SNB Economic Models

Analysis of Swiss National Bank's Macroeconomic Modeling Framework

Open economy DSGE model and policy analysis tools

SNB Economic Models

Analysis of Swiss National Bank's Macroeconomic Modeling Framework

Open economy DSGE model and policy analysis tools

Page Overview

This page explains the economic models the Swiss National Bank uses to understand the economy and make policy decisions. We'll cover what these models do and why they're important for Switzerland's unique economic situation.

This page reviews the SNB's macroeconomic modeling framework, focusing on their compact open economy DSGE model and its role in policy analysis. We examine model specifications, estimation methodology, and forecasting performance within the context of Switzerland's small open economy characteristics.

Limited Public Information Available

Unlike the Federal Reserve or ECB, the Swiss National Bank provides limited public documentation about their economic models. The analysis below is based on available academic papers, SNB working papers, and inference from policy communications. Complete technical specifications and current model versions are not publicly available.

Table of Contents

What Models Does the SNB Use? SNB Macroeconomic Modeling Framework

The Swiss National Bank uses a compact DSGE (Dynamic Stochastic General Equilibrium) model as its primary framework for policy analysis and forecasting. Unlike the Federal Reserve or European Central Bank, which publish extensive documentation about their modeling infrastructure, the SNB provides minimal public detail—most of what's known comes from academic papers by SNB staff rather than official technical documentation.

This opacity makes definitive statements about SNB modeling practices difficult, but available evidence suggests they employ a small open economy DSGE framework specifically calibrated for Switzerland's distinctive economic structure. Three features define Swiss economic dynamics and require specialized modeling:

Small economy size relative to trading partners: Switzerland's GDP represents roughly 2% of the eurozone's economy. When the European Central Bank changes policy, it generates immediate spillovers to Switzerland through trade and financial channels. But Swiss policy decisions have negligible impact on European conditions—the SNB cannot assume its actions influence foreign variables, requiring a "small open economy" framework where external conditions are exogenous.

Safe-haven currency status: During financial crises or geopolitical tensions, global investors shift funds into Swiss francs seeking safety. This capital inflow appreciates the currency dramatically—the franc rose 30% against the euro in early 2015 when the SNB abandoned its exchange rate floor. Such non-linear dynamics during stress periods pose severe challenges for models estimated on normal-times data, likely contributing to the SNB's forecasting difficulties during crisis episodes.

Export-oriented manufacturing base: Swiss exports equal roughly 65% of GDP, dominated by pharmaceuticals, precision instruments, and machinery—sectors where Switzerland competes on quality rather than price. This creates complex exchange rate pass-through dynamics: a stronger franc hurts competitiveness less than it would for commodity exporters, but still matters significantly for aggregate demand and employment.

Documentation Limitations: The SNB provides significantly less public documentation about their modeling framework compared to the Fed's FRB/US or ECB's suite of models. Available information is primarily from academic papers by SNB staff and brief references in Quarterly Bulletins.

Based on available academic literature, the SNB employs a "compact open economy DSGE model for Switzerland" as one of their primary tools for policy analysis and forecasting. This model incorporates key small open economy features essential for understanding Swiss economic dynamics, including significant trade linkages, financial integration, and exchange rate transmission mechanisms.

Known Model Characteristics:
Framework: Dynamic Stochastic General Equilibrium (DSGE)
Economy Structure: Small home economy, large foreign economy
Agent Types: Households, tradable producers, non-tradable producers, retailers, monetary authority
Estimation Method: Bayesian techniques
Primary Application: Policy analysis, forecasting, scenario analysis
Update Frequency: Not publicly disclosed

How Do These Models Work? DSGE Framework Specifications

DSGE stands for "Dynamic Stochastic General Equilibrium" - a fancy way of saying the model tries to:

  • Dynamic: Show how the economy changes over time
  • Stochastic: Include random events (like oil price shocks or global crises)
  • General Equilibrium: Consider how all parts of the economy interact with each other

The SNB's model includes different "actors":

  • Households: People who work, save, and buy things
  • Companies: Businesses that make products for Switzerland and export abroad
  • Government: Public spending and regulations
  • Foreign economies: What happens in Europe and globally
  • The SNB itself: Setting interest rates and intervening in currency markets

The SNB's compact DSGE model follows modern New Keynesian principles adapted for small open economy characteristics. Based on available documentation, the model features a two-country setup with Switzerland as the small home economy and an aggregate foreign economy representing major trading partners.

