The Laplace Transform

Solving a differential equation directly — chasing derivatives and integrals through a messy equation — is often much harder than it needs to be. The Laplace transform is a trick with an oddly powerful payoff: it re-expresses a function of time as a function of a new variable $s$, and under this re-expression, taking a derivative turns into ordinary multiplication. A hard calculus problem becomes an easy algebra problem in the new variable; you solve the algebra, then translate the answer back into a function of time. It's exactly like converting numbers to logarithms to turn multiplication into addition, solving the easy addition problem, then converting back.

The analogy: logarithms for calculus

Before calculators, multiplying two large numbers by hand was slow, but adding their logarithms and then exponentiating the sum was fast — logarithms turn a hard operation (multiplication) into an easy one (addition), at the cost of translating into and back out of "log space." The Laplace transform does the same trick one level up: it turns the hard operation of differentiation into the easy operation of multiplication by $s$, at the cost of translating into and back out of "$s$-space." You do the easy algebra in $s$-space, then transform back to get your answer in the original variable.

The mechanics, one symbol at a time

The Laplace transform of a function $f(t)$ defined for $t \ge 0$ is

$$F(s) = \mathcal{L}\{f(t)\} = \int_0^{\infty} e^{-st} f(t)\, dt,$$

an improper integral that converts the time-function $f(t)$ into a new function $F(s)$ of a complex variable $s$. In words: it weights every value of $f(t)$ by a decaying exponential and adds them all up — a kind of "spectral fingerprint" of the function that exists as long as $f(t)$ doesn't grow faster than the exponential decay can suppress.

The key property that makes this useful is what happens to a derivative:

$$\mathcal{L}\{f'(t)\} = sF(s) - f(0).$$

In plain English: differentiating in the time domain corresponds to multiplying by $s$ (and subtracting off the initial value) in the transformed domain. So a differential equation involving $f'(t)$, $f''(t)$, etc., becomes, after transforming, a purely algebraic equation in $s$ and $F(s)$ — no derivatives left to wrestle with. Solve for $F(s)$ algebraically, then look up (or compute) the function whose transform is $F(s)$ to get $f(t)$ back — the inverse transform.

Worked example 1: solving continuous discounting via transform

Take the simple ODE from continuous compounding, $y'(t) = ky(t)$, $y(0)=y_0$. Transforming both sides: $sY(s) - y_0 = kY(s)$, so $Y(s)(s-k) = y_0$, giving $Y(s) = \dfrac{y_0}{s-k}$. Looking this up against the standard table entry $\mathcal{L}\{e^{kt}\} = \dfrac{1}{s-k}$ immediately gives $y(t) = y_0 e^{kt}$ — the same exponential-growth answer as before, but reached by pure algebra in $s$-space rather than by directly integrating the ODE.

Worked example 2: a numeric transform value

For a cash flow model with $f(t) = 5e^{-2t}$ (a decaying rate of $5, decaying at rate 2 per year), the transform (using $\mathcal{L}\{e^{at}\}=\frac{1}{s-a}$ with $a=-2$) is $F(s) = \dfrac{5}{s+2}$. Evaluating at $s=3$: $F(3) = \dfrac{5}{3+2} = \dfrac{5}{5} = 1$. This single number is the "weighted total" of the cash flow stream as seen through a discount rate of 3 — in fact, setting $s=r$ (a constant discount rate) and $f(t)$ equal to a continuous cash flow makes $F(r)$ exactly the present value of that cash flow stream, which is why the Laplace transform and continuous-time discounting are the same mathematical object wearing different names.

The functions being transformed here are almost always exponentials or sums of exponentials — adjust the decay rate and notice this is the same family of curves the Laplace transform is built to handle exactly, term by term.

What this means in practice

The Laplace transform is the standard tool for solving linear ODEs and PDEs with known initial conditions — exactly the setup that appears in interest-rate models, queueing systems, and any linear system's transient response. In derivatives pricing it's a close cousin of the characteristic-function methods (Fourier transforms of the log-price density) used in fast option-pricing algorithms; both work by moving a hard problem into a transformed space where it becomes tractable algebra, then inverting.