MACROECONOMIC FACTORS THAT GENERATE FISCAL RISK IN ROMANIA 3

Full text

35
Adriana Elena Porumboiu1
Petre Brezeanu2 Volume 32(8), 2023
MACROECONOMIC FACTORS THAT GENERATE FISCAL RISK
IN ROMANIA3
Fiscal risk, under the form of unforeseen increases in public expenditures, can be
quantified by the (increasing) government debt level. This study aims to identify the
interdependencies between government debt, the number of people active on the labour
market (labour force), the level of the average salary in the economy, the harmonized
index of consumer prices and the RON-EURO exchange rate. The main objective is thus
to assess which ones of the last four variables influence in the short and long run the
government debt and constitute a fiscal risk. The analysis is focused on Romania, using
quarterly data starting from the 1st quarter of the year 2000 and up to the 2nd quarter
of 2021, the applied methodology being VECM. The most important conclusions show
that changes in the level of the average salary and active population are significant and
influence the government debt both in the short and long run and they can constitute in
this regard fiscal risk determinants, while HICP as an inflation indicator or HICP do
not exert significant impact on government debt either on short term or long term. While
the average salary constantly exerts a significant and positive influence on the
government debt, the change in the active population leads to a change of the same sign
in the government debt in the short term, but of the opposite sign in the long term. The
main recommendation for the government that derives from the results of the study is
to implement measures to increase the active population and increase the degree of
employment, in order to decrease the pressure on the government debt and fiscal risks,
both in the short term and long-term. This is in line with the results of numerous studies
that show that the decrease in the active population, combined with the rise in retired
people, lead to increasing indebtedness.
Keywords: fiscal risk; government debt; demographic tendencies; average salary
JEL: H63; J01; H50
1. Introduction
There are studies whose empirical results show that the evolution of government debt is
influenced in the same way by the evolution of social assistance expenditures (Porumboiu &
Brezeanu, 2022). Through this article we aim to identify if there is an interaction between the
1 Adriana Elena Porumboiu, PhD student, Department of Finance, The Bucharest University of
Economic Studies, +40 765 732 167, e-mail: porumboiuadriana12@stud.ase.ro.
2 Petre Brezeanu, PhD Professor, Department of Finance, Faculty of Finance and Banking, The
Bucharest University of Economic Studies, +402 1 3191 900, e-mail: petre.brezeanu@fin.ase.ro.
3 This paper should be cited as: Porumboiu, A. E., Brezeanu, P. (2023). Macroeconomic Factors that
Generate Fiscal Risk in Romania. – Economic Studies (Ikonomicheski Izsledvania), 32(8), pp. 35-56.

Porumboiu, A. E., Brezeanu, P. (2023). Macroeconomic Factors that Generate Fiscal Risk in Romania.
36
government debt (debt), the number of people active on the labour market (activepop), the
level of the average salary in the economy (sal), the harmonized index of consumer prices
(HICP) and the RON-EURO exchange rate (exchange). The reasoning for which we consider
it appropriate to estimate a VECM (vector error correction model) has foundations of
macroeconomic logic, explained in the following, but it is also suggested by the typology of
the series included in the analysis. This econometric model allows, in turn, the analysis of
the impact on each variable used, determined by changes in the other variables, and quantifies
the interdependencies between each of them. However, the main objective is to identify if
and how much the number of active people, the average salary in the economy, the HICP or
the RON-EUR exchange rate can lead to increases in the public debt, given the fact that an
increase in the public debt represents a risk fiscal, a potential threat to macroeconomic
sustainability.
The macroeconomic system is composed of components that evolve depending on the
measures that are taken with impact on themselves but also depending on the changes of the
others. For example, the public debt can be a consequence of some government measures
that do not find their financing in tax revenues, but then it can itself constitute a determinant
of the level of taxes and fees established, since the credit must be repaid.
The increase in government debt and the payment of the related interest can translate into a
reduced ability to cover social expenses and expenses with the salaries of the budget workers,
a fact that is quickly reflected in the level of the average salary in the economy and can also
influence the number of people active on the labour market. A government debt accompanied
by a budget deficit is most often the cause of adjustments to public spending, for example
through reforms with an impact on administration spending, which are felt in the incomes of
people employed by the state and even in the structure of the active population.
In the field of the labour market, governments propose to match vacant jobs with the existing
labour force (a complex process, which involves accompanying the individual in the long
term through education and training, for the acquisition of skills and even professional
reprofiling). In this way, the appropriate placement of the active population is ensured,
workers will obtain their own income and the need for financial support from the public
authorities will decrease (social assistance expenses). Placement of the labour force implies
the reduction of the number of unemployed beneficiaries and the increase of the number of
employed people. Of course, the situation will generate positive consequences on the average
salary in the economy.
The intervention of the state in the labour market by establishing minimum levels of salary
income should be carried out with caution, always taking into account the ability of
employers in the private sector to comply with the regulations without reducing the number
of employees. Setting a higher level of the economy's minimum salary is a welcome measure
for people employed on such a salary, but governments must ensure that this does not backfire
on the state by increasing unemployment and social spending. We appreciate that the main
objective of public policies on the labour market is to support the development of commercial
companies that offer new jobs, to facilitate the access of the active population to vacant
positions, to have a high level of employment and thus reduce the pressure on expenses public
with social assistance.

– Economic Studies Journal (Ikonomicheski Izsledvania), 32(8), pp. 35-56.
37
The changes regarding the active population are in turn the effect of several factors, among
which the demographic one deserves to be remembered as a priority. The low birth rate and
the increasing life expectancy are the phenomena that determine the working population
(people having the working age – from 15 to 64 years) in the long term and which also affect
the active population (employed and unemployed population). If we also take into account
the growing phenomenon of migration, we can say that the active population is fundamentally
determined by three major demographic phenomena: “reducing mortality, reducing birth
rates and external migration” (Rangelova & Bilyanski, 2019). The decrease in the number of
employed people, against the background of the increase in the number of retired and jobless
people, can lead to greater pressure on the component of social contributions of people who
earn income. Net salaries are negatively influenced, so there will also be changes in the
average salary in the economy. The decrease in the active population, against the background
of the two mentioned demographic phenomena, will require the addition of social expenses
from the governments, financing that could sometimes only be done through loans.
Part of the products and services that are sold, and even part of the salaries in Romania have
as a reference point the EUR value at a given moment, the price being represented by the
counter value in RON, paid in the national currency, but in fact also determined by the
evolution the exchange rate. Also, the structure of Romania's government debt shows a well-
defined component represented by loans in EUR currency. A possible change in the exchange
rate translates into lower/higher interest costs for loans in EUR.
The specialized literature abounds in studies that quantify the impact of demographic changes
on government debt, ex-ante analyses whose primary objective is to raise an alarm signal to
public authorities that the situation can only be partially resolved through reforms and
additional revenues (Afflatet, 2018). We therefore observe that there is a dynamic of the
variables: the government debt, the active population, the average salary in the economy, and
the complexity of the system is undoubtedly greater, given the correlations with other
macroeconomic variables. The use of a VECM model is justified.
2. Literature Review
Among the variables included in the analysis, the specialized literature most often concludes
that demographic factors are the ones that constitute fiscal risk and represent a vulnerability
of long-term fiscal sustainability.
In 2009, amid the economic crisis, the International Monetary Fund stated that, in the
conditions of increasing government debt and contingent liabilities, the main threat to fiscal
solvency is the ageing population trend. Although the countries of reference at that time were
advanced economies, as expected, this demographic problem extended to developing
economies as well. The estimated solution to prevent a government debt boom was to reduce
the level of government debt, spending and maintaining the level of taxes, in order to
subsequently have available fiscal space.
Increasing life expectancy does not only mean that people are living longer, but that they can
survive longer from diseases (Birg, 2015), and this reality of life entails increasing costs for