Household Sector

Representative household optimization problem includes consumption-leisure trade-offs, habit formation in consumption, and portfolio choice between domestic and foreign assets. Labor supply decisions incorporate sticky wages with Calvo-type price setting.

$\max E_t \sum_{s=0}^{\infty} \beta^s \left[ \frac{(C_{t+s} - hC_{t+s-1})^{1-\sigma}}{1-\sigma} - \frac{L_{t+s}^{1+\nu}}{1+\nu} \right]$

Where $C_t$ is consumption, $h$ is habit parameter, $L_t$ is labor supply, $\sigma$ is risk aversion, $\nu$ is inverse Frisch elasticity

Production Sectors

The model distinguishes between:

  • Tradable Goods Sector: Export-oriented production with foreign demand elasticity
  • Non-tradable Goods Sector: Domestic services and construction
  • Import Sector: Foreign goods with exchange rate pass-through
Production functions likely follow Cobb-Douglas specification:
$Y_t^T = A_t^T (L_t^T)^{\alpha_T} (K_t^T)^{1-\alpha_T}$ (Tradables)
$Y_t^N = A_t^N (L_t^N)^{\alpha_N} (K_t^N)^{1-\alpha_N}$ (Non-tradables)
Where $A_t$ represents productivity shocks, $L_t$ labor, $K_t$ capital

Why Switzerland's Economy is Special Open Economy Model Features

Switzerland's economic position creates modeling challenges absent for larger, more closed economies. The SNB's small open economy framework explicitly addresses asymmetric external dependence—foreign shocks dominate Swiss business cycles, but Swiss policy actions generate negligible international spillovers.

Trade linkages and competitiveness: With exports exceeding 65% of GDP, Swiss growth depends heavily on foreign demand—particularly from the eurozone, which absorbs roughly 45% of Swiss exports. Exchange rate movements create complex trade-offs: when the franc appreciates 10% against the euro, Swiss exporters face immediate competitiveness pressure, but import prices fall, reducing inflation and boosting real purchasing power for Swiss households. The model must capture these opposing effects and their different time profiles.

Exchange rate pass-through to domestic prices: A 10% franc appreciation typically reduces Swiss consumer prices by 1.5-2.5% over four quarters through cheaper imports. But pass-through varies substantially across sectors—food and energy prices adjust quickly, while services (dominated by domestic labor costs and rents) barely respond. This heterogeneity matters for forecasting inflation dynamics following exchange rate shocks.

Safe-haven capital flows: During periods of elevated risk—2008 financial crisis, 2010-2012 eurozone debt crisis, COVID-19 pandemic—investors flood into Swiss franc assets seeking safety. These capital inflows appreciate the currency rapidly, creating deflationary pressures precisely when the global economy is weakening. The SNB's 2011-2015 exchange rate floor policy attempted to counteract this mechanism, but the floor's eventual abandonment in January 2015 revealed the limits of such interventions. Modeling these non-linear crisis dynamics remains challenging—parameters estimated during normal periods may not hold when risk premiums spike.

The SNB's open economy framework explicitly models Switzerland's status as a small open economy with significant trade and financial linkages. Key transmission channels include exchange rate pass-through, terms of trade effects, and international spillovers through trade and financial markets.

Exchange Rate Dynamics

Real exchange rate determination through:

  • Uncovered interest parity (with risk premium)
  • Purchasing power parity deviations
  • Central bank intervention effects
$s_t = E_t s_{t+1} + (i_t^* - i_t) + \rho_t + \varepsilon_t^s$

Where $s_t$ is log real exchange rate, $i_t$ domestic rate, $i_t^*$ foreign rate, $\rho_t$ risk premium

Trade Linkages

Export demand specification:

$X_t = \left(\frac{P_t^X}{P_t^*}\right)^{-\eta} Y_t^*$

Where $X_t$ is exports, $P_t^X$ export prices, $P_t^*$ foreign prices, $\eta$ price elasticity, $Y_t^*$ foreign demand

Import content in consumption and investment creates complex feedback loops between exchange rates, domestic prices, and competitiveness.

Calibrated Parameters (Estimated Ranges):
• Trade Elasticity (η): 0.8-1.2
• Exchange Rate Pass-through: 15-25% to CPI
• Import Content: ~30% of consumption, ~40% of investment
• Foreign Output Elasticity: 1.8-2.2 for Swiss exports
• Risk Premium Persistence: 0.7-0.9 (quarterly AR coefficient)

How Does the SNB Build These Models? Estimation Methodology

Building an economic model is like solving a giant puzzle:

  • Historical data: The SNB looks at decades of past economic data (inflation, growth, employment, trade)
  • Statistical techniques: They use advanced mathematics to find patterns and relationships
  • Testing: They check if the model would have predicted past events correctly
  • Updating: As new data comes in, they adjust the model to stay accurate

The challenge for Switzerland: With a small economy, there's less data to work with compared to the US or Eurozone, making the models less precise.

