Data-driven insights into the effects of demographic ageing on Lithuania's labour market sustainability

The phenomenon of population ageing has come to the forefront as a pervasive global concern, exerting its influence on all nations. While advanced economies were the first to embark on this trajectory, experiencing a gradual increase in their older populations over decades, it is now low- and middle-income countries that are witnessing the most rapid and substantial demographic restructuring (Guillemot et al., 2024). A substantial body of scientific literature identifies declining birth and mortality rates as the primary factors propelling this trend (IMF, 2025; André et al., 2024; Pennings, 2022; Nagarajan et al., 2021; Colleran and Snopkowski, 2018; Wodon and de la Brière, 2018). Furthermore, population migration is frequently cited as a contributing factor by injecting younger populations into their workforce and reproductive age groups, but it often does so at the cost of accelerating ageing and increasing dependency ratios in the sending countries (Maestas et al., 2023; Bloom et al., 2015).

In essence, population ageing is not merely a demographic statistic; it is a fundamental transformation of societies worldwide. This shift is creating unprecedented challenges, particularly for labour markets, as the pool of working-age individuals shrinks relative to the growing number of retirees. This necessitates comprehensive planning and adaptation across various sectors—from healthcare and social security to economic policy and urban development—to mitigate potential risks and identify any opportunities that this evolving demographic landscape may present.

This global context underscores the critical need for country-specific analyses to understand the nuanced impacts of population ageing and to inform targeted policy interventions. Lithuania provides a compelling case study of a nation grappling with the accelerated effects of population ageing on its labour market. A confluence of factors, including an exceptionally low total fertility rate (1.18 in 2023), significant advancements in healthcare leading to increased life expectancy and a historical trend of outward migration until 2019, has rapidly increased the median age of the Lithuanian population to 44 years—four years higher than in 2010 (OECD, 2024). This rapid demographic transition is projected to have a profound impact on the country's old-age dependency ratio, which the European Commission (2024) anticipates will surge from 33% in 2022 to a peak of 72% by 2070—the highest level within the European Union. Concurrently, the proportion of the population aged 65 and over is expected to rise from the current 20%–35.6% by 2070. Consequently, Lithuania's labour market faces a significant depletion of its productive workforce, jeopardising its long-term economic development.

Despite the widely acknowledged societal challenge of an ageing population, there remains a notable scarcity of comprehensive, country-level analyses that quantitatively assess how an increased share of the elderly directly reconfigures labour market structures and performance using quantifiable indicators. While some studies have explored the broader economic implications of ageing in Lithuania, the specific and nuanced impacts on labour market dynamics remain underexplored. For instance, Kasnauskiene and Andriuškaitė (2017) found no statistically significant effect of ageing on key Lithuanian economic indicators, yet Pocius et al. (2019) warned that ageing could lead to serious workforce shortages.

Given Lithuania's projected trajectory to experience the highest old-age dependency ratio in the entire European Union, the absence of in-depth, quantitative research specifically on how this extreme demographic shift reconfigures its labour market constitutes a critical research gap.

The primary objective of this study is to systematically examine the impact of population ageing on the Lithuanian labour market. To this end, a rigorous methodology will be employed, utilising a set of carefully selected and quantifiable indicators at the country level. This approach provides a targeted understanding of Lithuania's situation and serves as an applicable case study for other countries exhibiting analogous socio-economic and demographic development patterns. Moreover, this study seeks to furnish policymakers with actionable insights and practical guidance to inform their decision-making processes, thereby enabling them to develop effective strategies to mitigate the adverse effects of population ageing on the labour market. Specifically, these could include customised employment market strategies, modifications to the retirement age in order to ensure the long-term viability of Lithuania's pension system, taking into account its disproportionately high dependency ratio, more precise projections of future healthcare and long-term care requirements and others. By rigorously detailing Lithuania's experience, the research results also aim to contribute significantly to the broader academic discourse on population ageing, offering new empirical evidence and conceptual advancements.

