Evaluating Mongolia’s Sugar-Sweetened Beverage Tax: A Cost-Effectiveness Analysis

Abstract

Background

Mongolia has approved a tax of 20% on sugar-sweetened beverages (SSBs) to take effect in 2027, yet while prior work has assessed potential health impacts, no peer-reviewed study has evaluated its cost-effectiveness.

Objective

This study evaluates the cost-effectiveness of Mongolia’s planned SSB tax by estimating its effect on health outcomes, healthcare costs, and changes in SSB consumption patterns.

Methods

Using a Markov cohort model, we project outcomes over 10-year, 20-year and lifetime horizons. The model incorporates changes in the prevalence of overweight and obesity, noncommunicable disease (NCD) incidence, and associated costs, with primary outcomes measured in disability-adjusted life years (DALYs) averted, and quality-adjusted life years (QALYs) gained. To improve precision, disease risk estimates are informed by biomarker data, such as blood sugar and lipid levels, minimizing reliance on self-reported health indicators.

Results

The ad valorem tax of 20% is projected to reduce consumption of SSBs by 31% through consumer price response alone. Accounting for additional calorie reductions from 20% industry reformulation, the combined effective reduction in SSB-related energy intake reaches approximately 45%. Across the three time horizons, the tax could avert 13, 260 and 2255 premature deaths; yield 2899, 14,362 and 33,613 QALYs; and avert US$6.13 million, US$30.34 million and US$71.02 million, respectively, in healthcare costs. From the healthcare perspective, the tax is cost-effective over 10 years (net cost US$1.77 million; incremental cost-effectiveness ratio (ICER) US$611.99/QALY) and becomes cost-saving within 20 years and over the lifetime horizon.

Conclusions

These findings support the implementation of the SSB tax as a strategy to reduce Mongolia’s NCD burden and generate long-term fiscal and health benefits.

Supplementary Information

The online version contains supplementary material available at https://doi.org/10.1007/s40258-026-01070-2.

Key Points for Decision Makers

Mongolia’s planned 20% tax on sugary drinks is projected to reduce sugary drink consumption and lower rates of diabetes, heart disease, and related illness.

The tax is expected to be cost-effective from the time it is introduced, and to start saving the health system money outright within about two decades.

Revenue raised by the tax could be reinvested in public health and nutrition programs, adding to its long-term benefits.

Introduction

The global increase in the consumption of sugar-sweetened beverages (SSB) is a serious public health issue, strongly related to chronic diseases such as obesity, type 2 diabetes, cardiovascular disease (CVD), and dental caries [1, 2]. Higher consumption of SSBs is associated with a higher incidence of type 2 diabetes, with each serving daily of SSBs increasing the risk by 18% [3]. These health burdens are especially significant in middle income countries, where the prevalence of obesity increases in parallel with increased SSB consumption [4].

Global studies consistently show that SSB taxes reduce consumption, improve public health, and generate revenue for health initiatives [5–8]. A meta-analysis of 62 global studies revealed that SSB demand is highly responsive to tax-induced price increases, with a price elasticity of −1.59 and an average reduction in SSB sales of 15% [9]. The majority of these studies were conducted in high-income countries; evidence from low- and middle-income countries (LMICs) remains more limited. In countries such as Mongolia, where demographic shifts are contributing to the transition from traditional diets to Western dietary patterns, understanding the impact of such fiscal policies is increasingly important [10].

Mongolia, like many LMICs, faces complex challenges as urbanization, shifts in dietary options and food environments, income disparities, and aggressive marketing contribute to rising SSB consumption. Limited access to fresh foods exacerbates the noncommunicable disease (NCD) burden, particularly among children and adolescents [11]. Recent nationally representative survey data indicate that overweight and obesity remain highly prevalent among Mongolian adults. The 2023 Sixth National Nutrition Survey reported that more than half of Mongolian adults were overweight or obese, including 53.0% of women and 57.4% of men, with obesity prevalence of 18.0% and 17.0%, respectively [12]. In 2017, Mongolia’s obesity-related death rate was triple the global average, with cardiovascular diseases accounting for 55% of hospitalizations and 37% of deaths [13]. Prior studies have documented rising SSB intake and associated dietary risks [12, 14]. Recent data show 66% of men and 64% of women reported consuming sugary beverages the previous day, and 42% of adolescents drank SSBs more than twice weekly [12]. To address this, Mongolia passed an excise tax on SSBs in June 2024, to take effect in 2027. This is the first peer-reviewed cost-effectiveness analysis of Mongolia’s SSB tax. Prior to the legislation, a World Bank commissioned study (2023) assessed potential impacts on specific diseases, health-adjusted life years, and healthcare costs, but did not evaluate cost-effectiveness.

