We consider two forms of the individual’s utility function. One version assumes that labor supply is elastic, while the other suggests it is inelastic. Furthermore, for both forms, the potential amount of “effective” labor that an individual can supply varies with their labor productivity, which is determined by their age and sex. In the case of elastic labor, an individual of gender s ∈ {f, m} has preferences given by the utility function u(c_t,j,s, n_t,j,s):

$u(c_{t,j,s},n_{t,j,s})=\log c_{t,j,s}-\frac{v}{1+\psi }{n}_{t,j,s}^{1+\psi }$, (1)

where c_t,j,s is consumption of an individual of gender s at age j in period t; n_t,j,s is labor hours of an individual of gender s at age j in period t; ψ is the inverse of the Frisch elasticity of labor supply; υ is the normalization parameter.

In the dynamic OLG model, the individual’s problem is solved in two steps. First, each individual identified by sex s, cohort birth year t₀, and age j = 1, ..., J at calendar year t = t₀ + j – 1 chooses consumption (c_t,j,s), labor supply (n_t,j,s) and savings for each actual retirement age π to maximize expected lifetime utility U^π_t0,s:

$\begin{array}{rl}{U}_{t_{0,s}}^{\pi }& =\sum _{j=1}^{\pi -1}{P}_{t_0+j-1,j,s}{\beta }^{j-1}[\log \left(c_{t_0+j-1,j,s}\right)-\frac{v}{1+\psi }{n}_{t_0+j-1,j,s}^{1+\psi }]+\\ & +\sum _{j=\pi }^J{P}_{t_0+j-1,j,s}{\beta }^{j-1}\log \left(c_{t_0+j-1,j,s}\right)\end{array}$, (2)

where β is the discount factor; P_t,j,s is the population size of gender s of age-j agents born at time t₀; π is the endogenously chosen actual retirement age.

The i n dividual supplies labor and receives wage income starting from labor market entry up to the statutory retirement age, at which point a person begins receiving pension payments. After reaching the statutory retirement age, the individual may continue working, simultaneously receiving both wages and pension benefits. Upon reaching the actual retirement age π, the individual ceases working and receives only pension income. Throughout both stages of life — before and after actual retirement — the individual also receives lump-sum transfers from the government and interest income from accumulated assets. Thus, the budget constraint takes the form:

$\{\begin{array}{l}{a}_{t+1,j+1,s}+(1+{\tau }_t^{VA})p_t^c{c}_{t,j,s}{P}_{t,j,s}=\\ =a_{t,j,s}(1+r_t)+(1-{\tau }_t^{\ast })w_t{Z}_{t,j,s}{n}_{t,j,s}{P}_{t,j,s}+t{r}_t{P}_{t,j,s}+b_{t,j,s}{P}_{t,j,s},j<\pi ,\\ a_{t+1,j+1,s}+(1+{\tau }_t^{VA})p_t^c{c}_{t,j,s}{P}_{t,j,s}=a_{t,j,s}(1+r_t)+t{r}_t{P}_{t,j,s}+b_{t,j,s}{P}_{t,j,s},j\ge \pi ,\end{array}$ (3)

where τ_t^VA is the value added tax rate in period t; p_t^c is the price of consumer goods at period t; a_t,j,s is assets of age-j agents of sex s in period t; rt is the world interest rate; τ_t^w is the labor income tax rate; Z_t,j,s is labor productivity of an individual of sex s from generation j in period t; w_t is wage at period t; tr_t is the net lump-sum transfer received by an individual from the government; b_t,j,s is the pension payment (= 0 if the individual has not met the requirements for a pension; > 0 otherwise). Moreover, the individual may decide to leave the labor market before reaching the statutory retirement age, in which case the individual will not receive pension payments until reaching the statutory retirement age.