Porumboiu, A. E., Brezeanu, P. (2023). Macroeconomic Factors that Generate Fiscal Risk in Romania.
38
the healthcare system that has to treat them (Lee & Tuljapurkar, 1998). Yared (2019) shows
that government debt fulfils three functions, as follows: it allows for lower taxation, it
represents an asset of safe value and determines the efficient use of capital over time. The
author raises the following fundamental question: does the fact that the use of the growing
government debt is done in order to increase the social benefits of the population compensate
for the risk it entails?
The way demographic phenomena affect the government debt of OECD countries has
attracted the attention of authors since the end of the last century, against the background of
the increase in life expectancy and the decrease in the birth rate. The fact that the working
population is declining, while the number of elderly people is high, has been flagged as a
challenge to the sustainability of public finances since 1995. Authors Jensen and Søren
opined that, for intergenerational equity to exist, in periods when there is a significant labour
force governments must focus on reducing the government debt so that, in times when the
ageing population requires financial support, there is not a large burden on workers
contributing to the social insurance fund. The two authors carry out an analysis regarding the
sustainability of public finances and the possibilities of adjusting the tax rate, starting from
the study of Blanchard, Chouraqui, Hagemann and Sartor (1990), in which the current tax
rate is compared with the permanent one (calculated in such a way as to ensure the
sustainability of public finances, the government debt to tend to 0, or at least at the time of
projection to be lower than the current one). Under normal circumstances, a current tax rate
that exceeds the permanent tax rate allows tax reduction, and vice versa. But Jensen and
Søren show that the ageing population makes the decision to cut tax levels unsustainable and
will actually lead to a more costly deferment of government debt service.
Other authors (Jensen & Søren, 1996; Balassone et al., 2009; Cecchetti et al., 2010) raised
the issue of the burden that the ageing population can constitute on the government debt and
sought to propose alternatives, but as observed relevant in repeated articles (Preston, 1984;
Sinn & Silke, 2002; Auerbach, 2009), possible fiscal consolidation reforms with an impact
on the elderly are all the more difficult to implement the larger their number, and implicitly,
and their political decision-making power is more significant. Although the private pension
system already operates in many countries, where the state only has the obligation to provide
a minimum level of pension (more like subsistence), pensions are still a concern for industrial
economies (Klyvienė, 2004).
The connection between the government debt and the active population is an obvious one:
the interest payment is based on the tax revenues collected by the public authorities, where
an essential place is occupied by labour taxation. People need a job, hence the rigidity of
labour as a factor of production. According to Menguy (2020), from a fiscal point of view,
this translates into greater efficiency in raising tax revenues by increasing labour taxation
than by raising other types of taxes. The same author develops an econometric model applied
to the euro states which confirms that states with a low level of labour taxation are those with
a lower debt ratio. Because of the need to finance interest, states with high government debt
have higher labour tax rates to meet their tax revenue targets. For employees, this means a
lower net salary.

– Economic Studies Journal (Ikonomicheski Izsledvania), 32(8), pp. 35-56.
39
However, an unjustified level of taxes on labour can also generate undesirable consequences,
such as increasing the unemployment rate and illegal work, that is, forms of tax evasion that
reduce the level of budget receipts (Genschel, 2001). The same idea emerges from the
recommendations made by Dieppe et al. (2015): to ensure economic growth in the medium
term in countries with high government debt and an ageing population, fiscal consolidation
requires prudent implementation, so as not to generate undesirable effects of decreasing
labour supply.
The fact that demographic phenomena affect government debt derives from the following
reasoning: starting from the last century, most European countries have used a system called
PAYG (pay-as-you-go) to obtain pension funds, a system that is based on social
contributions, determined by the number and salaries of active people. A larger pensionable
population and a decline in the working population are driving governments to identify new
sources of funding – even possibly borrowing. Precisely for this reason, the ageing of the
population is a challenge for the sustainability of finances and a matter of inter-generational
equity (Sánchez-Romero et al, 2019).
Empirical research with panel data models carried out on the Member States of the European
Union proved that there is a correlation between the active population represented by young
people who do not hold a job and macroeconomic conditions, in the sense that high
government debt, low GDP growth and poor development of the construction sector are
associated with a higher vacancy rate (Tomić, 2018). Hansson (2010) also explains with
reference to the EU that against the background of increased mobility of capital and people,
and in the context of an ageing population (he uses the example of Sweden), the use of the
PAYG system proves to be inefficient and cannot be corrected by increasing the level of
taxation of the work, but a possible source of financing the expenses with the elderly could
be the increase in property taxation. The objective of identifying sustainable sources to cover
these expenses results from the idea of not resorting to government debt to compensate for
this imbalance between the active population and the population benefiting from social
support. Bengtsson and Scott (2011) warn that it is unlikely that the problem of population
ageing will be solved only by increasing taxation since employees have (well-founded)
expectations that the evolution of labour productivity will be felt in the level of income and
living conditions. Bongaarts (2004) even expects a reduction in pension 'generosity' to be
inevitable.
Demmel and Keuschnigg (2000) opine that the alternative to the PAYG pension system and
the accumulation of excessive government debt is the development of the private pension
system, which would have a positive impact on the employment rate and capital
accumulation.
Moreover, there are studies that show not only that government debt negatively influences
the level of net income of the population, but even contributes to an unequal distribution of
wealth (Chatzouz, 2020). Topal et al. (2018) show that the government debt Granger causes
the unemployment rate, leads to an increase in the number of unemployed, as the
unemployment rate also leads to an increase in the government debt. Farmer and Kuplen
(2018) also confirm that higher government debt favours higher unemployment and slows
the growth rate of GDP and the interest rate.

Porumboiu, A. E., Brezeanu, P. (2023). Macroeconomic Factors that Generate Fiscal Risk in Romania.
40
Regarding the studies carried out on this topic and the methodology used, we note the analysis
by Afflatet (2018) which uses panel data for 18 European countries to identify whether
demographic factors (number of unemployed, age structure of the population) have
influenced the evolution of government debt. The regression results demonstrate that until
2015 the changes produced in the debt level were not mainly due to demographic factors, but
this does not exclude the future influence they will exert, a study that also confirms empirical
results previously obtained by Razin et al. in 2001, or Chen in 2004.
The study of Tanchev & Mose (2023) confirms the fact that government initiatives on the
labour market to develop human capital, although they involve additional expenses, have
measurable effects of increasing productivity and contribute to economic growth. However,
the recommendation is that these increases in expenses should be done at a reasonable cost,
since debt financing can translate into pressure on the economy and the private sector (in
order to obtain tax revenues intended for reimbursement) and, in the long term, in a delay of
economic development.
3. Methodology
Quarterly government debt (abbreviated debt) is, according to the European System of
Accounts ESA 2010, the ‘total gross debt at nominal value outstanding at the end of each
quarter for the general government sector”. The variable is expressed in million units of
national currency, RON. The active population (activepop) concerns both employed and
unemployed people, aged from 15 to 64 years, and it is synonymous to the labour force, in
accordance with the definition provided by the International Labour Organization. The active
population is expressed in thousands of persons. The average salary in the economy (sal)
refers to the average quarterly amounts paid by employers to their employees, and it is
expressed in the national currency, RON. The Harmonised Index of Consumer Prices
(abbreviated HICP) is an “economic indicator that measures the change over time of the
prices of consumer goods and services acquired by households” (definition provided by
Eurostat) and constitutes an indicator for inflation. The variable is expressed as an index, for
which the reference moment is represented by the first quarter of the year 2005. Finally, the
exchange rate between RON and EUR (exchange) represents the rate at which one currency
(RON) will be exchanged for another currency (EUR).
To proceed with the estimates for Romania, we used quarterly data starting from the 1st
quarter of 2000 and up to the 2nd quarter of 2021, the data being mostly provided by the
Eurostat database (for government debt, active population, HICP, exchange) and by the
website of the Ministry of Labour (only the evolution of the average salary in the
economy).We used log transformation for data, except for the exchange rate and the HICP.
Compared to the first quarter of 2000, the second quarter of 2021 shows very different values
for the analysed variables, as follows: the active population is 19 percentage points lower, in
absolute terms it is about 1.9 million people less for the workforce represented by employees
and the unemployed. The average salary in the economy experienced gradual increases,
reaching in 2021 a value 19 times higher than in 2000. The government debt increased
gradually, reaching in 2003 a value three times that of 2000, and this value was somewhat