This study evaluates the cost-effectiveness of Mongolia’s planned SSB tax using a Markov cohort model, an important tool for LMIC policymakers facing budget constraints [15]. Key outcomes include quality-adjusted life years (QALYs) gained, disability-adjusted life years (DALYs) averted, healthcare cost savings, and tax revenue generated. By quantifying health and economic outcomes, this study informs evidence-based SSB tax policy.

Methodology

Study Design Overview

This study uses a closed cohort Markov model—a macrosimulation approach—implemented in TreeAge Pro to evaluate the cost-effectiveness of an SSB tax in Mongolia. The aggregate cohort was initialized at the mean adult age of 40 years observed in the 2019 Mongolia STEPS survey [16]. The model simulates SSB-related health outcomes among Mongolia’s population aged 15 years and above using demographic and epidemiological data from the 2019 Mongolia STEPS survey and National Statistics Office of Mongolia. Age-specific body mass index (BMI) distributions, SSB consumption patterns, NCD prevalence, and transition probabilities were estimated separately across adult age groups (15–24 through 70+ years), and age-specific transition probabilities were applied annually throughout the simulation horizon as the cohort aged over time. Due to data limitations, 2019 epidemiological estimates were assumed to approximate 2023 conditions. Individuals younger than 15 years were excluded because the STEPS survey does not collect data for younger age groups. Young adults aged 15–24 years are included in the age-stratified parameter estimation.

A de novo Markov cohort model was developed rather than adapting existing public health tools such as DYNAMO-HIA [17], because the Mongolian STEPS 2019 survey provided age-specific NCD incidence rates, BMI distributions, and healthcare costs that could be directly incorporated into a custom TreeAge Pro model calibrated to the local context.

We model a 20% excise tax on SSBs (excluding milk-based products) implemented once and remaining continuously in force for the full duration of each modeled horizon (10 years, 20 years, and lifetime), consistent with WHO recommendations for sustained fiscal measures. The comparator is a no-intervention scenario. The study follows CHEERS 2022 guidelines [18]. We use elasticity estimates from our recent study based on Mongolian household data [19], as machine learning methods tested in the original protocol [20] were unsuitable for simulation. The model structure, outcomes, and policy scenarios are aligned with the protocol.

Model Structure

A closed cohort Markov model was developed and simulated using TreeAge Pro to estimate disease progression and estimate the cost-effectiveness of the tax over a lifetime horizon. The model projects changes in health outcomes, healthcare costs, and economic impacts. The Markov model includes three health states: healthy (individuals without an NCD), NCD (individuals diagnosed with an obesity-related disease such as diabetes or cardiovascular disease), and dead (an absorbing state). NCDs are assumed to be irreversible once diagnosed [21]

Figure 1 illustrates the relationships between the three health states. The BMI/weight pathway is shown as an intermediate modifier: reduced SSB consumption leads to a reduction in energy intake, a downward shift in BMI, and a lower transition probability from healthy to NCD. Hospitalization was not included as a separate state due to limited data on transition probabilities by hospitalization status.

Fig. 1.

Health state transition diagram. This diagram visualizes the relationships between the three health states for NCD. SSB sugar-sweetened beverage, BMI body mass index, NCD noncommunicable disease

Baseline characteristics, health costs, and transition probabilities are derived from Mongolian surveys and international sources. Individuals transition between these states annually based on age-specific probabilities. Transitions from Healthy to NCD are determined using age-specific incidence rates from national epidemiological data. Both the Healthy$\to US$NCD and NCD$\to US$Dead transition probabilities are age-dependent, drawn from age-stratified national data from the Health Development Center of Mongolia [22]. A half-cycle correction was applied by assigning 50% of the annual QALYs, costs, and other outcomes in the initial and final cycles.

Obesity was not modeled as a separate health state; instead, BMI-related risk changes were incorporated as modifiers of the Healthy$\to US$NCD transition probabilities.

QALY and DALY estimates for NCD cases are sourced from the Institute for Health Metrics and Evaluation (IHME) [23]. The disability weight of 0.26 was derived as a prevalence-weighted average across the obesity-related NCDs modeled (type 2 diabetes, cardiovascular disease, and chronic kidney disease), using condition-specific GBD 2024 disability weights and condition prevalence from the 2019 Mongolia STEPS survey. The utility weight of 0.74 ($=1-0.26$) was applied for QALY estimation.

Age-specific transition probabilities for obesity-related NCDs were primarily derived from the 2019 Mongolia STEPS survey using observed diabetes and cardiovascular disease prevalence across BMI and age groups. These epidemiological estimates were used to construct the healthy$\to US$NCD transition probabilities in the Markov model. IHME GBD 2024 data were used to inform disability, utility, and mortality-related parameters rather than to estimate a single pooled relative risk for all NCDs.

The effect of SSB taxation is modeled by adjusting transition probabilities from Healthy to NCD, based on reduced SSB consumption and associated disease risk. While younger individuals may accrue benefits over a longer horizon, older age groups face higher immediate NCD risk with less discounting erosion, making the distribution of gains age-dependent.

