Reading Time: 13 minutes

1. Introduction

Real gross domestic product (real GDP) measures the total value of goods and services produced in an economy, corrected for changes in prices. In the case of Slovenia, we are looking at a series of “chain-linked volumes, index 2020 = 100,” which allows comparisons over time without distortions from inflation. This index form makes it easier to track how much real economic activity has moved relative to the reference year 2020, while chain-linking takes into account structural changes in the economy over time. Because Slovenia is a small, export-oriented member of the EU and the euro area, movements in real GDP reflect both domestic cycles and external shocks transmitted through trade, investment and financial flows.

The analysis here is based on a quarterly series of Slovenia’s real GDP from 2000Q1 to 2025Q2, processed in two ways. First, a logarithmic transformation is applied, which stabilises the variance and turns levels into approximate growth rates; this log series is then seasonally adjusted and decomposed using the STL (Seasonal-Trend decomposition using Loess) method. Second, we consider the first differences and seasonal differences of the log series, which approximately correspond to quarterly and year-on-year growth rates and further reduce trend and seasonal patterns. Both perspectives, on the level (in logs) and on changes, provide complementary views of the dynamics of the Slovenian economy.

In the broader macroeconomic story, the period from the early 2000s to the mid-2020s covers several clearly distinct phases: a convergence phase towards the European Union and rapid pre-crisis expansion up to 2007; a deep recession during the global financial crisis (GFC) of 2008–2009; a second, domestically driven crisis related to the banking sector and public finances around 2012–2013; and a period of recovery and consolidation that was abruptly interrupted by the COVID-19 pandemic in 2020, followed by a rapid but uneven rebound. Macroeconomic data on annual GDP growth confirm these turning points: double-digit or near double-digit growth rates before the GFC were replaced by a sharp contraction in 2009, then a slow recovery and a new recession in the early 2010s, while the COVID-19 crisis in 2020 appears as a one-off shock without a prolonged collapse of the trend.

In the final years of the sample, especially after the pandemic downturn and the subsequent bounce-back, Slovenia’s real GDP returns to a level clearly above the reference year 2020, but the dynamics are more moderate than during the pre-crisis expansion. The log series suggests that the last four to five years have been marked by a combination of recovery, normalisation and new external challenges, including the energy shock and the slowdown in key trading partners, above all Germany. Changes in seasonal patterns are more subtle, but the graphs point to some adjustment over time, probably connected with industrial restructuring, the rising share of services and the growing integration into European markets.

2. Description of the time series and its components

2.1. Movements in the level and the seasonally adjusted series

Figure 1 shows the original and seasonally adjusted series of Slovenia’s real GDP in logarithmic form. The original series displays a clear four-quarter seasonal structure, with recurring patterns of increases and decreases within the year, which is typical for quarterly macroeconomic data. The seasonally adjusted curve smooths out these intra-year oscillations and allows the focus to shift to the underlying trend and business cycle.

The early 2000s are characterised by steady upward movement, reflecting real convergence and preparations for EU accession in 2004 and the later introduction of the euro in 2007. Economic activity accelerates, and the seasonally adjusted series shows an almost monotonic rise, with only minor and short-lived slowdowns. In this period there are no dramatic breaks, but rather a gradual strengthening of the GDP level in line with Slovenia’s entry into the European economic space.

A sharp turning point appears with the global financial crisis. In the graph there is a clearly visible drop in seasonally adjusted GDP during 2008–2009, which in the logarithmic index appears as a sudden fall in the level, a strong negative shock. This shock interrupts the previous upward trend and establishes a new, lower path. The short-term recovery after 2010 is not sufficient to fully offset the lost output, and around 2012–2013 we see another phase of weakening, consistent with domestic strains in the banking sector and public finances and with the European sovereign-debt crisis.

From the mid-2010s onwards, the seasonally adjusted series enters a period of more robust growth. A sequence of positive quarters gradually lifts GDP back towards and above its previous peak, suggesting that structural adjustments, including the clean-up of the banking sector and improvements in export competitiveness, have started to bear fruit. However, the recovery is not linear; there are episodes of mild deceleration, especially in years with heightened external uncertainty.

The COVID-19 pandemic in 2020 is clearly visible in both the original and seasonally adjusted series as a sudden, short-lived collapse in economic activity. The log series records a steep drop over a few quarters, followed by a strong rebound, with the seasonally adjusted curve highlighting a V-shaped recovery. In the years after 2020 the series enters a new phase: the level of GDP exceeds its pre-pandemic values, but the path is interrupted by a combination of solid domestic recovery and external shocks, including surging energy prices and a slowdown in the euro area.