We structure the rest of the article around several sections. The next section details the theoretical foundation of the impact of the population ageing on the labour market. Subsequently, the methodology and empirical framework are developed. This is followed by the presentation of the results. Finally, the discussion, conclusions, limitations and directions for future research are presented.

The phenomenon of demographic ageing exerts a direct influence on the labour market, manifesting in the process of workforce ageing. A close examination of statistical data reveals that population ageing has and will have a significant impact on workforce ageing in the future. As asserted by Mačiulytė-Šniukienė et al. (2019), the labour force represents a pivotal labour market indicator, rendering it imperative to undertake a thorough examination of the influence of ageing on the workforce. The phenomenon of workforce ageing, akin to population ageing, is subject to the influence of demographic factors, including increasing life expectancy, declining birth rates and migration patterns. A decline in birth rates will, by its very nature, have a detrimental effect on the size of the workforce (OECD, 2024). However, immigration can serve to counteract the adverse effects of ageing by contributing to higher birth rates, population growth and economic benefits through tax revenues (Bernstein et al., 2022; Harper, 2016; World Health Organization, 2002). It is imperative to consider the implications of out-migration of working-age individuals in sending countries. This phenomenon is characterised by an acceleration in the ageing process and an increasing representation within the total population. This, in turn, leads to an indirect contribution to the decline in birth rates in the respective origin countries. Workforce ageing is also determined by the concept of active ageing—a process related to maintaining quality of life in old age (World Health Organization, 2002). It is evident that workforce health constitutes a key factor; individuals who maintain their health over an extended period encounter a reduced number of obstacles to their continued participation in the labour market (IMF, 2025). In general, national legislation, including retirement age policies, pension amounts, the overall structure and the effectiveness of social security and healthcare systems, plays an important role in shaping the labour market participation of the elderly population (Eurofound, 2025; Barrela et al., 2024; Cylus and Al Tayara, 2021; Reddy, 2016). Consequently, these elements have the capacity to influence an individual's decision to continue their participation in the labour force.

This study investigates the impact of labour force ageing on the participation of the prime-age group (25–54 years), who constitute the majority of the workforce. While ageing could incentivise greater engagement in this group due to lower participation among older individuals, it might also lead to earlier retirement or increased caregiving responsibilities, creating an opposing effect. Conversely, the ageing process may also result in the decision of some individuals to exit the workforce and enter retirement, thereby exerting a countervailing effect on the labour force participation of the 25–54 age group. Subsequent to a comprehensive examination of the extant literature, the present study seeks to determine whether ageing exerts an influence on engagement in the specified age group. The following hypothesis is thus formulated:

The detrimental effect of ageing on productivity is a subject that has been extensively explored in academic discourse. As posited by Daniele et al. (2019) the phenomenon of demographic ageing exerts a deleterious effect on economic growth, owing to the decline in labour market participation and the decrease in employment levels among the elderly population. This assertion is corroborated by other researchers who have identified a negative relationship between ageing and productivity (Meng and Yu, 2024; Maestas et al., 2023; Kotschy and Bloom, 2023; Park et al., 2021; Poplawski-Ribeiro, 2020; Calvo-Sotomayor et al., 2019; Daniele et al., 2019; Mačiulytė-Šniukienė et al., 2019). Nevertheless, there is evidence that an ageing population can contribute to productivity growth. In contexts characterised by declining fertility rates and an ageing population, an increase in female labour force participation has been demonstrated to directly expand the available workforce. This expansion serves to mitigate the economic ramifications of ageing by contributing to tax bases and supporting pension systems. Furthermore, it serves to offset potential labour supply deficits, thereby enhancing productivity (Lee et al., 2020). Börsch-Supan (2013) posits that productivity undergoes minimal changes across the lifespan, attributing this to the superior management and social skills exhibited by older workers. The experience and knowledge accumulated over many years in a given field by older workers often places them at an advantage over their younger colleagues. In addition to this, some authors have proposed that the adverse effect may be mitigated by transitioning the economy towards knowledge- and capital-intensive sectors and the integration of automation, including artificial intelligence that compensates for worker shortages (Filippucci et al., 2024; Acemoglu and Restrepo, 2022; Tan et al., 2022).