Post-tax effects are simulated using the price elasticity of demand, linking lower SSB intake to reduced energy consumption, body-weight change, and improved health outcomes. The weight-change assumption follows the equilibrium model: a sustained reduction of 100 kJ/day corresponds to approximately 1 kg reduction in body weight at the new energy equilibrium [24, 25]—not a cumulative annual loss. The model applies this weight change once at the post-tax equilibrium and does not assume continued annual weight reduction.

We also account for anticipated industry reformulation, assuming 20% of SSB products will be modified to avoid taxation by reducing sugar content [26]. These reformulated products contribute to overall calorie reductions. Together, these pathways reflect both consumer and industry responses in the Markov model framework.

Figure B1 in Supplementary Material depicts the TreeAge Pro model framework. Figure 2 illustrates the logical pathway from tax implementation to health outcomes; the reformulation pathway is now shown as a parallel input to the energy-intake reduction step.

Fig. 2.

Pathway of SSB tax impact on BMI through changes in consumption and energy intake. The diagram shows how an SSB tax reduces energy intake through lower consumption and reformulation of SSBs, leading to changes in BMI and disease risk. SSB sugar-sweetened beverage, BMI body mass index, NCD noncommunicable disease

Data Sources and Assumptions

The model integrates data from multiple sources. The primary health dataset is the Mongolia 2019 STEPS survey, implemented by the Ministry of Health and National Center for Public Health [16] with support from the World Health Organization. It provides detailed information on disease prevalence, BMI distribution, healthcare costs, and productivity losses. As no updated nationally representative health surveys are currently available, we assume that the 2019 prevalence estimates are broadly applicable to the 2023 population, and all costs are inflation-adjusted to 2023 values and converted to USD using IMF 2023 purchasing power parity (PPP) conversion factors [27].

We incorporate the 2023 population structure from the National Statistics Office of Mongolia [28], baseline SSB consumption levels from the Sixth National Nutrition Survey [12], and SSB prices from the 2023 Household Socioeconomic Survey [29]. Disease epidemiology, relative risks of SSB-related NCDs, and utility values are sourced from international datasets including the Global Burden of Disease Study [23].

Intervention costs—tax administration and industry compliance—are conservatively assumed to equal 1% of annual SSB tax revenue each. Official Mongolian policy documents accompanying the proposed SSB tax suggest that no substantial additional implementation costs are expected. However, in the absence of formal costing studies, a small positive implementation cost was included to avoid understating policy costs. Lal et al. [6] reported implementation costs of a similar magnitude in the Australian context, although this was an empirical outcome rather than an explicit modelling assumption. Sensitivity analyses test alternative implementation cost assumptions of 0% and 2%.

The own-price elasticity of SSB demand ($-1.57$) was applied uniformly across age groups. An age gradient in elasticity (reflecting generally higher incomes in middle-aged groups) would be theoretically reasonable but age-stratified estimates were not available from the Mongolian household data. Cross-price elasticities on substitute beverages could not be incorporated because STEPS does not include data on substitute products.

The reduction in SSB consumption was assumed to be sustained throughout the model horizon, with no behavioral waning effect modeled. This is consistent with the assumption that the tax remains continuously in force.

Key model parameters and data sources are presented in Table 1, with additional details in Supplementary Tables B1–B5. Supplementary Table B2 presents the age-specific BMI distributions, NCD prevalence, SSB consumption patterns, and modeled reductions in NCD incidence that underpin the transition probabilities used in the Markov model.

Table 1.

Key model parameters and data sources (2023-adjusted values)

Parameter	Value/description	Source
Baseline characteristics
Target population (million)	2.4	[28]
Discount rate	5%	Assumed; [30]
Baseline demographics	Table C2	[16]
Mean daily SSB consumption	Table B2	[12]
Own-price elasticity of SSB demand	−1.57	[19]
Average SSB price	US$3.23	[28]
BMI distribution	Table C2	[16]
Disease epidemiology	Table C3	[16]
Relative risk of SSB-related diseases	See Table B4	[23]
Change in consumption to weight change	100 kJ/day $\approx US$ 1 kg at equilibrium	[6, 24, 25]
Intervention costs
Government tax collection costs	1% of SSB tax revenue	Conservative assumption (cf. [6])
Industry compliance and reformulation	1% of tax revenue	Conservative assumption (cf. [6])
Healthcare cost data
Government health insurance coverage	Table C4	[31]
Insurance fund for NCDs	US$917.41 per capita
Out-of-pocket (OOP) costs	Table C5	[16]
Formal healthcare cost$_{}^{†}US$	US$187.41
Informal care costs
Home care cost$_{}^{‡}US$	US$458.68	[16]
Productivity cost inputs
Average monthly wage (adult)	US$1680.18	[28]
Average annual work days missed per NCD case	18.16 days	[16]
Health weights
Disability weight (DALYs)	0.26 (prevalence-weighted average, GBD)	[23]
Utility weight (QALYs)	0.74 ($=1-0.26$)	Derived
Currency
USD conversion	IMF 2023 PPP conversion factors	[27]