Second, decisions on the optimal retirement age are made by evaluating the consumption equivalent variation (CEV), which measures the utility gain from retiring in the current versus the following period (4):

$CE{V}_{t_{0,s}}^{\pi }=\exp \left(\frac{U_{t_{0,s}}^{\pi +1}-U_{t_{0,s}}^{\pi }}{\sum _{j=1}^J{\beta }^{j-1}\cdot P_{t_0+j-1,j,s}}\right)$, (4)

where and are expected utilities with immediate labor market exit and next period labor market exit.

The model specifies a heterogeneous agent structure: each cohort is charac­terized by a distribution of threshold parameters λ reflecting differences in preferences. For each age, we calculate the share of individuals whose threshold λ^* exceeds the corresponding CEV, and therefore opts for actual retirement at that age. The cumulative distribution of retirements is described by a monotonic function G_j,t, with increments G_j,t – G_{j –1,t} representing the fraction retiring at each age in period t. This specification yields a flexible age profile of actual retirement that responds dynamically to economic, policy, and demographic shifts.

Pension payments are determined in accordance with Russian legislation. The minimum insurance period required for entitlement is at least 15 years. According to the Social Fund of Russia,² the number of people who postponed claiming an old-age insurance pension was negligible: 30.5 thousand persons, while 36.702 million­ persons were receiving an old-age insurance pension in the same period; therefore, in the model we assume that individuals start receiving a pension upon attaining the formal retirement age without postponement. The pension in Russia consists of two parts: an individual component and a fixed component. The individual component is calculated based on the amount of insurance contributions paid and the length of the employment record, while the fixed component represents a base payment and accounts for approximately one third of the pension benefit. However, given that we only model productivity heterogeneity by sex and age and consider low- or inelastic labor supply, individual pension differences are limited, and we therefore approximate pensions as equal across agents and starting at the formal retirement age. The pension benefit b_t,j,s in the model is set to match the observed aggregate replacement ratio in 2018–2024; for 2025–2026 we allow for increase toward 30% reflecting the restoration of indexation for working pensioners and the two-step indexation mechanism (indexation to inflation followed by an additional indexation conditional on Social Fund of Russia resources).

In the case of inelastic labor supply, the utility function depends only on consumption:

u(c_t,j) = logc_t,j, (5)

and individuals supply a fixed amount of labor until the actual retirement age, which is determined endogenously based on evaluating the consumption equivalent variation (4).

The Russian economy is modeled as a small open economy implying exogeneity­ of the world interest rate and prices of traded goods. The economy consists of two sectors S: domestic-oriented (N) and export-oriented (E). In addition, we consider the oil and gas sector, which exogenously produces output O_t at time t, which is sold on the global market at a price of p_t^O. The government receives a share of the profits of the oil companies τ_O,t.

Private final demand, which consists of household consumption and ­investment, is composed of domestically produced goods of the domestic-market-oriented sector and imported goods, as follows:

$D_t={\left(\frac{M_t}{\omega }\right)}^{\omega }{\left(\frac{Y_t^N-G{P}_t}{1-\omega }\right)}^{1-\omega }$, (6)

where D_t is private demand for final goods; M_t is imports; Y_t^N is output of the domestically oriented sector; GP_t is government purchases (assumed to be used only to purchase goods of the domestically oriented sector), ω is the share of imports in final private consumption.

The price of the final good p_t^c is determined as follows:

p_t^c = (p_t^M )^ω(p_t^N )^{1– ω}, (7)

where p_t^M is the price of imported goods; p_t^N is the price of domestic goods. The demand for domestically produced goods of the domestically oriented sector is:

$Y_t^N=G{P}_t+(1-\omega )\frac{p_t^c{D}_t}{p_t^N}$. (8)

And the demand for imports is given as:

$M_t=\omega \frac{p_t^c{D}_t}{p_t^M}$, (9)

The production function of a firm in each sector S (S ∈ {E, N}) has the Cobb–Douglas form:

Y_t^S = (K_t^S )^α(A_t^S L_t^S)^{1– α}, (10)

where Y_t^S is output in sector S in period t; K_t^S is capital in sector S in period t; A_t^S is productivity in sector S in period t; L_t^S is labor in sector S in period t; α is elasticity of output to capital.

