🍀 Clover

How Clover Works

Clover measures systemic economic pressure on mental health — how much current economic conditions, relative to a pre-pandemic baseline, are bearing on community wellbeing. Every weight and transmission pathway is grounded in published epidemiological evidence.

Baseline & Reference Period

All pressure indices are computed relative to the 2019–2020 pre-pandemic average — the last stable reference period before compound crises (COVID-19, supply chain disruption, inflation surge, interest rate increases) reshaped the economic landscape.

This baseline is scientifically clean: data is well-documented across all sources, and it represents conditions that most OECD economies would consider "normal." The purpose of Clover is to reveal accumulated systemic pressure — a pre-crisis baseline makes that visible.

The Calculation

Each stressor's pressure index is computed in six steps:

Step-by-step detail

Step 1–2: Indicator change. For each stressor, we track 2–4 key indicators. The percentage change from baseline captures how much conditions have deteriorated (or improved). For example, if the rent-to-income ratio was 0.29 in 2019–2020 and is now 0.34, that's a +17% change.

Step 3: Weighted composite. Each indicator within a stressor has a weight reflecting its relative importance, derived from the epidemiological literature. Weights within a stressor sum to 1.0.

Step 4: Sensitivity coefficient. The weighted change is scaled by a sensitivity coefficient derived from published effect sizes. Stressors with stronger evidence for mental health impact (e.g., unemployment, d=0.51) receive higher sensitivity than those with weaker or more indirect evidence (e.g., inequality, r=0.47 at the ecological level).

Step 5: Regional multiplier. The same economic stressor hits harder in countries without universal healthcare or with weak safety nets. See Regional Multipliers below.

Step 6: Uncertainty. We perturb all weights by ±20% across 500 Monte Carlo iterations. The reported pressure_low and pressure_high represent the 10th and 90th percentile of that distribution. These uncertainty bands are deliberately wider than those used in mechanical causal models (like commodity price tracking), reflecting the probabilistic nature of social determinant pathways.

The final value is clamped to 1–99% and classified into severity bands:

Stressors & Evidence

Phase 1 covers five economic stressors. Each is grounded in published, peer-reviewed evidence for its mental health impact.

🏠 Housing Cost Burden

Pathway: Housing cost >30% of income → financial anxiety → sleep disruption → reduced social participation → depression/anxiety risk

Effect size: OR 1.3–1.5 for anxiety/depression (Bentley et al., 2011)

Key evidence: Bentley, R. et al. (2011). Association between housing affordability and mental health: A longitudinal analysis of 35 Australian surveys. Social Science & Medicine, 73(12), 1771–1780. The WHO (2014) identifies housing cost burden as a primary social determinant of mental health.

📉 Unemployment

Pathway: Job loss → income loss + identity disruption + social isolation → depression/anxiety. In countries without universal healthcare, job loss also means insurance loss, compounding the effect.

Effect size: d=0.51 for depression/anxiety (Paul & Moser, 2009 meta-analysis, N=237 cross-sectional and 87 longitudinal studies)

Key evidence: Paul, K.I. & Moser, K. (2009). Unemployment impairs mental health: Meta-analyses. Journal of Vocational Behavior, 74(3), 264–282. The relationship is causal: longitudinal studies show mental health deteriorates after job loss and recovers after re-employment, controlling for prior health status.

🔥 Inflation Pressure

Pathway: Inflation above wage growth → declining purchasing power → perceived financial insecurity → anxiety/stress → sleep disruption → depression

Effect size: OR 1.1–1.3 for anxiety/depression (Rohde et al., 2016); effect concentrated in lowest income quintiles

Key evidence: Rohde, N. et al. (2016). The effect of economic insecurity on mental health: Recent evidence from Australian panel data. Social Science & Medicine, 151, 250–258. Frasquilho, D. et al. (2016). Mental health outcomes in times of economic recession: A systematic literature review. BMC Public Health, 16, 115.

💳 Consumer Debt

Pathway: Rising debt → persistent financial worry → rumination → sleep disruption → depression/anxiety. Debt-related shame compounds the effect through social withdrawal.

Effect size: OR 1.2–1.4 for depression; OR 1.9 for suicidal ideation among those with unmanageable debt (Richardson et al., 2013)

Key evidence: Sweet, E. et al. (2013). The high price of debt: Household financial debt and its impact on mental and physical health. Social Science & Medicine, 91, 94–100. Richardson, T. et al. (2013). The relationship between personal unsecured debt and mental and physical health: A systematic review and meta-analysis. Clinical Psychology Review, 33(8), 1148–1162.

⚖️ Income Inequality

Pathway: High inequality → status anxiety + reduced social trust + underinvestment in public services → population-level mental health burden

Effect size: r=0.47 between inequality and mental illness prevalence at the ecological level (Pickett & Wilkinson, 2015)

Key evidence: Pickett, K.E. & Wilkinson, R.G. (2015). Income inequality and health: A causal review. Social Science & Medicine, 128, 316–326. Note: inequality operates primarily at the ecological (society) level — it's the distribution shape, not individual income, that predicts population mental health.

Regional Multipliers

The same economic shock has different mental health impacts depending on a country's structural buffers. A 2% rise in unemployment means something very different in Sweden (strong safety net, universal healthcare, multiplier ~0.6×) vs. the United States (employer-linked insurance, weak safety net, multiplier ~1.2×).

Each country's impact multiplier is derived from four factors:

Current multipliers (Phase 1 countries):

A multiplier of 1.0× means the raw pressure index is used directly. Values above 1.0 amplify the stressor (weaker buffers); values below 1.0 dampen it (stronger buffers). The United States has the highest multiplier in this dataset due to employer-linked health insurance (job loss = healthcare loss) and comparatively low social expenditure.