A significant factor determining whether the process of ageing exerts a beneficial or detrimental influence on the labour market is human capital. A productive workforce can offset its effects on economic growth, so investing in human capital is a crucial strategy. Sodirjonov (2020) also says education is important for human capital because it improves health and productivity, which helps labour market stability. Nevertheless, educational initiatives need state funding, which is limited and affected by population ageing.

It is imperative to acknowledge the intricate nature of the impact of ageing on labour productivity. This phenomenon exhibits considerable variability across various sectors, geographical regions and job categories. For instance, industries relying on physical labour may experience negative impacts from an ageing workforce, while sectors valuing experience and cognitive skills might see stable or increased productivity (IMF, 2016). The variation in outcomes can also be attributed to national differences in industrial structure, policy responses, health and education levels (European Commission, 2024). Furthermore, research indicates that manual labour jobs often exhibit a decline in productivity with age, while non-manual, cognitively demanding roles frequently demonstrate more stable or improving productivity (André et al., 2024; van Ours and Stoeldraijer, 2010).

It is broadly acknowledged that labour productivity plays a pivotal role in the promotion of economic growth. Nevertheless, the influence of the ageing population on this indicator may exert a detrimental effect, necessitating careful consideration of its implications within the broader economic landscape. Conversely, other authors posit that older workers possess a distinct set of knowledge, experience and skills that contribute to their labour productivity. Consequently, the specific effect of ageing on a nation's labour productivity remains uncertain.

A comprehensive review of the literature yielded the following hypothesis:

The phenomenon of population ageing exerts a considerable influence on levels of employment, not solely through the numerical magnitude of retirees, but also through the dynamic challenges encountered by older individuals within the workforce. While older workers tend to exhibit lower rates of unemployment than their younger counterparts, their vulnerability is acutely exposed should they become jobless; they face considerably greater difficulty in securing new employment (Chan and Huff Stevens, 2001). This challenge is exacerbated by the continuous development of technology and profound shifts in the demographic structure of society, which often necessitate rapid alterations in job roles and qualification requirements. It has been demonstrated that older workers often encounter substantial obstacles when attempting to adapt to the evolving demands of their occupations. This makes it significantly more challenging for them to transition into new roles or re-enter the labour market once they have been displaced (Serban, 2012). This assertion is corroborated by the findings of other researchers, including Akanni and Čepar (2015), who have documented a positive correlation between the ageing population and unemployment rates among older cohorts. This decline in adaptability and re-employability, therefore, contributes to an increased probability of older workers exiting the workforce entirely, whether through early retirement or discouraged worker effects (Dixon, 2003).

Beyond the challenges of re-employment, the process of ageing is directly associated with a growing proportion of the population reaching traditional retirement age and, consequently, exiting the labour market. The decision to enter retirement is a complex, multifaceted process that is shaped by a dynamic interplay of different factors. Individuals often exit the labour market when they possess sufficient financial resources to maintain their desired standard of living without earned income. This includes public and private pensions, as well as personal savings and investments. The main “push” factor is declining physical or mental health or the onset of chronic conditions or disability, which makes continuing to work difficult, painful or impossible. Low job satisfaction, inability of flexible work arrangements, a need to provide care for family members and the desire to have more free time for hobbies, travel, volunteering or simply relaxation are the key factors that can contribute to an individual's transition into retirement. This phenomenon gives rise to a secondary, yet closely related, challenge: the impending shortage of workers (Flodberg and Österberg, 2025; Scott, 2023; Lisenkova et al., 2010). Furthermore, even for those who maintain employment, older workers demonstrate diminished adaptability in transitioning between employment opportunities in comparison to their younger counterparts. This renders them more vulnerable to job downgrading and reduced working hours (Tan et al., 2022; Visser et al., 2018).

