Announcement

Collapse
No announcement yet.
X
  • Filter
  • Time
  • Show
Clear All
new posts

  • correct solution not found error message with xtdpdbc estimator

    Hi,

    I want to compute a difference in difference using the xtdpdbc function.
    I have t = 4 with 3 observations per year. Overall I have 510 observations

    con = vote share
    cont_fund = treatment condidtion
    post = post treatment dummy
    const = constituencies (fixed effects)

    When I run the code below I run into an error message.

    Code:
    xtdpdbc con i.cont_fund##i.post i.const population age_50 if conwin == 1, lags(1) vce(robust)
    Code:
    incorrect solution -- reinitialize
    Iteration 0: F(b) = 905.59617
    Iteration 1: F(b) = .04219247
    Iteration 2: F(b) = 7.589e-06
    Iteration 3: F(b) = 2.570e-11
    Iteration 4: F(b) = 2.903e-22
    incorrect solution -- reinitialize
    Iteration 0: F(b) = 808.64146
    Iteration 1: F(b) = .04752317
    Iteration 2: F(b) = 3.916e-06
    Iteration 3: F(b) = 6.815e-12
    Iteration 4: F(b) = 2.043e-23
    incorrect solution -- reinitialize
    Iteration 0: F(b) = 3736.4966
    Iteration 1: F(b) = .63437999
    Iteration 2: F(b) = .00186345
    Iteration 3: F(b) = 1.947e-06
    Iteration 4: F(b) = 1.655e-12
    Iteration 5: F(b) = 1.205e-24
    correct solution not found -- try alternative initial values
    
    Group variable: const Number of obs = 340
    Time variable: t Number of groups = 170
    
    Fixed-effects model Obs per group: min = 2
    avg = 2
    max = 2
    
    (Std. err. adjusted for clustering on const)
    ---------------------------------------------------------------------------------------------------------
    | Robust
    con | Coefficient std. err. z P>|z| [95% conf. interval]
    ----------------------------------------+----------------------------------------------------------------
    con |
    L1. | .9714512 .159612 6.09 0.000 .6586175 1.284285
    |
    1.cont_fund | 0 (omitted)
    1.post | -3.051654 1.154322 -2.64 0.008 -5.314083 -.7892248
    |
    cont_fund#post |
    1 1 | 2.636553 .6869516 3.84 0.000 1.290153 3.982953
    |
    const |

    When I try to solve this issue with the from() command:

    Code:
    xtdpdbc con i.cont_fund##i.post i.const population age_50 if conwin == 1, lags(1) vce(robust) from(0.99 0 0 0, copy)
    I get this error message: from(): matrix must be dimension 295
Working...
X