Data Sources

All data sources are free and publicly accessible:

Clover's data pipeline runs weekly (Monday 06:00 UTC) via GitHub Actions. If any data source is unavailable, the pipeline retains the previous values and continues — ensuring the tool remains available even during source outages.

Uncertainty & Confidence

Every pressure index is reported with an uncertainty band (the lighter shaded range on each stressor card). These bands are derived from 500 Monte Carlo iterations with ±20% random perturbation of all weights.

This uncertainty margin is deliberately wider than what you'd see in a mechanical causal model (like tracking commodity prices through a supply chain). Mental health pathways are mediated, moderated, and confounded — the wider bands are an honest reflection of that complexity.

Data confidence tiers

Each country is assigned a data confidence level that affects precision of reporting:

• High confidence — raw values used (countries with comprehensive OECD/FRED data)
• Medium confidence — values rounded to nearest 5% (partial data coverage)
• Low confidence — values rounded to nearest 10% (limited or delayed data)

All 10 Phase 1 countries currently have high-confidence data.

Limitations

• Ecological inference: Clover operates at the country level. Within-country variation (urban vs. rural, regional economies) can be substantial. A national pressure index of 42% does not mean every community faces 42% pressure.
• Temporal lag: Economic indicators update monthly; mental health outcomes in populations may take weeks to months to manifest. Clover tracks leading indicators (economic data), not lagging outcomes (mental health surveys).
• Baseline sensitivity: The 2019–2020 baseline was chosen for scientific cleanliness, but it wasn't perfect everywhere — some countries were already experiencing housing stress, for example. The baseline is a reference point, not a claim that 2019 was "fine."
• Effect size variation: Published effect sizes are population averages from specific study contexts. Actual transmission from stressor to mental health outcome varies by sub-population, co-occurring stressors, and protective factors.
• Confounding: Economic stressors rarely occur in isolation. Housing cost burden, inflation, and debt are correlated. Clover does not attempt to decompose shared variance — each stressor is measured independently, which means the total pressure across all stressors is not simply additive.
• Phase 1 scope: Only economic stressors are included. Environmental stressors (air quality, extreme weather) and social stressors (isolation, digital overload) are planned for future phases.

Full References

• Bentley, R. et al. (2011). Association between housing affordability and mental health: A longitudinal analysis of 35 Australian surveys. Social Science & Medicine, 73(12), 1771–1780.
• Chapman, B. & Dearden, L. (2017). Conceptual and empirical issues for alternative student loan designs. Economics of Education Review, 57, 42–54.
• Chetty, R., Hendren, N. & Katz, L.F. (2016). The effects of exposure to better neighborhoods on children. American Economic Review, 106(4), 855–902.
• Clark, A.E. & Oswald, A.J. (1994). Unhappiness and unemployment. The Economic Journal, 104(424), 648–659.
• Cooper, K. & Stewart, K. (2013). Does money affect children's outcomes? A systematic review. Joseph Rowntree Foundation.
• Drentea, P. (2000). Age, debt and anxiety. Journal of Health and Social Behavior, 41(4), 437–450.
• Fitch, C. et al. (2011). Debt and mental health: The role of psychiatrists. Advances in Psychiatric Treatment, 17(2), 85–92.
• Frasquilho, D. et al. (2016). Mental health outcomes in times of economic recession: A systematic literature review. BMC Public Health, 16, 115.
• Giorgi, G. et al. (2020). Economic stress in the workplace: The impact of fear of the crisis on mental health. Work, 65(1), 187–197.
• Liu, S. et al. (2014). Effectiveness of job search interventions: A meta-analytic review. Psychological Bulletin, 140(4), 1009–1041.
• Lusardi, A. & Mitchell, O.S. (2014). The economic importance of financial literacy. Journal of Economic Literature, 52(1), 5–44.
• Milner, A. et al. (2013). Employment status and mental health: A meta-analysis. Social Science & Medicine, 95, 275–283.
• Norström, T. & Grönqvist, H. (2015). The Great Recession, unemployment and suicide. Journal of Epidemiology and Community Health, 69(2), 110–116.
• Patel, V. et al. (2018). The Lancet Commission on global mental health and sustainable development. The Lancet, 392(10157), 1553–1598.
• Paul, K.I. & Moser, K. (2009). Unemployment impairs mental health: Meta-analyses. Journal of Vocational Behavior, 74(3), 264–282.
• Pickett, K.E. & Wilkinson, R.G. (2015). Income inequality and health: A causal review. Social Science & Medicine, 128, 316–326.
• Pleasence, P. & Balmer, N.J. (2012). Caught in the web? Legal information, advice and support. Legal Education Review, 22(1), 201–226.
• Puig-Barrachina, V. et al. (2020). The impact of Active Labour Market Policies on health outcomes. BMC Public Health, 20, 1799.
• Richardson, T. et al. (2013). The relationship between personal unsecured debt and mental and physical health: A systematic review and meta-analysis. Clinical Psychology Review, 33(8), 1148–1162.
• Rohde, N. et al. (2016). The effect of economic insecurity on mental health: Recent evidence from Australian panel data. Social Science & Medicine, 151, 250–258.
• Sweet, E. et al. (2013). The high price of debt: Household financial debt and its impact on mental and physical health. Social Science & Medicine, 91, 94–100.
• WHO (2014). Social Determinants of Mental Health. World Health Organization & Calouste Gulbenkian Foundation.

Open Source

Clover is open source under the Apache 2.0 license. The full codebase, data pipeline, and historical snapshots are available on GitHub. We welcome scrutiny of our methodology — if you find an error or have a suggestion, please open an issue.