– Economic Studies Journal (Ikonomicheski Izsledvania), 32(8), pp. 35-56.
41
maintained at a similar level until 2008. Then, given the context of the financial crisis, the
government debt increased, and the COVID-19 pandemic period caused that during 6
quarters (2020Q1 – 2021Q2) accumulate a surplus of government debt, equivalent to 9 times
the value of the entire government debt from the year 2000.
The HICP had a gradual evolution, reaching in 2021 to have 4.66 times the value of the year
2000, which shows that inflation had a slower growth, there were no periods of inflationary
boom. As for the RON-EUR exchange rate, it increased the least among all the other variables
included in the study, compared to its own value from the year 2000. With an increase of
2.66 times compared to the year 2000, we can say that the exchange rate had a controlled
flotation and therefore, even at first glance, it does not seem to constitute a fiscal risk.
Figure 1. The evolution of the analysed variables compared to their value from the year
2000, quarter I
Source: own representation using Eurostat and Romanian Ministry of Labour data.
4. Modelling and Findings
To check if there are links between the five variables that are included in the analysis, we
started with the following hypotheses:
H1: l_debt = F( exchange, HICP, l_debt, l_activepop, l_sal )
H2: l_sal = F( exchange, HICP, l_debt, l_activepop, l_sal )
H3: l_activepop = F( exchange, HICP, l_debt, l_activepop, l_sal )
H4: HICP = F( exchange, HICP, l_debt, l_activepop, l_sal )
H5: exchange = F( exchange, HICP, l_debt, l_activepop, l_sal )
0
500
1000
1500
2000
2500
3000
3500
4000
4500
2000-Q3
2001-Q3
2002-Q3
2003-Q3
2004-Q3
2005-Q3
2006-Q3
2007-Q3
2008-Q3
2009-Q3
2010-Q3
2011-Q3
2012-Q3
2013-Q3
2014-Q3
2015-Q3
2016-Q3
2017-Q3
2018-Q3
2019-Q3
2020-Q3
Average salary index Active population index
Government debt index

Porumboiu, A. E., Brezeanu, P. (2023). Macroeconomic Factors that Generate Fiscal Risk in Romania.
42
Table 1 includes the main statistical parameters of the variables included in the analysis:
Table 1. Descriptive statistics
EXCHANGE HICP L_DEBT L_ACTIVEPOP L_SAL
Mean 3.994993 125.1977 11.60796 9.123588 7.073937
Median 4.257000 137.9650 11.92217 9.111310 7.269255
Maximum 4.924000 175.9800 13.17346 9.278064 8.169336
Minimum 1.850600 37.76000 9.511326 8.993986 5.187386
Std. Dev. 0.734462 37.38058 1.026696 0.057982 0.761393
Observations 86 86 86 86 86
Source: own results obtained through Eviews software
The correlation coefficients are presented in Table 2 and illustrate a strong correlation
between the variables under analysis.
Table 2. Correlation matrix
EXCHANGE HICP L_DEBT L_ACTIVEPOP L_SAL
EXCHANGE 1.000000 0.934750 0.925912 -0.888426 0.920761
HICP 0.934750 1.000000 0.970304 -0.881426 0.986515
L_DEBT 0.925912 0.970304 1.000000 -0.885611 0.945281
L_ACTIVEPOP -0.888426 -0.881426 -0.885611 1.000000 -0.855317
L_SAL 0.920761 0.986515 0.945281 -0.855317 1.000000
Source: own results obtained through Eviews software.
To verify the stationarity of the series, we applied unit root tests. In order to determine
whether the series are integrated of order 0, we used two unit root tests, both the Augmented
Dickey-Fuller test and the Phillips-Perron test in Eviews.
The ADF test gave us a probability of 65% for government debt, 43% for the labour force,
and 70% for the average salary (greater than 5%), which means that we will not reject the
null hypothesis, so these series are non-stationary. As for the exchange rate and the HICP,
the probabilities indicated by the test are less than 5%, so we will accept the null hypothesis
that there is a unit root and the series are stationary. Similar results are also indicated by the
PP test, except for the series corresponding to the average salary in the economy, which is
suggested to be a stationary series (obtained probability of 0%).
Table 3. ADF-level stationarity test for the studied variables
Variable exchange l_debt HICP l_activepop l_sal
Test critical
values t-statistic t-statistic t-statistic t-statistic t-statistic
ADF test
statistic -3.105720 -1.243046 -6.990468 -1.691584 -1.128088
1% -3.510259 -3.511262 -3.509281 -3.513344 -3.513344
5% -2.896346 -2.896779 -2.895924 -2.897678 -2.897678
10% -2.585396 -2.585626 -2.585172 -2.586103 -2.586103
Interpretation Stationary
series
Non-stationary
series
Stationary
series
Non-stationary
series
Non-stationary
series
Source: own results obtained through Eviews software

– Economic Studies Journal (Ikonomicheski Izsledvania), 32(8), pp. 35-56.
43
Table 4. PP-level stationarity test for the studied variables
Variable exchange l_debt HICP l_activepop l_sal
Test critical values t-statistic t-statistic t-statistic t-statistic t-statistic
PP test statistic -4.613771 -1.609957 -11.65858 -2.171072 -5.233594
1% -3.509281 -3.509281 -3.509281 -3.509281 -3.509281
5% -2.895924 -2.895924 -2.895924 -2.895924 -2.895924
10% -2.585172 -2.585172 -2.585172 -2.585172 -2.585172
Interpretation Stationary
series
Non-
stationary
series
Stationary
series
Non-
stationary
series
Stationary
series
Source: own results obtained through Eviews software
Since the ADF test indicates three of the variables as non-stationary, we applied the first-
order difference for them. We have obtained stationary series for all the variables in question,
which are now integrated in the first order. Consequently, we expect the relationships
between the variables under analysis to manifest themselves in the long run.
Table 5. ADF stationarity test - differences of the first order for the studied variables
Variable exchange l_debt HICP l_activepop l_sal
Test critical values t-statistic t-statistic t-statistic t-statistic t-statistic
ADF test statistic -6.560735 -3.585188 -5.372754 -17.60305 -3.249879
1% -3.510259 -3.511262 -3.510259 -3.511262 -3.513344
5% -2.896346 -2.896779 -2.896346 -2.896779 -2.897678
10% -2.585396 -2.585626 -2.585396 -2.585626 -2.586103
Interpretation Stationary
series
Stationary
series
Stationary
series
Stationary
series
Stationary
series
Source: own results obtained through Eviews software
Table 6. Stationarity test PP- differences of the first order for the studied variables
Variable exchange l_debt HICP l_activepop l_sal
Test critical values t-statistic t-statistic t-statistic t-statistic t-statistic
PP test statistic -6.549692 -7.569184 -5.204739 -12.76910 -9.470411
1% -3.510259 -3.510259 -3.510259 -3.510259 -3.510259
5% -2.896346 -2.896346 -2.896346 -2.896346 -2.896346
10% -2.585396 -2.585396 -2.585396 -2.585396 -2.585396
Interpretation Stationary
series
Stationary
series
Stationary
series
Stationary
series
Stationary
series
Source: own results obtained through Eviews software
Given the expert recommendation to use the Schwartz criterion for quarterly series that have
more than 20 observations, we will take this into account to determine the number of lags of
the model. According to this criterion, the optimal number of lags is 3.

