Reading Time: 15 minutes

1. Introduction

Real gross domestic product (real GDP) is the broadest single measure of economic activity. It summarises how much value an economy produces, adjusted for inflation, and thus captures the evolution of living standards, productive capacity and cyclical conditions over time. For a small, open, post-transition economy such as North Macedonia, real GDP is also a convenient lens through which to view the interaction of domestic reforms, external shocks and the long process of European integration.

The quarterly real GDP series for North Macedonia from 2000Q1 to 2025Q2, measured in chain-linked volumes with 2020=100 as the base year and analysed in logarithms, spans a quarter century of profound change. In the early 2000s, the country was still consolidating its institutions and markets after the dissolution of Yugoslavia and the 2001 security crisis. The mid-2000s brought a phase of stronger growth, supported by foreign direct investment, remittances from the diaspora and gradual integration into European value chains. North Macedonia became an EU candidate in 2005 and later joined NATO in 2020; negotiations on EU accession were formally launched in 2022, symbolising a long-term geopolitical and economic anchoring in Europe.

The graphs summarising the behaviour of real GDP over this period show a pattern familiar from other Western Balkan economies: a broadly rising path interrupted by a sharp contraction during the Global Financial Crisis (GFC) of 2008–2009 and an even more dramatic, but shorter-lived, collapse during the COVID-19 shock in 2020, followed by a rebound and a more moderate, uneven expansion in the early 2020s. Recent European Commission and IMF assessments confirm that real GDP fell by around 4–5% in 2020, recovered by roughly 4–5% in 2021 and then grew more modestly, around 2% per year, in 2022–2024, as the economy grappled with high inflation, energy-price shocks and tighter global financial conditions.

The ten figures accompanying this analysis provide a detailed view of both the level and growth-rate behaviour of real GDP in North Macedonia, before and after seasonal adjustment and under two key transformations: the logarithm of the index, and the first differences of that log series. They reveal not only the broad trend of convergence and interruption, but also the regular seasonal rhythm of the economy, linked to tourism, construction and agriculture, and the irregular shocks that periodically disturb it.

Viewed from the vantage point of 2025Q2, real GDP in North Macedonia stands moderately above its pre-pandemic peak, but the path to that level has been uneven. The growth burst that followed the lifting of COVID-19 restrictions has given way to a slower, more fragile expansion. Seasonal patterns remain clearly visible but appear somewhat better behaved in recent years, and volatility in growth rates, while still present, is lower than during the most turbulent episodes of the past two decades.

2. Description of the time series and components

2.1 Overall patterns in levels and adjusted series

The Figure 1, based on the log-transformed series, compares the original quarterly real GDP index with its seasonally adjusted counterpart. In logarithms, the underlying trend becomes easier to read: a rising line in the early and mid-2000s, a sharp kink during the GFC, a renewed ascent in the 2010s, and then a pronounced “V-shape” around 2020 as the pandemic abruptly interrupts activity and is followed by a swift, if incomplete, recovery.

In the pre-crisis years, roughly from 2002 to 2008, the seasonally adjusted series slopes upwards at a relatively steady pace, indicating robust but not explosive growth. This corresponds to a period when annual real GDP growth frequently exceeded 3–4%, reflecting catch-up dynamics in a still-reforming economy and the gradual integration into EU markets. The GFC leaves a visible imprint: around 2009 the adjusted series flattens and briefly dips, signalling a recession driven by collapsing external demand and tighter financial conditions.

The subsequent decade is characterised by more modest, but generally positive, growth. The seasonally adjusted series climbs, but with stretches of notable plateauing, especially around the euro-area sovereign-debt crisis and during episodes of domestic political uncertainty. Growth remains sufficient to keep the trend pointing upwards, but the slope is less steep than in the pre-2008 years, hinting at limits to the pace of convergence and structural bottlenecks in productivity and investment.