A firm maximizes the present value of its cash flow:

$\begin{array}{rl}\underset{L_t^S,I_t^S,K_{t+1}^S}{max}& \sum _{m=0}^{\infty }\left(\prod _{i=0}^m\frac{1}{1+r_{t+i}}\right)\left[\right(1-{\tau }_{t+m}^{PR})(\left(p_{t+m}^S\right){\left(K_{t+m}^S\right)}^{\alpha }{\left(A_{t+m}^S{L}_{t+m}^S\right)}^{1-\alpha }-\\ -& (1+{\tau }_{t+m}^f)w_{t+m}{L}_{t+m}^S)+{\tau }_{t+m}^{PR}\delta p_{t+m}^c{K}_{t+m}^S-p_{t+m}^c{I}_{t+m}^S],\end{array}$ (11)

where τ_t^PR is the corporate income tax rate; I^S_t is investments in sector S; w_t is wages, δ is the depreciation rate.

Capital accumulation is given as follows:

$K_{t+1}^S=(1-\delta )K_t^S+I_t^S(1-\frac{\phi }{2}{(\frac{I_t^S}{I_{t-1}^S}-1)}^2)$, (12)

where φ is a parameter of investment adjustment costs.

For simplicity, the production function of the oil and gas export sector takes the form of a Leontief function, which allows the output level in this sector to be set exogenously:

Y_t^O = min{φ₁L_t°, φ₂ K_t°}. (13)

The government raises revenue from the labor tax (τ_t^w), social security contributions (τ_t^f), value added tax (τ_t^VA), corporate income tax (τ_t^PR), and oil and gas industry tax (τ_t^O) and spends it on government purchases (GP_t), pensions (B_t) and transfers (TR_t) to the population and the interest on public debt (Debt_t). The govern­ment budget constraint may be written as:

Debt_{t +1} = (1 + r_t )Debt_t + p_t^N GP_t + TR_t + B_t – T_t^VA – T_t^w – T_t^f – T_t^O –

– τ_t^PR (p_t^N Y_t^N – (1 + τ_t^f)w_t L_t^N ) + τ_t^PR δ p_t^I K_t^N –

– τ_t^PR(p_t^E Y_t^E – (1 + τ_t^f)w_t L_t^E ) + τ_t^PR δ p_t^I K_t^E, (14)

where T_t^VA, T_t^w, T_t^f, T_t^O are tax revenues from VAT, labor income tax, social contri­butions, and mineral extraction tax.

Debt cannot grow indefinitely: if the debt-to-GDP ratio deviates from the target level, the government is forced to increase taxes in the next period. In the model, the value-added tax is endogenous; hence, the government’s response function to an increase in the debt-to-GDP ratio can be written as:

${\tau }_t^{VA}=(1-\rho ){\overline{\tau }}^{VA}+\rho {\tau }_{t-1}^{VA}+\xi (\frac{Deb{t}_{t-1}}{GD{P}_{t-1}}-\overline{d})$, (15)

where τ ^{̅ VA} is the target tax rate; ξ is the sensitivity of tax policy to the deviation of debt from the target value; d^̅ is the target debt-to-GDP ratio.

The model is calibrated to 2018, the year the pension reform was announced. The minimum age of entry into the labor market is assumed to be 21; before that, agents are supported by their parents and neither work nor save. An individual’s labor productivity is specified as a second-degree polynomial in age and sex:

log Z_j,s = k_s,2 j² + k_s,1 j + k_s,0, (16)

where j is the age of the agent, the coefficients of the equation are taken from (Zamnius et al., 2023): k_m,2 = –0.00036, k_m,1 = 0.025, k_f,2 = –0.00043, k_f,1 = 0.039. To account for differences in wages between men and women, the coefficients k_m,0, k_f,0 were chosen so that the model reproduces the ratio of average wages of men and women at age 25.