Porumboiu, A. E., Brezeanu, P. (2023). Macroeconomic Factors that Generate Fiscal Risk in Romania.
44
Table 7. Criteria for determining the number of lags
Lag LogL LR (sequential
modified LR test
statistic)
FPE (Final
prediction
error)
AIC (Akaike
information
criterion)
SC (Schwarz
information
criterion)
HQ (Hannan-
Quinn
information
criterion)
0 88.03907 NA 2.27e-05 -2.180489 -2.089846 -2.144203
1 472.3199 729.1483 1.50e-09 -11.80308 -11.44050 -11.65793
2 502.2979 54.57530 8.78e-10 -12.34097 -11.70647 -12.08697
3 529.8995 48.12578 5.47e-10 -12.81793 -11.91151* -12.45508
4 541.1222 18.70448 5.19e-10 -12.87493 -11.69657 -12.40321
5 565.3879 38.57627 3.54e-10 -13.26636 -11.81607 -12.68578
6 576.1907 16.34269 3.42e-10 -13.31258 -11.59037 -12.62315
7 593.1593 24.36520* 2.84e-10* -13.51690* -11.52277 -12.71861*
8 601.4855 11.31506 2.95e-10 -13.49963 -11.23356 -12.59248
* indicates lag order selected by the criterion
Source: own results obtained through Eviews software
However, it must be taken into account that the macroeconomic variables are both stationary
and non-stationary, which also implies checking the cointegration of the time series, in order
to determine which estimate can be accepted.
Table 8. Cointegration test
Unrestricted Cointegration Rank Test (Trace)
Hypothesized Trace 0.05
No. of CE(s) Eigenvalue Statistic Critical Value Prob.**
None * 0.365652 84.77993 69.81889 0.0020
At most 1 0.289998 47.00181 47.85613 0.0600
At most 2 0.104009 18.57533 29.79707 0.5238
At most 3 0.088938 9.459821 15.49471 0.3246
At most 4 0.020614 1.728802 3.841466 0.1886
Trace test indicates 1 cointegrating eqn(s) at the 0.05 level
* denotes rejection of the hypothesis at the 0.05 level
**MacKinnon-Haug-Michelis (1999) p-values
Source: own results obtained through Eviews software
Table 9. The coefficients of the cointegration equation
1 Cointegrating Equation(s): Log-likelihood 541.4410
N
ormalized cointegrating coefficients (standard error in parentheses)
L_DEBT L_SAL L_ACTIVEPOP EXCHANGE HICP
1.000000 1.993308 -6.072202 -0.875232 -0.076134
(0.49530) (4.76459) (0.29863) (0.01263)
Source: own results obtained through Eviews software
The fact that there is cointegration between the variables under analysis, and there are both
stationary and non-stationary series, denotes that the VECM model can be estimated, as
indicated by Shrestha, M.B & Bhatta, G.R. (2018).

– Economic Studies Journal (Ikonomicheski Izsledvania), 32(8), pp. 35-56.
45
Figure 2. Determination of the right model according to the specifics of the analysed
series
Source: Shrestha, M.B & Bhatta, G.R. (2018)
The cointegration equation of the ECT considering debt as the interest variable, as announced
in the objectives of the paper, and the long-term model obtained by running the VECM in
Eviews is as follows:
ECT  1.000𝐿󰇛󰇜 2.234𝐿󰇛󰇜 5.580𝐿󰇛󰇜
0.952𝐸𝑋𝐶𝐻𝐴𝑁𝐺𝐸  0.079𝐻𝐶𝑃𝐼  37.331 (1)
Therefore, in the long term, we note that the value of the average salary in the economy exerts
a significant influence in the same direction on the government debt, while the active
population exerts an influence in the opposite direction, but with a larger value impact. the
economic foundation that explains these relationships is the following: the increase in the
average salary in the economy also includes the salary level of budget workers, therefore, the
increase in the average level would mean an increase in public expenses for the payment of
employees employed by public institutions, hence an increased need to obtain income public,
even by the instrumentality of indebtedness. On the other hand, the opposite relationship
between the active population and long-term government debt confirms the fact that a
growing labour force can generate additional sources of tax revenues and social contributions
for the government, thus new public revenues that would decrease the need to obtain credits.

Porumboiu, A. E., Brezeanu, P. (2023). Macroeconomic Factors that Generate Fiscal Risk in Romania.
46
Compared to the other variables, the HICP and the exchange rate exert a less important
influence on the public debt in the long run.
If we consider the debt as the target variable, we can say that in the short term, estimates
show that the deviations of previous quarters from the long-term equilibrium are corrected in
the current period at a speed of 0,8%. If we refer to sal as the main variable, the previous
period deviations from the long run equilibrium are corrected in the current period at an
adjustment speed of 4.9%, while for activepop the speed is of 1.4%, for exchange is of 5%,
and for HICP it is of 138%.
In conditions of caeteris paribus, as the coefficients estimated in Table 10 reveal, among the
variables under analysis, the change in the average salary and the change in the active
population are more able to produce effects on the government debt. 1% shocks in the level
of the exchange rate or the prices of consumer goods are not likely to generate an impact on
the government debt. Thus, considering the number of lags and the error correction estimates
presented in table 10, we will present the main consequences in the short run between the
variables in the following: 1% change in sal is associated to a 0.33% decrease in debt, on
average caeteris paribus, in the short run; in the same time, a 1% modification of activepop
leads to a 0.55% decrease in debt; 1% change of exchange would lead to 0.06% increase in
debt, while 1% modification of HICP conduces to only 0.01% increase of debt.
If we focus on the government debt as a possible indicator of fiscal risk, we observe that if
the number of active persons and the average salary increase, this will diminish the value of
the government debt, which confirms the economic principles which show that these positive
changes on the labour market mean an increase in public revenues, thus a greater capacity to
finance public expenses, to the detriment of the government debt.
Table 10. VECM estimates for the variables under analysis
Vector Error Correction Estimates
Sample (adjusted): 2001Q1 2021Q2
Included observations: 82 after adjustments
Standard errors in ( ) & t-statistics in [ ]
Cointegrating Eq: CointEq1
L_DEBT(-1) 1.000000
L_SAL(-1) 2.234770
(0.37584)
[ 5.94610]
L_ACTIVEPOP(-1) -5.580016
(4.01443)
[-1.38999]
EXCHANGE (-1) -0.952348
(0.24455)
[-3.89436]
HICP(-1) -0.079593
(0.00981)
[-8.11309]