The combination of these factors can perpetuate a perception that ageing is associated with a decline in innovation and business dynamism (Hopenhayn et al., 2022). These structural shifts in employment and participation, driven by both retirement and re-employment difficulties, contribute directly to economic slowdown. The retirement of a growing number of individuals, coupled with the difficulties faced by older displaced workers, leaves the working-age population with an increasingly heavy responsibility of supporting the rest of society. Moreover, it is imperative to acknowledge that the challenges associated with an ageing population may over time necessitate an extension of the retirement age beyond 65. There is a broad consensus that enhancing labour market participation and increasing the productivity of workers are fundamental strategies for mitigating the adverse economic consequences of ageing.

The employment rate is widely regarded as one of the most fundamental labour market indicators with a significant impact on the overall economy. Fluctuations in the employment levels of individuals within the 15–24 age bracket may also be associated with an ageing population. Older workers have been shown to have a lower likelihood of securing new employment following the loss of a position and are more likely to exit the labour market (Dixon, 2003). Consequently, employment growth among younger individuals may be observed. However, if older workers remain in the labour force for longer, the impact on youth employment could be negative. The third hypothesis is thus formulated as follows:

The majority of scientific articles that assess the impact of ageing on the labour market most commonly utilise panel regression methods when analysing multiple countries (Maestas et al., 2023; Park et al., 2021; Lee et al., 2020; Calvo-Sotomayor et al., 2019; Tang and MacLeod, 2006). However, when evaluating a single country, methodologies such as trend analysis, decomposition methods (Mačiulytė-Šniukienė et al., 2019; Lisenkova et al., 2010) and various time series models (Maity and Sinha, 2021; Monteiro and Simões, 2019; Frini and Ben Jedidia, 2018; Kasnauskiené and Andriuškaitė, 2017) are more commonly used. These include vector autoregression (VAR), the error correction model and ARIMAX.

This study uses VAR methodology to assess the impact of ageing on the Lithuanian labour market. This choice is driven by several critical analytical considerations that align precisely with the nature of our research questions and the characteristics of the data. A core strength of the VAR model is its capacity to treat all variables as endogenous, meaning it accounts for the mutual feedback and dynamic relationships among them. This is crucial for capturing the reality of an evolving labour market under demographic pressure, unlike simpler models that might impose restrictive exogeneity assumptions. Given that both demographic ageing and labour market indicators are inherently time-series phenomena, VAR is a well-established and robust econometric technique for analyzing their co-movement and causality over time. Its ability to handle multiple interdependent time series makes it a superior choice for our comprehensive assessment of the Lithuanian case. It identifies the relationship between ageing and the Lithuanian labour market variables and extends the analytical scope using tools such as impulse response functions and forecast error variance decomposition (FEVD). However, structural VAR modelling is not used due to a lack of theoretical assumptions regarding the relationships between variables that would allow for restricting the VAR model.

The State Data Agency (2024)has provided data on selected variables for the period from the first quarter of 2002 to the third quarter of 2024. The statistical software RStudio is utilised for the purpose of data analysis.

In accordance with the reviewed literature and to capture key aspects of ageing and labour market dynamics, the following seasonally adjusted indicators are examined in the study:

1. Participation rate (25–54) – the first difference of the labour force participation rate among individuals aged 25–54;

2. Labour Force Share (55–64) – the first difference of the share of the labour force aged 55–64

3. Labour Productivity – the first difference of the log-transformed labour productivity, expressed as GDP per employed person;

4. Employment Rate (15–24) – the first difference of the employment rate among individuals aged 15–24.

The baseline VAR model equation with the variables used in this study can be written as:

where $xt$​ =  $[Participation Rate(25–54)tLabour Force Share (55–64)tLabour ProductivitytEmployment Rate (15–24)t]$⁠, $α$ – $4×1$ vector of model constants,

$Ai$​ – $4×4$ coefficient matrices,

and $ϵt$ – $4×1$ vector of model residuals (Lütkepohl, 2005).

