Page 504 - ISC PROCEEDINGS 21.4

Resource-Based View (Barney, 1991) and Digital Business Strategy literature (Bharadwaj
et al., 2013) predict that heterogeneous resource endowments - particularly in digital
infrastructure and data capabilities - produce persistently differentiated competitive
positions, justifying a cluster-based rather than continuous-scale analytical approach.
Empirically, the classification reflects the structural segmentation observable in Vietnam’s
banking sector: state ownership (C1) creates fundamentally different investment
incentive structures, capital allocation mechanisms, and performance mandates relative
to private banks (C2–C4); within private banks, business model orientation (enterprise
lending vs. retail mass market vs. small institution) drives systematically different digital
investment priorities and capability profiles. This typology is consistent with the
segmentation adopted in the SBV’s own regulatory and supervisory classifications,
enhancing its policy relevance. A key limitation is that cluster assignment is theory-guided
rather than data-driven (e.g., k-means clustering), which may introduce subjectivity. To
mitigate this, all cluster boundaries are anchored to publicly verifiable thresholds (state
ownership status; IT expenditure intensity ≥18% vs. below), reducing discretionary
classification. Future research employing latent class or hierarchical cluster analysis could
provide a more fully data-driven validation of this typology.
While data-driven clustering methods such as k-means or hierarchical clustering
could be applied, they require fully standardised and continuous input variables, which
are not consistently available in the Vietnamese banking context due to heterogeneous
disclosure practices. The threshold-based classification adopted in this study ensures
transparency, replicability, and policy relevance, as all classification criteria are anchored
in publicly observable indicators. The full distribution of 27 listed banks across the four
clusters is as follows: C1 - 4 state-owned banks (BIDV, CTG, VCB, Agribank); C2 - 6
enterprise-oriented banks (TCB, MBB, VPB, MSB, SSB, LPB); C3 - 8 retail-oriented banks
(ACB, STB, VIB, TPB, HDB, EIB, BVB, KLB); C4 - 9 small and medium-sized banks (NVB, ABB,
SGB, BAB, PGB, VBB, BID, NAB, OCB and peers). This distribution is consistent with
publicly available listings on HOSE and HNX as of end-2024.
3. Results and discussion
3.1. Digital infrastructure investment: scale, structure, and development trajectory
Vietnam’s banking sector significantly increased digital infrastructure investment
over the study period. IT expenditure as a share of operating costs, which stood at 5–7%
before 2020, reached 14.85% by 2024, equivalent to approximately VND 32,437 billion
system-wide (MIC, 2018–2024). This level approaches the 15–20% international
benchmark applied by digital-first banks such as DBS (Singapore) and KakaoBank (South
Korea) (World Bank, 2022; Frost, 2020).
Table 1. Technology Expenditure by Bank Cluster in Vietnam, 2024
IT expenditure Share of operating costs (%) - Key
Bank cluster
(VND billion) characteristics
State-owned banks 8,415 11.44% - Focus on core infrastructure,
(BID, CTG, VCB, AGR) cybersecurity, large-scale projects

Enterprise-oriented 19.89% - AI, Big Data, cloud computing,
(TCB, MBB, VPB, 13,743 enterprise data applications
MSB…)

Retail-oriented (ACB, 13.72% - Digital banking apps, super-apps,
STB, VIB, TPB, 7,234 customer experience

503