– Economic Studies Journal (Ikonomicheski Izsledvania), 32(8), pp. 35-56.
47
C 37.33150
Error Correction: D(L_DEBT) D(L_SAL) D(L_ACTIVEPOP) D(EXCHANGE) D(HICP)
CointEq1 -0.008460 0.049314 -0.014786 0.050870 1.389985
(0.01969) (0.00895) (0.00718) (0.03992) (0.40359)
[-0.42954] [ 5.50942] [-2.05829] [ 1.27430] [ 3.44408]
D(L_DEBT(-1)) 0.165748 -0.186429 -0.087998 0.422005 -1.542798
(0.12571) (0.05713) (0.04585) (0.25480) (2.57600)
[ 1.31854] [-3.26320] [-1.91917] [ 1.65622] [-0.59891]
D(L_DEBT(-2)) 0.437101 -0.078100 0.002745 -0.021710 2.797173
(0.12714) (0.05778) (0.04637) (0.25770) (2.60531)
[ 3.43806] [-1.35167] [ 0.05919] [-0.08425] [ 1.07364]
D(L_DEBT(-3)) 0.085575 -0.103780 -0.020775 -0.164704 -1.432176
(0.13825) (0.06283) (0.05043) (0.28023) (2.83308)
[ 0.61899] [-1.65171] [-0.41196] [-0.58775] [-0.50552]
D(L_SAL(-1)) 0.461707 -0.513574 0.058346 0.220677 -8.134823
(0.25530) (0.11603) (0.09312) (0.51748) (5.23167)
[ 1.80849] [-4.42628] [ 0.62655] [ 0.42645] [-1.55492]
D(L_SAL(-2)) 0.364534 -0.043958 0.132890 -0.316072 -5.985569
(0.25197) (0.11451) (0.09191) (0.51072) (5.16336)
[ 1.44676] [-0.38387] [ 1.44592] [-0.61887] [-1.15924]
D(L_SAL(-3)) -0.330139 -0.254706 0.120570 0.191603 -4.929696
(0.22231) (0.10103) (0.08109) (0.45061) (4.55559)
[-1.48506] [-2.52099] [ 1.48690] [ 0.42521] [-1.08212]
D(L_ACTIVEPOP(-1)) -0.242940 0.354343 -0.289748 -1.122037 0.273715
(0.36239) (0.16470) (0.13218) (0.73454) (7.42616)
[-0.67039] [ 2.15147] [-2.19200] [-1.52753] [ 0.03686]
D(L_ACTIVEPOP(-2)) 0.440340 0.580881 -0.745456 0.326987 12.22438
(0.25258) (0.11479) (0.09213) (0.51198) (5.17606)
[ 1.74333] [ 5.06016] [-8.09112] [ 0.63867] [ 2.36172]
D(L_ACTIVEPOP(-3)) -0.555423 0.598075 -0.113337 -1.415589 7.292355
(0.34915) (0.15868) (0.12735) (0.70770) (7.15483)
[-1.59080] [ 3.76906] [-0.88994] [-2.00025] [ 1.01922]
D(EXCHANGE (-1)) -0.047531 0.016170 -0.004308 0.232696 1.290988
(0.06116) (0.02780) (0.02231) (0.12398) (1.25340)
[-0.77710] [ 0.58168] [-0.19312] [ 1.87692] [ 1.02999]
D(EXCHANGE (-2)) 0.015183 0.042301 0.023696 -0.177239 -1.098480
(0.06327) (0.02875) (0.02308) (0.12824) (1.29645)
[ 0.23998] [ 1.47118] [ 1.02683] [-1.38213] [-0.84730]
D(EXCHANGE (-3)) 0.063231 0.022750 0.014235 0.201903 0.318334
(0.06013) (0.02733) (0.02193) (0.12187) (1.23214)
[ 1.05163] [ 0.83254] [ 0.64905] [ 1.65665] [ 0.25836]

Porumboiu, A. E., Brezeanu, P. (2023). Macroeconomic Factors that Generate Fiscal Risk in Romania.
48
D(HICP(-1)) -0.005994 0.001841 -0.001623 -0.008450 0.180321
(0.00569) (0.00259) (0.00208) (0.01153) (0.11660)
[-1.05349] [ 0.71181] [-0.78181] [-0.73269] [ 1.54651]
D(HICP(-2)) -0.002156 0.003045 0.004444 0.003388 0.074334
(0.00569) (0.00259) (0.00208) (0.01154) (0.11663)
[-0.37874] [ 1.17714] [ 2.14063] [ 0.29363] [ 0.63733]
D(HICP(-3)) 0.008979 -0.000359 -0.003206 -0.005100 0.156379
(0.00559) (0.00254) (0.00204) (0.01134) (0.11460)
[ 1.60559] [-0.14143] [-1.57183] [-0.44993] [ 1.36459]
C -0.008053 0.068733 -0.012700 0.021928 1.580580
(0.02344) (0.01065) (0.00855) (0.04751) (0.48032)
[-0.34356] [ 6.45234] [-1.48549] [ 0.46154] [ 3.29071]
R
-squared 0.423411 0.679892 0.715836 0.355649 0.579748
A
dj. R-squared 0.281482 0.601097 0.645888 0.197040 0.476301
S
um sq. resids 0.149954 0.030973 0.019951 0.616094 62.97116
S
.E. equation 0.048031 0.021829 0.017520 0.097357 0.984270
F
-statistic 2.983248 8.628548 10.23384 2.242295 5.604317
L
og likelihoo
d
142.1170 206.7822 224.8155 84.18112 -105.5272
A
kaike AIC -3.051633 -4.628833 -5.068670 -1.638564 2.988469
S
chwarz SC -2.552679 -4.129879 -4.569716 -1.139610 3.487423
M
ean dependen
t
0.041037 0.031907 -0.002841 0.033524 1.552683
S
.D. dependen
t
0.056664 0.034562 0.029441 0.108647 1.360108
D
eterminant resid covariance (dof adj.) 2.29E-12
D
eterminant resid covariance 7.16E-13
L
og likelihoo
d
564.7770
A
kaike information criterion -11.57993
S
chwarz criterion -8.938407
N
umber of coefficients 90
Source: own results obtained through Eviews software
In addition to estimating the VECM model, it is important to determine whether there are
indeed interdependencies between the variables under analysis. We will apply the Granger
causality test (Appendix 1). It studies pairs of variables (X and Y) and identifies whether the
evolution of one variable (Y) produces a change in the other analysed variable (X). The null
hypothesis of the causality test shows that the evolution of one variable does not produce
results on the other variable, and vice versa. When the probability of the F-Statistic estimated
by the test has a value lower than 0.05 (5%), then the null hypothesis is rejected, in which
case there is a Granger causality between the analysed variables. This means that a change in
one variable will also influence the evolution of the other variable. The existence of
cointegration also confirms the validity of the Granger causality test. Ac cording to t he author s
Sims et al. (1990), if the variables are non-stationary and the cointegration condition would
not have been met, the results provided by the Granger test could not be accepted. The results
of the Granger causality test can be found in Appendix 1 and show that:

