Notice: On April 23, 2014, Statalist moved from an email list to a forum, based at statalist.org.

# Re: st: RE: disequilibrium model

 From Matt To statalist@hsphsun2.harvard.edu Subject Re: st: RE: disequilibrium model Date Fri, 10 May 2013 02:18:34 +0100

```Mark,

Thanks for advice and explanations of the movestay command.
But i'm still a bit confused about the code. Please correct me if i wrong.

I'm trying to estimate whether a firm faces bank borrowing constraint or unconstraint,
in other words, whether short-term bank loans an observation receive (as on balance sheet) is the quantity demanded (unconstraint because the demand is fulfilled), or quantity supplied (constraints because the bank issue only amount less than demanded).

qDemand = x1b1 + e1
qSupply = x2b2 + e2
qTransacted = min (qDemand, qSupply)

the dependent variables qDemand and qSupply are unobserved,
what observed dependent variable is only qTransacted, which is the short-term bank loans that actually transacted with banks.

I'm trying to find whether qTransacted belongs to qDemand (unconstraints), or belongs to qSupply (bank constraints)

So, in the movestay command, is that correct if:
depvar1 and depvar2 = short-term bank loans
x1 = those factors affecting depvar1
x2 = those factors affecting depvar2

i'm very confusing on the selection part:  because my rationale is that
1) an observation belongs to regime 1 (face borrowing constraint) if the demand for bank loans is greater than supply.
2) an observation belongs to regime 2 (unconstraint) if demand for bank loans is EQUAL TO / NOT greater than supply.

Matt

On 22 Apr 2013, at 12:21, Schaffer, Mark E wrote:

> Matt,
>
> -movestay- by Lokshin and Sajaya estimates the kind of "switching-regression" model you are working with - see description below.  Playing with the -findit- turned up some other options, e.g., switchr (Zimmerman) and  ssm (Miranda and Rabe-Hesketh).
>
> HTH,
> Mark
>
>    movestay uses the maximum likelihood method to estimate the endogenous
>    switching regression model.  It is implemented using the d2 evaluator to
>    calculate the overall log likelihood together with its first and second
>    derivatives.
>
>    movestay estimates all of the parameters in the model:
>
>        (regression equation for regime 1: y1 is depvar1, x1 is varlist1)
>                y1 = x1 * b1 + e_1
>
>        (regression equation for regime 2: y2 is depvar2, x2 is varlist2)
>                y2 = x1 * b2 + e_1
>
>        (selection equation: Z is varlist_s)
>                y1 observed if Zg + u > 0
>                y2 observed if Zg + u <= 0
>
>        where:
>                e_1 ~ N(0, sigma1)
>                e_2 ~ N(0, sigma1)
>                u ~ N(0, 1)
>                corr(e_1, u) = rho_1
>                corr(e_2, u) = rho_2
>
>    Here depvar1, depvar2 and varlist1, varlist2 are the dependent variables
>    and regressors for the underlying regression models (y1, y2 = xb), and
>    varlist_s specifies the variables Z thought to determine which regime is
>    observed.
>
>> -----Original Message-----
>> From: owner-statalist@hsphsun2.harvard.edu [mailto:owner-
>> statalist@hsphsun2.harvard.edu] On Behalf Of Matt
>> Sent: 22 April 2013 11:05
>> To: statalist@hsphsun2.harvard.edu
>> Subject: st: disequilibrium model
>>
>> Hi,
>>
>> I'm current working on the disequilibrium model as in the Maddala and
>> Nelson (1974).
>>
>> Quantity of demand (qDt) = B1X'1t + u1t
>> Quantity of supply (qSt) = B2X'2t + u2t
>> Quantity observed (qt) = min (qDt, qSt)
>>
>> the model consists of a demand equation qDt, a supply equation qSt, and a
>> transaction equation qt.
>> the vectors X'1t and X'2t are exogenous, independent variables, B1 B2 are
>> their coefficients, u1t and u2t are their disturbances.
>>
>> qDt and qSt in this model are the amount of bank debt demanded and
>> supplied, but they are not observed by any external party.
>> Only the amount of bank debt which was actually received, the transaction
>> amount qt can be perceived.
>> But we don't know if this transaction amount of debt is the agrees to the
>> amount demanded by the firm or whether it is limited by bank. We don't
>> know whether the firms in the sample face credit rationing - unknown
>> sample separation.
>>
>> So, to avoid writing the complicated ML programme, i uses 3-stage least
>> square to estimate the demand and supply model, but i'm confused about
>> the dependent variable should be used for estimation of demand and supply
>> equation. Indeed, for the quantity observed (qt) is the short-term bank loans
>> of the firm as in the balance sheet. But what should be the dependent
>> variable for qDt and qSt?
>>
>> Also, how can i do conditional probability in stata?
>>
>> Many thanks if you can answer my query.
>>
>> Matt
>> *
>> *   For searches and help try:
>> *   http://www.stata.com/help.cgi?search
>> *   http://www.stata.com/support/faqs/resources/statalist-faq/
>> *   http://www.ats.ucla.edu/stat/stata/
>
>
> -----
> Sunday Times Scottish University of the Year 2011-2013
> Top in the UK for student experience
> Fourth university in the UK and top in Scotland (National Student Survey 2012)
>
> We invite research leaders and ambitious early career researchers to
> to apply.
>
> Heriot-Watt University is a Scottish charity
> registered under charity number SC000278.
>
>
> *
> *   For searches and help try:
> *   http://www.stata.com/help.cgi?search
> *   http://www.stata.com/support/faqs/resources/statalist-faq/
> *   http://www.ats.ucla.edu/stat/stata/

*
*   For searches and help try:
*   http://www.stata.com/help.cgi?search
*   http://www.stata.com/support/faqs/resources/statalist-faq/
*   http://www.ats.ucla.edu/stat/stata/
```