BMI body mass index, SE standard error, NCD noncommunicable disease, OOP out-of-pocket, PPP purchasing power parity

$_{}^{†}US$Formal healthcare cost refers to the average annual cost per person with NCD, including provider, medicine, test, transport, and health worker fees

$_{}^{‡}US$Home care cost refers to the annual informal care cost per household related to NCDs. All monetary values are reported in 2023 USD adjusted using IMF purchasing power parity (PPP) conversion factors

Cost-Effectiveness Analysis

Outcomes

The primary outcome of this study is the Incremental Cost-Effectiveness Ratio (ICER), expressed as the cost per QALY gained. The ICER is calculated as:

$$\begin{array}{c}\hfill \text{ICER}=\frac{\mathrm{\Delta }\text{Cost}}{\mathrm{\Delta }\text{QALY}}\end{array}$$ 1

where

• $\mathrm{\Delta }\text{Cost}$ = Difference in total costs between the intervention (SSB tax) and the baseline (no tax) scenarios.

• $\mathrm{\Delta }\text{QALY}$ = Difference in QALYs gained between the intervention and baseline scenarios.

Secondary outcomes include a range of health, economic and social indicators. Health outcomes encompass DALYs averted, QALYs gained, life years gained, premature deaths averted, NCD cases averted, healthcare cost savings, tax revenue generated, and productivity gains.

Assessment of Costs

Costs related to the SSB tax included implementation, healthcare, and productivity components. Implementation costs are described in Section 2.3. Healthcare costs included government and out-of-pocket expenditures associated with obesity-related NCD treatment. Productivity losses were estimated using a human capital approach based on NCD-related missed workdays reported in STEPS 2019 and monetized using average national wage estimates. The base-case analysis assumed 100% tax pass-through to retail prices, while sensitivity analyses examined alternative pass-through rates of 50% and 150% (Table D5). Mongolia-specific pass-through data are not yet available, as the tax has not taken effect; full pass-through is retained as a conservative upper-bound base case in the absence of local evidence, with the tax’s cost-effectiveness shown to be robust to more conservative assumptions (see Sect. 4).

Assessment of Benefits

The benefits of the SSB tax are assessed in terms of health and economic outcomes. Health gains include reductions in obesity, type 2 diabetes, cardiovascular diseases, and other NCDs, measured as DALYs averted, QALYs gained, and premature deaths averted.

Discount Rate

A 5% discount rate is applied to both costs and health outcomes from year 2 onward, consistent with LMIC economic evaluation guidelines [30].

Sensitivity Analysis

A deterministic sensitivity analysis (DSA) was conducted over a 20-year horizon to test the robustness of the SSB tax’s cost-effectiveness. Key assumptions were varied across scenarios: discount rates (3%, 5%, 7%, 9%), pass-through rates (50%, 100%, 150%), tax rates (20%, 30%, 40%, 50%), and product reformulation levels (0%, 20%, 40%). An extended DSA additionally varied the disability weight (±20%), composite relative risk of NCD (±20%), and maximum age for DALY calculations (70, 75, and 80 years); results are provided in Table D9. A probabilistic sensitivity analysis (PSA) was conducted using Monte Carlo simulation with 1000 iterations over a 20-year horizon. The following parameter distributions were assigned: own-price elasticity (Normal: mean $-1.57$, SE 0.12) [19]; disability weight (Beta: mean 0.26, SD 0.02); utility weight ($=1-$ disability weight); composite relative risk of NCD (Log-normal: SD 0.10 on log scale); healthcare costs (Gamma: mean equal to base-case values, SD 15% of mean); and pass-through rate (Uniform: 50–150%). Results are presented as a scatter plot on the cost-effectiveness plane and a cost-effectiveness acceptability curve (CEAC) in Supplementary Figs. E1 and E2.

Model Validation

The model was validated through three approaches. First, face validity was assessed through consultation with two health economists and one Mongolian public health expert, who reviewed the model structure, parameter choices, and outputs for plausibility. Second, internal consistency was confirmed by verifying that model outputs under no-tax conditions were broadly consistent with observed national NCD prevalence and mortality estimates from STEPS 2019. Third, cross-validation compared 10-year projections against similar SSB tax models from Mexico [32] and Australia [6], with per-capita outcomes falling within the ranges reported in those studies.

Ethics

This study used only publicly available secondary data and involved no human participants. As per the 2023 National Statement on Ethical Conduct, it is exempt from further ethical review.

