Why the Units of a Rate Matter
A rate can describe how quickly a measured quantity changes, but the units tell you what quantity is changing. If a statistic is reported as a percentage, its derivative may be measured in percentage points per year. If a rate is reported per person, its units include a population denominator. Neither should be mistaken for a change in the total number of people.
For example, an unemployment rate rising from 5% to 6% has increased by 1 percentage point. That is a different statement from saying that the rate increased by 1% of its previous value. Likewise, a population increase of 600 people in a year is an absolute change, while a growth rate of 0.012 people per person per year adjusts that change for the population size.
Rates of Percentage-Valued Quantities
Let \(r(t)\) be a percentage reported in percentage units, such as \(r(t)=5\) for a rate of 5%. Then \(r'(t)\) describes the instantaneous change in that percentage-valued quantity. If \(t\) is measured in years, the derivative’s units are percentage points per year.
A percentage-point rate is not the same as a relative percentage rate. The relative rate compares the change with the current value. When \(r(t)\ne 0\), its instantaneous relative percentage rate is \(100\,r'(t)/r(t)\) percent per unit of time. This answers a different question: how large is the instantaneous change compared with the current percentage?
The distinction also depends on how the percentage is recorded. If \(u(t)\) is the corresponding decimal proportion, then \(u(t)=r(t)/100\). Thus \(u'(t)=r'(t)/100\). For example, a change of 0.4 percentage points per year corresponds to a change of 0.004 in the decimal proportion per year.
Why Relative Percentage Rates Use a Quotient
The relative change over a short time interval compares the change in \(r\) with the starting value \(r(t)\). Dividing that relative change by the length of the interval gives an average relative rate. Taking the limit as the interval shrinks gives the instantaneous relative rate.
Over an interval of length \(h\), the change in \(r\), divided by its current value, is \(\frac{r(t+h)-r(t)}{r(t)}\), provided \(r(t)\ne 0\).
The relative change per unit time is \(\frac{r(t+h)-r(t)}{h\,r(t)}\).
If \(r\) is differentiable at \(t\), taking the limit as \(h\to 0\) gives \(\frac{1}{r(t)}\lim_{h\to 0}\frac{r(t+h)-r(t)}{h}=\frac{r'(t)}{r(t)}\). Multiplying by 100 expresses this relative rate as a percentage rate.
The factor of 100 converts a relative rate into a percentage rate. This formula requires \(r(t)\ne 0\); in contexts where \(r(t)\) is a positive percentage, the denominator is positive. The result does not replace \(r'(t)\): it reports a different, relative measure.
Worked Example: A Rate Rising in Percentage Points
Worked Example: A Rate Rising in Percentage Points
Suppose an unemployment rate \(r(t)\) is modeled by \(r(t)=3+0.4t\), where \(r\) is recorded in percent and \(t\) is the number of years after a chosen starting date. Interpret the rate of change at \(t=5\). If the labor force is a constant 250,000 people, also find the corresponding rate of change in the number of unemployed people.
First find the unemployment rate at \(t=5\):
Thus the rate is 5% at that time. Differentiate the model:
So the unemployment rate is increasing at 0.4 percentage points per year. Its instantaneous relative percentage rate at \(t=5\) is:
These are distinct descriptions: 0.4 percentage points per year is the change in the reported rate, while 8% per year compares that change with the current value of 5%.
With a constant labor force \(L=250{,}000\), the number unemployed is \(U(t)=Lr(t)/100\). Therefore:
Differentiating gives \(U'(t)=1{,}000\) people per year. This agrees with the decimal-rate calculation: \(r'(t)=0.4\) percentage points per year is \(0.004\) per year as a decimal proportion, and \(250{,}000(0.004)=1{,}000\) people per year.
Check: Differentiating the expanded expression \(U(t)=7{,}500+1{,}000t\) gives \(U'(t)=1{,}000\), confirming the calculated absolute rate. The conversion assumes the labor force remains constant; if it changes, the number of unemployed cannot be found from the rate’s derivative alone.
Per-Capita Rates
A per-capita rate adjusts a change in a count for the size of a population. If \(P(t)\) is a population count, then \(P'(t)\) is an absolute rate, measured in people per year when \(t\) is measured in years. The ratio \(P'(t)/P(t)\) is a per-capita growth rate: it measures the instantaneous population increase per person currently in the population, per year.
