Please use this identifier to cite or link to this item: https://hdl.handle.net/10419/327971 
Year of Publication: 
2024
Citation: 
[Journal:] CBN Journal of Applied Statistics [ISSN:] 2476-8472 [Volume:] 15 [Issue:] 2 [Year:] 2024 [Pages:] 37-70
Publisher: 
The Central Bank of Nigeria, Abuja
Abstract: 
This study investigates the effects of fiscal deficits on human development in Nige- ria from 1990 to 2021, employing standard econometric and machine learning ap- proaches. Specifically, the paper applies autoregressive distributed lag (ARDL) mod- els to examine linear relationships, and multivariate adaptive regression splines (MARS) to capture linear and nonlinear interactions among key variables influenc- ing human development. Results from the ARDL analysis reveal that fiscal deficit does not significantly impact human development in Nigeria both in the short- and long run. However, MARS results indicate that fiscal deficits only enhance human development when kept below 1.34% of GDP but become detrimental beyond this threshold, suggesting fiscal sustainability constraints. Nevertheless, this adverse im- pact can be mitigated, or even reversed, when accompanied by stronger economic capacity. The study also finds that access to electricity is a key driver of human de- velopment, underscoring its role in improving healthcare, education, economic pro- ductivity, and overall living standards. To harness the benefits of fiscal deficits for human development, the study recommends directing deficit spending toward critical sectors, enforcing fiscal rules to maintain sustainable deficit levels, conducting reg- ular fiscal policy assessments, and prioritising long-term infrastructure investments, particularly electrification projects.
Subjects: 
Fiscal deficit
human development
health
education
machine learning
JEL: 
E62
H62
I15
I25
Persistent Identifier of the first edition: 
Document Type: 
Article

Files in This Item:
File
Size





Items in EconStor are protected by copyright, with all rights reserved, unless otherwise indicated.