Results

Table 2 presents the baseline characteristics of the population. Among Mongolian adults aged 15 years and older at baseline, the mean age was 40 years, and approximately 52% were female. The mean BMI was 26.4 kg/m². Overall, 46.4% had a BMI $<25$ (underweight or normal weight), 33.2% had overweight (BMI 25.0–29.9), and 20.4% had obesity (BMI $\ge US$ 30). Obesity prevalence was higher among people in the age groups of 45–54 years (29.8%) and 55–69 years (29.8%) (Table B2). While only 11% of adults self-reported having a noncommunicable disease (NCD), 30% of those with an NCD were hospitalized during the last 12 months. However, biomarker-based estimates indicate that 23.1% of adults had at least one NCD, including diabetes, cardiovascular disease, or chronic kidney disease. This approach reduces self-report bias and enhances the accuracy of policy-relevant evidence. At baseline, the average daily intake of sugar-sweetened beverages (SSBs) was 0.64 cups (0.16 l) [12], contributing 116.6 kilocalories (kcal) per person per day. A 20% SSB tax is projected to reduce consumption by 31.4%, based on observed price responsiveness in Mongolia [19]. This estimate reflects the consumer price response alone (own-price elasticity of −1.57 applied to the 20% price increase). Additionally, 20% of SSBs are expected to be reformulated over 10 years to contain less than 5 g of sugar per 100 ml. Accounting for both consumer and industry responses, the combined effective reduction in SSB-related energy intake reaches approximately 45%; in absolute terms, the tax is projected to reduce average SSB consumption by 0.072 l per day and reduce daily energy intake by 46.2 kcal per adult.

Table 2.

Baseline health and SSB intake of adults aged over 15 years in Mongolia

Characteristic	Value
Number of Mongolian adults represented (age 15+ years)	2.40 million
Age, mean (SD)	40 years (est.)
Age group distribution, %
15–24	20.03
25–34	22.16
35–44	22.15
45–54	16.68
55–69	14.98
70+	4.00
Sex, %
Male	48.02
Female	51.98
SSB consumption
SSB consumption, cups/day (SD)	0.64 (0.47)
SSB consumption, l/day (SD)	0.16 (0.12)
SSB energy intake, kcal/day	116.6
Health at baseline, %
Having obesity (BMI $\ge 30$), %	20.4
Having overweight (BMI 25–29.9), %	33.2
With at least one NCD	23.1
Systolic blood pressure, mmHg (SD)	122.90 (18.57)
Diastolic blood pressure, mmHg (SD)	79.17 (12.05)
Total cholesterol, mmol/l (SD)	5.09 ( 6.58)
Triglycerides, mmol/l (SD)	4.47 (14.46)
Urinary sodium, mmol/l (SD)	110.93 (57.36)
Body mass index, kg/m² (SD)	26.39 (5.14)
Current smoker, %	24.23 (1.35)
Prevalent NCD (incl. diabetes, CVD), %	10.96
Hospitalized due to NCD (past 12 months, % of NCD cases)	30.18

The hospitalization figure (30.18%) is the percentage of those with a diagnosed NCD who were hospitalized; it is not a percentage of the total adult population. NCD prevalence includes diabetes (fasting blood glucose $\ge US$7.0 mmol/L) [33], cardiovascular disease (systolic blood pressure > 140 mmHg or total cholesterol $\ge US$ 5.65 mmol/L) [34, 35], chronic kidney disease (urine protein excretion > 220 mg/L) [36], and self-reported NCDs. Classification of health conditions follows WHO and clinical diagnostic guidelines. Indicators in this table are drawn from multiple sources, including national demographic, nutrition, and health surveys. See Table 1 for detailed data sources

BMI body mass index, NCD noncommunicable disease, SD standard deviation

Implementation of a 20% SSB tax is projected to yield health benefits across all time horizons (Table 3). Over a lifetime (mean simulated follow-up of 36 years), the policy is estimated to generate 33,613 QALYs, avert 11,810 DALYs, prevent 9452 NCD cases, and avert 2255 premature deaths. The larger number of QALYs gained relative to DALYs averted reflects the 75-year reference age ceiling applied in YLL calculations, which limits the mortality component of DALYs for older cohorts, combined with 5% annual discounting; QALYs capture utility gains over remaining life years without this age ceiling. These improvements are driven by sustained reductions in sugar intake and associated body-weight change.

Table 3.

Cost-effectiveness outcomes of a 20% SSB tax in Mongolia across 10-year, 20-year, and lifetime horizons^*

