Diagnosing Timing, Sign, and Rate-Convention Errors

Why Integrated Problems Need a Diagnostic Check

A multi-date project can be evaluated with correct arithmetic and still produce a wrong answer if a receipt is assigned to the wrong date, an outlay’s sign is reversed, or a quoted interest rate is applied over the wrong period. These errors often appear together: a candidate may compute a familiar-looking present value while discounting with a rate that does not match the cash-flow timeline.

A reliable check starts with the dates and signs, then identifies the rate applying to each interval or maturity. Only after those choices are clear should the present values be added. This is especially important under a nonflat term structure: different dated cash flows generally require different spot rates, so a single yield or duration may not answer the question being asked.

Definition: With the investor’s outflows recorded as negative amounts and inflows as positive amounts, the net present value at time 0 is the sum of each dated cash flow discounted to time 0. If $C_t$ is the net cash flow at time $t$ and $s_t$ is the effective annual spot rate for maturity $t$, then, for annual payment dates and annual compounding, $$ \operatorname{NPV}_0=\sum_{t=0}^{n}\frac{C_t}{(1+s_t)^t}, $$ where the time-0 term is $C_0$. This form assumes the stated spot rate applies to the full interval from time 0 to time $t$. If the rates are quoted on a different compounding basis or cash flows occur at other dates, first use a discount factor consistent with that basis and date.

Start With Dates and Signs

Write each net cash flow on a timeline before evaluating it. A receipt at the end of year 2 is discounted for two years, not one. A cash flow at time 0 is not discounted. With the convention above, an initial investment is negative; a positive NPV means the discounted inflows exceed the discounted outflows at the rates used. If a problem instead defines outlays as positive, the reported NPV sign is reversed, so the convention must be stated before interpreting the result.

Worked Example: A Receipt Placed One Year Too Early

A proposed project requires an investment of $1,000 now, pays $500 at the end of year 1, and pays $600 at the end of year 2. The effective annual spot rate is 5% for one year and 8% for two years. Find the time-0 NPV, then quantify the error from discounting the year-2 receipt for only one year.

Method. Use negative signs for the investment and positive signs for receipts. Discount each receipt according to its actual maturity.

The cash flows are $C_0=-1000$, $C_1=500$, and $C_2=600$. The one-year receipt has present value $500/1.05=476.190476$. The two-year receipt has present value $600/(1.08)^2=600/1.1664=514.403292$, to six decimal places.

\operatorname{NPV}_0=-1000+476.190476+514.403292=-9.406232

The project’s NPV is approximately $-\$9.41$ to the nearest cent. If the year-2 receipt is incorrectly treated as a year-1 receipt, its present value is $600/1.05=571.428571$, and the erroneous NPV is $-1000+476.190476+571.428571=47.619047$. The timing mistake changes the result by $47.619047-(-9.406232)=57.025279$, using the displayed rounded amounts. The calculation avoids the common mistake of using the one-year rate or one-year discount period for every receipt.

Check. The year-2 discount factor is smaller than the year-1 discount factor, so the year-2 receipt’s present value must be below $600/1.05$. The computed value, 514.403292, satisfies that check.

A sign error is different from a timing error. If all the cash-flow signs are reversed, the computed NPV is multiplied by $-1$, but the underlying project has not changed. A useful final check is to state what a positive or negative answer means under the chosen convention, rather than reporting only a number.

Match the Rate Convention to the Cash-Flow Interval

An effective annual rate applies over one year. A nominal annual rate convertible $m$ times per year is divided by $m$ to obtain the effective rate per conversion period, provided the cash-flow interval is measured in those periods. Thus a nominal annual rate $i^{(m)}$ gives a period rate $i^{(m)}/m$, and a cash flow $k$ such periods away is discounted by $(1+i^{(m)}/m)^k$. Do not divide a rate by the number of periods unless the rate is stated to be nominal with that conversion frequency.

Worked Example: Semiannual Compounding and a One-Year Receipt

A project costs $450 now and pays $500 at the end of one year. The discount rate is a nominal annual rate of 10% convertible semiannually. Calculate the NPV, and compare it with the result from incorrectly treating 10% as an effective annual rate.

Method. Convert the nominal rate to the semiannual period rate, then count the number of periods to the payment. A one-year interval contains two half-year periods.