In the model with inelastic labor, labor supply is normalized to 1. In the model with elastic labor the labor supply elasticity is 1/ψ = 0.2. The discount factor β is 0.965.

The distribution of heterogeneous retirement thresholds (λ) across agents is modeled using a lognormal distribution, with parameters chosen so that the bulk of agents’ optimal actual retirement matches observed actual retirement ages — 60.2 for women and 62.4 for men in 2019 (Lyashok and Varshavskaya, 2022).

The global market prices (prices for imported goods p_t^M and prices for exported goods p_t^E) are fixed and equal to 1. The world interest rate r is equal to 0.02. The oil price p_t^O is scaled to 1 at the initial year and then changes exogenously. Three oil price scenarios are based on the U.S. Energy Information Administration forecasts (Fig. 3).³ Historically until 2022 prices for Russian oil (Urals) were on average 2.0% below Brent. However, after 2022, the Urals price was significantly lower than the Brent price. We assume that this negative shock is short-term. In 2024, the Urals price was trading at a discount of 10%. We assume that by 2028, the impact of the negative shock will become negligible and Urals oil will trade at a discount of 2% in the long term.

According to the Federal Customs Service, fuel and energy products accounted­ for 62.1% of Russian exports in 2019 and 49.6% in 2020. Given the new economic conditions that emerged after February 2022, we assume that fuel and energy exports in the model account for half of total exports.

Accordingly, the share of the oil and gas export sector is about 15% of GDP. Parameters φ₁, φ₂ from the production function of the oil and gas sector are selected to target the share of investments in the oil and gas sector — 12% and the share of the oil and gas sector wage fund in the total wage fund — 5% (Martyanova and Polbin, 2023). In our simulations, we will assume a scenario in which the level of oil and gas production does not change, and only energy prices change. That is, output in the export oil and gas sector Y_t^O is fixed at the 2018 level. As a result, in a growing economy, this modeling assumption implies that the share of oil and gas revenues in the government budget will gradually decline over time, reflecting the increasing contribution of non-resource sectors to GDP and government revenues.

The productivity in both sectors of the economy A_t^S is set exo­genously to 1.0 for the initial year. Furthermore, we assume that this parameter, representing labor productivity, grows at a rate of 2%.

The share of imported goods in the production of final consumption goods ω corresponds to the ratio of imports to GDP in the initial year and equals 21% according to Rosstat. The elasticity coefficient of the production function with respect to capital α equals 0.33. The investment adjustment costs parameter φ is 1.2 (Polbin, 2013). The depreciation rate δ is 0.06.

Items of the government budget are calibrated as follows. The average pension replacement rate in 2018 was 0.33. Rates for corporate tax, labor income tax, value added tax, social security contributions are calibrated at the values: τ_t^PR = 20%, τ_t^w = 13%, τ_t^VA = 20%, τ_t^f= 30%. Note that in the model, consumer investment income is not taxed. This simplification is due to the non-linearity of personal income tax calculations, which stems from a non-taxable threshold. The target debt‑to-GDP ratio (d^̅ ) was assumed to be 0.1. The value of the lagged variable ρ is set at 0.8, and the value of the variable ξ is set at 1.2.

In general, the values of the model parameters were selected to reproduce the key economic indicators of Russia in 2018 (Table 1).

To estimate the effectiveness of the reform raising the statutory retirement age, we first discuss the baseline scenario, which assumes a medium change in demographics­ and oil prices. First, we compute the values of the endo­genous variables­ assuming that the statutory retirement age has not been increased. Then, we compute the values of the endogenous variables assuming that the statutory retirement age has increased. Finally, we analyze the differences between these estimates to isolate the effect of increasing the statutory retirement age.

