Bond Indices as Strategic Asset Allocation Benchmarks

Central-bank and sovereign reserve managers face a sharper version of the problem this workspace is built around: turn an observed or modelled yield curve into a defensible allocation. A presentation from the National Bank of Georgia’s Risk Management and Control Division, "Bond Indices as Strategic Asset Allocation Benchmarks," walks through that process end to end, and several of its building blocks map directly onto the models here. Its central chart shows how a return distribution’s risk changes with holding period: a 1–3Y government index is high-risk over one month but low-risk over five years, while a 10Y+ index stays risky even three years out. That is the same mechanic behind the Forecast and Scenario Returns tools — simulate the curve forward under a chosen drift regime, then read the terminal distribution off the holding period, rather than quoting volatility as one fixed number. Its risk/return table across five currencies — expected return, volatility, and CVaR at both 95% and 99% for maturities from three months to five years — is close to what the Scenario Returns tool now produces for a set of bond durations at once, with confidence levels you choose yourself. The deck’s own numbers show the gap: for the USD government index, 5Y CVaR runs –2.98% at the 95% level but –4.82% at 99%, a reminder that a single VaR figure hides how much worse the tail gets. The most transferable finding is on correlation: within one currency, adjacent maturities move almost as one (0.94–0.99 for adjacent USD points), so stacking bonds along a single curve adds little diversification — currency, not duration, is the real diversifier. That is exactly what the Portfolio tool’s covariance and correlation matrices are built to surface: price every duration bucket off the same simulated curve, then see what actually moves independently versus together. Its portfolio-construction slide sweeps Conservative → Balanced → Aggressive allocations across five maturity buckets to trace a full risk-return trade-off — the same efficient frontier the Portfolio Optimization tool solves via Markowitz mean-variance, here on a Nelson-Siegel-simulated curve rather than a historical bootstrap. The full 18-slide deck, including the reserve-tranche benchmark framework, active-return attribution, and a worked example converting a GEL portfolio’s return into USD, is attached below.