– Economic Studies Journal (Ikonomicheski Izsledvania), 32(8), pp. 35-56.
49
• Government debt Granger causes the active population;
• Government debt Granger causes the exchange rate;
• The active population Granger causes the average salary;
• The exchange rate Granger causes the active population;
• The average salary Granger causes the HICP;
• There is bidirectional Granger causality between the HICP and the labour force.
Therefore, the VECM model is representative to explain the interdependencies that are
created between the government debt variables, the level of the average salary in the
economy, the active population, the HICP and the RON-EURO exchange rate and above all,
to quantify the evolution of the government debt in case of possible changes to the level of
the studied variables. The impulse responses that evaluate the shock on each variable of
interest to a change in the reference variable are shown in Figure 3. Some of the estimates
that can be formulated starting from the graphical representation would be the following:
• A 1% shock to the average salary in the economy is quickly reflected in the government
debt, which experiences considerable growth in the medium term. The government debt
also responds to the shock produced by an evolution of the active population, but the
response is more moderate than in the case of the evolution of the average salary. This
practically confirms the economic theory since the determination of the price on the
labour market is not a simple result of the meeting between the demand and the supply of
labour, and the salary expenses are one of the key components of public expenses, less
flexible expenses due to their social characteristics;
• Broadly speaking, a 1% shock on the RON-EURO exchange rate does not generate
significant changes on the government debt. The interdependence between the exchange
rate and the government debt deserves to be studied separately, as several other factors
must be considered: the structure of the government debt portfolio in currencies, payment
terms, the history of the country regarding the payment of foreign interests, etc.
• The impulse response of the average salary level on the active population and vice versa
suggests, as expected, that a change in the average salary causes an increase in the active
population in the short term, but the trend is not maintained in the medium term, given
the fact that the labour market knows numerous imperfections, and the employee status
is the result of several demographic, social, cultural conditions in a context that exceeds
the country's borders.
Appendix 2 captures the variance decomposition of the forecast error. The evolution of the
government debt is weakly influenced during the next 10 quarters by the evolution of the
other analysed variables, the only variable that exerts an influence greater than 2% is the
average salary in the economy. Therefore, government debt is a strongly endogenous
variable, or whose change is generated by other variables than those included in the present
study.

Porumboiu, A. E., Brezeanu, P. (2023). Macroeconomic Factors that Generate Fiscal Risk in Romania.
50
The variables labour force, HICP and exchange rate are explained to a proportion of 50% or
more by their own evolution, although there is also an important influence on them from the
other variables.
During the evaluation period, we observe that the average salary is the variable least affected
by its own evolution. Although it was expected that the salary level would be correlated with
the evolution of the exchange rate and with the HICP for example, the estimate denotes the
fact that it is more strongly influenced by the evolution of the government debt even than by
its own change.
Figure 3. Impulse response estimates
.00
.02
.04
.06
.08
2 4 6 8 10
Response of L_DATORIE to L_DATORIE
.00
.02
.04
.06
.08
2 4 6 8 10
Response of L_DATORIE to L_P OPULATIE
.00
.02
.04
.06
.08
2 4 6 8 10
Respon se of L_DA TORIE t o L_S A LAR IU
.00
.02
.04
.06
.08
2 4 6 8 10
Response of L_DATORIE to HCPI
.00
.02
.04
.06
.08
2 4 6 8 10
Response of L_ DATORIE to C URS _DE _SCH IMB_RON _EU R
-.005
.000
.005
.010
.015
2 4 6 8 10
Response of L_POP ULATIE to L_DA TORIE
-.005
.000
.005
.010
.015
2 4 6 8 10
Resp onse of L_P OP ULA TIE to L_P OP ULA TIE
-.005
.000
.005
.010
.015
2 4 6 8 10
Respons e of L_PO PULA TIE to L_SA LA RIU
-.005
.000
.005
.010
.015
2 4 6 8 10
Response of L_POPULATIE to HCPI
-.005
.000
.005
.010
.015
2 4 6 8 10
Response of L_P OP ULA TIE to CU RS_D E_S CHI MB_RON_E UR
-.02
-.01
.00
.01
.02
.03
2 4 6 8 10
Respons e of L_S ALA RIU to L_DA TORIE
-.02
-.01
.00
.01
.02
.03
2 4 6 8 10
Respons e of L_SA LA RIU to L_P OPULA TIE
-.02
-.01
.00
.01
.02
.03
2 4 6 8 10
Response of L_S ALA RIU to L_SA LAR IU
-.02
-.01
.00
.01
.02
.03
2 4 6 8 10
Response of L_S A LARI U to HCP I
-.02
-.01
.00
.01
.02
.03
2 4 6 8 10
Response of L_SA LA RIU t o CUR S_DE _S CHIMB_R ON_E UR
-0.4
0.0
0.4
0.8
2 4 6 8 10
Response of HCPI to L_DATORIE
-0.4
0.0
0.4
0.8
2 4 6 8 10
Respons e of HC PI to L_POP ULA TIE
-0.4
0.0
0.4
0.8
2 4 6 8 10
Response of H CP I to L_SA LAR IU
-0.4
0.0
0.4
0.8
2 4 6 8 10
Response of H CP I to HCP I
-0.4
0.0
0.4
0.8
2 4 6 8 10
Response of HC PI to CUR S_DE _SC HIMB_R ON_EU R
.00
.05
2
4
6
8
10
Response of C URS _DE _SC HIMB_R ON_E UR to L_D ATORIE
.00
.05
2
4
6
8
10
Response of C URS _DE _SC HIMB_R ON_E UR to L_P OPULA TIE
.00
.05
2
4
6
8
10
Response of CU RS_D E_S CHIMB _RON_E UR to L_SA LAR IU
.00
.05
2
4
6
8
10
Response of CU RS_D E_S CHIMB _RON_E UR t o HCP I
.00
.05
2
4
6
8
10
Response of CUR S_DE _SC HIMB_R ON_EU R to C URS _DE _SC HIMB_RON _EU
R
Respons e to Cholesk y O ne S. D. (d.f. ad
j
ust ed) I nnovations
Source: own results obtained through Eviews software

– Economic Studies Journal (Ikonomicheski Izsledvania), 32(8), pp. 35-56.
51
5. Conclusions
This article aims to identify the interdependencies between the government debt, the number
of people active on the labour market, the level of the average salary in the economy, the
harmonized index of consumer prices and the RON-EURO exchange rate, for Romania
during the period starting from the 1st quarter of the year 2000 and up to the 2nd quarter of
the year 2021. Given the specificity of the series under analysis, the economic model used
for estimations is VECM.
The focus of the analysis is still related to fiscal risks, which can be measured by government
debt. Thus, the main objective of the study is to identify which of the other variables used
exert a greater impact on the government debt: the size of the average salary and the number
of employed people that can bring increases in public revenues (or, on the contrary, increases
in social assistance expenses and budget employees' salary expenses), HICP which is a form
of measuring inflation (inflation is a known means of reducing of the value of the government
debt), or the RON-EUR exchange rate that can influence the government debt depending on
the predominant currency of the loans and the evolution of the national currency?
In the long term, the average salary exerts a significant impact in the same direction on the
government debt, whilst the labour force (active population) exerts an influence in the
opposite direction, but with a larger value impact. Two very important conclusions emerge
from this. The first is that an increase in the average salary that does not have a corresponding
increase in productivity only increases the government's indebtedness and fiscal pressure for
the next generations. The salary increases of budget officers require prudence in order not to
become a risk for fiscal sustainability in the long term. The second conclusion is that
government measures to increase the active population prove to be beneficial in the long term
and make the public debt decrease faster than the number of active people increases.
Thus, it is recommended that the government implement measures to increase the labour
force by: discouraging the phenomenon of early retirement, better visibility of existing jobs,
attracting foreign investors to open workplaces in Romania, encouraging local businesses,
the implementation of professional training and conversion programs (including the fruition
of human capital development opportunities through projects financed by the European
Union).
We note that HICP and RON-EUR exchange rate don’t have as much influence on the
evolution of government debt as the other two variables mentioned before, from which it
follows that there were generally neither large fluctuations in prices, nor major changes in
the exchange rate that would require the refinancing of the public debt at high costs.
Instead, in the short term, the modifications in the average salary and in active population are
more able to produce effects on the government debt: 1% change in sal is associated to a
0.33% decrease in debt, on average caeteris paribus, in the short run; in the same time, a 1%
modification of activepop leads to a 0.55% decrease in debt; while 1% change in HICP or in
exchange rate conduct to less than 0.10% change in government debt.
Not least, according to the variance decomposition of the forecast error, the most important
conclusions show that the government debt is influenced, among the variables included in