The labour force participation and employment rate indicators are segmented into specific age categories in order to reflect the impact of ageing on different labour market age groups, while avoiding multicollinearity issues arising from the structural similarity of these indicators.

The proportion of the labour force aged 55–64 has been selected as a metric for labour market ageing. This group is of particular relevance as it includes individuals nearing retirement age, thereby enabling a precise assessment of the impact of ageing on the labour market (Mačiulytė-Šniukienė et al., 2019; Temple and McDonald, 2017). In addition, this indicator is utilised to assess the impact of ageing on productivity, as outlined by Milanez (2020). A similar age threshold is applied in other studies analysing ageing through the employment of 55–64-year-olds (Poplawski-Ribeiro, 2020; Cristea et al., 2020).

The research process is delineated through a series of sequential steps, each employing specific tools vital for the validity and interpretation of our VAR model:

1. Graphical data analysis: employed for the purpose of evaluating the dynamics of the variables over the study period. This process enables a crucial visual inspection of time series properties, such as seasonality, long-term trends and volatility. It is an indispensable part of exploratory data analysis (EDA), informing subsequent statistical tests and model specification (Chatfield, 2003).

2. Stationarity tests: Stationarity is a fundamental assumption for time series analysis, particularly for VAR models. Non-stationary time series can lead to spurious regressions, where variables appear to be related even if they are not, yielding unreliable inference (Granger and Newbold, 1974). The augmented Dickey–Fuller (ADF) (Dickey and Fuller, 1979) and Zivot–Andrews (Zivot and Andrews, 1992) tests are employed to formally assess the stationarity of the variables. The ADF test checks for the presence of a unit root, while the Zivot–Andrews test is particularly important as it allows for an endogenous structural break in the series, which is crucial given the economic shocks experienced during the study period.

3. Cointegration analysis: the Johansen procedure is utilised to evaluate the cointegration of variables, thereby identifying long-term relationships between the examined indicators (Johansen, 1988). Cointegration implies that even if individual time series are non-stationary, a linear combination of them is stationary, indicating a stable long-term equilibrium relationship.

4. Model specification: in the event of cointegration being present, a vector error correction model (VECM) is constructed for the purpose of assessing both long-term and short-term relationships between variables. In instances where cointegration is not detected, a VAR model is employed to analyse the interaction between variables without making the assumption of a long-term relationship. Additionally, Granger causality tests (Granger, 1969) are conducted to determine the predictive significance of variables. Furthermore, impulse response functions (IRF) (Lütkepohl, 2005) are utilised to assess the impact of a shock in one variable on others, while FEVD (Lütkepohl, 2005) is employed to quantify the contribution of each variable to the forecast of the target variable. These tools move beyond simple correlation to illuminate dynamic interdependencies, crucial for policy relevance.

5. Model diagnostics: This final stage involves evaluating the estimated model's stability and testing the assumptions regarding the error terms. This encompasses the implementation of checks for autocorrelation (testing whether residuals are correlated over time), homoscedasticity (checking if the variance of the residuals is constant across observations) and the normality of residuals. Satisfying these diagnostic assumptions ensures the reliability and efficiency of the parameter estimates and the validity of statistical inference (Brooks, 2019; Lütkepohl, 2005).

In accordance with the research process plan, the study commenced with an initial analysis of the variables. The graphical assessment of the variables enabled the conclusion to be drawn that the stationarity of the variables should be assessed by statistical tests. The implementation of the Andrew-Zivot and ADF tests resulted in the revelation that the variables are stationary only in the first difference forms. Subsequently, five dummy variables were incorporated in order to assess the impact of a level shift following the financial crisis and additive outliers during the pandemic, the financial crisis and periods of significant labour force activity. The next step was to evaluate the co-integration of the variables. However, no statistically significant relationship was identified in the constructed VECM model. Consequently, a reversion to the VAR model was necessitated. The final results of the VAR model are presented in Table 1.