The COVID-19 shock in 2020 is unmistakable. In Figure 1, the original log series shows a steep fall over a single or a couple of quarters, followed by a rebound as restrictions ease and global conditions normalise. The seasonally adjusted series strips out the usual quarterly pattern and presents this as a stark downward spike, underscoring how exceptional the shock was relative to normal fluctuations. Official data confirm that real GDP dropped sharply in mid-2020, with a double-digit year-on-year contraction at the trough, before rebounding strongly in 2021 as travel, tourism and domestic demand partially recovered.

In the most recent part of the sample, from roughly 2022 onwards, the seasonally adjusted line continues to edge upwards but with a noticeably flatter slope. This is consistent with the macroeconomic environment described by recent IMF and European Commission reports: a period of subdued growth in the low-single-digit range, as high inflation, tighter financial conditions and external uncertainty offset some of the positive impulses from investment and exports. The original, unadjusted series continues to oscillate around this trend in a regular pattern, reflecting the seasonal structure of the economy.

2.2 Trend, seasonal and irregular components

The STL decomposition of the log series in Figure 2 separates the observed data into three components: a smooth trend, a recurring seasonal pattern and an irregular residual. The trend component depicts the long-run trajectory of real GDP. In North Macedonia, the trend rises steadily from the early 2000s, indicating sustained gains in real output per quarter. The slope steepens in the mid-2000s, when growth was strongest, then flattens around the GFC and again during the euro-area crisis, before resuming a moderate upward climb. The pandemic appears as a short but deep notch in the trend, reflecting the fact that STL, while smoothing over purely transitory noise, still captures major, persistent shocks as changes in the underlying level. After 2020, the trend resumes its upward path, but at a slightly lower level than the pre-COVID extrapolation would have implied, suggesting a permanent loss of output relative to the pre-pandemic trajectory.

The seasonal component in Figure 2 exhibits a stable, repeating annual pattern. In many years, the first quarter is visibly weaker, reflecting post-holiday slowdowns and the combined effect of winter weather on construction, transport and parts of agriculture. The middle quarters, especially the third, tend to be stronger, consistent with tourism, construction activity and harvest-related output contributing more in the warmer months. Tourism plays a meaningful, if not dominant, role in North Macedonia’s economy; pre-pandemic estimates suggest that travel and tourism accounted for around 6–7% of GDP in 2019, before dropping sharply in 2020 when international and domestic travel collapsed. This helps explain why the seasonal pattern is clearly visible but not as dramatic as in tourism-dependent economies such as Montenegro or Croatia. Over time, the amplitude of the seasonal swings appears reasonably stable, with no strong evidence of either a secular intensification or fading of seasonality.

The irregular component captures what is left after removing trend and seasonal effects. For much of the sample, it fluctuates within a relatively narrow band, behaving like small, short-lived surprises around the predictable pattern. However, there are several episodes where the irregular component spikes notably. The GFC years show a cluster of larger negative and subsequent positive residuals, reflecting the abruptness of the downturn and rebound that cannot be fully explained by gradual trend movements or regular seasonality. The pandemic shock stands out even more: the irregular component displays a very large negative outlier during the initial lockdown phase, followed by unusually strong positive values as the economy reopens. This highlights the value of decomposition: by isolating the irregular component, the analyst can clearly see just how exceptional these episodes were compared with the normal “noise” of quarterly data.

Overall, the main features of the decomposed series indicate an economy that has steadily expanded over the past two and a half decades, but with growth punctuated by global crises and domestic uncertainties, a moderate and quite regular seasonal pattern, and sporadic large shocks, most notably the COVID-19 collapse, that temporarily push real GDP far away from its usual path.

2.3 Seasonal diagnostics and detailed patterns

The seasonal plot and seasonal subseries plot for the log level of real GDP (Figures 3 and 4) provide a more granular view of how the seasonal pattern behaves within and across years. In the seasonal plot, each quarter is traced across the entire sample, revealing that first-quarter values typically lie below the annual average, while third-quarter observations are frequently above it. This reinforces the impression from the decomposition: North Macedonia’s economy tends to start the year relatively slowly and picks up momentum in the middle of the year, when tourism, construction and export-oriented manufacturing are at their busiest.