The effective rate per half-year is $0.10/2=0.05$. Therefore, the year-1 payment has present value $500/(1.05)^2=500/1.1025\approx\$453.514739$, or $453.51 to the nearest cent.

\operatorname{NPV}_0=-450+453.514739=3.514739

The NPV is approximately $3.51 to the nearest cent. Equivalently, the effective annual rate implied by 10% nominal convertible semiannually is $(1.05)^2-1=0.1025$, or 10.25%. If 10% is incorrectly treated as an effective annual rate, the payment’s present value is $500/1.10=\$454.545455$, giving an erroneous NPV of $4.545455. The discrepancy arises because the two annual rates represent different accumulation over a year.

Check. The correct effective annual rate, 10.25%, is greater than 10%; consequently, its one-year discount factor is larger and its present value is smaller than $500/1.10$. The computed value of $453.514739 is smaller than $454.545455, as required.

IRR and Spot-Curve NPV Answer Different Questions

The internal rate of return (IRR) is a single effective rate per period that makes the NPV of a specified sequence of cash flows equal to zero. For annual cash flows, it solves $\sum_t C_t/(1+r)^t=0$. When the term structure is nonflat, this single rate is not a replacement for the maturity-specific spot rates in a spot-curve valuation. The IRR describes the rate that balances the project’s own cash flows under a flat-rate equation; the spot-curve NPV discounts each cash flow using the rate for its maturity.

Worked Example: Positive Spot-Curve NPV Despite a Different IRR Comparison

A project costs $200 now and pays $100 at the end of year 1 and $132 at the end of year 2. The effective annual spot rates are 5% for one year and 10% for two years. Find the spot-curve NPV and the project’s IRR. Explain why the IRR alone should not be used to calculate the spot-curve NPV.

Method. First value the dated receipts on the given spot curve. Then solve the separate flat-rate equation defining the IRR.

At the one-year spot rate, the first receipt has present value $100/1.05=95.238095$. At the two-year spot rate, the second receipt has present value $132/(1.10)^2=132/1.21=109.090909$. Thus $\operatorname{NPV}_0=-200+95.238095+109.090909=4.329004$, to six decimal places. The project has a positive spot-curve NPV of about $4.33.

For the IRR, test $r=10\%$ in the flat-rate equation: $100/1.10=90.909091$, and $132/(1.10)^2=109.090909$. Their sum is exactly $200 before rounding, so $-200+100/1.10+132/(1.10)^2=0$, and the IRR is 10%.

The positive spot-curve NPV does not contradict the 10% IRR. The spot-curve calculation discounts the year-1 cash flow at 5%, whereas the IRR equation discounts both receipts at the same 10% rate. The common mistake is to substitute an IRR for a nonflat curve and then claim to have calculated a spot-curve value. The two calculations use different discounting assumptions and answer different questions.

Check. The first receipt is discounted at 5% in the spot-curve valuation but at 10% in the IRR equation, so its spot-curve present value is higher. The second receipt is discounted at 10% in both calculations. This explains why the spot-curve NPV is positive even though the IRR equation has zero NPV at 10%.

Duration Does Not Make Nonparallel Changes Disappear

For a bond or other fixed set of positive cash flows, a duration measure summarizes price sensitivity under a specified rate-change model. Modified duration based on a single yield is a local sensitivity to a small parallel change in that yield, with the cash flows and yield convention held fixed. Under a nonflat spot curve, a change in only one maturity rate is not a parallel shift. The effect depends on which cash flows are exposed to that maturity rate.

For annual cash flows $C_t$ discounted at annual spot rates $s_t$, price is $P=\sum_t C_t(1+s_t)^{-t}$. If every spot rate shifts by the same small amount $\varepsilon$, differentiation gives $\left.dP/d\varepsilon\right|_{\varepsilon=0}=-\sum_t tC_t/(1+s_t)^{t+1}$. This derivative is the first-order price change per unit parallel shift. If only $s_k$ changes, the derivative instead is $\partial P/\partial s_k=-kC_k/(1+s_k)^{k+1}$ for a single cash flow at time $k$. The two sensitivities need not be equal.