Porumboiu, A. E., Brezeanu, P. (2023). Macroeconomic Factors that Generate Fiscal Risk in Romania.
52
the study, by changes in the average salary in the economy. The increase in the average salary
implies (and derives from) also increases in expenses with the salaries of budget workers.
The latter is a rigid public expenditure, that needs financing funds, obtained even through
borrowing. Fluctuations in the number of active persons do not have a medium-term impact
on the government debt, just as the change in the average salary does not have the ability to
maintain in the medium and long term an evolution in the same direction of the active
population. As for the ability of the RON-EUR exchange rate to influence the government
debt, the estimates do not prove a sensitivity of the latter determined by the euro currency,
which denotes, at least at first glance, a prudent structure of the government debt portfolio.
References
Afflatet, N. (2018). The Impact of Population Ageing on Government debt: A Panel Data Analysis for 18 European
Countries. – International Journal of Economics and Financial Issues, 8 (4), pp. 68-77.
Auerbach, A. J. (2009). Long-Term Objectives for Government Debt. – FinanzArchiv / Public Finance Analysis, 65
(4), pp. 472-501.
Balassone, F., Cunha, J., Langenus, G., Manzke, B., Pavot, J., Prammer, D., Tommasino, P. (2009). Fiscal
Sustainability and Policy Implications for the Euro Area. – Working Paper Series No 994, Frankfurt am
Main, European Central Bank.
Bengtsson, T., Scott, K. (2011). Population Aging and the Future of the Welfare State: The Example of Sweden. –
Population and Development Review, 37, pp. 158-170.
Birg, H. (2015). Die alternde Republik und das Versagen der Politik: Eine demographische Prognose. Münster,
Berlin: LIT-Verlag.
Blanchard, O., Chouraqui, J., Hagemann, R., Sartor, N. (1990). The sustainability of fiscal policy: new answers to
an old question. – OECD Economic Studies, 15, pp. 7-36.
Bongaarts, J. (2004). Population Aging and the Rising Cost of Public Pensions. – Population and Development
Review, 30(1), pp. 1-23.
Cecchetti, S. G., Madhusudan, M., Zampolli, F. (2010). The Future of Government debt: Prospects and Implications.
– BIS Working Papers No. 300.
Chatzouz, M. (2020). Government debt and Wealth Inequality. [Online], [Retrieved April 24,
2022],http://dx.doi.org/10.2139/ssrn.3531199.
Chen, D. H. C. (2004). Population age structure and the budget deficit. – Journal for Insurance and Financial
Management, 1(4), pp. 32-67.
Demmel, R., Keuschnigg, C. (2000). Funded Pensions and Unemployment. – FinanzArchiv/Public Finance
Analysis, 57(1), pp. 22-38.
Dieppe, A., Guarda, P., Albani, M., Gonzalez Pandiella, A., Grech, O., Irac, D., Kilponen, J., Kulikov, D., Marchiori,
L., Félix, R., Papadopoulou, N., Rodano, M. L., Sideris, D., Vogel, E., Gordo Mora, E. (2015). Government
debt, Population Ageing and Medium-Term Growth. – ECB Occasional Paper No. 165.
Farmer, K., Kuplen, S. (2018). Involuntary Unemployment in a Neoclassical Growth Model with Government debt
and Human Capital. – Studia Universitatis Babes-Bolyai Oeconomica, 63(2), pp. 3-34.
Genschel, P. (2001). Globalization, tax competition, and the fiscal viability of the welfare state. – MPIfG Working
Paper, No. 01/1, Cologne, Max Planck Institute for the Study of Societies.
Hansson, A. (2010). In This World Nothing Is Certain but Death and Taxes: Financing the Elderly. – Population
Ageing – A Threat to the Welfare State?, Heidelberg, Springer Berlin, pp. 23-45.
International Monetary Fund. (2009). The IMF on the Fiscal Impact of Population Aging. – Population and
Development Review, 35(2), pp. 437-442.
Jensen, S. E. H., Søren, B. N. (1995). Population Ageing, Government Debt and Sustainable Fiscal Policy. – Fiscal
Studies, 16(2), pp. 1-20.
Klyvienė, V. (2004). The Government debt and the Problem of Population Ageing in Lithuania. Available at
http://dx.doi.org/10.2139/ssrn.2070664, referred on 20/04/2022.
Lee, R., Tuljapurkar, S. (1998). Uncertain demographic futures and social security finances. – The American
Economic Review, 88(2), pp. 237-241.

– Economic Studies Journal (Ikonomicheski Izsledvania), 32(8), pp. 35-56.
53
Menguy, S. (2020). Tax Competition, Fiscal Policy, and Government debt Levels in a Monetary Union. – Journal
of Economic Integration, 35(3), pp. 353-395.
Porumboiu, A. E., Brezeanu, P. (2022). Determinants of Government Debt in the Member States of the European
Union: Sources of Fiscal Risk. – Proceedings of the International Conference on Business Excellence,
16(1), pp. 707-721.
Preston, S. H. (1984). Children and the elderly: Divergent paths for America’s dependents. – Demography, 21(4),
pp. 435-457.
Rangelova, R., Bilyanski, V. (2019). Economic Aspects of Demographic Changes in the European Union and in
Bulgaria. – Economic Studies (Ikonomicheski Izsledvania), 28(5), pp. 25-54.
Razin, A., Efraim, S., Phillip, S. (2001). The Aging Population and the Size of the Welfare State. – NBER Working
Paper Series 8405.
Sánchez-Romero, M., Abío, G., Botey, M., Prskawetz, A., Sambt, J., Juvés, M. S., Souto, G., Vargha, L., Patxot, C.
(2019). Welfare State Winners and Losers in Ageing Societies. – Vienna Yearbook of Population Research,
17, pp. 9-36.
Shrestha, M. B., Bhatta, G. R. (2018). Selecting appropriate methodological framework for time series data analysis.
– The Journal of Finance and Data Science, 4(2), pp. 71-89.
Sinn, H. W., Silke, Ü. (2002). Pensions and the path to gerontocracy in Germany. – European Journal of Political
Economy, 19, pp. 153-158.
Tanchev, S., Mose, N. (2023). Fiscal Policy and Economic Growth: Evidence from European Union Countries. –
Economic Studies Journal (Ikonomicheski Izsledvania), 32(3), pp. 19-36.
Tomić, I. (2018). What drives youth unemployment in Europe? Economic vs non‐economic determinants. –
International Labour Review, 157(3), pp. 379-408.
Topal, M. H., Bölükbaş, M., Bostan, M. K. (2018). The Government debt and Unemployment Growth Nexus in the
PIIGS (Portugal, Ireland, Italy, Greece and Spain) Countries and Turkey: Empirical Evidence. – Studies on
Balkan and Near Eastern Social Sciences, 2, pp. 1-16.
Yared, P. (2019). Rising Government Debt: Causes and Solutions for a Decades-Old Trend. – The Journal of
Economic Perspectives, 33(2), pp. 115-140.