2.2. Trend, seasonal and irregular components

Figure 2 shows the STL decomposition of the log series of real GDP into the three standard components: trend, seasonal and irregular. In the trend component, the main phases hinted at by the seasonally adjusted series are clearly delineated. The first phase, from the early 2000s to 2008, is characterised by a stable upward trajectory with only mild oscillations. This is followed by a steep decline during the GFC, which breaks the previous path and establishes a new, lower level. After a short recovery, a second, milder downward phase around 2012–2013 reflects domestic financial tensions and the restructuring of the banking system. Only in the middle of the decade does the trend stabilise and shift into a period of gradual but persistent growth that continues up to the pandemic shock.

The pandemic downturn and recovery appear less dramatic in the trend component than in the original series because part of the extreme movement is absorbed by the seasonal and irregular components. Even so, it is evident that the pandemic temporarily interrupted the rising trajectory, after which the trend resumes its upward movement but with a somewhat lower slope, suggesting that the long-run growth potential is more moderate than during the pre-crisis expansion.

The seasonal component in the log series shows a relatively stable but not entirely unchanging pattern. In the early years the amplitude of seasonal oscillations is somewhat more moderate, while in the periods before and after the GFC we can see a more pronounced rhythm of quarterly swings. This may reflect changes in the structure of the economy, for example, a larger share of industry and exports that are strongly linked to global cycles, as well as the rising importance of tourism and services with clear seasonal peaks. After the pandemic, seasonality is still present, but its shape may be slightly altered, especially if the tourist season and consumption patterns have adjusted to new circumstances.

The irregular or residual component is relatively modest in most periods, suggesting that the bulk of variation in the series is explained by the trend and seasonal effects. Nevertheless, there are periods of heightened volatility, especially during 2008–2009, 2012–2013 and 2020, when the residual component displays large positive and negative spikes. These episodic shocks are typically associated with unexpected events such as global financial disruptions, the domestic banking crisis or public-health measures that temporarily shut down activity in entire sectors.

2.3. Seasonal diagnostics and detailed patterns

The seasonal diagnostic plots deepen our understanding of intra-year patterns and their stability over time. Seasonal plots in levels (the log series, Figure 3) organise the observations by quarter, so that for each quarter we can see how it typically behaves relative to the others.

For Slovenia, these plots reveal a recognisable pattern: some quarters, for example the second and fourth, systematically show higher levels of economic activity, which may be linked to stronger industrial production and the tourist season, while the first and sometimes the third quarter tend to be weaker, partly due to seasonal slowdowns in construction and post-holiday consumer spending.

Seasonal subseries plots (Figure 4) track the same quarter across many years and reveal how the seasonal profile changes over time. In the Slovenian series, we can observe that before the GFC, for instance, the second and third quarters followed a clear upward trajectory, whereas after the crisis there is some reduction or “flattening” of the seasonal peaks, suggesting that some seasonal growth drivers have lost strength. In the period after 2014, the seasonal peaks re-emerge, but with a somewhat altered structure, consistent with a larger role for services and tourism.

Lag plots (Figure 5) provide insight into autocorrelation, the extent to which GDP values in one quarter help predict values in the next. In the log series for Slovenia there is a clear positive relationship between adjacent quarters, meaning that periods of expansion tend to follow one another, and downturns also cluster. At the same time, the lag plots reveal how extreme shocks such as the GFC or the pandemic break the usual pattern, creating points that stand out from the normal cloud of observations.

Seasonal and subseries diagrams of the differenced series (Figures 6 and 7) show that, even after aggressive filtering, some intra-year patterns are still present, for example, somewhat higher growth rates in certain quarters, but they are less pronounced and more variable over time.

Lag plots for the differenced data (Figure 8) show reduced but still noticeable short-run autocorrelation, consistent with the idea that changes in GDP are partly serially correlated (quarterly growth is often followed by similar growth or contraction in the next quarter), but that differencing weakens this link and emphasises the unpredictable, shock-driven elements of the series.

When the log GDP series is further transformed by combining first differences, we obtain a series that can be interpreted as a mixture of quarterly and annual growth rates. In this form, the long-run trend disappears. The seasonally adjusted differenced series in Figure 9 shows a sequence of episodes of positive and negative growth, with particularly striking falls during the GFC and the pandemic, and smaller but noticeable negative movements during the sovereign- and banking-crisis period in the early 2010s. Recoveries after these episodes appear as spikes above the usual range, especially after the pandemic, when the strong rebound temporarily generates very high growth rates.