The seasonal subseries plot adds a second dimension by following each quarter across time. For example, the line for the third quarter shows how Q3 output has evolved from 2000 to 2025. Here, we see that all quarters participate in the long-run rise in real GDP, but with some interesting nuances. In the mid-2000s, all quarters move up together, reflecting broad-based growth. During crisis periods, the drop is evident in every quarter, but the magnitude can differ: the pandemic shock appears particularly severe in the second and third quarters of 2020, when international travel restrictions and domestic lockdowns most directly affected activity. The subseries plots also suggest that the seasonal pattern has remained broadly stable, with no obvious structural break in the ranking of quarters, although the pandemic temporarily compresses differences as all quarters are hit hard.

Lag plots for the log level (Figure 5) shed light on persistence and cyclical dynamics. When each quarter’s value is plotted against its lag (for example, current GDP against GDP one quarter earlier), the points tend to line up along an upward-sloping cloud, indicating that strong quarters are typically followed by relatively strong ones and weak quarters by weaker ones. This positive autocorrelation is characteristic of macroeconomic series, in which shocks propagate gradually through demand, income and expectations. The tighter the cloud along the 45-degree line, the more persistent the process; in North Macedonia’s case, the pattern suggests a meaningful degree of persistence but with enough dispersion to allow for occasional sharp reversals, such as those observed during the GFC and COVID-19.

Figures 6–8 repeat the seasonal and lag-plot diagnostics for the first- and seasonal-differenced log series, that is, for quarter-to-quarter changes that have also been adjusted for recurring annual patterns. Here, the seasonal plot (Figure 6) is much flatter, as differencing adjustment have effectively removed most of the predictable intra-year pattern. The remaining variation reflects genuine growth surprises: quarters in which the economy expands or contracts by more than usual for idiosyncratic reasons.

The seasonal subseries plots for the differenced series (Figure 7) show that, while there are still some quarters that tend to be slightly more volatile, the dominant impression is one of irregular fluctuations around a roughly zero mean.

The lag plots for the differenced series (Figure 8) are more dispersed than for the level series, indicating that growth rates are less persistent than levels. There is still some tendency for positive growth to be followed by positive growth, and for contractions to cluster, but the relationship is weaker. This is consistent with the idea that while the level of output is highly persistent, the deviations from trend and seasonality are more short-lived, driven by shocks that fade relatively quickly.

2.4 Differenced series and decomposition

The first-differenced log series, effectively capturing trend adjusted growth rates, offer a complementary view of North Macedonia’s real GDP dynamics. In Figure 9, the original and seasonally adjusted differenced series oscillate around zero, with most observations falling within a moderate band. Periods of steady expansion show up as clusters of modest positive growth rates, while recessions appear as sequences of negative values.

Several episodes stand out. During the GFC, the differenced series records a run of negative quarters, with the trough showing a particularly pronounced contraction. The COVID-19 shock is even more dramatic: the differenced series plunges to a large negative value in the initial lockdown quarter, illustrating how rapidly activity collapsed. The subsequent rebound appears as a strong positive spike, reflecting the mechanical arithmetic of bouncing back from a low base when restrictions are eased. These swings are far larger than the typical quarterly movements, underlining the exceptional nature of the pandemic episode.

The STL decomposition of the differenced series in Figure 10 confirms that, once the long-run trend has been removed, the residual trend component is close to zero for much of the sample. This is exactly what one would expect from a series of growth rates rather than levels: aside from temporary cycles, there is no strong “trend in the growth rate” itself. What remains is a relatively noisy irregular component, with occasional large spikes during the GFC and COVID-19 and smaller fluctuations in more tranquil periods.

Taken together, the differenced-series figures show that, while North Macedonia’s real GDP growth is subject to sizable shocks, especially during global crises, the underlying process displays a degree of mean reversion: very weak quarters tend to be followed by stronger ones, and vice versa, as the economy adjusts back towards its longer-run path.

3. Economic outlook

The recent segment of the data, covering roughly 2022–2025, suggests that North Macedonia has moved past the acute phase of the pandemic shock but has not returned to the brisk growth rates of the mid-2000s. The level series shows real GDP edging above its pre-COVID peak, while the differenced series indicates that quarterly growth has settled into a modest, uneven pattern. This is consistent with the macroeconomic assessments of international institutions, which project real GDP growth in the low-single-digit range in 2023–2024, supported by investment and exports but constrained by weak external demand, high energy costs and lingering uncertainty.