Fig. 4 illustrates the projected dynamics of the proportion of people of statutory retirement age before the reform (left panel) and the budget-balancing value-added tax (VAT) rate (right panel) for three demographic scenarios: median, low fertility, and high fertility.

The left panel shows that, regardless of fertility assumptions, the share of the population at statutory retirement age will increase substantially in the coming decades. Under the medium variant, this fraction rises from approximately 0.34 in 2019 to nearly 0.44 by the late 2040s. This demographic shift entails higher expenditures on pension benefits and intensifies fiscal pressure on the pension system.

The right panel indicates that, as the demographic burden grows, the VAT rate required to balance the government budget must also be risen markedly. Specifically, under the low fertility scenario, the VAT rate approaches 0.27 by 2050, compared with about 0.24 under high fertility. The medium variant requires­ a VAT rate of roughly 0.26 in the same period. The VAT adjustment reflects the need to finance expanding pension outlays and maintaining fiscal sustainability as the ratio of retirees to the working-age population rises.

These results demonstrate that the policy reform enacted in Russia in 2018 — raising the statutory retirement age — can be justified as a response to anticipated demographic challenges. The greatest demographic and fiscal strain is projected to occur around 2050, when pension expenditures peak and the VAT rate must be set at a level of about 0.26 absent further reforms or alternative financing measures.

The seemingly divergent dynamics of the share of people at statutory retirement age and the budget-balancing VAT rate arise from the model’s fiscal block rather than from demographics alone. In the first half of the century, the rapid increase in the share of retirees indeed worsens the pension system’s implicit ­liabilities, but the government partly absorbs this pressure by allowing public debt to grow and by drawing on still relatively high non-pension revenues. As a result, the VAT rate adjusts only moderately during this period.

In the second half of the century, the share of the elderly stabilizes, but by this time two cumulative effects become crucial. First, the stock of public debt will have already approached its target upper bound, which triggers a much stronger tax response in the government’s feedback rule. Second, the tax base will have shrink due to weaker labor force dynamics and declining resource-based revenues, so that maintaining intertemporal budget balance requires a sharper increase in the VAT rate. Thus, the VAT “hike” reflects the delayed fiscal reaction to accu­mulated demographic and debt pressures under the assumed policy rule, rather than a one-to-one contemporaneous link with the age structure in any single year.

To assess the potential impact of raising the statutory retirement age on key economic aggregates, we compare the pension reform scenario with a baseline model with two labor supply specifications: inelastic and elastic. Fig. 5 presents the dynamics of GDP, aggregated labor, consumption, investment, exports, and government purchases in both cases. Here and hereafter, the dynamics of all macroeconomic variables are considered in real terms (at constant prices).

The GDP and aggregated labor plots show that after the reform, both output and total labor input increase, especially when labor supply is elastic. More people remain in the workforce, boosting employment and economic growth. This effect is particularly visible in the years following the reform, when the lines for the elastic labor scenario are consistently above those for the inelastic one.

Consumption declines temporarily following the reform, as older workers lose immediate access to pension benefits, reducing their permanent income. However, in the long run, aggregate consumption recovers and even exceeds baseline levels, especially in the elastic labor scenario, though the effect remains relatively modest, fluctuating between 0 and 0.5%.

Investment and government purchases are higher after the reform, indicating greater economic activity and fiscal stability. The increased labor supply and reduced pension burden allow for stable public spending, even as demographic pressure from an aging population grows. Exports also benefit from increased labor resources, with the elastic labor scenario consistently outperforming the inelastic alternative.

Overall, the main outcome of the pension reform is a more robust and resilient economy in the face of population aging. By encouraging later retirement and higher employment, the reform alleviates fiscal pressure — especially pension expenditures — and supports stable or rising consumption, investment, and public spending. The scenario with elastic labor supply yields the strongest positive effects, showing that labor market flexibility is central to mitigating demographic and fiscal challenges.