The STL decomposition of the differenced series (Figure 10) shows that, after removing the trend, the residual component dominates the movements, while the remaining trend-like elements become relatively small.

3. Economic outlook

When we compare the last part of the period, roughly from 2021 to 2025, with earlier phases, the picture of Slovenia’s real GDP suggests an economy that has successfully offset the pandemic losses but now operates in a context of higher uncertainty. The seasonally adjusted log series shows that the level of activity is above its pre-pandemic values, but the slope of the trend in the recent period appears somewhat flatter than in the early 2000s. This points to a more mature phase of convergence, in which the scope for exceptionally high growth rates is gradually fading, and issues of productivity, innovation and structural reform move to the forefront.

Seasonal patterns in the most recent period still reflect the importance of tourism, industry and exports, but they are shaped by new factors as well. Energy-price shocks after 2021, disruptions in global supply chains and the slowdown in key European partners create occasional negative impulses in quarterly growth rates, visible in the differenced series as a sequence of short but sometimes sharp declines. At the same time, active monetary and fiscal policy, including measures to cushion the impact of higher energy prices and support firms, has dampened the amplitude of these shocks, so the series does not show the same destructive pattern as during the GFC or the domestic banking crisis.

In the regional context of the former Yugoslavia, Slovenia’s real GDP series differs along several dimensions. First, it starts at a higher initial level and shows faster early convergence towards EU averages, as seen in the long-term upward trend and in the fact that Slovenia has long been among the most developed countries of Central and Eastern Europe in terms of GDP per capita. Second, integration into the EU and the euro area means that the economy is strongly tied to the cycles of the eurozone, particularly Germany’s, which heightens its sensitivity to external shocks but simultaneously provides access to a large market and financial resources. Third, the double crisis, the global financial crisis and the domestic banking and fiscal crisis, left a deeper and more prolonged imprint on Slovenian GDP than in some other countries in the region, but also triggered structural adjustments that strengthened resilience in later phases.

The most recent quarters in the graphs point to an economy moving between two extremes: it does not experience the spectacular growth rates of the pre-crisis boom, but nor does it suffer the dramatic renewed recessions that marked 2009 or 2012–2013. Instead, we see a mixture of moderate growth, occasional pauses and differentiated effects of shocks across sectors. In such an environment, the macroeconomic narrative becomes less about “big crashes” and more about the ability of policy and firms to adapt to gradual changes in global trade, technology and demographics.

For the short term, the time series of real GDP does not suggest an imminent severe recession or a sustained “out-performance”. Moderate growth, with occasional quarterly fluctuations, remains the most plausible scenario, provided that external shocks remain contained and that the gradual resolution of structural bottlenecks continues, from productivity and innovation to labour-market functioning and institutional effectiveness.

4. Methodological appendix

4.1. The role of graphical analysis

Graphical analysis of time series is a basic step in understanding how GDP and other macroeconomic indicators evolve. Simple plots of the level of the series make it possible to discern trends, long-run growth or stagnation, as well as business cycles, phases of expansion and recession, structural breaks and changes in volatility. Comparing the original and seasonally adjusted series helps to distinguish the underlying economic signal from recurring seasonal patterns.

At the same time, caution is needed. Visually striking changes do not always have a unique explanation; it is easy, for instance, to mistake stronger seasonality for a change in trend, or to over-interpret individual extreme observations that may be due to statistical revisions or one-off technical factors. Graphs should therefore be treated as a starting point for reasoning, not as definitive proof.

4.2. Transformations: logs and differencing

For series such as real GDP, which trend upwards over time and often exhibit increasing variance, it is common to apply a logarithmic transformation. If () denotes the original GDP series, the log-transformed series () approximately turns differences into relative changes, so differences in logs can be interpreted as growth rates. This transformation also reduces differences in volatility between periods with low and high levels of GDP, which facilitates modelling and interpretation.

First differencing, defined as (), focuses on quarter-to-quarter changes and helps stabilise the mean of the series, especially when there is a strong trend. Seasonal differencing, , where () is the seasonal period (for quarterly data typically four), removes recurring seasonal patterns so that the focus falls on changes relative to the same quarter of the previous year. In Slovenia’s case, the combination of logging, first differences differences made it possible to separate long-run levels, persistent seasonal patterns and short-term growth rates.