As demonstrated in Table 1, ageing has little to no impact on the variables in the equation. The assessment of individual model coefficients does not provide as much benefit as the analysis of the results of the analytical tools of the VAR model. However, it is evident that exogenous variables have contributed to enhancing the accuracy of the models. While these variables may not be statistically significant across all models, their incorporation facilitates a comprehensive evaluation of the outliers manifested in the variable graphs. Notably, each outlier is deemed statistically significant in at least one model. The period preceding the financial crisis was statistically insignificant with respect to the variable of the share of the labour force aged 55–64, which reflects the ageing trend. This indicates that this period did not differ from the others. With regard to the impact on other variables, the period preceding the crisis was characterised by a positive impact on labour productivity and a negative impact on labour force activity among the 25–54 age group and on the employment rate among the 15–24 age group. This suggests that labour productivity grew at a faster rate prior to the crisis, and the negative effect can be attributed to higher labour force participation and employment growth among these age groups in the post-crisis period. The remaining variables encompass single-period periods, the coefficients of which elucidate the shocks experienced during the financial crisis and the emergence of the novel coronavirus.

The Granger causality test is first evaluated using the constructed VAR model, which will demonstrate whether the variables enhance the accuracy of the forecasts of the remaining variables in the VAR system (Table 2).

The p-values for each variable are too large to reject the null hypothesis of the Granger test. The largest p-value is observed in the 55–64 age group within the labour force, indicating that fluctuations in the demographic composition of the labour force, associated with the ageing process, may not be adequately reflected in alternative labour market indicators in the near term. Consequently, the efficacy of this indicator in predicting the remaining variables is questionable. Conversely, other indicators, such as the labour force participation rate among 25–54-year-olds and the employment rate among 15–24-year-olds, exhibited p-values of approximately 0.2. This finding suggests some possible relationship with other variables, but this relationship is not strong enough to be considered statistically significant. The inability of labour productivity to predict other indicators may be attributable to the fact that changes in labour productivity in the short term are unrelated to changes in employment among the 15–24-year-old group or the 55–64-year-old labour force. In summary, the results of the Granger test suggest that all variables in the VAR system are not suitable for predicting the indicators of the remaining system equations.

The impulse response function (IRF) plot provides information about how the variables in the VAR system respond to a shock in ageing. The results of this analysis are presented in Figure 1.

The impulse response function plot demonstrates that the response of the labour force share aged 55–64 to labour productivity in the initial periods exhibits a weak positive effect, which becomes negative after four periods. The effect is also not long-lasting and approaches zero in ten periods, which means that the function stabilises quickly. However, the 95% significance intervals surrounding the function line suggest that the results do not differ significantly from zero. The response of labour force activity among individuals aged 25–54 to the impulse of the labour force share aged 55–64 exhibits fluctuations ranging from negative to positive. It is observed that the response weakens in subsequent periods, yet it remains analogous to its lags in preceding periods, ultimately converging to zero over the span of ten periods. The fourth period is statistically significant, indicating a temporary and delayed response. However, the overall response is deemed to be insignificant. The final IRF function demonstrates that the unemployment rate's impulse response is initially negative, fluctuates and eventually approaches zero. These outcomes suggest that the system under scrutiny is stable. The impulse response function and the Granger causality test yielded analogous results, with the former failing to identify a substantial contribution of the ageing indicator to the predictions of other variables. Furthermore, the values of the IRF are minimal, reaching only decimal percentages, which indicates an extremely limited response.

Finally, the FEVD method was employed to assess the contribution of various factors to the forecast variance (see Figure 2).