The refor m increases the average effective retirement age by about one year for women and by around half a year for men. This moderate shift in the effective retirement age is consistent with existing empirical estimates for OECD countries. A number of studies show that an increase in the statutory retirement age leads only to a partial adjustment of the actual retirement age (Turner and Morgavi, 2021). The reported estimates indicate that, in developed countries, a one-year increase in the statutory retirement age raises the average effective retirement age by approximately 1.1–1.4 months. It should be noted that most of the studies reviewed focused on male labor force participation.

Statistical data and empirical evidence indicate that the effective retirement age in Russia showed substantial heterogeneity even prior to the pension reform. Some workers exited the labor market before reaching the statutory retirement age, while others continued working beyond it. This is illustrated by Rosstat data on employment rates by age group (Fig. 6). For example, prior to the pension reform, in 2018, employment among women aged 50–54 (with the statutory retirement age set at 55 before 2019) stood at approximately 84%, implying that up to 16% of women in this age group were already out of employment. At the same time, employment among men of pre-retirement age stood at 77.7%. The data presented in the figure point to substantial heterogeneity in the effective retirement age and confirm that the formal statutory retirement age did not represent a rigid boundary for labor market participation. It should be noted that the observed decline in labor force participation at pre-retirement ages may be driven not only by individuals’ voluntary decisions to cease working, but also by institutional and demand-side labor market factors. In particular, some workers may exit the labor market as a result of layoffs or job displacement and subsequently face difficulties in finding new employment at older ages. In such cases, the final decision to stop working may be made only after a prolonged period of unsuccessful job search. Distinguishing between voluntary and ­in­voluntary exits from the labor force using aggregated data is challenging and lies beyond the scope of this study; this issue is left for future research.

Additional empirical evidence of substantial heterogeneity in retirement ­behavior is provided by Lyashok and Varshavskaya (2022). The authors find that the average effective retirement age in 2019 was 62.4 years for men and 60.2 years for women, already exceeding the statutory retirement thresholds in force at that time (60 and 55 years, respectively). Over the period from 2010 to 2019, the average retirement age remained virtually unchanged for men and increased by only 0.7 years for women.

It is important to emphasize that, as of 2019, the average retirement age varied substantially across Russian regions: in some regions it was below the statutory retirement age, while in others it was comparable to or exceeded the statutory threshold. Since these figures refer to average values, they also imply the presence of within-group variance in actual retirement ages, toward both earlier and later exits from the labor market. This regional variation confirms strong household heterogeneity even under uniform institutional rules.

Moreover, Lyashok and Varshavskaya (2022) find that a one-percentage-point increase in the pension-to-wage ratio reduces the average effective retirement age by 0.06 years for women and by 0.04 years for men, pointing to asymmetric behavio­ral responses across genders and differences in sensitivity to financial incentives.

Taken together, these findings indicate that, in reality, the decision to exit the ­labor market is not a mechanical consequence of reaching the statutory ­retirement age, but is instead determined by individual characteristics, financial incentives, and pre­ferences. To capture these patterns, the paper employs a model with an endogenous­ retirement age choice and heterogeneous preferences. Within the framework of the model’s assumptions, it reproduces the distribution of retirement decisions observed in the data: for some individuals, an increase in the statutory retirement age leads to a rightward shift of the effective retirement age, while for others, pension benefits, as before the reform, do not constitute a decisive incentive to continue working up to the new statutory threshold. Finally, there exists a group of individuals who had already been retiring later than the previous statutory retirement age before the reform and therefore respond only weakly to the change in institutional rules.

The oil price is normalized to one in the initial year and then evolves exo­genously according to three scenarios based on U.S. Energy Information Administration (EIA) forecasts: baseline (median price), high price, and low price. Historically, the price of Russian Urals oil averaged 2% below Brent prior to 2022, with a temporary shock leading to a 10% discount in 2024. The long-term assumption is stabilization at a 2% discount by 2028.

