Who Did Better: UPA or NDA?

The complete report has seven parts and one 204-page master PDF. The research cutoff for this part is 3 October 2026. Every figure retains its own observation year. Page references in the original reading guide below refer to this volume’s printed pages within the master PDF; use the contents here to move through the web edition.

I wanted an answer I could defend.

A final judgment by Lovepreet Singh.

After six volumes, I did not want to leave you with a pile of numbers and no answer. The question was which government did better. This last volume gives my answer, explains how I reached it and keeps the strongest objections beside it.

On MisterLove, I try to make a difficult subject readable without asking you to learn a new language first. Here, I start with a few ordinary questions. Did daily services reach more people? Did incomes grow faster? Did people become healthier? Could they find decent work? Could they question those in power?

My overall judgment gives the BJP-led NDA a qualified advantage when I put greatest weight on infrastructure execution, basic-service reach and financial inclusion. That is a declared choice of priorities. Congress-led UPA has the stronger observed ten-year GDP-growth record in the common historical series, and the stronger case on the independent freedom assessments examined here.

I cannot turn those different outcomes into a scientific score proving that one government was better for every person. I can tell you where the evidence is strong, where the comparison is weak and why I still make the judgment I make.

COVID belongs in this assessment. So do the global financial crisis, oil-price changes, policy disruptions and unfinished work inherited from earlier years. A government's response deserves scrutiny even when the original shock came from outside.

This is my reading of published evidence. It is not an official finding, a claim of personal fieldwork or a request that you support a party. I want the reasoning to remain useful even when you disagree with my priorities.

Research cut-off: 3 October 2026. Final volume prepared 4 October 2026. Observations, calculations and editorial judgments are distinguished throughout.

Sources: 1, 3, 43.

Who did better? My answer.

A clear conclusion, with the conditions attached.

My overall choice in this report is the BJP-led NDA after 2014, with a qualified lead on material delivery. By material delivery, I mean useful public networks and wider access to everyday services: transport, power, taps, sanitation, bank accounts and payment systems. The later record has substantial, documented expansion in these areas.

I give those outcomes the largest weight because they affect daily life and support later improvements. I count infrastructure and connected services as one broad group, rather than treating every road, pipe and bank-account row as a separate political point.

Congress-led UPA did better on observed real GDP growth across the equal ten-year periods. The annual compound rates were 6.80% in 2004-05 to 2013-14 and 6.06% in 2014-15 to 2023-24. The NDA figure includes the pandemic. NDA's first six years, mostly before COVID, compounded at 6.79%, almost the same as UPA's full-decade rate.

UPA also has a stronger case on the freedom assessments considered here. Neither that advantage nor the growth result proves that every UPA policy was good. They do mean that 'BJP was better in everything' is unsupported.

Health outcomes and poverty improved across both periods. The available jobs, wages, learning, safety and environmental evidence does not establish a clean overall winner. These areas limit how confidently I can call either era a complete success.

My final sentence is therefore specific: NDA has the stronger case for delivery and expansion; UPA has the stronger measured full-decade growth record and a stronger case on institutional freedom. I choose NDA overall only under the delivery-first priorities disclosed in this report.

The overall choice is editorial judgment, not a statistically estimated causal effect or an official national score.

Sources: 3, 16, 19, 22, 43.

Congress before 2014 needs dates.

The main comparison is UPA 2004-2014 versus NDA from 2014.

'Before 2014' is not the same as 'under Congress'. India had other governments, including an earlier BJP-led NDA. An asset already present in 2004 cannot automatically be credited to UPA. Equally, an asset present in 2026 cannot automatically be credited entirely to Modi.

Both UPA and NDA are alliances. Their central policies also operated alongside governments in the states, local authorities, RBI, public institutions, businesses and citizens. I use the alliance names when making government-level judgments.

FY means financial year. FY2013-14 runs from April 2013 to March 2014. Governments changed in May 2004 and May 2014, so both starting financial years contain a short period under the predecessor. Annual data cannot isolate the exact handover day.

For growth, ten years are compared with ten years. For the latest condition of India, I also use the newest suitable observations checked by the cut-off. Those observations are not all from 2026: poverty may describe 2023-24, crime 2024, and health a multi-year period.

This volume adds a judgment to the six completed reports. It does not replace their detailed tables, and it does not claim to have verified every post ever published by The Matrix India.

