Let's start with units
Let the conditional yield of the security be equal to 9,5%, and that of the comparable benchmark to be 6,5%. The difference is 3 percentage points, or 300 basis points. One basis point is equal to one hundredth of a percentage point. You cannot call this difference “three percent relative to the benchmark”: the relative increase has a different denominator and will give a different number. Specifying bps is convenient specifically for absolute differences in rates. All values in the article are fictitious and do not represent the current profitability of any company or state.
After the selected conditional interval, let the yield of the security increase to 10%, and the benchmark - to 8%. The new spread is equal to 200 basis points, that is, it has narrowed by 100. At the same time, the profitability of the paper itself increased. The example shows why the phrase “spread has decreased” is not the same as “yield has decreased.” A payment calendar and other conditions will also be required to explain the price. It is impossible to reconstruct the full movement of a quote from one difference between two rates without information about their individual changes.
The reference point must answer the question
It is necessary to agree on the currency, term or interest sensitivity, type of return and observation date. Sometimes the calculated point of the curve is used for the required period, rather than a separate tradable security. Then the interpolation technique is important. For an instrument with a callable right, a simple comparison of the yield to final maturity may confuse the credit consideration with the value of the right. The SEC describes a variety of corporate debt terms;[14](#source-cm-corporate) the same issuer does not make all its flows and rights the same.
The credit premium is also not a pure estimate of the expected loss. The observed difference may include liquidity, supply and demand, collateral features and other factors. FINRA looks at how spreads change when both the security and the benchmark change.[19](#source-cm-spread) Therefore, dividing the spread by an arbitrarily chosen loss at default does not provide a reliable probability of the event without a separate model and assumptions. If such a calculation is made, the limitations, horizon and origin of all parameters must be clearly shown.
Different spreads are not interchangeable
Bid–ask spread refers to the difference in counter trading prices and is of a different nature. A credit spread expresses a relative estimate of yields. Floating coupon surcharge is a payment formula parameter. They can be found on the same card, but cannot be added or compared as one indicator just because of the common word. [Coupon and Yield Guide](/guides/coupon-versus-yield/) addresses the basis of interest flow, and [Interest Risk](/guides/bond-rate-risk/) helps separate base rate changes from other reasons for debt valuation.
The verification record must contain the security's yield, benchmark, date, methodology, units and justification for comparability. It is then useful to show which of the two rates has changed and by how much. If the comparison is historical, the methodology must remain consistent. [Article about charts](/guides/chart-literacy/) warns about false conclusions for different series and periods. The material does not promise that a wide spread will compensate for all losses, while a narrow spread confirms safety. The number only becomes useful in conjunction with a definition, not in place of an analysis of the terms of the debt.
Checking the Spread Change Explanation
If the spread has increased, first find out which rate has changed: the paper, the benchmark, or both. The direction of the difference alone is not enough to confidently attribute the entire movement to issuer deterioration. For example, the benchmark may change for reasons unrelated to the specific company. This is not to say that credit quality doesn't matter, but rather to avoid passing off one possible cause as an established one.
Another test is to compare two values calculated relative to different landmarks. Their arithmetic difference is not necessarily a change in credit compensation. The same base or an explicit transition between techniques is required. If the paper life is shortened, the use of the previous curve point also requires explanation. A valid record retains all original bets, dates and properties, not just the final spread. If there is no data for comparability, the conclusion is limited. This approach is also useful when reading a beautiful historical graph: a continuous line does not guarantee a continuous methodology. The spread itself remains a measure of relative valuation rather than a contractual promise of return or a credible probability of default.