📄 Abstract
Firms treat training expenditure as an investment recovered over a horizon, but they choose that horizon administratively while the asset it buys depreciates at a rate set by the technology frontier. This study measures the distance between the two. The training amortisation horizon (A) is the period over which a firm expects to recover a training outlay, observed through four verifiable practices: the certification currency window in the staffing system, the curriculum refresh cycle, the management-accounting payback horizon and the skills-inventory decay rule. The skill half-life (H) is the observed time until half the economic value of the trained capability has decayed, estimated from exponential fits to skill-value trajectories. Their ratio, Γ = A / H, is the study's central quantity, and the resulting stranded share of expenditure is directly interpretable as misallocated HRD budget. Evidence comes from a retrospective longitudinal cohort of 1,240 training completions in 38 organisations, tracked across four waves over 36 months, with 55 learning and finance leader interviews. Half-lives range from 1.1 years for generative-AI prompt and application patterns to 11.5 years for leadership capability, while the horizon firms apply ranges only from 4.2 to 5.3 years. Half-life varies 10.4 times more across skill clusters than the horizon applied to it, and 84.0 per cent of HRD spend sits on the over-amortised side of the crossover at H ≈ 4.6 years. The spend-weighted horizon is 4.65 years against a spend-weighted half-life of 3.12 years, a gap of 1.53 years and a ratio of 1.49. Discounting the decaying benefit flow at 12 per cent against the recovery multiple firms themselves assume (2.84×) yields a realised multiple of 1.60× and a stranded share of 43.7 per cent — ₹43.70 of every ₹100 committed. Realised return falls by 0.284 for each unit increase in Γ. Half-life is not invariant to use: skills deployed on more than sixty per cent of billable hours decay 2.9 times more slowly than those below fifteen per cent, so obsolescence is partly an allocation outcome rather than a purely technological one. The eight organisations that differentiate horizon by cluster strand ₹27.90 against ₹48.60 for those applying a uniform horizon. The study supplies a cluster-differentiated amortisation schedule and argues that skills with half-lives beyond eight years should be reclassified from amortising training capital to capability endowment.
🏷️ Keywords
📚 How to Cite:
Dr. N Subbukrishna Sastry, Dr. Manjula Mallya M , HRD INVESTMENT AND THE ECONOMICS OF SKILL OBSOLESCENCE UNDER GENERATIVE AI , Volume 13 , Issue 9, September 2026, EPRA International Journal of Socio-Economic and Environmental Outlook(SEEO) , Pages: 38 - 57 ,