The six evidence volumes total 168 pages. This additional final volume uses their findings plus newly checked earlier baselines and growth calculations.

Sources: 2, 59.

A larger number is only a start.

I ask about pace, use, quality, fairness and responsibility.

First, I separate the level from the pace. A later government should normally inherit a larger economy and more roads. The useful question is how much changed during its term, over how many years, and what the change meant.

Second, I compare the same thing. Capacity is what equipment could produce; output is what it actually produced. A sanctioned hospital is an approved project; a fully functioning hospital treats people with working staff and facilities. An account or tap connection is not proof of regular use.

Third, I remove inflation when judging purchasing power. Rs 100 becoming Rs 200 does not double what a family can buy if prices also doubled. Physical units, such as tonnes or route kilometres, do not need that same price correction.

Fourth, I keep shared achievements together. Roads, railways, ports and logistics are connected. Counting every indicator as a separate vote would let a long infrastructure chart overpower health or freedom simply because it contains more rows.

Fifth, I distinguish a recorded result from its cause. A government can influence a result without causing all of it. Weather, technology, inherited projects, private investment, state decisions and world conditions can also matter.

Finally, I disclose my priorities. Delivery receives the largest weight in my overall choice. Growth, human outcomes, work quality, safety and rights remain independent checks. I use no invented percentage of political credit, no fake marks out of 100 and no claim that the evidence forces everyone to choose the same winner.

A verdict is an interpretation of several measures. None of the sources supplies a universal Congress-versus-BJP performance score.

Sources: 59, 5.

The equal ten-year comparison.

Real GDP measures output after removing price changes.

Across the decade, UPA's real GDP rose about 93.0%; NDA's first-decade increase was about 80.1%. The compound annual rates were 6.80% and 6.06%. A compound rate is the steady annual pace that would produce the same start-to-end change.

The simple averages of the annual rates are 6.81% and 6.15%. They differ from the compound result because a fall followed by a rise must multiply, rather than simply cancel. This report uses the compound measure for its main growth verdict.

The table retains one published historical framework: the Economic Survey's January 2026 snapshot of the 2011-12 GDP series. The new 2022-23-base estimates are not spliced into these older observations.

Verdict on the observed full-decade growth pace: UPA leads. COVID accounts for an important part of NDA's weaker result; the next page shows how much that qualification matters.

Annual rates are calculated from published GDP levels and rounded to two decimals. FY2023-24 is revised. These are observed national outcomes, not estimates of a government's isolated contribution.

Sources: 3, 5.

Keep the fall and the rebound together.

A recovery percentage can look large because the earlier year was weak.

The pandemic fall was 5.78%, followed by 9.69% growth. Taken together, output rose about 3.35% over the two years, or 1.66% a year compounded. The recovery year cannot be used as if the loss had never happened.

For a sensitivity check, I remove both pandemic years from NDA's first-decade annual-rate list. The eight retained rates compound to 7.19% a year. If I also remove FY2008-09 and FY2009-10, the financial-crisis-and-recovery block, from UPA's list, its eight retained rates compound to 7.14%. That tiny gap does not establish a decisive winner.

These retained-year calculations are not continuous historical periods. They do not reconstruct an India without either crisis. Choosing different affected years can change the answer. They show how sensitive a verdict is to the way shocks are handled.

My conclusion: UPA's observed decade was faster. The evidence does not support claiming a large underlying growth advantage for either government once the major shock blocks are selectively removed. NDA's slowdown was already underway before the main pandemic years; FY2019-20 recorded 3.87% growth.

FY2019-20 includes early COVID effects and the final week's lockdown. Both shocks affected other years too. Deleting years is a sensitivity exercise, not a causal adjustment or forecast.

Sources: 3, 10.

The income race is much closer.

National income per person is an average, not a typical salary.

UPA's ten-year increase was 59.49%, equivalent to 4.78% a year compounded. NDA's first-ten-year increase was 58.64%, or 4.72% a year. The difference is small enough that I treat these two average-income growth records as essentially close, rather than a dramatic victory.

Through FY2024-25, NDA's eleven-year increase was 67.28%, or 4.79% a year. That extra year changes the comparison length. It is useful recent context, not an equal-ten-year win.

These amounts use constant 2011-12 prices. They are tools for comparing buying power over time, not the money you would see deposited into your account in those years.