The seasonal diagnostics continue to show a clear, but not excessive, within-year rhythm: softer first quarters and stronger mid-year quarters. This pattern fits the structure of the economy, where tourism, construction and agriculture make important contributions but do not dominate the overall picture. Studies of tourism in North Macedonia highlight its growing role in GDP and employment, particularly in destinations such as Ohrid and the southwestern lake regions, but they also underscore that the sector’s share of GDP, while significant, remains below that of more heavily tourism-dependent neighbours. This helps explain why seasonal effects are visible yet moderate.

In the broader regional context, North Macedonia’s real GDP dynamics resemble those of other Western Balkan economies: a long-run convergence process punctuated by global downturns and domestic shocks, with a high degree of synchronisation with the euro area and EU business cycles. Empirical work suggests that North Macedonia’s real GDP growth is moderately correlated with that of the EU-27, reflecting trade, financial and remittance links. The GFC, the euro-area crisis and the pandemic all arrive via external channels, and the Figures show that each episode leaves a visible imprint on the series.

At the same time, the decomposition and diagnostic plots hint at some resilience. The trend component, while dented by crises, resumes its upward path after each major shock. The irregular component, though volatile during turbulent periods, does not display persistent high volatility; rather, it spikes and then recedes. The lag plots of the differenced series suggest that growth shocks are not strongly self-perpetuating, which may reflect the stabilising role of policy responses, the cushioning effect of remittances and the flexibility of small, open economies to adjust to changing external conditions.

Looking ahead from mid-2025, the real GDP pattern portrayed in the figures points to a cautious but constructive economic outlook. The recovery from COVID-19 is largely complete in level terms, but the economy has not yet re-established a strong, sustained growth momentum. The combination of incremental EU-accession progress, ongoing investment in infrastructure and tourism, and potential improvements in governance and the business environment could support a gradual strengthening of the trend. At the same time, the series reminds us how exposed North Macedonia remains to external shocks, whether financial, health-related or geopolitical, and how important it is for policymakers to build buffers and enhance resilience.

4. Methodological appendix (short)

4.1 Graphical exploration

The analysis in this blog post rests heavily on graphical methods. Plotting the level of real GDP over time makes it possible to see at a glance where the economy is expanding, stagnating or contracting, and to spot the impact of major shocks such as the GFC and COVID-19. Comparing original and seasonally adjusted series helps disentangle underlying movements from regular intra-year patterns. Decomposition figures further separate the smooth trend from recurring seasonal swings and irregular noise, making it easier to interpret the economic significance of each component. Seasonal plots, subseries plots and lag plots offer additional insight into how the seasonal pattern behaves and how persistent growth or contraction tends to be.

Graphs are powerful, but they also invite over-interpretation. A single outlier can look dramatic yet reflect a temporary measurement issue or one-off shock. Apparent changes in trend may in fact be the result of shifting seasonal patterns, and visually striking movements may not be statistically significant. For this reason, graphical analysis is best seen as a starting point, guiding more formal statistical work rather than replacing it.

4.2 Transformations (logs and differencing)

The real GDP series is analysed in logarithms and, in some graphs, in first and seasonal differences of the log series. Taking logs is a standard step in macroeconomic time-series analysis when the variance of a series increases with its level: the transformation compresses large values more than small ones, helping to stabilise variability and making proportional changes easier to interpret. If () denotes the original index, the log-transformed series is (); differences in () can be interpreted as approximate growth rates.

First differences, defined as (), focus on quarter-to-quarter changes rather than levels and are often used to analyse short-run dynamics and to work with stationary series. Seasonal differences, such as () with () for quarterly data, remove recurring annual patterns by comparing each quarter with the same quarter a year earlier. In this analysis, first differences of the log series is used to highlight growth behaviour net of long-run trend.

