Periodic Rates And Accumulation Drift

Introduction

A rate can repeat its pattern while its accumulation does something different. For example, a rate that oscillates around a positive average repeats every cycle, but its accumulated total continues to grow. A rate whose positive and negative contributions cancel over a cycle can instead produce an accumulation that repeats.

The distinction is captured by one number: the signed integral of the rate over a period. It determines the accumulation’s change per cycle, regardless of where that cycle begins.

Periodic Rates and Their Period Integral

A function \(f\) is periodic with period \(T>0\) if \(f(x+T)=f(x)\) for every \(x\) in its domain. The period need not be the smallest positive period. In this tutorial, assume \(f\) is bounded, integrable on every finite interval, and defined on all of \(\mathbb{R}\).

Positive and negative values both contribute to this integral, with their signs. Thus, the integral is a net change, not the total amount accumulated without regard to sign. The average value of the rate over one period is \(\mu=P/T\).

Fix a basepoint \(a\), and define the accumulation function \(A(x)=\int_a^x f(t)\,dt\). The interval-increment identity says that \(A(v)-A(u)=\int_u^v f(t)\,dt\). Applying it to an interval of length one period gives the key result.

The Same Increment on Every Cycle

The first identity says that the accumulation changes by the same amount over every interval of length \(T\). The amount does not depend on the starting point \(x\). The second identity extends the result to any integer number of cycles: positive \(n\) counts cycles forward, and negative \(n\) counts them backward.

Why the Cycle Integral Does Not Depend on the Starting Point

Here is a proof of the central fact. It is enough first to consider a starting point \(x\) with \(0\leq x<T\). Split the interval from \(x\) to \(x+T\) at \(T\). In its second part, translate the variable by one period and use \(f(t+T)=f(t)\):

For an arbitrary real \(x\), write \(x=kT+r\), where \(k\) is an integer and \(0\leq r<T\). Translating the interval by \(kT\) leaves the integrand unchanged, so \(\int_x^{x+T}f=\int_r^{r+T}f=P\) as well. The interval-increment identity now gives \(A(x+T)-A(x)=P\).

Applying this identity repeatedly gives \(A(x+nT)=A(x)+nP\) for positive integers \(n\). For negative integers, apply the same identity backward: \(A(x-T)=A(x)-P\), and repeat. For \(n=0\), the formula is immediate. This proves the theorem for every integer \(n\).

Worked Example: A Positive Drift Over Each Cycle

When Is the Accumulation Periodic?

The cycle identity gives an exact condition for the accumulation to repeat with the rate’s period \(T\). It is \(T\)-periodic precisely when \(P=0\). Indeed, \(P=0\) makes \(A(x+T)=A(x)\) for every \(x\). Conversely, if \(A(x+T)=A(x)\) for every \(x\), the cycle identity forces \(P=0\).

In fact, under the assumptions above, the condition also tells us whether \(A\) can have any positive period at all. If \(P\neq0\), the values \(A(x+nT)=A(x)+nP\) grow without bound in magnitude as \(n\) increases through the positive integers. But a continuous periodic function is bounded: on one period it is bounded, and periodicity repeats those values everywhere. The accumulation \(A\) is continuous because \(f\) is bounded and its integral over a short interval has magnitude at most the bound for \(|f|\) times the interval length. Thus, if \(P\neq0\), \(A\) cannot be periodic with a different positive period either. If \(P=0\), \(A\) is already \(T\)-periodic.

Worked Example: Zero Net Change Gives a Repeating Accumulation

Separating Linear Drift from Periodic Fluctuation

Even when \(A\) is not periodic, its non-repeating part is especially simple: it is linear. Let \(\mu=P/T\) be the average rate per unit of time, and remove this average from the rate by defining \(g(t)=f(t)-\mu\). The function \(g\) is still \(T\)-periodic, and its integral over one period is \(P-\mu T=0\).

Define \(B(x)=\int_a^x g(t)\,dt\). By the zero-period-integral result, \(B\) is \(T\)-periodic. Linearity of accumulation gives the decomposition:

If \(\mu=0\), there is no linear drift, so the accumulation itself repeats. If \(\mu\neq0\), the periodic fluctuation rides on a steadily increasing or decreasing linear trend.

Worked Example: Identifying the Trend and Fluctuation

A Common Mistake: Assuming Every Periodic Rate Has Periodic Accumulation

Periodicity of the rate alone does not guarantee periodicity of the accumulation. For example, consider \(f(t)=1+\sin t\), which has period \(2\pi\). A tempting but incorrect claim is that its accumulation must also repeat every \(2\pi\). But its period integral is:

For \(A(x)=\int_0^x f(t)\,dt\), this means \(A(2\pi)-A(0)=2\pi\), not zero. The accumulation cannot repeat after one cycle. The correction is to check the net signed integral over a full period: the accumulation repeats exactly when that integral is zero, not merely when the rate repeats.

Why Cycle Drift Is Useful

The cycle integral lets you predict long-term accumulation without recalculating every cycle. If the net integral is positive, each cycle adds the same positive amount; if negative, each cycle subtracts the same amount; and if zero, each cycle returns the accumulation to its value one period earlier. For an integer number of cycles, the total change follows immediately from \(A(x+nT)-A(x)=nP\).

This is useful whenever a repeated rate pattern represents repeated inflow and outflow, demand and supply, or another quantity that accumulates over time. The within-cycle shape may matter for values between cycle boundaries, but the period integral alone determines the net drift from one cycle boundary to the next.