Setting Thresholds for Variables

Each variable used in the analysis will have an implicit "threshold", where risk begins to accelerate.  For example, in Emerging Markets, current account deficits become worrisome when they exceed 4% of GDP.  Often these deficits are not financeable above 6% of GDP. Similarly, BIS research shows that rapid credit growth is one of the best indicators of looming financial crisis. If private credit growth exceeds 15% per annum, the likelihood of future financial crisis escalates sharply.

To enable variables of different types to be compared, the analyst muct convert the numeric score for each variable (e.g. rate of inflation) into a standardized Risk Scale (e.g. from High to Medium to Low Risk, or from 9 to 1, High to Low.) These scaled-scores can then be compared across countries, to show - for each variable - where risk is higher.

Thresholds for Scoring

Weighting and Scaling the Variables

Most rating agencies place heavy emphasis on Economic and Political Structure indicators (e.g. Per Capita Income or Governance measures) as Credit Rating Agencies tend to have a longer term (3-7 year) time horizon. Consequently, the CRAs tend to give less emphasis to short-term liquidity or funding issues which can be fatal to borrowers.  This was a serious weakness during the European sovereign crises -- one that yet to be resolved.

Sensitivity to Sector Risk

Scholars have shown that banking, fiscal and exchange rate  crises have different causes and etiologies.  At the same time, the past three decades shows that a crisis erupting in one sector may quickly spread to others, or to the entire region. Thus, a careful weighting of risk variables is essential.

Risk Begins in Sectors

Risk Scores and Asset Values

While most rating scales are linear (e.g. AAA to C, or 1 to 9),  ACTUAL RISK IS NOT LINEAR.  At lower risk levels (AAA or AA),  modest changes in risk -- indicated by a one or two notch rating move -- can increase borrowing costs by a few basis points. At higher risk levels, the impact of a rating move is significant. During non-crisis periods, it can lift borrowing costs by hundreds of basis points, i.e. several percentage points.  At the highest risk levels, the score reflects expected losses from default, restructuring or long-term failure to pay.  This is often measured in terms of the "recovery value" of the bond or loan, once the borrower begins to repay.  Historically, recovery values tend to fall in the 30% range (30 cents on the dollar), with severely troubled sovereign assets priced at 15 cents or less.  This highlights the need for care when using Country Risk Scores for applications like regulatory minimum capital calculations. 

applications of risk scores

---