Worked Example: Repricing a Single-Maturity Spot-Rate Change

A hypothetical asset pays $100 at the end of year 1 and $100 at the end of year 2. The annual spot rates are 5% for year 1 and 10% for year 2. Calculate its price, then reprice it after only the year-1 spot rate rises to 6%. Separately reprice it after only the year-2 spot rate rises to 11%.

Method. Apply each maturity’s spot rate to its matching cash flow. Revalue after each isolated change instead of treating either change as a parallel shift.

Initially, the year-1 payment is worth $100/1.05=95.238095$. The year-2 payment is worth $100/(1.10)^2=100/1.21=82.644628$. The initial price is $95.238095+82.644628=\$177.882723$, to six decimal places.

If only the year-1 rate rises to 6%, the year-1 payment is worth $100/1.06=94.339623$, while the year-2 payment remains worth $82.644628. The new price is $176.984251, so the change is approximately $-\$0.898473$, or $-\$0.90$ to the nearest cent. Using unrounded values, the direct change is $100/1.06-100/1.05\approx-0.898473$.

If only the year-2 rate rises to 11%, the year-1 payment remains worth $95.238095. The year-2 payment is now worth $100/(1.11)^2=100/1.2321\approx81.162243$. The new price is approximately $176.400338, a change of approximately $-\$1.482385$, or $-\$1.48$ to the nearest cent. The small difference that may arise from subtracting displayed rounded prices is avoided by calculating each change from unrounded cash-flow values.

Check. Both rate increases reduce the present value of the affected positive payment. The year-2 change has the larger effect here, but that conclusion follows from the amounts, maturity, and rates; it is not a rule that every later-rate change must have greater impact. A single parallel-shift duration does not, by itself, identify either isolated change. The correction is to use maturity-specific sensitivity or to reprice the dated cash flows on the changed curve.

A Practical Diagnostic Sequence

Before accepting an integrated valuation, use a short sequence of checks. It separates conceptual errors from arithmetic errors and makes the calculation auditable.

Write the timeline.
List each cash-flow date and amount. Distinguish time 0 from end-of-period payments, and count the number of periods to each payment.

State the sign convention.
For example, record investments as negative and receipts as positive. Interpret the NPV only after checking that convention.

Identify each rate basis.
Record whether a rate is effective or nominal, its conversion frequency, and the interval to which it applies. Convert it only as needed to match the cash-flow dates.

Match rates to maturities.
Under a spot curve, discount each cash flow using the rate associated with its maturity. Do not replace a nonflat curve with an IRR unless the question specifically asks for the IRR.

Check the risk interpretation.
Ask whether a rate change is parallel or affects only selected maturities. Use duration only for the rate-change model it represents, or reprice when the curve change is not captured by that model.

These checks matter because a plausible final number is not evidence that the timeline or rate convention was correct. In particular, rounding should be postponed until the final reporting step: adding rounded present values can change the last reported digit. Keep full precision in intermediate calculations, then state the requested precision and check the result against the direction of the cash flows and rates.

Key takeaway: A sound integrated calculation makes three things explicit before discounting: when each cash flow occurs, what sign it carries, and which rate convention applies to it. IRR and duration are useful measures, but neither substitutes automatically for maturity-specific valuation under a nonflat term structure.

Check Your Understanding

For each problem, write a dated timeline and keep unrounded values until the final answer.

A project costs $1,000 now, pays $400 at the end of year 1, and pays $700 at the end of year 2. The effective annual rate is 5% for both maturities. Find the NPV using negative signs for outflows and positive signs for inflows. (Answer: $15.87 to the nearest cent.)
A payment of $1,000 is due in two years. The discount rate is 8% nominal convertible quarterly. Find its present value. (Answer: $1000/(1.02)^8\approx\$853.49$.)
A project costs $500 now and pays $300 at the end of year 1 and $250 at the end of year 2. The annual spot rates are 4% for year 1 and 8% for year 2. Find its spot-curve NPV. (Answer: approximately $2.80.)
A project costs $200 now and pays $100 at the end of year 1 and $132 at the end of year 2. Find its IRR, then find its NPV using spot rates of 5% for year 1 and 10% for year 2. (Answer: IRR $=10\%$; spot-curve NPV $\approx\$4.33$.)

What You'll Learn