An average can rise while many people struggle. If one person's income grows much faster than everyone else's, it pulls up the average. The middle person's income, wages adjusted for prices, and dependable days of work would tell us more about ordinary households.

Verdict on the measured income pace: the two equal decades are close. Verdict on today's income level: it is higher. Neither statement establishes that every family became better off, or that one party alone caused the increase.

Calculations use per capita net national income, not GDP per capita or a household salary series. FY2024-25 is provisional in the retained snapshot.

Sources: 4, 34.

UPA did not govern in calm weather.

Outside shocks explain part of its record. Domestic weaknesses also matter.

The global financial crisis of 2008-09 damaged world demand, finance and business confidence. India faced tighter credit, weaker exports and slower activity. UPA used additional public spending, tax reductions and sector support; RBI also responded through monetary measures.

The common historical GDP series shows growth falling to 3.09% in FY2008-09 and recovering to 7.86% in FY2009-10. Contemporary reports used older methods and published different figures. I use those reports for policy history, and the common series for numerical comparison.

Later, international financial volatility put pressure on India. In 2013, the expected withdrawal of exceptionally easy US monetary policy encouraged money to leave vulnerable emerging markets. That helps explain currency and financial-market pressure, but does not remove domestic responsibility.

The IMF's February 2014 assessment identified persistent inflation, weak investment, domestic supply constraints and deteriorating corporate and bank finances alongside external pressures. Those weaknesses made the outside shock harder to absorb.

UPA deserves credit for substantial growth and poverty reduction despite a major world crisis. It also deserves criticism for its late-period slowdown, investment obstacles and weak financial conditions. Calling the whole record a disaster ignores the decade; calling every late problem an outside event ignores the domestic diagnosis.

My treatment: the crisis counts as a real external disadvantage. UPA's response is part of its governing record. So are the vulnerabilities that remained when it left office.

No attempt is made to estimate how India would have grown under a different government during the same crisis.

Sources: 6, 3.

NDA had shocks and advantages.

Fairness requires counting help as well as harm.

The large fall in world oil prices after 2014 helped India, a major oil importer. The IMF described benefits for economic activity, the external account, the budget and inflation. Monetary and supply measures also contributed; the exact size of the oil effect is debated.

Weather and harvest conditions could work in the other direction. Poor rainfall, heat and floods affect food supplies and rural incomes. They do not originate in a prime minister's office, but preparedness, irrigation, relief and adaptation remain matters of policy.

Demonetisation in November 2016 was a government decision. The resulting shortage of cash was described by the IMF as an adverse temporary shock. It belongs in the assessment of NDA policy, rather than in the same category as a virus or a foreign financial crisis.

GST began in July 2017 after years of groundwork. Its national framework is an important reform; its design, transition and administration also deserve assessment. A chosen transition cost cannot simply be deleted from the record as bad luck.

COVID in 2020-22 was a severe outside shock. Russia's invasion of Ukraine in 2022 added global commodity and inflation pressure. Conversely, inherited systems, private investment and technological improvements helped India expand services and recover.

My treatment: credit NDA's response and delivery where the evidence supports it; recognise the difficult circumstances; keep policy choices and beneficial outside conditions visible. The result should not depend on remembering only the shocks that help one side's argument.

This covers major relevant disruptions and enabling conditions; it is not an exhaustive list of every event from 2014 to 2026.

Sources: 7, 8, 9, 10, 50.

The virus was external. The response was policy.

Both belong in the final assessment.

A national lockdown was announced on 24 March 2020, beginning at midnight, initially for 21 days. Restricting movement aimed to slow infections. It also interrupted ordinary work, travel and spending in a country where many livelihoods depend on daily activity.

The economic damage came from several things at once: the disease itself, fear of infection, restrictions, world disruptions and weaknesses already present. The annual GDP fall cannot identify the exact contribution of each.

India's response included food support, financial measures, monetary easing and later vaccination. IMF analysis recognised support to vulnerable groups and policies that helped recovery. Health workers, state administrations, manufacturers and households also contributed.

Recovery deserves credit, but a rebound is not the same as catching up with the path India might have followed without COVID. That imagined path is uncertain. The report therefore gives the actual two-year change and avoids inventing a precise 'COVID-adjusted' performance figure.