Indicator	10 years	20 years	Lifetime
Health cost savings
Government health insurance savings (US$ million)	3.59	17.81	41.67
Out-of-pocket cost savings (US$ million)	2.53	12.54	29.35
Total healthcare costs averted (US$ million)	6.13	30.34	71.02
Healthcare costs averted per capita (US$)	2.55	12.65	29.62
Health impact
QALYs gained (years)	2899	14,362	33,613
DALYs averted (years)	1019	5046	11,810
NCD cases averted (years)	1119	7109	9452
Premature deaths averted (cases)	13	260	2255
Economic impact
Generated tax revenue (US$ million)	268.4	421.28	526.08
Tax intervention cost (US$ million)	5.37	8.43	10.52
Time loss averted (days)	71,299	353,231	826,708
Cost-Effectiveness outcome
Net costs (US$ million)$_{}^{‡}US$ (Healthcare perspective)	1.77	−9.38	−31.15
Net costs (US$ million)$_{}^{†}US$ (Government perspective)	−266.63	−430.66	−557.23
Net costs (US$ million)§ (Societal perspective)	−0.76	−21.92	−60.50
ICER—healthcare perspective (QALYs)$_{}^{‡}US$	US$611.99/QALY	Dominant	Dominant
ICER—healthcare perspective (DALYs)$_{}^{‡}US$	US$1741.81/DALY	Dominant	Dominant

QALY quality-adjusted life year, DALY disability-adjusted life year, ICER incremental cost-effectiveness ratio, NCD noncommunicable disease

^*Assumptions for all scenarios: 20% tax rate, 100% pass-through, 20% product reformulation, 5% discount rate, and a simulated population of 2.4 million

$_{}^{‡}US$Healthcare perspective includes government health insurance and out-of-pocket cost savings; net cost and ICERs throughout the paper are calculated from this perspective unless stated otherwise

$_{}^{†}US$Government perspective additionally includes generated tax revenue

^§Societal perspective additionally includes productivity gains from averted time loss. Where the intervention is both cost-saving and more effective, it is reported as “Dominant”; a negative ICER is uninterpretable and is not reported [37]

From a healthcare perspective, it is cost-effective over 10 years (ICER: US$611.99/QALY) and becomes cost-saving within 20 years, with net savings of US$9.38 million and US$31.15 million over 20 years and lifetime, respectively. Total healthcare costs averted were US$6.13 million (10 years), US$30.34 million (20 years), and US$71.02 million (lifetime). The government perspective yields net savings of US$266.63 million at 10 years and US$557.23 million over a lifetime. The societal perspective also shows consistent savings, including out-of-pocket reductions and productivity gains. Over a lifetime, the tax is projected to avert over 9400 NCD cases, 2255 premature deaths, and yield 33,613 QALYs and 11,810 DALYs. Full cost-effectiveness thresholds and scenario results are in Table D1.

A probabilistic sensitivity analysis (PSA), conducted from the healthcare perspective over a 20-year horizon using Monte Carlo simulation with 1000 iterations, confirmed the robustness of these results (Supplementary Figures E1 and E2). The intervention remained cost-saving and health-improving in nearly all simulations. Mean incremental QALYs gained were 14,263 (95% CrI 6859–23,833), and mean DALYs averted were 5008 (95% CrI 2392–8302). Mean incremental healthcare savings were US$17.73 million (95% CrI US$8.00–$32.04 million), with mean net savings of US$9.38 million (95% CrI US$2.45–$20.54 million). The intervention was dominant in all iterations, and the probability of cost-effectiveness was 100% at both the Mongolia GDP-per-capita threshold (US$4370/QALY) and the US$50,000/QALY threshold. These results confirm that the cost-effectiveness conclusion is robust to joint uncertainty across key model parameters.

The tax becomes more effective with higher tax rates, both in terms of health gains and cost offsets. As shown in Table 4, increasing the tax from 20% to 50% nearly triples the number of NCD cases and deaths averted, while increasing QALYs gained and healthcare cost savings. Full sensitivity analysis results are provided in Tables D5–D8. This robustness held consistently across alternative tax-rate scenarios (Table 4; Supplementary Figures E1–E2).

Table 4.

Sensitivity analysis: key outcomes of SSB tax under alternative tax rates over 20 years

Indicator	20%	30%	40%	50%
NCD cases averted (thousand)	7.1	10.7	14.2	17.8
Premature deaths averted	260	390	521	651
QALYs gained	14,362	21,543	28,724	35,905
Healthcare cost savings (US$ million)	17.81	26.71	35.61	44.51
Tax revenue (US$ million)	421.28	631.92	842.56	1053.20
Net cost (US$ million)$_{}^{‡}US$	−21.92	−32.88	−43.84	−54.80
Cost-effectiveness result $_{}^{‡}US$	Dominant	Dominant	Dominant	Dominant

Results reflect 20 year projections assuming full pass-through and 20% industry reformulation

NCD noncommunicable disease, QALY quality-adjusted life year

$_{}^{‡}US$Calculated from the healthcare perspective. Full analysis is available in Table D7. The similar cost-effectiveness results across tax-rate scenarios reflect modelling simplifications, including the assumption of constant price elasticity and proportionally scaled intervention costs

Discussion

This is the first study to evaluate the cost-effectiveness of an SSB tax in Mongolia, assessing its impact on BMI, NCDs, healthcare costs, and mortality. Our findings show that an SSB tax is consistently cost-effective and is dominant (cost-saving with greater health gain) from the 20-year and lifetime healthcare perspectives, and from all time horizons under the government and societal perspectives. The intervention offers a relatively low-cost strategy to reduce BMI in the short term while generating fiscal revenue that could fund public health programs. The estimated annual per capita tax burden is modest (US$14.23), lower than figures reported in high-income settings such as Australia (US$30) [6].

