Burning ₹168 to Earn ₹100

Tamil Nadu’s civil service hopefuls give up nearly 1.68 times as much in lost earnings (and coaching fees) as the state will ever pay in salary (see Table 2.5 in Mangal, 2023 (PDF); see MR as well). At first pass, this looks like over-dissipation: candidates appear to spend 68 percent more than the prize. As Alex notes, "Classical rent-seeking logic predicts full dissipation: if a prize is worth a certain amount, rational individuals will collectively spend resources up to that amount attempting to win it." So what causes over-dissipation?

A government post offers far more than its paycheck. Lifetime tenure, predictable hours, social status, lighter monitoring, subsidized housing, all add what labor economists call amenity value. A recent World Bank study of Western India applicants estimates that amenity value is at least two-thirds of the compensation (Mangal 2024). Add to it the possibility of kickbacks, and the prize rises further. Thus, ₹168 could simply be full dissipation calculated against the true private prize. (Or it could be that we have under-dissipation!)

Could something else explain the overspend? Classically, overoptimism can explain over-dissipation. Kunal’s survey shows that many candidates believe their chance of selection on retaking the exam is several times the historical rate.

One might object that treating exam prep as pure social waste overlooks two potential benefits: improved employer-employee matching and human capital development. However, data refute both possibilities. If exams effectively differentiated candidates by ability, we would expect meaningful score variation among those selected. Instead, Kunal finds extreme score compression at the selection margin. In the Group 4 exam, only 51 marks separate the top-ranked candidate from the candidate ranked 20,000th, meaning each candidate in this range shares the same score as over 1,000 other candidates on average. Such compression in rankings signals poor measurement reliability, offering little value to employers.

The human capital argument fares no better. Mangal exploits a hiring-freeze shock to track male college graduates over a decade. Cohorts most exposed to the freeze—and thus forced into extended exam preparation—cut their early-career paid work by 13 percent. Even ten years later, they remain nine percentage points less likely to be employed, earn less on average, and delay household formation. Macro shocks muddy the causality, but the pattern hardly screams "skill upgrade."

Policy Levers

1. Shrink the cash-plus-perk bundle ($\tau$). Cut salary premia, publish perk ledgers, and improve outside options; bids fall automatically when the prize shrinks.
2. Bust the halo ($\lambda$). Information campaigns that show the unglamorous side of the job trim imagined perks. (Or, make the amenity value close to 0 and put all the perks in wages, assuming people are rational.)
3. Debias success odds ($\omega$). Release precise pass-rate data; mock tests with honest feedback collapse over-optimism.
4. Decouple effort from probability ($\mu$). After a competence threshold, pick by lottery, or require a refundable bond so extra cramming converts into a cash transfer, not a deadweight study.

None is a silver bullet—bonds could deter poor but talented candidates; lotteries invite litigation, but each targets a distinct jet of fuel.

Modeling Waste

$$B \;=\; (\tau-1)\; + \;(1-\mu)\,\omega\,\tau\,(1+\lambda)$$

Justification:

• $\tau = 2.0$: Cash wages alone carry a 105 % premium (Finan, Olken & Pande 2017).
• $\lambda = 0.66$: Exam candidates value the job at ₹250 k vs ₹81 k salary—amenities = 66 % of total pay (Mangal 2024).
• $\omega = 25$: Surveyed re‑applicants overestimate pass odds by 25×+. "Candidates are way too optimistic about the likelihood of success when re-applying, by a factor of 25 to 500 times."(Mangal 2024).
• $\mu = 0.05$: Evidence of a hiring‑freeze cohort shows prolonged prep barely improves long‑run labor‑market outcomes, suggesting < 5 % of the study is transferable.

Interpretation. The government overspends by τ − 1. Candidates then fully dissipate the perceived prize τ(1 + λ) W; only the productive slice μ survives as human capital.

With the baseline numbers
$B_₀ = 79.85 ×$ outside‑option wage.

Four levers – each pulled in isolation

All four levers together

$$\tau=1,\;\lambda=0,\;\omega=5,\;\mu=0.60 \;\Longrightarrow\; B_{\text{joint}} = 2.00$$

Joint reduction: 97.5 % relative to the status quo.