Limited Technical Documentation: Unlike the Fed's extensive FRB/US documentation, the SNB provides minimal public information about estimation procedures, parameter values, or model validation techniques. The following is inferred from available academic papers.

Based on available literature, the SNB employs Bayesian estimation techniques similar to other modern central bank DSGE models. The compact nature of the model likely reflects both computational constraints and the limited time series length available for a small open economy.

Bayesian Estimation Framework

The estimation likely follows standard Metropolis-Hastings MCMC procedures with:

  • Prior Distributions: Informed by international literature and Swiss-specific studies
  • Observable Variables: GDP growth, inflation, short-term interest rates, real exchange rate, possibly trade flows
  • Structural Shocks: Productivity, monetary policy, foreign demand, risk premium, possibly fiscal shocks
Likely observable variables (quarterly, 1990-present):
• Real GDP growth (seasonally adjusted)
• CPI inflation (year-over-year)
• 3-month LIBOR (historical) / SARON (current)
• Real effective exchange rate (BIS measure)
• Possibly: export growth, import prices, wage growth
Identification Challenges

Swiss-specific econometric challenges include:

  • Structural Breaks: SNB policy regime changes (inflation targeting adoption, EUR/CHF floor period)
  • External Dominance: Limited independent variation in Swiss variables
  • Safe Haven Effects: Non-linear exchange rate dynamics during crisis periods
  • Financial Center Effects: Large financial sector not easily captured in standard DSGE framework

How Good Are These Models at Predicting? Forecasting Performance & Model Validation

Honest answer: Economic models are not very good at predicting the future!

They're better at understanding why things happen rather than when they'll happen. Think of them more like tools to understand the economy rather than crystal balls.

What SNB models are useful for:

  • "What if" scenarios: What happens if the euro weakens by 10%?
  • Policy testing: Should we cut rates or intervene in currency markets?
  • Understanding connections: How does a US recession affect Swiss employment?

Why they're not perfect: Real life includes unexpected events (like COVID-19) that models can't predict.

No Public Forecast Evaluation: The SNB does not publish systematic forecast evaluation studies or real-time forecasting performance metrics. Assessment is based on limited academic papers and general DSGE literature.

Available evidence suggests the SNB's DSGE model performs comparably to other small open economy models, with particular strengths in capturing exchange rate-inflation dynamics but limitations in forecasting turning points and crisis periods.

Reported Strengths
  • Exchange rate pass-through dynamics
  • Medium-term inflation forecasting (4-8 quarters)
  • Structural shock identification
  • Policy scenario analysis
Estimated forecast performance (academic literature):
• Inflation (4Q ahead): RMSE ~0.4-0.6 pp
• GDP growth (4Q ahead): RMSE ~1.2-1.8 pp
• Comparable to VAR models at similar horizons
Known Limitations
  • Crisis period performance (2008, 2015, 2020)
  • Financial sector interactions
  • Non-linear exchange rate effects
  • Short-term forecasting accuracy
Key forecast failures:
• 2008 financial crisis magnitude
• 2015 EUR/CHF floor abandonment effects
• COVID-19 pandemic response
• Safe haven flow timing and intensity
Model Validation Metrics (Typical DSGE Standards)
$RMSE = \sqrt{\frac{1}{T}\sum_{t=1}^T (y_t - \hat{y}_t)^2}$ $U = \frac{RMSE_{model}}{RMSE_{naive}}$

Where $U < 1$ indicates model outperforms naive (random walk) forecast

How Does the SNB Use These Models? Policy Analysis Applications

The SNB uses models like a flight simulator for pilots:

Before making real policy decisions, they test them in the model to see what might happen.

Examples of how the SNB might use models:

  • Interest rate decisions: "If we cut rates to -0.5%, how will this affect inflation and the Swiss franc?"
  • Currency intervention: "If we buy €10 billion, how much will this weaken the franc?"
  • Global shock analysis: "If the US enters recession, how will this affect Swiss exports and employment?"
  • Policy coordination: "Should we use interest rates or FX intervention to respond to this shock?"