Fig. 7 presents the trajectories of key macroeconomic aggregates — GDP, aggregate labor, consumption, investment, exports, and government purchases — expressed as the ratio of post-reform outcomes to their respective values in the no-reform scenario. Each series corresponds to a distinct oil price assumption (median, high, and low), thereby isolating the effect of the pension reform under identical external price conditions.

Under all oil price scenarios, the pension reform yields a stable positive impact on GDP and aggregate labor over the simulation horizon. The effect is particularly pronounced under low and median oil prices, highlighting the role of the reform as a stabilizing mechanism for output and labor supply when external resource revenues are under strain. In these cases, the increase in labor supply driven by delayed retirements translates into sustained GDP growth relative to the baseline.

In the scenario with persistently high oil prices, the budget-balancing VAT rate may fall significantly in the near term, as elevated oil revenues reduce the government’s reliance on broad-based consumption taxation for fiscal stability. A lower VAT rate raises the purchasing power of households by reducing the tax-inclusive price of consumer goods, thereby increasing the real return to labor. In our model, this effect is captured through the higher marginal utility of consumption relative to leisure, implying that a decline in the VAT rate raises individuals’ incentives to participate and supply more labor. Consequently, the positive effect of pension reform on aggregate indicators is amplified in the initial years following the reform under high oil prices, as reflected by the stronger response in the corresponding trajectories.

The response of consumption is more muted, with the ratio fluctuating around 0.5% in the long run, regardless of oil prices. Investment and exports exhibit the largest positive deviations from the baseline in scenarios with low or median oil prices, especially during the periods immediately following the reform. This reflects the fiscal relief achieved by postponing pension outlays and the associated reallocation of resources toward capital formation and export-oriented production. Under high oil prices, the relative impact of the reform on these aggregates diminishes, as abundant resource revenues independently support aggregate demand and investment.

Government purchases, a proxy for the public sector’s capacity to provide goods and services, also tend to rise compared to the no-reform scenario, again most noticeably when oil prices are low or median. This result demonstrates the fiscal space created by reduced pension expenditures, allowing for more robust public spending even in less favorable external environments.

Importantly, under no scenario does the pension reform lead to a long-term deterioration in any of the reported macroeconomic aggregates. Rather, its impact is countercyclical: it is most beneficial when fiscal conditions are tight and remains at least neutral when resource revenues are buoyant.

These findings confirm that the pension reform is a genuine and effective stabilization instrument for public finances and overall macroeconomic performance, particularly in the face of adverse external shocks such as prolonged periods of low oil prices. The results underscore the importance of structural demographic measures alongside traditional macro-fiscal policy tools in ensuring the long-term resilience of the Russian economy.

Fig. 8 illustrates the joint effect of the statutory r etirement age reform and changes in oil prices, relative to the baseline scenario. The figure presents the dynamics­ of key macroeconomic indicators under scenarios where both the pension reform and alternative oil price trajectories are implemented, expressed as a ratio to the baseline case with the median oil price and no reform. This allows for assessing the combined impact of the reform and external shocks on the main aggregates of the economy.

Under the high oil price scenario, all major macroeconomic aggregates — GDP, aggregate labor, consumption, investment, and government purchases — display significantly higher values relative to the median scenario. The positive revenue shock from oil stimulates economic activity, increases incomes, and supports stronger public expenditure, resulting in a more robust response to demographic pressures. Importantly, the pension reform amplifies these effects: more people remain in the workforce, and the increase in public revenues helps moderate fiscal adjustment needs. In this scenario, the pronounced decline in real exports is consistent with the symptoms of “Dutch disease,” namely a contraction of the export sector when the economy becomes more dependent on the resource sector.