The shock does not explain all earlier weakness: growth was already slowing before it, and FY2019-20 recorded 3.87%. That year includes early COVID effects and the final week's lockdown. Comparing the main pandemic decline with an ordinary UPA year would be unfair; UPA's financial-crisis period must remain visible too.

Verdict: COVID materially weakens any simple claim that NDA's lower decade-average growth proves worse economic management. It does not justify removing the recession, ignoring pre-COVID slowdown or giving the recovery to one leader alone.

GDP calculations describe national output. They do not measure all lost work, learning, health, or the distribution of hardship.

Sources: 11, 10, 3.

A pandemic record cannot end with GDP.

Deaths, care, migration and vaccination also count.

WHO's 2022 modelling estimated 4.7 million excess deaths in India during 2020-21, with an interval of 3.31 to 6.48 million. In Indian units, the central estimate is 47 lakh. This is a disputed estimate of deaths above an expected baseline, including direct and indirect effects. It is not a verified list of 47 lakh deaths all medically recorded as COVID.

The Indian government objected to WHO's modelling and the use of selected data. This report does not independently resolve that dispute. Uncertainty about the total cannot turn the pandemic's human cost into zero, or justify assigning every excess death to one government's action.

The Supreme Court's May 2021 oxygen proceedings and its National Task Force show that oxygen assessment and allocation required urgent attention. NHRC's migrant-worker intervention similarly records serious hardship after the lockdown. These concerns belong alongside the recovery story.

The vaccination programme delivered more than 220 crore doses by 6 January 2023. That was a major logistics achievement. Doses are not unique people: the same person could receive more than one. The achievement also depended on scientists, manufacturers, health workers and state delivery.

There is no observed UPA-managed COVID pandemic against which to test a winner. Any statement that Congress definitely would have done better or worse is an untested claim.

My verdict on NDA's pandemic response is mixed: large vaccination and relief delivery, alongside serious problems in emergency care and protection of vulnerable workers. COVID is an important qualification in the overall verdict, and a major human outcome in its own right.

WHO figure refers to a 2022 estimate for January 2020-December 2021, not the latest all-period registered COVID death total. No claim is made about an individual death's cause.

Sources: 13, 12, 11.

NDA has its strongest case here.

Network expansion is substantial. Earlier work was real too.

Highway stock rose about 25,718 km in 2004-14, or 2,572 km a year. It then rose 55,285 km over roughly twelve years, or about 4,607 km a year. This is network-stock change, which includes designation of existing roads; it is not a new-road construction speed.

The broader later expansion supports NDA's delivery case. Rail electrification and metro growth add to it. Inherited projects, state governments, private contractors and users remain part of the story.

The record also has limits. Railway passenger journeys were about 839.7 crore in FY2013-14 and 741 crore in FY2025-26, while freight loading increased. More infrastructure did not mean a higher figure for every usage measure.

Verdict: NDA has the stronger case on transport-network expansion and execution in the evidence assembled. The precise efficiency of each rupee, and the separate government contribution to each kilometre, are not established.

The two NH stock-change speeds use approximately 10 and 12 years. Detailed category and commissioning qualifications remain in Part 2.

Sources: 14, 15, 16, 17.

More connections can change a day.

The next test is whether the service works and is affordable.

Rural tap-water coverage reported under Jal Jeevan Mission rose from 16.72% in August 2019 to 82.24% in August 2026. That is a large later-period reach improvement. The starting date is 2019, not a Congress-era 2014 baseline.

A recorded tap connection is not proof of safe water at reliable pressure every day. Maintenance, source quality, local capacity and drought resilience determine the service a household actually receives.

The NFHS household measure of improved sanitation rose from 48.5% in 2015-16 to 70.2% in 2019-21. This is more informative about household conditions than treating the administrative declaration of universal open-defecation-free status as an identical survey measure.

Power networks, generation equipment and cooking-gas connections expanded. They can make life easier, but equipment capacity differs from dependable electricity, and a gas connection differs from being able to afford regular refills. I retain these distinctions when awarding delivery credit.

UPA and earlier governments also built rural roads, electricity systems and service foundations. NDA continued and enlarged many of them. A government deserves credit for useful expansion even when it inherited the starting system; inheritance does not make execution automatic.

Verdict: NDA has a strong case on wider basic-service reach. Confidence is lower on uniform service quality and affordability. Those limits matter because the intended result is usable daily access, rather than a completed administrative target alone.