The model assumes full pass-through to retail prices. Through consumer price response alone, the 20% tax is projected to reduce SSB consumption by 31%. The estimated 31% reduction reflects the direct consumer response to the tax, while the larger overall reduction additionally incorporates assumed reformulation-related calorie reductions. This combination of consumer and industry responses reduces energy intake and subsequent improvements in BMI and NCD risk. The projected average daily calorie reduction among adults with overweight or obesity (46 kcal) aligns with previous models and represents approximately one-third of the effect seen in controlled interventions [6].

The similar scaling of outcomes across tax-rate scenarios reflects simplifying assumptions within the model. Specifically, own-price elasticity and the BMI-to-NCD risk relationship were assumed linear across the tax range tested. In addition, intervention costs were modeled as a fixed proportion of tax revenue, causing costs and benefits to scale proportionally across scenarios. In practice, diminishing returns and nonlinear behavioral responses may emerge at higher tax levels. Future modelling with greater individual-level heterogeneity may better capture these dynamics.

When placed in the context of international literature, several of our estimates align with modelling results from other middle and high income countries. For instance, over 10 years, Mongolia’s projected healthcare cost savings of US$2.55 per capita are higher than US$1.40 in Mexico [32], while lifetime per capita savings of US$29.62 are above the US$4.50–$27.10 range reported for Australia [38]. Our modeled gain of 33,613 QALYs over a lifetime is lower in absolute terms than the 175,000–24 million QALYs reported in large-population studies in the USA and Australia [25, 26], but proportionally similar when scaled per capita. Likewise, the number of deaths averted in our study (2255) corresponds well to estimates from Mexico and South Africa, adjusting for population size and tax design [32, 39]. These comparisons suggest that the projected health and economic impacts of Mongolia’s 20% SSB tax are consistent with international experience, reinforcing the policy’s potential relevance for LMIC settings. Tables D2–D4 present these international comparisons over 10-year, 20-year, and lifetime horizons.

The intervention is dominant at the 20-year and lifetime healthcare perspectives, generating net healthcare savings of US$9.38 million (20-year) and US$31.15 million (lifetime) alongside substantial QALYs and DALYs gained.

These findings suggest the tax is a dominant strategy. Even in lower-middle-income settings with limited fiscal space, SSB taxation can yield health and economic benefits comparable to those observed in wealthier countries when adjusted for disease burden and population size.

The modelling incorporated both consumer responses to price changes and industry-level reformulation, offering a comprehensive view of how the tax could influence behavior and health outcomes. Sensitivity analyses confirmed that the cost-effectiveness of the proposed SSB tax remains robust across a range of policy-relevant scenarios. Varying discount rates, pass-through assumptions, tax rates, and levels of product reformulation produced consistent results supporting the intervention. Even under conservative assumptions, such as 50% pass-through and 0% reformulation, the tax remained dominant over the 20-year horizon, with net healthcare savings of US$10.96 million and US$10.61 million respectively, compared with US$21.92 million in the base case (Tables D5 and D8). The base-case assumptions therefore represent the upper end of the estimated benefit rather than a precondition for cost-effectiveness. The probabilistic sensitivity analysis confirmed the intervention remained dominant in nearly all simulations, reinforcing this robustness. This supports the growing consensus that SSB taxes can be a strategic lever for preventing NCDs, particularly in settings with increasing obesity and diabetes rates, as seen in Mongolia.

Beyond health gains, the SSB tax offers fiscal benefits. The revenue generated provides an opportunity to reinvest in public health programs, potentially amplifying the policy’s long-term impact. However, public acceptance may depend on how the tax is framed and implemented. Evidence from other countries indicates that taxes perceived as revenue-raising rather than health-promoting often face greater resistance and reduced effectiveness [40]. To ensure sustainability, it is essential to maintain transparency in revenue use, clearly communicate the health rationale, and allocate funds toward visible health initiatives [41, 42]. As noted by Firdaus et al. [7], the success of SSB taxation depends not only on fiscal design but also on complementary actions such as public education, promotion of healthy lifestyles, and incentives for industry reformulation. These integrated strategies can enhance behavioral change and sustain long-term health and economic benefits.