It demonstrated that the errors of each indicator under study are most affected by the indicator itself, which means that these indicators have the greatest impact on the forecast errors. However, it was found that the labour force participation rate among people aged 25–54 was the main factor determining forecast inaccuracies, demonstrating an exogenous nature and low dependence on other system variables. Meanwhile, the labour productivity indicator is more endogenous, as its error variance is also more influenced by the activity rate. The employment rate among individuals aged 15–24 is notable for its heightened sensitivity to system variables, as evidenced by the increasing influence of other indicators on its errors during the observed period.

Overall, the consistent findings from the Granger causality tests and our VAR model, further supported by complementary analytical tools, indicate a statistically insignificant short-term causal relationship between the measured ageing variable and key labour market indicators in Lithuania. From an economic perspective, this suggests that, contrary to our initial hypotheses, the direct demographic shift of ageing, as captured by our chosen variable, is not currently acting as a significant independent driver of immediate changes in labour market dynamics in Lithuania. This leads to the rejection of all three of our study's hypotheses, implying that other economic or structural factors may be more dominant in shaping the short-run evolution of the Lithuanian labour market.

In order to provide a rationale for the observed absence of significant relationships, it is first necessary to recall the salient points from the theoretical section of this article – there is no consensus on the nature of this impact – whether it is positive or negative. The results of the Lithuanian case study are broadly in line with some previous findings in this field. For instance, Acemoglu and Restrepo (2022) – focusing on population rather than workforce ageing – find no negative effect of population ageing on GDP per capita. The findings of this study indicate that the youth employment rate in Lithuania is predominantly influenced by structural and economic factors within the labour market, rather than by population ageing. This is consistent with broader research indicating that factors such as skills mismatches, the alignment of educational systems with market needs, economic growth and policies promoting entrepreneurship are critical drivers of youth employment outcomes, often outweighing direct demographic shifts (World Bank, 2025; ILO, 2022). The negligible impact of ageing on the labour force participation of individuals between the ages of 25 and 54 could be attributable to this age group's professional maturity and active participation in the labour market. This demographic group, typically in their prime working years, exhibits high labour force attachment due to established careers, familial responsibilities and fewer immediate incentives for retirement. Participation rates in the labour force are more sensitive to business cycles, educational attainment and evolving social norms than to population ageing itself (IMF, 2018).

In contrast, fluctuations in the proportion of the labour force aged 55–64 do not exert a significant influence on the indicator. This finding warrants further exploration but could potentially be explained by a complex interplay of factors within this age cohort. While retirement incentives and health status are commonly cited influences on older worker participation (Eurofound, 2025; IMF, 2025), the Lithuanian context might exhibit strong labour retention policies, a prevalence of less physically demanding service sector jobs or a relatively high proportion of individuals in this age group choosing to remain employed for financial stability or personal fulfilment, thus counteracting any potential negative impact of increasing age within this bracket (Konle-Seidl, 2018). The negligible impact of ageing on the labour force participation of 25–54-year-olds could be attributable to this age group's professional maturity and active participation in the labour market. Furthermore, this age group does not face significant risks of early retirement, as most individuals in this age group have not yet reached retirement age.

A distinctive feature of this study is the assessment of ageing through a labour force ageing in the context of limited data by applying a VAR methodology. This facilitates the utilisation of a reliable quarterly indicator, in contrast to the annual demographic indicators frequently employed in other studies. The findings of this study align with previous research, underscoring that ageing is a long-term and continuous process, raising doubts about the short-term relationship between the variables under consideration.

The present analysis is therefore intended to provide a valuable addition to the existing corpus of research by way of a more meticulous examination of the statistical significance of the relationship between the two factors. While the analysis of quarterly data allows for a larger sample size, certain external factors, such as the financial crisis or the SARS-CoV-2 pandemic, have further complicated the assessment of the impact.