In contrast, the low oil price scenario places considerable strain on the economy. GDP and aggregate labor in this case remain below the baseline throughout the projection period, particularly in the years following the negative price shock. Consumption, investment, aggregate labor, and government purchases decline. The pension reform remains beneficial, as it boosts labor supply and alleviates pressure on the pension system; however, these positive effects are insufficient to fully offset the impact of persistently low oil prices. Fiscal challenges intensify, and overall economic performance is notably weaker. Real exports, by contrast, increase in the low oil price scenario: lower domestic absorption and a reallocation of resources away from the domestic sector toward non-resource exports lead to higher export volumes.

Notably, the reform’s effectiveness is highly sensitive to the prevailing oil price regime. With higher oil prices, the economy enjoys higher output, investment, and public spending, and the reform’s positive effects are amplified. In a persistently adverse external environment, the reform provides fiscal relief but cannot fully counter the macroeconomic drag from lower oil revenues.

We now consider how the pension reform scenario changes under alternative demographic assumptions. In the high-fertility scenario, higher birth rates increase the size of the working-age population relative to retirees, which alleviates pressure on the pay-as-you-go pension system and reduces the need for additional parametric reforms. By contrast, in the pessimistic low-fertility scenario, persistently low birth rates lead to a contraction of the working-age population and an increase in the old-age dependency ratio, which amplifies fiscal pressure on the pension system and makes the effects of the reform more pronounced.

Thus, under the high-fertility demographic scenario, the need for reform and, accordingly, its effect are less pronounced. Fig. 9 presents macroeconomic indicators for the economy facing the pension reform under alternative demographic scenarios. A sharp divergence between the demographic forecasts affects aggregate macroeconomic indicators, such as labor supply and GDP.

Conversely, the pessimistic low-fertility scenario results in significant demographic contraction. The population decreases by 23% by 2100, accompanied by a 30% reduction in aggregate labor. These declines translate into lower GDP, consumption, investment, exports, and government finances.

Fig. 10 shows that the effect of the pension reform — in terms of its capa­city to improve macroeconomic indicators and reduce the budget-balancing VAT rate — is lower in the long run under the high-fertility scenario than under the low‑fertility scenario. This result arises because, although the high-fertility scenario produces a larger population and a growing workforce, these trends mean that the pension system remains more sustainable even without the reform. In other words, when the demographic outlook is favorable, fiscal pressures are already less acute, and raising the statutory retirement age provides less additional relief. By contrast, under low fertility and population aging, the baseline (no-reform) scenario faces severe fiscal strain: the effect of the reform is much stronger, with a larger reduction in the necessary VAT rate and more visible improvements in macroeconomic aggregates in the long run.

Fig. 11 presents trajectories of the VAT rate required to balance­ government budget under alternative demographic scenarios: median, low, and high fertility. Each panel compares two paths: the solid line shows the VAT rate that would be required if the pension reform (raising the statutory retirement age) had not been implemented, while the dashed line reflects the VAT rate after the reform.

Across all scenarios, the VAT rate rises steadily over the projection horizon, reflecting growing fiscal pressures from population aging and increased pension obligations. However, the pension reform clearly mitigates the need for higher taxes: in every demographic scenario, the dashed line lies below the solid line, indicating that raising the statutory retirement age reduces the VAT rate required for budget balance. Importantly, implementing the pension reform allows the government to avoid increasing the VAT rate until around 2040.

The magnitude of this fiscal relief varies depending on the demographic assumptions. In the low-fertility scenario, where population aging is most pronounced, the gap between the post-reform and no-reform VAT rates is widest­, indicating that the reform delivers especially strong fiscal benefits when demographic pressure is severe. In the high-fertility scenario, the gap is narrower, because a larger, younger population places less strain on the pension system, and the marginal fiscal benefit of reform is correspondingly smaller (Fig. 12).

The results underscore that, although pension reform is effective in containing future tax increases, its relative impact depends heavily on the underlying demographic trajectory — not only the total population size but also its age structure and rate of aging.