Tap and sanitation observations mostly fall within the NDA period. They show improvement, but do not create matched UPA-era rates of improvement where those measurements are missing.

Sources: 19, 20, 16.

Foundations under UPA. Scale under NDA.

A useful system can have more than one builder.

The account-ownership increase is 35.9 percentage points. Imagine 100 adults: roughly 53 had an account in the earlier survey and 89 in the later one. The 2014 survey is a handover-year observation, not a clean measurement taken immediately before the new government.

UPA's digital institutions and identity groundwork deserve recognition. NDA's account-opening drive, UPI-era expansion and integration into services also deserve recognition. Banks, RBI, NPCI, private firms, cheaper technology and users contributed.

Access must be distinguished from control and use. An account may be dormant; a person may rely on someone else's phone; a payment can fail. Authentication trouble and fraud can impose new costs. Large transaction totals are not the same as new national income.

Verdict: shared foundational credit, with NDA having the stronger case for the later scale and financial-inclusion delivery. Saying digital India began from nothing in 2014 would erase documented earlier work.

The Findex dates are survey years; the latest cited ownership observation is 2024, published in the 2025 database.

Sources: 21, 22.

Poverty fell in both periods.

Do not change the poverty line halfway through the race.

The first row shows substantial UPA-era improvement. The later rows show substantial improvement mostly during NDA, although the World Bank starting year precedes 2014. The Tendulkar and World Bank rates use different lines and methods, so subtracting across their rows would be wrong.

Multidimensional poverty counts several disadvantages, such as health, education and living conditions. It differs from a money-based poverty line. Survey redesign also requires care in the later consumption comparison.

The National Food Security Act of 2013 created a major UPA-era entitlement. NDA provided additional pandemic food support and made existing NFSA grain entitlements free from January 2023. The later free entitlement is not the same as continuing the earlier pandemic's additional quantity.

Verdict: strong poverty reduction and food-security contributions in both eras. NDA has a strong later delivery case, but the available comparisons do not establish a clean winner on the poverty-reduction rate from precisely 2014. Moving just above a poverty line also does not guarantee financial security.

The national Tendulkar baseline differs from the later international PPP lines. NITI's extrapolated 2013-14 estimates are not substituted for an observed survey at the handover.

Sources: 24, 25, 26, 23.

Health improved before and after 2014.

Long-running progress, more capacity, and unfinished access.

The infant-death decline was 20 points over ten years before the handover and 16 points over eleven years afterward. Proportional declines were about 33% and 40%. A lower starting death rate makes later reductions a different task. Maternal estimates average three-year windows rather than single years. Neither comparison isolates a government's exact effect.

The spending shares show a larger public role and a smaller household-payment share. They do not prove that every family pays less. The direct-payment share rose from 39.4% in 2021-22 to 43.4% in 2022-23; progress was not continuous.

States supplied about 63.7% of government health spending in 2022-23. Medical colleges and training seats expanded after 2014, but approved AIIMS projects differ in operational stage. More seats do not instantly become doctors available in every village.

Verdict: both periods made meaningful health progress. NDA has a stronger later capacity-and-coverage case; the mortality evidence does not support awarding all health improvement to one government. COVID and continuing household costs limit a sweeping victory claim.

Earlier SRS national rates excluded some small-state coverage; the IMR comparison is indicative, not perfectly identical coverage. Health spending includes states. Detailed definitions and stages are in Part 3.

Sources: 27, 28, 29, 20, 30.

A school place is not the same as learning.

Pandemic damage and recovery must both remain visible.

Reading improved before COVID, fell sharply in 2022 and recovered by 2024. Class V reading remained slightly below its 2018 level. Arithmetic improved more clearly in the latest observation. ASER measures rural children; it is not an urban-and-rural national census.

Children's nutrition also improved unevenly. Stunting fell from 35.5% in NFHS-5, surveyed in 2019-21, to 29.3% in NFHS-6, surveyed in 2023-24. Underweight changed much less, from 32.1% to 31.8%. One better nutrition figure does not make all nutrition problems solved.

Schools, teachers, families, state policy, income, food access and health all influence these outcomes. The pandemic caused an exceptional interruption, but the learning gap was already large before it.

Verdict: credit NDA-period recovery, arithmetic gains and newer nutrition improvement. The assembled evidence does not establish superior overall learning performance against the full UPA decade. Human development needs better outcomes as well as buildings, seats and enrolment.