4.3 Seasonal adjustment and decomposition using STL

Seasonal-Trend decomposition using Loess (STL) is the main tool used to separate the observed real GDP series into trend, seasonal and irregular components. STL treats the series as the sum of three parts,

where () is a smooth trend (or trend-cycle), () is the seasonal component and () is the remainder. The method uses local regression (Loess) to estimate the trend and seasonal components iteratively, allowing for flexible shapes and slowly evolving patterns.

For quarterly real GDP, STL is particularly attractive because it can handle changes in seasonal patterns over time and is robust to outliers, which is crucial when the series includes large shocks such as the GFC and COVID-19. By working on the log-transformed series and, where relevant, on its differenced versions, STL provides a clear, visually intuitive breakdown of how much of the observed movement reflects long-run growth, how much is due to recurring seasonal forces and how much is left as irregular noise.

5. Conclusion

The quarterly real GDP series for North Macedonia from 2000Q1 to 2025Q2, examined through the lens of logarithmic transformation, seasonal adjustment and STL decomposition, tells a coherent story of gradual convergence interrupted by major external shocks. The level and trend components show an economy that has expanded substantially over the past quarter century, albeit with slower growth in the post-GFC and post-pandemic periods than in the early 2000s. The GFC and, even more starkly, COVID-19 stand out as deep, sudden breaks in the series, followed by rebounds that restore much of the lost ground but leave the long-run trajectory somewhat below its pre-crisis path.

Seasonality is a persistent but manageable feature of North Macedonia’s real GDP. The seasonal component and diagnostic plots reveal a stable pattern in which the first quarter is typically weaker and the middle of the year stronger, in line with the influence of tourism, construction and agriculture. This pattern has remained broadly consistent over time, suggesting that the structure of the economy’s seasonal drivers has not undergone radical change, even as tourism and services have grown in importance. The use of STL to extract and remove this seasonal component clarifies the underlying trend and helps ensure that cyclical analysis is not distorted by predictable intra-year fluctuations.

The irregular component and the behaviour of the differenced series highlight the role of shocks and short-term dynamics. Most of the time, real GDP growth oscillates within a moderate range, with only modest persistence in quarter-to-quarter changes. However, during crisis episodes, the irregular component spikes and growth rates exhibit extreme values, testifying to the vulnerability of a small, open economy to global disruptions. The fact that these spikes are relatively short-lived and that the trend resumes its upward path after each shock points to a degree of resilience, supported by macroeconomic policy responses, external support and the economy’s ability to adapt.

From a policy and analytical perspective, the patterns observed in North Macedonia’s real GDP series underscore the importance of three themes. First, long-run growth remains crucial: sustaining and strengthening the trend requires continued structural reforms, investment in productivity and progress towards EU accession, which can anchor expectations and attract capital. Second, managing seasonality and sectoral dependence, especially in tourism and related activities, remains a key challenge for stabilising employment and income across the year. Third, building resilience to shocks, whether financial, health-related or geopolitical, is essential, as the GFC and COVID-19 episodes demonstrate how quickly external events can push the economy off its usual path.

The decomposition and diagnostic toolkit applied here is not specific to real GDP. The same methods can be fruitfully applied to other macroeconomic series for North Macedonia, industrial production, employment, wages, prices or tourism indicators, to build a richer, more nuanced picture of how different parts of the economy move over time. As the country continues its path towards deeper integration with the European Union and the global economy, such careful, data-driven analysis will remain indispensable for understanding where the economy has been, where it is now and what the recent patterns might imply for the road ahead.

References

Aqifi, E., & Kadriu, A. (2021). Contribution of tourism on GDP in North Macedonia. Knowledge – International Journal, 49(1), 139–143. https://www.researchgate.net/publication/368461831_CONTRIBUTIONOFTOURISMONGDPINNORTHMACEDONIA

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

International Monetary Fund. (2023). Republic of North Macedonia: 2023 Article IV Consultation and First Review under the Precautionary and Liquidity Line—Staff Report. https://www.imf.org/en/publications/cr/issues/2024/01/25/republic-of-north-macedonia-2023-article-iv-consultation-and-first-review-under-the-544112