11
$\begingroup$

I ask because those functions are on the TI BA II Plus financial calculator.

I saw some interesting answers but I don't think a calculator would be practical in their environment.

$\endgroup$
  • 8
    $\begingroup$ The question is weird because it comes from a funky reason. But I'm fine having somebody giving an example of trigonometric functions in quantitative finance. $\endgroup$ – SRKX May 9 '14 at 10:38
  • 2
    $\begingroup$ sinh and cosh are used in some formulations of the Heston model of stochastic volatility, but you're not going to be doing those on a calculator. $\endgroup$ – experquisite May 9 '14 at 15:53
  • $\begingroup$ @SRKX Now that I think about it, I don't think this reason is funky at all. Who else but the non-mathematically inclined would ask such? $\endgroup$ – BCLC May 11 '14 at 9:18
  • 1
    $\begingroup$ You can also refer to the similar post in Economics.stackexchange: economics.stackexchange.com/questions/19172/… $\endgroup$ – agassi Oct 24 at 12:47
  • $\begingroup$ All the answers below totally gloss over exactly what @experquisite is getting at. Trig functions in finance are mostly used inside the guts of other, considerably more complicated, calculations. It's extremely unlikely unlikely you'd see any use in a calculators trig functions working in a finance role. $\endgroup$ – will Oct 25 at 6:15
8
$\begingroup$

Fourier methods use sine and cosine functions, and are used in calculating option prices, VaR, time series analysis etc. It is an alternative process for doing many things in finance. Some links Fourier Methods in trading on StackExchange and Wiki

$\endgroup$
7
$\begingroup$

One can use the Karhunen–Loève expansion to approximate a trajectory of a Wiener Process, which can be used to model the evolvement of returns in time. (http://en.wikipedia.org/wiki/Karhunen%E2%80%93Lo%C3%A8ve_theorem#The_Wiener_process)

Though the Karhunen–Loève expansion has theoretical advantages to other variants to generate a trajectory of a Wiener Process, many users will use different methods because on computers evaluation of trignometric is very expensive in terms of calculation time.

$\endgroup$
5
$\begingroup$

You can use $\sin$ or $\cos$ to model seasonality. If all you have is a calculator it might be the most practical way.

$\endgroup$
  • $\begingroup$ But it would be such a crude way to do it that you might as well draw a wiggly line on a bit of paper and eyeball it... $\endgroup$ – will Oct 25 at 6:15
4
$\begingroup$

When you do Monte Carlo simulation and would like to draw sample from the normal distribution $\mathcal{N}(\mu,\sigma^2)$, you may use Box-Muller transform and come up with formulas using $\sin$ and $\cos$.

$\endgroup$
  • $\begingroup$ But in reality, noone samples normal random numbers like this because its very inefficient... $\endgroup$ – will Oct 25 at 6:16
3
$\begingroup$

Trigonometric functions show up in econometric models for business cycles. For example: the average length of a cycle of an AR(2) process is

$ k = \frac{2 \pi}{\cos^{-1}( \phi_1/ (2 \sqrt{-\phi_2}))}$

For an AR(2) model given by $ r_t = \phi_0 + \phi_1 r_{t-1} + \phi_2 r_{t-2} + a_t$

with complex roots, $\phi_1^2 + 4\phi_2 <0 $

$\endgroup$
3
$\begingroup$

Trigonometric functions are WAVE phenomena. As such, they are best used to model so-called periodic functions, that is, functions with cycles of a fixed period in length. That's why they are good for modelling, seasonal, annual, "blue moon" (once every two and half years), or other functions with set "periods."

$\endgroup$
2
$\begingroup$

Solving some heat/diffusion equations under certain conditions needs trigonometric functions.

Black-Scholes reduces to a heat/diffusion equation by a change of variables.

$\endgroup$
  • 1
    $\begingroup$ But the initial conditions in Black Scholes lead to the normal distribution as a solution (i.e. not trig functions)! $\endgroup$ – vonjd May 11 '14 at 11:33

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Not the answer you're looking for? Browse other questions tagged or ask your own question.