Newey-West t-statistics in a Fama MacBeth regression using asreg, fmb

[Johannes Koehlerr]

Hi together,

Is there a way to come up with Newey-West t-statistics during the last step of the FMB regression? I performed the FMB regression via the following code:

*standardization of betas 
        foreach var of varlist TNIC_ew_peer_ret_t0 FF48_vw_peer_ret_t0 delta_Mkt_RF SMB HML ln_lagged_BtM SIZE_prev_june ret_t1_t12 ret_t1 {
        egen z_`var' = std(`var'), mean(0) std(1)
        }
    
    *time series regression: regression for each firm over time -> beta estimate  
    bys PERMNO: asreg excess_ret z_TNIC_ew_peer_ret_t0 z_FF48_vw_peer_ret_t0 z_delta_Mkt_RF z_SMB z_HML z_ln_lagged_BtM z_SIZE_prev_june dummy_neg_BtM z_ret_t1_t12 z_ret_t1
    
    drop _Nobs _R2 _adjR2
    
    *FMB regression 
    asreg excess_ret _b_z_TNIC_ew_peer_ret_t0 _b_z_FF48_vw_peer_ret_t0 _b_z_delta_Mkt_RF _b_z_SMB _b_z_HML _b_z_ln_lagged_BtM _b_z_SIZE_prev_june _b_dummy_neg_BtM _b_z_ret_t1_t12 _b_z_ret_t1, fmb newey(2) 

    *Newey t-statistics 
    tsset PERMNO month    
    newey excess_ret _b_z_TNIC_ew_peer_ret_t0 _b_z_FF48_vw_peer_ret_t0 _b_z_delta_Mkt_RF _b_z_SMB _b_z_HML _b_z_ln_lagged_BtM _b_z_SIZE_prev_june _b_dummy_neg_BtM _b_z_ret_t1_t12 _b_z_ret_t1,lag(2) force

Knowing that the last step probably has not the intended effect since it results in the t stats of a regular newey regression, not as intedend in the newy t stats of atwo step fmb regression.


Thank you in advance for your support. I would be glad if someone might have some helpful insights

Johannes