Selection Patterns and Economic Variables in the Civil Services Examination: A Probabilistic Assessment of Service Allocations and Opportunity Costs (2021–2024)

UPSC CSE: A Probabilistic Analysis of Selection Patterns (2021–2024)

UPSC Civil Services Examination: A Probabilistic Analysis of Selection Patterns (2021–2024)

First-Timers vs. Candidates "Selected at Least Once" — What the Data Reveals

⚠ Disclaimer

This analysis is generated with AI assistance. The data has been parsed from official UPSC service allocation lists (2020–2024) using automated name-matching algorithms. The analysis is intended purely for educational and informational purposes — to help candidates understand examination patterns, probabilities, and trends. This document should NOT be used to make career decisions, financial decisions, or life choices. Candidates are advised to consult mentors, career counsellors, and official UPSC resources before making any decisions regarding their examination strategy. The Author has also written the exam thrice in years 2002, 2003, 2004 and got selected for IPoS, IPS and IAS, and firmly believes all civil services are created equal, but perceived differently because of various notions in society and culture. All services provide equal opportunity to keep the bureaucratic machine of the Indian Government running, which in turn is subservient to the will of the people.

Index of Chapters

§ Assumptions and Methodology5 UPSC CSE 2026 Notification — Restricting Re-Attempts1 The Examination Funnel — Apply, Write, Get Selected6 Economics of Repetition — Opportunity Costs & Superannuation2 The "Selected at Least Once" Phenomenon7 Why Give Your Best Shot — The Repeated First Attempt3 Adjacent-Year Re-Selection — Immediate Return Cycle 4 Service Selection Probability — FTS vs. SAO

§ 3.4 covers adjacent-year (Y→Y+1) transitions only  |  § 4.3 covers lifetime first-to-last trajectories across the full 2020–2024 window

Assumptions and Methodology

1- Name-based matching: Candidates appearing across multiple years are identified by matching cleaned, standardised names (uppercase, special characters removed, whitespace collapsed). Minor spelling variations may cause missed matches (false negatives); common names may produce false positives.

2- Data cleaning: Two categories of data quality issues were identified and addressed:

• True duplicates: Same name + same service appearing twice in the same year; 10 such rows removed.
• Ambiguous common names: Same name with different services in the same year — different individuals; 33 such names excluded from all cross-year matching.

3- Reference year vs. analysis years: CSE 2020 (673 usable candidates) is used solely as a reference year. The 2020 cohort is not itself analysed. All analysis covers CSE 2021, 2022, 2023, and 2024.

4- "Selected at Least Once" (SAO): Candidates whose names also appear in any prior year's allocation list (2020 onward).

5- "First-Time Selectees": Names that do NOT appear in any prior year's list. They may not be first-time exam takers — only first-time selectees within our detection range.

6- "Adjacent-year re-selector": Selected in year Y and also in Y+1 — actual name-matching.

7- Aggregate statistics: Data is drawn from PIB press releases and UPSC annual reports; figures are approximate where exact figures were unavailable.

8- No category or reservation data: No category or reservation data was used. All figures are all-category aggregates.

9- Service grouping: IAS, IFS, IPS, and "Other Central Services" (IRS-IT, IRS-C&IT, IRMS, IDAS, IA&AS, ICLS, IIS, IPoS, DANICS, DANIPS, IDES, ITS, IRPFS, IP&TAFS, ICAS, AFHQ, PONDICS, etc.).

10- Service hierarchy: IAS > IFS > IPS > Other Central Services.

11- Financial calculations: Calculations use the 7th CPC pay matrix and standard investment return assumptions.

Data Cleaning Summary

33 ambiguous names (common names like ASHISH KUMAR, SANDEEP KUMAR, NIDHI, etc. appearing with different services in the same year) were excluded from all cross-year matching. Full list documented in the accompanying Excel workbook.

Chapter OneThe Examination Funnel — How Many Apply, Write, and Get Selected?

1.1 Year-Wise Examination Statistics

Sources: PIB press releases, UPSC Annual Reports.

1.2 Stage-Wise Elimination Probabilities

~10.9 Lakh AppliedAverage across 2020–2024▼ ~50–55% appearance rate~5.8 Lakh Appeared for Prelims▼ ~1.5–2.5% qualification rate~14,000 Qualified for Mains▼ ~19–22% interview rate~2,800 Called for Interview▼ ~33–40% selection rate~1,000 Finally Selected

The Preliminary Examination eliminates approximately 97–98% of candidates who appear. Once a candidate reaches the interview stage, the selection rate improves to roughly one-in-three — the Mains examination is the true differentiator.