Percentages refer to the stated skill, not an examination pass rate or all forms of learning. NFHS-6 was released in May 2026; its survey period was 2023-24.

Sources: 31, 20.

Low unemployment is not the whole answer.

People need enough paid, dependable work.

PLFS 2025 reports usual-status unemployment of 3.1% among people aged 15 and above, and 9.9% among people aged 15-29. The youth figure is a separate age group. The denominator is people in the labour force, rather than the entire population.

In the same survey, 56.2% of employed people were self-employed, 23.6% had regular wage or salaried work and 20.2% were casual workers. Self-employment includes very different situations, from a successful business to poorly paid work or unpaid help in a household enterprise.

Survey design changed in 2025. Earlier NSS employment surveys and later PLFS results also require care. I do not create a clean 2004-2026 jobs series by joining different methods, or count every payroll registration as a newly created job.

The registered-factory survey shows more persons engaged after 2014, a real positive within that sector. It does not cover every informal worker or tell us whether the increase would have happened without the policies adopted.

Average income growth was close across the equal decades. Research estimates also show substantial concentration of income and wealth. These are reasons to ask about pay, days worked and who received the gains, rather than assume the rising national total was shared evenly.

Verdict: the evidence does not establish a decisive UPA-versus-NDA winner on good jobs and ordinary workers' buying power. NDA's delivery advantage cannot be treated as proof that secure employment is solved. UPA's faster GDP growth is not proof that every worker benefited either.

Usual status summarises work over a longer reference period; it is not the same as unemployment during one week. The report makes no invented estimate of unique net jobs created.

Sources: 32, 33, 38, 4, 34.

Later stability has a stronger case.

The case is clearest for bank health; other measures need caution.

Late UPA had persistently high inflation and financial vulnerabilities. Later inflation moderated, helped by monetary and supply policy as well as favourable commodity conditions. Inflation of 4.82% in August 2026 cannot be directly compared as a whole-era average with FY2013-14's annual 9.5%; the dates and CPI framework differ.

Lower inflation means prices rise more slowly. It does not mean groceries or rent return to old prices. A government's price-stability record must therefore be read alongside wages, affordability and food-price volatility.

Scheduled commercial banks' gross bad-loan ratio fell from about 4.1% in March 2014 to 1.8% in March 2026. Between those dates, recognition of stressed loans first increased, including RBI's 2015 Asset Quality Review. Later recovery, restructuring, write-offs and loan-book growth all affected the ratio.

Recognising inherited problems and improving bank balance sheets deserve credit. A write-off is not the same as recovering all the money, cancelling every borrower obligation or proving corruption disappeared.

Central capital expenditure rose from Rs 1.88 lakh crore in FY2013-14 to Rs 10.69 lakh crore in FY2025-26 provisional accounts. That is a nominal change. Interest payments also consumed about 37.6% of central revenue receipts in the later year. Higher investment and continuing fiscal pressure coexist.

Verdict: NDA has a stronger case on bank cleanup and later macroeconomic stabilisation than late UPA. I do not extend this to a proven full-decade inflation win, uniformly cheap living, or complete debt sustainability.

RBI, banks and market conditions share responsibility. Budget GDP vintages and central-versus-all-government definitions are retained; forecasts are not treated as completed outcomes.

Sources: 6, 7, 35, 36, 37.

India produces more. Judge the type of gain.

Factory activity and useful capacity expanded; quality of work still matters.

Persons engaged increased about 55%, or 74.59 lakh. This is a difference between two survey estimates, not a list of unique people who received a new permanent job. The sector is registered manufacturing, rather than every workshop or informal producer.

Gross value added is production value after subtracting bought-in inputs. Its rupee increase includes inflation. It cannot be called a 153% increase in physical goods without an appropriate price adjustment.

Steel capacity rose from 102.26 million tonnes in FY2013-14 to more than 220 million tonnes in June 2026. Capacity differs from actual annual production. Startups, patents, electronics and defence achievements in Part 4 add evidence of industrial and technological expansion, but their registration and money-value definitions must remain attached.

The latest Global Innovation Index ranks India 38th of 139 economies, compared with 66th of 142 in 2013. It supports a better relative innovation position under changing models, rather than a precise measurement of government-caused innovation.