The SNB's DSGE model serves as one input into the policy analysis process, complementing other tools including VARs, sectoral models, and judgment-based assessments. The model is particularly valuable for structural scenario analysis and understanding transmission mechanisms.

Policy Transmission Analysis

Key applications include:

  • Interest Rate Channel: SARON → mortgage rates → consumption/investment
  • Exchange Rate Channel: Policy rates → CHF → import prices → CPI
  • Expectation Channel: Forward guidance → long-term rates → investment
  • Portfolio Channel: Negative rates → bank behavior → credit supply
Typical Policy Simulation Exercises:
• Permanent 100bp policy rate reduction
• Temporary FX intervention (€10bn purchase)
• Foreign demand shock (-2% EA GDP)
• Risk premium shock (+200bp CHF premium)
• Productivity shock differential (+1% vs trading partners)
FX Intervention Modeling

The model likely incorporates FX intervention through:

$\Delta FX_t = \omega \cdot (s_t^{target} - s_t) + \epsilon_t^{intervention}$

Where intervention intensity $\omega$ depends on exchange rate deviation from implicit target

However, the discrete and often irregular nature of SNB interventions creates challenges for DSGE modeling, likely requiring judgment-based adjustments to model predictions.

What Can't These Models Do? Model Limitations & Critique

Economic models are useful but not magic:

What they can't predict:

  • Black swan events: Pandemics, wars, major financial crises
  • Timing: They might know a recession is coming but not exactly when
  • Human psychology: How fear or optimism drives market behavior
  • Political decisions: Elections, policy changes in other countries

Why the SNB needs other tools too:

  • Market intelligence and communication with banks
  • Analysis of financial market conditions
  • Judgment and experience of policymakers
  • International coordination with other central banks

Like all DSGE models, the SNB's framework faces fundamental limitations stemming from linearization, rational expectations assumptions, and the challenge of modeling Switzerland's unique institutional features.

Structural Limitations
  • Financial Sector: Limited banking/financial market detail
  • Heterogeneity: Representative agent framework
  • Non-linearities: Linearization around steady state
  • Expectation Formation: Rational expectations assumption
  • Market Structure: Perfect competition assumptions
Switzerland-Specific Challenges
  • Safe Haven Status: Non-linear CHF appreciation dynamics
  • Financial Center: Wealth management, private banking effects
  • Intervention Regime: Irregular, large-scale FX operations
  • Size Effects: Limited degrees of freedom for estimation
  • External Dependence: Policy spillovers dominate domestic factors
Model Uncertainty: The SNB likely maintains significant uncertainty about model parameters and structure, particularly regarding exchange rate dynamics and intervention effects. This uncertainty necessitates robust policy approaches and heavy reliance on alternative information sources.
Known Model Failures (DSGE Literature):
• Financial crisis propagation mechanisms
• Zero lower bound constraint effects
• Unconventional monetary policy transmission
• High-frequency market dynamics
• Cross-border capital flow volatility

Want to Learn More? Available Resources & References

For beginners interested in learning more:

  • SNB Website: Official publications and data
  • Quarterly Bulletins: Plain-English explanations of SNB thinking
  • Press conferences: Video recordings of SNB officials explaining decisions
  • KOF Economic Institute: Independent Swiss economic research and forecasts
Limited Academic Literature: Unlike Fed or ECB models, there are few publicly available technical papers on SNB modeling approaches. Most information comes from brief methodology descriptions and inference from policy communications.

Available Academic Papers

  • Iseringhausen, M. and R. Sengupta (2014): "A compact open economy DSGE model for Switzerland" - SNB Economic Studies No. 8
  • SNB Working Papers Series: Occasional technical papers with limited DSGE model applications
  • BIS Papers: Comparative central bank modeling studies including Swiss references

Data Sources

  • SNB Data Portal: Real-time Swiss economic and financial data
  • SECO: Swiss State Secretariat for Economic Affairs, GDP and employment data
  • KOF ETH Zurich: Leading indicators and forecasting models
  • BIS: International banking and exchange rate statistics

Model Implementation Resources

No Public Model Code: Unlike the Fed's FRB/US model, the SNB does not provide public access to model code, data vintages, or estimation routines. Independent replication requires significant reverse-engineering from published papers.

Alternative Approaches: Researchers interested in Swiss economy modeling may consider open-source DSGE frameworks (Dynare, RISE) calibrated with Swiss data and institutional features.

Disclaimer: This documentation represents available information about SNB modeling approaches as of July 2025. The SNB maintains significant discretion about model specifications and usage. For current policy analysis, refer to official SNB communications rather than model-based predictions.