1.3 Parsed Allocation Data

The 3,303 usable allocations across the four analysis years represent 2,939 unique individuals. An additional 673 usable individuals in the 2020 reference year provide the baseline against which "Selected at Least Once" status is determined.

• • •

Chapter TwoThe "Selected at Least Once" Phenomenon — Repetition Across Years

2.1 Year-Wise Composition: First-Time Selectees Vs. Selected at Least Once

For each analysis year, a candidate is classified as "Selected at Least Once" (SAO) if their name appears in any prior year's allocation list (starting from 2020). Only usable (non-ambiguous) names are classified.

SAO Proportion Growing Steadily (2021–2024)2021 SAO (8.9%) based on 1-year lookback; 2024 (20.8%) based on 4-year lookback — most reliable figure.

The proportion of SAO candidates has grown steadily from 8.9% in 2021 to 20.8% in 2024. By 2024, approximately one in five usable selectees had already been selected in at least one prior year.

Note: The 2021 SAO figure (8.9%) is based on only one year of lookback (2020). The true SAO proportion in 2021 is likely higher, as candidates selected before 2020 are invisible to our detection. The 2024 figure (20.8%), with four years of lookback, is the most complete and reliable.

2.2 Unique Candidate Summary (2021–2024)

2,939Unique Individuals (usable)456Selected at Least Once (15.5%)2,483First-Time Selectees (84.5%)364Re-selection Instances

2.3 Frequency of Re-Selection (Within 2021–2024)

Selection Frequency Distribution

Total re-selection instances (usable allocations minus unique individuals) = 364 across the four analysis years. These 364 "seat-opportunities" were consumed by candidates who had already been allocated a service at least once before.

2.4 Year-Pair Overlap Matrix (Usable Names Only)

Cross-Year Overlap — Candidates Appearing in Both YearsAdjacent years show highest overlap; diminishes with distance as candidates exit the cycle.

2.5 Implications

For every 1,000 usable seats in 2024, approximately 208 were taken by returning selectees, leaving roughly 792 for new entrants. Across the four analysis years, 364 seat-opportunities were consumed by re-selection.

Chapter ThreeAdjacent-Year Re-Selection — The Immediate Return Cycle and Service Improvement Probability

This chapter examines candidates selected in year Y who re-appear in year Y+1 — the most consequential form of re-attempting. All figures are from actual name-matching on usable (non-ambiguous) names only.

3.1 Adjacent-Year Re-Selection Rates

Approximately 8.5–11.3% of each year's selectees re-appear in the very next cycle, consuming 7.3–10.9% of the next year's seats. The rate has been trending upward.

3.2 Where Do Adjacent-Year Re-Selectors Come From?

Prior Service of Adjacent-Year Re-Selectors~66–76% consistently from Central Services; ~19–32% from IPS

3.3 Service Improvement Probability

Outcome of Adjacent-Year Re-Selection: Did They Upgrade?

~58–68% of adjacent-year re-selectors achieve a service upgrade. But ~21–31% remain in the same tier, and ~7–11% move to a different tier in the opposite direction.

3.4 Service Transitions in the Immediate Next Year — Aggregated Across All Four Adjacent Pairs

The table below shows where candidates moved when re-selected in the very next examination cycle (Y → Y+1 only). This captures only the immediate one-year return pattern.

Adjacent-Year Service Transitions Only (Y → Y+1, All Four Pairs Combined)

The most frequent transition pathways: Central Services → IAS (68), Central Services → IPS (61), IPS → IAS (36) — accounting for 165 of 186 total service changes to a higher-ranked group.

3.5 IAS Selection Rate Among Re-Selectors

One-third to two-fifths of adjacent-year re-selectors are allocated IAS in their subsequent selection — a proportion notably higher than the ~18% share IAS represents in the overall allocation.

3.6 The Non-Adjacent Gap: Skipping a Year

Substantial numbers skip a year and return — the re-attempt cycle is a persistent, multi-year phenomenon.

3.7 Summary

8.5–11.3%of selectees re-appear next cycle~70%come from Central Services~62%achieve service upgrade~36%are allocated IAS• • •

Chapter FourService Selection Probability — First-Time Selectees vs. Selected at Least Once

4.1 Service-Wise Breakdown by Year

CSE 2021

CSE 2022

CSE 2023

CSE 2024

4.2 Proportion of SAO Candidates in IAS Over Time

SAO Proportion by Service Group (2021–2024)

By 2024, nearly four out of every ten IAS seats (among usable names) went to candidates who had been selected in a prior year. The steady acceleration from 13.7% to 37.6% over four years is consistent.