Moreover, this study makes a significant contribution to the existing literature by underscoring the utility of high-frequency labour market data in comprehending demographic shifts, even in circumstances where comprehensive demographic statistics are limited. This approach provides a foundation for analogous analyses in settings with limited data availability, thus expanding beyond the utilisation of conventional, lower-frequency demographic indicators frequently employed in ageing studies (Bloom et al., 2015; Lee and Mason, 2010). By directing attention toward the phenomenon of labour force ageing, the present study offers a more immediate and economically relevant perspective on the ongoing demographic transition, thereby aligning with recent calls for more granular analyses of population dynamics and their economic consequences (Maestas et al., 2023; Cutler et al., 2022).

In view of the manifest trends in ageing, the Lithuanian government is strongly recommended to employ a range of proactive policy measures to mitigate the consequences of ageing. In order to achieve this objective, the following measures are recommended: firstly, an adjustment of pension schemes should be considered, with a view to allowing individuals to extend their participation in the labour market; secondly, efforts should be made to integrate both younger and older people into the labour market by promoting flexible working conditions; thirdly, there should be investment in human capital through the improvement of the education system, promoting employability and skills of older workers; fourthly, to improve productivity, the government should increase investment in research and development of new automation technologies.

The impact of ageing can also be mitigated by promoting fertility and formulating a favourable immigration policy. Concurrently, the state should provide incentives for families desiring to have children, thereby reducing their financial strain through various forms of compensation. While acknowledging the limitations of immigration as a panacea, it is recognised that it plays a crucial role in supplementing the labour market with specialists in areas of greatest need (André et al., 2024).

Another important measure is the policy of active ageing, which includes the participation of older people in society, ensuring their health, well-being and safety, enhancing the efficiency of health care provision and expanding the coverage of social security systems. Addressing age discrimination and modifying societal attitudes towards the elderly is also crucial, with a focus on fostering their involvement in public life. The implementation of these strategies, encompassing both direct labour market assistance and measures to mitigate the impact of ageing, is poised to foster a more sustainable Lithuanian labour market.

It is important to note that the modelling process was determined by several key limitations:

Firstly, the study sample encompasses periods of considerable uncertainty and economic shocks within the time series (e.g., the 2008–2009 global financial crisis, the COVID-19 pandemic and the ongoing geopolitical instability since 2022 due to the war in Ukraine), thereby rendering the estimation process more arduous and potentially influencing variable dynamics. While the Zivot–Andrews test helps account for one structural break, multiple or more complex breaks could impact results.

Secondly, the utilisation of quarterly data, as opposed to longer-period annual data, is employed to augment the sample size, though this may be less reliable. However, quarterly data may sometimes be more prone to short-term noise or measurement issues compared to annual aggregates, potentially making long-term trends less smooth.

Thirdly, it should be noted that the Granger causality test solely reflects the predictive capability of variables; it does not provide an accurate representation of the true causal relationships between them (Granger, 1969).

Also, the possibility of omitted variable bias – the exclusion of important predictors – may lead to biased results (Wooldridge, 2013).

In view of these limitations, the results of the study are to be approached with a degree of caution and subject to critical interpretation, recognising that they offer insights within the defined methodological framework and data constraints. Nevertheless, it is imperative that subsequent research take these constraints into consideration.

The relatively brief study period may be a contributing factor to the inability to identify significant relationships between the analysed indicators. Despite the utilisation of quarterly data, the assessment of the correlation between ageing and the labour market over short periods may be more difficult than the analysis of annual or even longer data aggregates. Given the gradual and long-term nature of ageing, the employment market's response to it can be more accurately gauged by analysing long-term data. Furthermore, some authors have utilised methodologies that facilitate cross-country or cross-region comparisons, thereby expanding the analytical database and enhancing the reliability of the results. In the future, as more detailed statistical information on the interaction of ageing and the Lithuanian labour market is accumulated, it will be necessary to further analyse the impact of this demographic phenomenon on the national labour market, using the latest data and advanced statistical methods.