Context: although I'm not a mathematician, I'm a curious software engineer which remember high school Math and had my time with calculus at college.
I'm trying to manually calculate the amortization rate of a loan (so really none of the fields i now something about). given that I've already deduced the amortization formula, i was given a problem like this:
principal = U\$$1000$, monthly payment = U\$$100$, $12$ monthly payments in total. What is the monthly rate of this loan?
I've ended at this: $$10x = 1 - (1+x)^{-12}$$
from all possible roots, there's a single positive real root ~0.0292 that matches the solution (the loan monthly rate should be 2.92% per month, so this is right towards the correct answer).
my problem is: if i had to solve this with pen and paper without any computer to aid me (like wolfram alpha or any online solver), how should i manipulate this to get that $0.0292$ root? is this possible?
It is quite straightforward to set up a simple recursion relation to obtain the unique positive real root. Such a recursion can be efficiently implemented in certain kinds of hand calculators; in particular, the calculator needs to have an
Ans("answer") key, and the ability to input algebraic expressions that are functions of the previous output.In your case, consider the polynomial $$f(x) = (10x-1)(1+x)^{12} + 1. \tag{1}$$ Its derivative is $$\begin{align} f'(x) &= \frac{d}{dx}[10x-1](1+x)^{12} + (10x-1) \frac{d}{dx}[(1+x)^{12}] \\ &= 10(1+x)^{12} + (10x-1)(12)(1+x)^{11} \\ &= (1+x)^{11}(10+10x+120x-12) \\ &= 2(65x-1)(1+x)^{11}. \tag{2} \end{align}$$ So the Newton's method recursion is $$\begin{align} x_{n+1} &= x_n - \frac{f(x_n)}{f'(x_n)} = x_n - \frac{(10x_n-1)(1+x_n)^{12} + 1}{2(65x_n-1)(1+x_n)^{11}} \\ &= \frac{(1-11x_n+120x_n^2) - (1+x_n)^{-11}}{2(65x_n-1)}. \tag{3}\end{align}$$ Now all that is needed is a suitable initial guess $x_0$. The original equation suggests that if there is no interest, i.e. $x = 0$, then the total paid is $1200$, which exceeds the present value of the loan by $20\%$. So if we assume that interest is not compounded (i.e., simple interest), the monthly rate would be $0.20/12 \approx 0.016667$. This will serve as our choice of $x_0$.
Then to implement this recursion on a calculator, we type in
which outputs $0.016666667$ or something like that. Then we input
which will return $0.0975032$. Then you recall the last input (like the
Ansbutton, the name of this button might vary, but it is typically just an upward arrow button), and press=orAnsagain, which causes the calculator to apply the previous output as the argument to the previous function you typed. You will get $0.0664072$. If you do this repeatedly--recall the previous input, then compute its value based on the answer you obtained in the previous iteration--then you should get a sequence of numbers like this:$$\begin{array}{c|c} n & x_n \\ \hline 0 & 0.0166667 \\ 1 & 0.0975032 \\ 2 & 0.0664072 \\ 3 & 0.0460899 \\ 4 & 0.03476 \\ 5 & 0.0301476 \\ 6 & 0.0292615 \\ 7 & 0.0292286 \\ 8 & 0.0292285 \\ 9 & 0.0292285 \\ \end{array}$$ and we stop when the output no longer changes. This represents, to within the display precision of the calculator, the fixed point of the recursion relation, and therefore the desired root of $(1)$.