4.3 Lifetime Service Trajectory of All 456 SAO Candidates — First Selection to Most Recent (Full 2020–2024 Window)

Unlike Section 3.4 (which tracked only adjacent-year Y → Y+1 transitions), this section traces each SAO candidate's entire journey from their first-ever selection to their most recent selection across the full five-year window. A candidate selected in 2020, 2022, and 2024 is represented here as a single trajectory from their 2020 service to their 2024 service — regardless of what happened in between.

Among the 456 SAO candidates (usable, 2+ selections, present in 2021–2024):

Overall Trajectory Outcomes — 456 SAO Candidates

163 of 456 SAO candidates (35.7%) ultimately ended up in IAS. This means that nearly two-thirds of SAO candidates did not achieve the IAS allocation despite being selected multiple times.

4.4 What This Means for a First-Time Selectee Targeting IAS

A first-time selectee targeting IAS in 2024 is competing for approximately 106 first-time selectee IAS seats out of 170 usable total — the remaining 64 were taken by SAO candidates. The effective competition for first-time selectee IAS seats is substantially more intense than the headline numbers suggest.• • •

Chapter FiveThe UPSC CSE 2026 Notification — Restricting Re-Attempts by Serving Officers

5.1 What Changed

The UPSC CSE 2026 notification (February 4, 2026) introduced landmark changes:

Pre-2026 Rules

IAS/IFS officers: Permitted to attempt exam while serving.

IPS officers: Permitted to attempt and opt for IPS again.

Central Services: Unlimited attempts within age/attempt limits.

Post-2026 Rules

IAS/IFS officers: Completely barred. Resignation mandatory.

IPS officers: Can attempt, but cannot be allocated IPS again.

Central Services: One-time improvement window only. Non-repeatable.

2028 cutoff: Resignation mandatory for any serving officer to re-attempt.

5.2 Seats Not Available for First-Time Selectees Due to "Selected at Least Once" Candidates

Our cleaned data shows that 20.8% of 2024 selections and 37.6% of IAS seats went to SAO candidates. The adjacent-year data (Chapter 3) shows ~8.5–11.3% of selectees immediately re-appear, with ~66–76% from Central Services and ~19–32% from IPS. Each repeated selection creates a cascading displacement chain through the service ladder.

5.3 Projected Impact

If the 2026 rules had been in effect, approximately ~125 additional seats per year (averaging across the four analysis years) would have been freed for first-time selectees. For IAS, approximately ~44 additional first-time selectee seats per year — a meaningful improvement in selection probability.

5.4 The Debate

Proponents argue for a more level playing field for first-time selectees. Critics counter that restrictions may limit legitimate career choices. UPSC's stated rationale: administrative efficiency, training resource optimisation, and fairness to new entrants.

Chapter SixThe Economics of Repetition — Opportunity Costs, Near-Term Salary Loss, and Superannuation Penalties

Engaging in repeated examination cycles carries a profound financial penalty — the opportunity cost. This manifests in two devastating phases: the immediate hemorrhage of near-future earnings, and the compounding destruction of terminal wealth and pension benefits at superannuation.

6.1 The Financial Architecture of the Civil Services (7th Pay Commission)

Promotions are bound to years of service, not age or rank. This makes the year of entry the single most consequential financial variable in an officer's career.

Level 18Cabinet Secretary (37+ yrs)₹2,50,000Level 17Apex Scale (34–36 yrs)₹2,25,000Level 15Above Super Time Scale (25–30 yrs)₹1,82,200Level 14Super Time Scale (16–24 yrs)₹1,44,200Level 12Jr. Administrative Grade (9–12 yrs)₹78,800Level 11Senior Time Scale (4–9 yrs)₹67,700Level 10Junior Scale — Entry Level (0–4 yrs)₹56,100

Gross monthly at entry scales to ₹90,000–₹1,00,000 with DA, HRA, and TA.

6.2 The Near-Term Salary Haemorrhage

6.3 The Mathematics of Compounding: Wealth Destruction Through Delay

Candidate A — Enters workforce at 23. Invests ₹40,000/month (10% return) for 5 years (age 23–28), then stops adding money.

Candidate B — Prepares full-time ages 23–28. Zero income. Enters service at 28.

Candidate A's ₹24 lakh grows to over ₹5 crore purely from time-in-market. The five years of lost compounding destroys roughly ₹2 crore from this tranche alone.