4.3. Seasonal adjustment and decomposition using STL

Seasonal adjustment of GDP in this analysis was carried out using the STL (Seasonal-Trend decomposition using Loess) method, which is a standard tool for decomposing time series into trend, seasonal and irregular components. In this framework, the observed log GDP ,

where () is the trend-cycle, () the seasonal component and () the irregular remainder. The trend represents the slowly changing baseline of growth, the seasonal component describes recurring within-year patterns (for example, typically stronger second and third quarters), and the residual captures the remaining, unpredictable or short-term shocks.

STL uses local regression smoothing (Loess) to estimate the trend and seasonal components. The method iteratively passes through the series, estimating a smooth trend curve over longer windows and a seasonal component by comparing the same positions within the seasonal cycle (for example, all second quarters across years), while employing robust procedures to limit the influence of extreme values. Because of its flexibility, the ability to adjust window lengths and the robustness of the fitting, STL is particularly well-suited to series such as real GDP, where seasonality may be stable but not perfectly constant over time.

Applied to the log series and its differenced versions, STL made it possible to clearly distinguish the long-run trends of Slovenia’s economic convergence, the seasonal patterns associated with the structure of the economy (industry, tourism, construction) and irregular shocks such as the GFC, the domestic banking crisis and the pandemic. The methodological framework promoted in Eurostat and ESS guidelines further underlines the importance of systematic and transparent seasonal adjustment in the analysis of key macroeconomic series.

5. Conclusion

The time series of Slovenia’s real GDP from 2000Q1 to 2025Q2 offers a compact view of twenty-five years of economic successes, challenges and adjustments. After a stable and rapid expansion in the early 2000s, driven by convergence towards the EU and the adoption of the euro, the economy experienced a sharp contraction during the global financial crisis and an additional break caused by the domestic banking and fiscal crisis in the early 2010s. The trend component in the STL decomposition clearly marks these breaks, while the seasonally adjusted series shows how the recovery was uneven but eventually lifted GDP to levels above its pre-crisis peaks.

The seasonal characteristics of real GDP remained relatively stable, with repeated patterns of stronger and weaker quarters reflecting a combination of industrial, tourism-related and institutional cycles. Gradual changes in the amplitude and shape of these patterns point to adjustments in the structure of the economy, a growing role for services, shifts in export markets and seasonal effects of new sectors. The differenced log series further highlights the key episodes of volatility, where the GFC, the domestic banking crisis and COVID-19 left the most pronounced marks, but also shows that these shocks, although dramatic, were temporally concentrated.

Compared with the previous decade, the last years of the sample show an economy that has gone through several rounds of stress and restructuring and now operates in a complex external environment shaped by the slowdown of major trading partners, energy shocks and geopolitical tensions. Moderate but positive growth, with occasional oscillations, suggests that macroeconomic policies and institutional frameworks are strong enough to absorb shocks, but that further progress depends on accelerating structural reforms and strengthening innovative capacity.

From the perspective of future analysis, the same toolkit, logarithmic transformations, differencing, seasonal adjustment with STL and a rich set of graphical diagnostics, can be applied to other key Slovenian series such as industrial production, employment, wages, inflation or tourism indicators. This opens the way to constructing a broader “atlas” of time series that would systematically document how different segments of the economy move through the same crises and recoveries. Such an approach does not offer a crystal ball for forecasting the future, but it does provide a solid empirical basis for discussing where the Slovenian economy stands today, how it got here and what the key risks and potentials are in the years ahead.

References

Cleveland, R. B., Cleveland, W. S., McRae, J. E., & Terpenning, I. (1990). STL: A seasonal-trend decomposition procedure based on Loess. Journal of Official Statistics, 6(1), 3–73. https://www.math.unm.edu/~lil/Stat581/STL.pdf

Eurostat. (2025). National accounts and GDP – Statistics explained. European Commission. https://ec.europa.eu/eurostat/statistics-explained/index.php/National_accounts_and_GDP

Eurostat. (2024). Gross domestic product (GDP) and main components (output, expenditure and income), quarterly – Chain linked volumes, index 2020 = 100 (namq_10_gdp). Retrieved from Eurostat database.

Hyndman, R. J., & Athanasopoulos, G. (2025). Forecasting: Principles and practice (3rd ed.). OTexts. https://otexts.com/fpp3