Serbia Equity Risk Premium Decomposition (2015–2025)
Description
This dataset provides quarterly estimates of Serbia’s equity risk premium (ERP) decomposed into four structural components: sovereign risk (CRP), liquidity risk (LP), behavioral or sentiment risk (BP), and institutional quality risk (IQP). The series cover the period 2015Q1–2025Q3 and are expressed in percentage points, capturing how each premium contributes to Serbia’s total cost of equity. The data were constructed from multiple reliable sources, including CEIC Data, the Belgrade Stock Exchange (BELEX), Bloomberg, Damodaran’s country risk spreadsheets, and governance indicators from the World Bank (WGI) and Transparency International (CPI). Each component is estimated using a transparent, data-driven empirical framework: The sovereign risk premium (CRP) follows Damodaran’s volatility-adjusted spread method. The liquidity premium (LP) is derived from principal component analysis of bid–ask spreads, turnover ratios, and Amihud illiquidity metrics. The behavioral premium (BP) is estimated from an EGARCH-in-Mean model capturing volatility feedback in BELEX15 returns. The institutional quality premium (IQP) aggregates standardized governance and corruption indices through principal component analysis. The dataset supports the study “Decomposing Serbia’s Equity Risk Premium: Liquidity and Institutional Constraints in a Frontier Market” and enables replication or extension of the analysis for other emerging and frontier markets. It provides insight into how structural factors—rather than macroeconomic volatility alone—shape equity pricing in small, underdeveloped capital markets. The accompanying figure (“Decomposition of Serbia’s ERP, 2015–2025”) visualizes the quarterly evolution of the four components. Variables: Date (Quarter, 2015Q1–2025Q3) CRP – Sovereign Risk Premium (%) LP – Liquidity Premium (%) BP – Behavioral Premium (%) IQP – Institutional Quality Premium (%)
Files
Steps to reproduce
1) Collect data (2015Q1–2025Q3). Sovereign USD-bond yields for Serbia and 10-year U.S. Treasury (Bloomberg/CEIC). BELEX15 index levels/returns, bid–ask spreads, and turnover (BELEX). Governance indicators (WGI sub-indices) and CPI (Transparency International). Global equity series for GERP context (MSCI World/S&P 500). Damodaran (2024) country-risk sheet for base parameters and spreads. 2) Prepare series (quarterly). Convert all inputs to quarterly frequency (calendar quarter end). Compute BELEX15 excess returns (USD or domestic with consistent risk-free). Align time indexes; forward-fill within quarter only when the source is lower frequency (e.g., annual WGI/CPI → quarter-constant). 3) Construct components. CRP: CRP = Sovereign Spread × (σ_equity / σ_bond), where σ are rolling volatilities consistent across assets. LP: Build a composite liquidity index via PCA of Amihud illiquidity, bid–ask spread, and (1/turnover). Rescale to mean 0, SD 1; map to % premium using the calibration in step 5. BP: Estimate an EGARCH-in-Mean on quarterly BELEX15 excess returns; use the conditional variance term in mean (and/or a standardized volatility-of-vol proxy) as the behavioral premium. IQP: PCA of standardized WGI sub-indices (rule of law, regulatory quality, government effectiveness) and CPI; first component (higher = weaker governance) mapped to % premium. 4) Baseline parameters. Adopt Rf = 4% and GERP = 5% (Damodaran, 2024) for comparability. Ensure all component series are in percentage points. 5) Calibration & scaling. Scale LP, BP, IQP to economically reasonable magnitudes using in-sample alignment to valuation anchors (inverse P/E, earnings yield), targeting aggregate ERP consistent with observed yields. Document any linear scaling factors. 6) Total ERP & regression checks. Compute rₑ = Rf + β(GERP) + CRP + LP + BP + IQP. Validate with: (i) fit to inverse P/E and earnings yield; (ii) lagged OLS with Newey–West errors: ERPₜ = α + ϕ₁CRPₜ₋₁ + ϕ₂LPₜ₋₁ + ϕ₃BPₜ₋₁ + ϕ₄IQPₜ₋₁ + εₜ; (iii) rolling-window robustness. 7) Visualization. Produce a quarterly stacked-area chart of CRP, LP, BP, IQP contributions and optionally a cross-country bar comparison (Hungary, Turkey, Romania, Bosnia) using the same methodology. 8) Files & software. All final series are in Serbia_FMVM_Quarterly.xlsx (single sheet: Date, CRP, LP, BP, IQP). Replication can be done in Python (pandas, statsmodels, matplotlib) or R (data.table, rugarch, FactoMineR/psych).