Porumboiu, A. E., Brezeanu, P. (2023). Macroeconomic Factors that Generate Fiscal Risk in Romania.
54
Appendix 1
Granger causality test
Null Hypothesis: Obs F-Statistic Prob.
L_ACTIVEPOP does not Granger Cause L_DEBT 83 2.44671 0.0702
L_DEBT does not Granger Cause L_ACTIVEPOP 7.15844 0.0003
L_SAL does not Granger Cause L_DEBT 83 1.48340 0.2258
L_DEBT does not Granger Cause L_SAL 1.39540 0.2507
HICP does not Granger Cause L_DEBT 83 1.19855 0.3161
L_DEBT does not Granger Cause HICP 1.82251 0.1502
EXCHANGE does not Granger Cause L_DEBT 83 1.08992 0.3586
L_DEBT does not Granger Cause EXCHANGE 3.35675 0.0231
L_SAL does not Granger Cause L_ACTIVEPOP 83 2.06470 0.1119
L_ACTIVEPOP does not Granger Cause L_SAL 11.7493 2.E-06
HICP does not Granger Cause L_ACTIVEPOP 83 3.74250 0.0145
L_ACTIVEPOP does not Granger Cause HICP 5.66151 0.0015
EXCHANGE does not Granger Cause
L_ACTIVEPOP 83 4.02693 0.0103
L_ACTIVEPOP does not Granger Cause EXCHANGE
2.68983 0.0522
HICP does not Granger Cause L_SAL 83 1.29596 0.2820
L_SAL does not Granger Cause HICP
2.92603 0.0391
EXCHANGE does not Granger Cause L_SAL 83 0.66056 0.5789
L_SAL does not Granger Cause EXCHANGE 1.48654 0.2249
EXCHANGE does not Granger Cause HICP 83 0.98309 0.4053
HICP does not Granger Cause EXCHANGE 1.26722 0.2917
Source: own results obtained through Eviews software

– Economic Studies Journal (Ikonomicheski Izsledvania), 32(8), pp. 35-56.
55
Appendix 2
Decomposition of forecast error variance
Variance Decomposition of L_DEBT:
Period S.E. L_DEBT L_SAL L_ACTIVEPOP EXCHANGE HICP
1 0.048031 100.0000 0.000000 0.000000 0.000000 0.000000
2 0.071697 97.06874 2.091219 0.145329 0.197629 0.497081
3 0.102940 94.50962 3.672110 0.770145 0.212689 0.835438
4 0.129501 95.55705 3.181306 0.511160 0.166162 0.584322
5 0.159034 96.02821 2.947297 0.340292 0.296745 0.387452
6 0.186213 96.49320 2.612465 0.248735 0.336586 0.309016
7 0.213345 96.69091 2.552267 0.191250 0.311748 0.253824
8 0.238281 96.86589 2.446594 0.161430 0.303961 0.222124
9 0.263491 96.89309 2.466633 0.148255 0.289585 0.202437
10 0.287002 96.96159 2.440336 0.129872 0.259773 0.208434
Variance Decomposition of L_SAL:
Period S.E. L_DEBT L_SAL L_ACTIVEPOP EXCHANGE HICP
1 0.021829 1.189430 98.81057 0.000000 0.000000 0.000000
2 0.027543 12.34495 85.92132 0.327328 0.889564 0.516840
3 0.034752 15.12480 78.80928 4.773992 0.861504 0.430425
4 0.041987 28.44087 62.28280 6.807582 0.978428 1.490316
5 0.049716 33.30501 58.07903 4.877168 1.679675 2.059117
6 0.058551 40.33927 49.79460 3.567126 2.839019 3.459981
7 0.068842 40.46952 45.93426 3.157552 4.538871 5.899802
8 0.079905 43.19655 39.77462 3.078150 6.332991 7.617696
9 0.090397 44.30046 37.59241 2.501413 7.314248 8.291471
10 0.100665 45.89219 34.93076 2.024022 8.004720 9.148311
Variance Decomposition of L_ACTIVEPOP:
Period S.E. L_DEBT L_SAL L_ACTIVEPOP EXCHANGE HICP
1 0.017520 1.465850 0.310026 98.22412 0.000000 0.000000
2 0.022605 2.606718 0.544643 96.66057 0.152974 0.035090
3 0.023836 3.170772 0.565670 88.06258 3.172622 5.028357
4 0.025176 3.391752 0.580837 82.32175 8.179525 5.526132
5 0.029330 2.965058 0.443631 85.66082 6.749332 4.181163
6 0.032461 5.335667 0.374519 84.78302 5.937477 3.569321
7 0.033754 7.372661 0.360304 79.74551 7.991042 4.530480
8 0.034847 8.139082 0.338887 77.04927 9.865719 4.607038
9 0.037682 8.329344 0.291694 78.49993 8.935112 3.943921
10 0.040294 10.01171 0.263643 77.91921 8.279349 3.526087
Variance Decomposition of EXCHANGE
Period S.E. L_DEBT L_SAL L_ACTIVEPOP EXCHANGE HICP
1 0.097357 17.97981 0.146384 0.159123 81.71468 0.000000
2 0.159174 24.02357 0.099340 1.322598 73.99798 0.556515
3 0.194940 28.91296 0.118503 1.557413 68.18984 1.221281
4 0.228593 29.69810 0.601428 1.654309 64.96015 3.086010

Porumboiu, A. E., Brezeanu, P. (2023). Macroeconomic Factors that Generate Fiscal Risk in Romania.
56
5 0.262869 30.88257 0.764920 2.153226 61.98742 4.211865
6 0.293384 32.26979 1.049924 2.428865 59.06119 5.190226
7 0.320131 33.21993 1.403103 2.354517 56.26570 6.756750
8 0.344985 33.88148 1.819987 2.431621 53.78860 8.078309
9 0.368760 34.29200 2.213809 2.667903 51.64911 9.177180
10 0.391548 34.68398 2.656722 2.742265 49.47846 10.43857
Variance Decomposition of HICP:
Period S.E. L_DEBT L_SAL L_ACTIVEPOP EXCHANGE HICP
1 0.984270 1.739118 2.355293 2.443715 0.299583 93.16229
2 1.478220 1.500530 4.577816 5.056199 0.300388 88.56507
3 1.866653 3.710542 5.343116 4.452801 2.216124 84.27742
4 2.202938 3.786926 5.650759 3.628472 5.595714 81.33813
5 2.509342 4.334810 4.852778 4.160208 8.583813 78.06839
6 2.810398 4.858230 4.228982 5.517831 11.45524 73.93972
7 3.060681 5.754609 3.611752 5.329459 15.34158 69.96260
8 3.288317 6.108548 3.154579 4.886057 19.95577 65.89504
9 3.522318 6.336670 2.755468 5.013423 23.78140 62.11304
10 3.746673 6.501581 2.461009 5.513684 27.13522 58.38850
Cholesky Ordering: L_DEBT L_WAGE L_ACTIVEPOP EXCHANGE HICP
Source: own results obtained through Eviews software