Calcite | Level 5

## Estimating implied cost of capital

I am trying to calculate the implied cost of capital (k_ct in the model below) of four models following Hail and Leuz (2006, 2009) with SAS using IBES dataset.

I have tried using 'proc model' but SAS keeps saying that I need to define the parameters.

So I have tried generating 10000 rates by using do i=1 to 10000 and setting rate = t/10000 and plugging the rates into the model to come up with 10000 prices (left hand variable).

But this seems to take too long and I still cannot get the result with required precision.

Can anyone please recommend me better way to solve this problem?

3 REPLIES 3
SAS Super FREQ

## Re: Estimating implied cost of capital

Please post your PROC MODEL code so we can see how you've implemented the model.

Calcite | Level 5

## Re: Estimating implied cost of capital

```data ae_parm; set ae;
do t=1 to 1000;
prc=prc;
bkvlps=bkvlps;
medest=medest;
r_cl=rand('Uniform');
medest2=medest2; bkv1=bkv1; medest3=medest3; bkv2=bkv2;medest4=medest4; bkv3=bkv3;medest5=medest5; bkv4=bkv4;
growth=growth;
output;
end;
run;

data goal; set ae;
prc=prc;
bkvlps=bkvlps;
medest=medest;
medest2=medest2; bkv1=bkv1;	medest3=medest3; bkv2=bkv2;medest4=medest4; bkv3=bkv3;medest5=medest5; bkv4=bkv4;
growth=growth;
run;

proc model data = ae_parm noprint outmodel = r_cl;
endogenous r_cl;
prc = bkvlps + (medest - bkvlps*r_cl)/(1+r_cl) + (medest2 - bkv1*r_cl)/(1+r_cl)**2 + (medest3 - bkv2*r_cl)/(1+r_cl)**3 +(medest4 - bkv3*r_cl)/(1+r_cl)**4 + (medest5 - bkv4*r_cl)/(1+r_cl)**5
+ ((medest5 - bkv4*r_cl) + (1 + growth))/((r_cl - growth) * ((1+r_cl)**5));
fit prc/ outest = prcest;
solve r_cl  /estdata = prcest data = goal;
run;```

This is the code I have used trying to estimate r_cl.

I am new to using SAS so I just tried to follow the sample code from online but it doesn't seem to be working as intended.

Calcite | Level 5

## Re: Estimating implied cost of capital

I have got the same problem. Did you manage to solve it?

Discussion stats
• 3 replies
• 4107 views
• 0 likes
• 3 in conversation