Table 2 Descriptive statistics: large...

Table 2

Descriptive statistics: large economy simulation

Variable	Count	Unique	Top	Freq	Mean	Std	Min	25%	50%	75%	Max
True_Income	1,000				49.676	45.987	3.243	21.650	36.872	62.274	446.949
Observed_Education	1,000	3	Medium	344
Observed_Sector	1,000	3	Industry	351
Observed_Region	1,000	4	South	274
Predicted_Income_GNN	1,000				51.958	48.033	3.131	22.158	38.092	64.034	496.511
Tax_GNN	1,000				12.989	12.008	0.783	5.539	9.523	16.009	124.128
Tax_Benchmark	1,000				10.929	10.117	0.714	4.763	8.112	13.700	98.329
Post_Tax_GNN	1,000				36.687	34.132	2.461	16.168	27.164	45.844	322.821
Post_Tax_Benchmark	1,000				38.748	35.870	2.530	16.887	28.760	48.573	348.620

Variable	Count	Unique	Top	Freq	Mean	Std	Min	25%	50%	75%	Max
True_Income	1,000				49.676	45.987	3.243	21.650	36.872	62.274	446.949
Observed_Education	1,000	3	Medium	344
Observed_Sector	1,000	3	Industry	351
Observed_Region	1,000	4	South	274
Predicted_Income_GNN	1,000				51.958	48.033	3.131	22.158	38.092	64.034	496.511
Tax_GNN	1,000				12.989	12.008	0.783	5.539	9.523	16.009	124.128
Tax_Benchmark	1,000				10.929	10.117	0.714	4.763	8.112	13.700	98.329
Post_Tax_GNN	1,000				36.687	34.132	2.461	16.168	27.164	45.844	322.821
Post_Tax_Benchmark	1,000				38.748	35.870	2.530	16.887	28.760	48.573	348.620

Note(s): This table mirrors Table 1 but corresponds to the larger, more fragmented developing economy. The higher variance and kurtosis of the income variables suggest greater structural heterogeneity and network modularity, which, in turn, constrain the inferential precision of the GNN model

Source(s): Author's simulations based on synthetic data calibrated to household survey and administrative records

[ViewLarge]

Sharing Unavailable