6.4 Full Career Compounding Impact

Compounding Curves: Candidate A (37 yrs) vs Candidate B (32 yrs)10% CAGR, ₹5L starting annual savings, 5% annual increment. A ends at ~₹31Cr, B at ~₹18Cr. Delta: ~₹13 crore at age 60.

6.5 The Superannuation Penalty — The Career Ceiling Effect

Because pensions are derived from the last drawn salary, years spent preparing for exams directly amputate the highest-paying years — a lifelong financial penalty extending into retirement. The pension differential between Level 15 and Level 17 is approximately ₹40,000–₹65,000 per month, or ₹5–8 lakh per year. Over a 20–25 year retirement, this exceeds ₹1–2 crore.

6.6 The Private Sector Comparison

A premier-institution graduate taking a private sector job at 23, investing ₹4 lakh/year at 10% for 37 years with 8% annual increment, accumulates approximately ₹12–15 crore — comparable to or exceeding civil services, with more career flexibility and no examination stress. The point is not that private employment is superior — civil services offer irreplaceable public impact and constitutional authority — but the financial cost of delay is substantial, mathematically irreversible, and almost universally underestimated.

6.7 Total Cost of a 5-Year Delay

₹55–65LDirect salary foregone₹2–13CrLost compounding1–2 levelsReduced terminal pay₹1–2CrPension loss (20 yrs)

The years between 22 and 28 are the most financially consequential years of an officer's career — they sit at the base of the compounding curve.

Chapter SevenWhy Candidates Should Give Their Best Shot at the First Attempt — or the Repeated First Attempt

7.1 The Data Makes the Case

The analysis converges on a single insight: the examination rewards decisive, well-prepared attempts over prolonged, incremental re-tries.

7.2 The Odds Do Not Improve with Repetition

Syllabus evolution: Current affairs are entirely new each cycle. Additional years add the burden of staying current alongside deepening existing knowledge.

Diminishing returns: The first 6–12 months yield the steepest improvement. Subsequent years show diminishing marginal gains.

Psychological erosion: Self-doubt, comparison anxiety, and burnout can reduce performance in subsequent attempts, creating a vicious cycle.

7.3 The "Repeated First Attempt" Philosophy

The most successful candidates treat every attempt as if it's their only one — 100% preparation intensity each cycle. The data supports this: 88.3% of unique individuals in our analysis years appear only once in the allocation lists.

7.4 The Changing Regulatory Landscape

CSE 2026 introduces strict re-attempt restrictions. The "join a lower service, upgrade later" strategy now carries: mandatory resignation from CSE 2028 onward, loss of seniority, a single one-time improvement window, and the psychological burden of having something to lose.

7.5 The Adjacent-Year Evidence (Chapter 3 Revisited)

Even among the most determined re-selectors — those returning in the very next year — only ~58–68% achieve a service upgrade. ~33–42% are allocated IAS in the subsequent selection. The remaining ~32–42% stay in the same tier or move to a different one.

7.6 A Framework for Decision-Making

Before re-attempting, evaluate honestly:

Marginal benefit: Is the target service significantly different in career satisfaction and lifestyle? IPS, IFS, and Central Services offer deeply fulfilling careers.

Marginal cost: A 3-year delay costs several crores in terminal wealth. A 5-year delay can destroy ₹4–17 crore in lifetime financial value — including lost compounding, reduced terminal pay, and diminished pension.

Marginal probability: Of 456 SAO candidates, only 163 (35.7%) ultimately landed in IAS — nearly two-thirds did not achieve the IAS allocation despite multiple selections.

Changing rules: CSE 2026 norms introduce resignation risk, training disruption, and irreversible administrative consequences.

7.7 The Bottom Line

Every attempt is precious. Every year of career is irreplaceable. The numbers are clear: prepare well, execute decisively, and make your attempt count.❧

Analysis: Purushottam B.  |  Data: UPSC Service Allocation Lists 2020–2024  |  AI Assistance: Claude (Anthropic)

Methodology: Official UPSC service allocation files for CSE 2020–2024 parsed using Python (pandas). Names standardised to uppercase with special characters removed. Data cleaning: 10 true duplicate rows removed; 33 ambiguous common names excluded from all cross-year matching. CSE 2020 (673 usable candidates) used as reference year. Analysis covers CSE 2021–2024 (3,303 usable allocations, 2,939 unique individuals). Adjacent-year and service trajectory analyses based on actual name-matching of usable names only. Aggregate examination statistics from PIB press releases.

Published on Stories Through Data (darkorange-squirrel-769791.hostingersite.com)