Mata (command mm_root) within a Stata loop does not work

[Marco Seruset]

Dear Statalist members,

I'm trying to solve the unknown Lambda using mm_root function of the moremata package. The original formula can be found in O'Kane & Turnbull (2003): Valuation of CDS, Lehman Brothers, page 13 (http://www.javaquant.net/papers/ValuationCDSLehman.pdf)

Following the post of Attaullah Shah, I could fix it for one company over 48 months (http://www.statalist.org/forums/forum/general-stata-discussion/mata/1345852-mm_root-function-to-solve-for-unknows). However, I need to find the unknown variable lambda for 30 companies over 48 months.

Here is the code for one company (I have just created a fake dataset in order to reduce the complexity).

clear all
set obs 3 // number of months (in my dataset 48)
gen fraction_0 = 0
gen fraction_1 = 0.25
gen fraction_2 = 0.5
gen fraction_3 = 0.75
gen fraction_4 = 1
gen Rec = 0.4
gen Gen_fraction = 0.25
gen CDS_1Y = 0.021
gen lambda = .

mata
v = J(1,1,.)
st_view(v, . , "lambda CDS_1Y fraction_0 fraction_1 fraction_2 fraction_3 fraction_4 Rec Gen_fraction")
function y(lambda, a, b, c, d, e, f, g, h) {
return (a/(1-g)*((h*0.99822*exp(-lambda*c))+(h*0.99648*exp(-lambda*d))+(h*0.99476*exp(-lambda*e))+(h*0.99301*exp(-lambda*f)))-(0.99822*(exp(-lambda*b)-exp(-lambda*c))+0.99648*(exp(-lambda*c)-exp(-lambda*d))+0.99476*(exp(-lambda*d)-exp(-lambda*e))+0.99301*(exp(-lambda*e)-exp(-lambda*f))))
}
for (i=1;i<=rows(v);i++) {
r=mm_root(lambda=., &y(),0,1,0.01,1000,v[i,2], v[i,3], v[i,4], v[i,5], v[i,6], v[i,7], v[i,8], v[i,9])
v[i,1] = lambda
}
end
l, noo

I have tried to write a loop in Stata, which includes the Mata calculation, but it doesn't work.
The idea with curly brackets around mata code was originated from Niels Henrik Bruun (http://www.statalist.org/forums/foru...ile-loop-error).

Here is my Stata-loop

clear all
local currentfile C:\...
cd "`currentfile'"
foreach name company1 company2 company3 company4 ... {
use "`currentfile'\OriginalDataset.dta", clear
collapse CDS_1Y_`name' fraction_0 fraction_1 fraction_2 fraction_3 fraction_4 Rec Gen_fraction, by(month)
rename CDS_1Y_`name' CDS_1Y // that was done for the mata-code to avoid any changes in the mata code above.
gen lambda = .
mata {
... code above (but without the "end" command)
}
rename CDS_1Y CDS_1Y_`name'
rename lambda lambda_`name' // lambda for each company can be uniquely identified
merge 1:1 month using "`currentfile'\OriginalDataset.dta", nogen
save "`currentfile'\OriginalDataset.dta", replace
}

The error message is as follows.

'function' found where almost anything else expected
r(3000)

As I mentioned above, the code works perfectly for one company, however, I don't know how to modify the code, that it will work for all of my companies.
Additionally, I read some proposals to put every calculation step in Mata and call the Mata function within the Stata loop - but again, I wasn't able to apply it for my problem.

I'm looking forward to hearing from you!

Thank you very much in advance!
Kind regards from Germany, Marco