factors influencing investment during the financial crisis and deep economic
recession: the European experience since 2008” Investment and Investment Finance in
Europe.
EIB (2015). “Investment in Tangible Capital” Investment and Investment Finance in
Europe.
EIB (2016). “Investment and Investment Finance in Europe: Financing productivity
growth”, Annual Report 2016.

47
 Giannone, D., M. Lenza and G. Primiceri (2012). “Prior selection for Vector Autoregres-
sions”, ECB Working Paper Series №1494/November 2012.
Girardi, A., A. Reuter (2017). “New uncertainty measures for the euro area using survey
data”, Oxford Economic Papers, Volume 69, Issue 1: 278–300.
Havik, K., K. Morrow, F. Orlandi, C. Planas, R. Raciborski, W. Röger, A. Rossi, A.  Thum-
Thysen and V. Vandermeulen (2014). “The Production Function Methodology for Calcu-
lating Potential Growth Rates & Output Gaps”, European Commission, Economic Papers
535.
IMF (2007). “Bulgaria: Selected Issues”, IMF Country Report No. 07/390.
IMF (2014). “Article IV Consultations: Euro Area Policies”. Country Report 14/199, Wash-
ington, D.C.: IMF.
Ivanov, E. (2018). “Constructing an Uncertainty Indicator for Bulgaria”, BNB,
DP/109/2018.
Ivanova, N. and E. Ivanov (2017). “The Role of Bulgaria in Global Value Chains”, BNB,
DP/105/2017.
Karamisheva, T. (2016). “Analysis of the Bank Lending Survey Results for Bulgaria (for the
2003–2014 Period)”, Bulgarian National Bank DP/102/2016.
Kose, M., F. Ohnsorge, L. Ye, E. Islamaj (2017). “Weakness in Investment Growth. Causes,
Implications and Policy Responses”, World Bank Policy Research Working Paper 7990.
Kotseva, P. and M. Yanchev (2017). “Analysis of the Housing Market Developments and the
Underlying Macroeconomic Fundamentals in Bulgaria”, BNB, DP/103/2017.
Manteu, C., S. Serra (2017). “Impact of uncertainty measures on the Portuguese economy”,
Working Papers 9 (2017), Banco de Portugal.
Meinen, P. and O. Rohe (2016) “On Measuring Uncertainty and its Impact on Investment:
Cross-country Evidence from the Euro Area”, Deutsche Bundesbank Discussion Paper No
48/2016.
Ministry of Finance (2016). “EU funds in Bulgaria: assessment of the macroeconomic
effects from the implementation of EU co-financed programmes “, available only in Bulgar-
ian, Sofia 2016.
Mitra, P. (2011). “Capital Flows to EU New Member States: Does Sector Destination
Matter?”, IMF Working Paper WP/11/67.
D P / 110/2018

Mizen, P. and P. Vermeulen (2005). “Corporate Investment and Cash-Flow Sensitivity:

What Drives the Relationship”, ECB Working Paper Series, No. 485 / May 2005.
Narodowy Bank Polski (2017). “The absorption of EU structural funds in 2016”, Box in the
NBP inflation report, Issue March 2017.
Paliova, I. and T. Lybek (2014). “Bulgaria’s EU Funds Absorption: Maximizing the Poten-
tial!”, IMF Working Paper 14/21, International Monetary Fund, Washington, DC.
Vashakmadze, E., G. Kambou, D. Chen, B. Nandwa, Y. Okawa, D. Vorisek (2017).
“Regional Dimensions of Recent Weakness in Investment. Drivers, Investment Needs and
Policy Responses” World Bank Policy Research Working Paper WPS 7991.
48
ISBN 978-619-7409-07-9
Elements of the 1 lev banknote, issue 1999, are used in cover design.

View publication stats