This study has several limitations. The macrosimulation approach, with a cohort initialized at mean age of 40 years, does not capture individual-level heterogeneity in SSB consumption, income, or NCD risk. This aggregation means the model’s point estimates are best interpreted as population-level, order-of-magnitude projections rather than precise forecasts; the probabilistic sensitivity analysis (Section 3), in which the intervention was dominant in approximately 100% of 1000 Monte Carlo iterations, provides a more informative indication of the certainty attached to these results than the point estimates in Tables 3 and 4 alone. The model does not simulate childhood trajectories (under 15 years), as STEPS does not cover this age group. As a result, cumulative long-term health gains are likely underestimated. Obesity was not modeled as an explicit health state, meaning obesity-specific healthcare costs and quality-of-life impacts beyond the included NCDs (e.g., musculoskeletal disorders and certain cancers) were not captured. The analysis did not estimate broader welfare effects, including changes in consumer surplus and producer surplus associated with reduced SSB consumption. As a result, the reported cost-effectiveness estimates reflect health system, patient, and productivity impacts rather than the full societal welfare consequences of the tax.

The study relies on secondary data sources, which may not fully capture all factors influencing SSB consumption and health outcomes in Mongolia. As with any simulation model, our projections represent approximations based on available evidence. Due to data constraints, we used age-based averages from the WHO STEPS survey and assumed uniform price elasticities across age groups, as income-specific data were unavailable. Costs in life years gained (e.g., from age-related conditions such as heart failure or Parkinson’s disease in years of life saved) were not modeled; this may lead to a slight overestimate of net healthcare savings. Baseline SSB intake may also be affected by reporting bias inherent in survey data. Cross-price elasticities were not incorporated into the base-case model. While the Household Socio-Economic Survey (HSES) provides information on alternative beverage categories, previous Mongolia-specific demand-system analyses suggested that substitution effects were relatively small and did not substantially offset reductions in SSB-related calorie intake [19]. Nevertheless, excluding substitution effects remains a limitation and may lead to a modest overestimation of health benefits. We relied on price data from the Household Socioeconomic Survey, which may differ slightly from actual market prices. Additionally, broader social and industry-level effects, such as employment changes in the beverage sector, are not included in the analysis. The model also does not account for benefits related to improved oral health or broader indirect cost savings, such as increased productivity from reduced absenteeism or disability. As a result, the full societal benefits, particularly for disadvantaged groups, are likely underestimated. Additionally, baseline prevalence and risk factor data were taken from 2019, assuming stability to 2023, which may slightly understate current NCD burden. The model also does not account for population ageing and population growth over time, which may affect long-term projections. Sensitivity analyses address some of these limitations by testing the robustness of results under different scenarios.

Conclusions

This study used economic modelling and simulation to evaluate the health, economic, and fiscal outcomes of a proposed 20% sugar-sweetened beverage (SSB) tax in Mongolia. The findings suggest that the policy would be cost-effective from a healthcare perspective and cost-saving from both government and societal perspectives over 10-, 20-year, and lifetime horizons. The tax is projected to substantially reduce SSB consumption, avert thousands of noncommunicable disease (NCD) cases and premature deaths, and generate healthcare savings and fiscal revenue, over US$526 million in tax revenue and nearly 33,600 QALYs gained across a lifetime horizon. These results support the implementation of the tax and highlight the potential for similar policies to improve health and generate revenue in other low- and middle-income countries (LMICs) facing rising diet-related disease burdens. Allocating a portion of the tax revenue to health promotion or nutrition programs may further increase public support and enhance the policy’s impact. Future research could explore optimal revenue use, implementation challenges, and behavioral responses to maximize long-term public health gains.

Supplementary Information

Below is the link to the electronic supplementary material.

Author Contributions

N.B. conceived the study, developed the methodology, conducted the formal analysis, prepared the software and data curation, and drafted the original manuscript. R.B. contributed to conceptualization, validation, and manuscript review and editing. U.M. contributed to conceptualization, investigation, acquisition of resources, and manuscript review and editing. K.S. provided supervision, contributed to conceptualization and methodology, and supported manuscript reviewing and editing. All authors read and approved the final manuscript.

Funding

Open Access funding enabled and organized by CAUL and its Member Institutions. This research was supported by Macquarie University via the first author’s PhD scholarship and HDR support funding. No external funding was received. The authors declare no conflicts of interest.

Data Availability

This study used secondary datasets obtained from the National Statistics Office of Mongolia and the National Center for Public Health. These datasets are accessible for research purposes upon request through the official data portals: https://web.nso.mn/nadamn/index.php/catalog/HSES/dataset and https://extranet.who.int/ncdsmicrodata/index.php/catalog/836.

Code Availability

The TreeAge Pro model used to generate the cost-effectiveness results is available from the corresponding author upon reasonable request.

Declarations

Conflict of interest

The authors declare that they have no competing interests.

Ethics approval and consent to participate

Not applicable. This study used only publicly available, de-identified secondary data and involved no human participants. Under the 2023 National Statement on Ethical Conduct in Human Research, the analysis is exempt from ethics review.

Consent for publication

Not applicable.

Declaration of generative AI and AI-assisted technologies in the writing process

During the preparation of this work, the authors used ChatGPT (OpenAI) to improve the readability and clarity of the manuscript. After using this tool, the authors reviewed and edited the content as needed and take full responsibility for the content of the published article.