The numerator and denominator answer different questions. \(P'(t)\) tells how many people are being added per unit of time. \(P'(t)/P(t)\) scales that change by the current population. Multiplying the per-capita rate by the population recovers the absolute rate: \(P(t)\bigl(P'(t)/P(t)\bigr)=P'(t)\), when \(P(t)>0\).
Worked Example: Converting Population Growth to a Per-Capita Rate
A town’s population is modeled by \(P(t)=40{,}000+500t\), where \(t\) is in years. Find its absolute and per-capita growth rates at \(t=8\), and express the per-capita rate as a percentage per year.
The population at that time is:
Differentiate the model:
The absolute growth rate is therefore 500 people per year. Divide by the current population to find the per-capita rate:
As a percentage rate, this is:
The approximation is rounded to two decimal places. The population is growing by 500 people per year, which is about 1.14% of its size per year at \(t=8\).
Check: Multiply the per-capita rate by the population: \(44{,}000(1/88)=500\). The original absolute rate is recovered, with units \((\text{people})(\text{people per person per year})=\text{people per year}\).
When the Denominator Also Changes
Some per-capita statistics are ratios of two changing counts rather than a total population’s growth rate divided by itself. For example, a dependency ratio may be the number of dependent people divided by the number of working-age people. Its derivative describes the change in the ratio, not the change in either count by itself.
For differentiable counts \(D(t)\) and \(W(t)\), with \(W(t)\ne 0\), the quotient rule gives:
The second term matters because a changing denominator can raise or lower the ratio even when the numerator is increasing. The units of the derivative are units of the ratio per unit of time—for population counts, people per working-age person per year.
Worked Example: A Changing Number of Dependents per Worker
Suppose a community has \(D(t)=6{,}000+120t\) dependent people and \(W(t)=30{,}000+300t\) working-age people, with \(t\) measured in years. Find the instantaneous change in dependents per working-age person at \(t=0\).
At \(t=0\), the two counts and their rates are:
The ratio is initially \(D(0)/W(0)=6{,}000/30{,}000=0.2\) dependent people per working-age person. Apply the quotient rule:
The ratio is increasing at 0.002 dependent people per working-age person per year. The dependent count itself is increasing by 120 people per year, and the working-age count by 300 people per year; neither of those absolute rates is the derivative of the ratio.
Check: For any \(t\), the quotient-rule numerator simplifies as follows:
At \(t=0\), dividing this by \(30{,}000^2=900{,}000{,}000\) gives \(0.002\), confirming the result. The derivative’s units are dependents per working-age person per year.
A Common Mistake: Confusing Points with Percent
Suppose a rate rises from 5% to 6%. A common incorrect calculation is to say, “It increased by 1%.” That treats the difference of 1 percentage point as a 1% relative increase. The correction is to identify which comparison is intended.
So the rate increased by 1 percentage point, which is a 20% increase relative to its original value. For derivatives, the same distinction applies: \(r'(t)\) gives percentage points per unit of time when \(r\) is recorded in percent, while \(100r'(t)/r(t)\) gives the instantaneous relative percentage rate when \(r(t)\ne 0\).
Another mistake is to treat a per-capita rate as a total count change. A rate of 0.01 people per person per year does not mean the population increases by 0.01 people per year. For a population of 10,000 people, it corresponds to an absolute rate of \(10{,}000(0.01)=100\) people per year.
How to Interpret These Rates
When a rate is given in a contextual problem, first identify the quantity being differentiated and how that quantity is recorded. Then state the rate using those units. In particular, check whether the rate describes a percentage-valued variable, a total count, or a count normalized by a population.
- For a percentage-valued variable, report its derivative in percentage points per unit of time.
- If asked for relative percentage change, divide the derivative by the current value and multiply by 100, provided that value is nonzero.
- For a population count, report its derivative as an absolute change in people per unit of time.
- For per-capita growth, divide the population’s absolute rate by its current size.
- For a ratio of changing counts, account for the change in both numerator and denominator.
Check Your Understanding
For each question, state what the rate measures and include appropriate units.
- A percentage-valued rate has derivative \(0.3\) percentage points per year and current value \(6\%\). What is its instantaneous relative percentage rate?
- A population of 80,000 people is increasing at 960 people per year. What is its per-capita growth rate, in people per person per year and as a percentage per year?
- A statistic rises from 4% to 5%. How many percentage points did it increase, and what was its relative percentage increase?
- Why is the derivative of a ratio of two changing population counts not generally equal to the numerator’s derivative divided by the denominator?
- A per-capita growth rate is \(0.02\) people per person per year for a population of 15,000. What absolute growth rate does this correspond to?