Verdict: NDA has a positive later production-and-capacity record. Without matched full UPA-period baselines for every sector, I do not call this proof of faster industrial transformation across the whole economy or better employment for every worker.

ASI 2024-25 was released 30 September 2026. Index rank, installed capacity, annual production and administrative recognition are different measures.

Sources: 38, 39, 53.

A larger harvest is not a richer farmer.

Shared scientific work deserves more than a party label.

Foodgrain production rose from 265.05 million tonnes for crop year 2013-14 to an estimated 376.563 million tonnes for crop year 2025-26. This is about 42% more physical food. The later number is a third advance estimate, so it remains revisable.

The farmer's result also depends on prices received, inputs, debt, land size, weather and days of work. A national harvest total does not demonstrate doubled real farm-household income. The available household surveys are not a clean 2014-to-today comparison using one unchanged definition.

R&D spending rose substantially in rupees. Part 4 records a GDP share of 0.71% in 2013-14 and 0.84% in 2023-24, while explaining the different GDP-base treatment. Projected later spending is not completed research. More patents or papers also need questions about quality and use.

India launched Chandrayaan-1 in 2008. The Mars Orbiter Mission launched in November 2013 and entered Mars orbit in September 2014. Chandrayaan-3's successful landing came in August 2023. Scientists and institutions carried knowledge across the political handover.

These dates show why a mission's arrival year cannot allocate all its credit. Earlier preparation, continuing budgets, testing and final execution matter. The same principle applies to freight corridors and medical institutions.

Verdict: NDA's later farming-output and scientific achievements are substantial. UPA's earlier missions and foundations also matter. A clean winner on farmer prosperity or the entire science system is not established by the assembled comparisons.

Foodgrain quantity is not farm income. R&D ratios retain their stated denominator vintage; mission history is not a calculation of political credit.

Sources: 40, 41, 42.

Here, UPA has the stronger case.

Material delivery does not cancel the need to question power.

Freedom House assessed India as Free in its 2013 edition. Its 2026 country assessment rates India Partly Free, with a score of 62 out of 100. These are the publisher's judgments under its framework, rather than an official Indian government classification. Kashmir is assessed separately by that publisher.

Reporters Without Borders places India 157th of 180 in 2026, compared with 140th of 179 in 2013. Its method changed in 2022, including the direction of its score. I do not calculate a deterioration percentage by splicing those scores. The recent 2025-to-2026 score also declined within the newer framework.

These assessments use choices about indicators, expert views and evidence. They can be challenged and should be read in detail. Their concerns still deserve consideration; improved infrastructure does not answer a question about press freedom or civil rights.

UPA's Right to Information Act of 2005 created an important accountability framework. Later central-authority rejection rates fell, but fewer recorded rejections do not automatically mean fuller answers, shorter delays or better accountability across every state.

My verdict on the evidence examined: UPA has a stronger case on institutional freedom. This is more qualified than declaring every institution healthier in 2004-14. India's earlier record also had serious shortcomings, and the indicators do not isolate one party's contribution.

This is a substantial debit in NDA's overall assessment. A delivery-first conclusion can favour NDA while still treating rights as an independent public good. A reader who gives institutional freedom the greatest weight may reasonably choose UPA overall.

Civil-rights assessments and administrative RTI disposal statistics measure different things. Freedom House's old ratings are not converted into an invented historical score out of 100.

Sources: 43, 44, 45.

The weaker results remain in the record.

Progress in one area does not erase harm elsewhere.

The murder decline is a positive recorded outcome. Higher road deaths and the higher suicide rate are serious adverse outcomes. Reporting, population and exposure matter, and policing is largely a state responsibility. These figures do not isolate Union-government performance.

The ST-crime chart illustrates how averages can mislead. National recorded cases fell by 2,994 in 2024; Manipur's fall alone was 3,110. Other places together rose by 116. This does not establish causes, but it prevents a simple nationwide improvement story from that one total.

Clean-air progress remains uneven. Under the later NCAP comparison, many monitored cities improved against FY2017-18, but only 14 met the stated Indian annual PM10 standard. That starting year is not 2014, and PM10 improvement does not establish safe PM2.5 everywhere.

Renewable capacity and environmental programmes are meaningful achievements. Heat and changing weather also reflect a global climate problem. Governments remain responsible for local emissions control, safe transport, adaptation and enforcement.

Verdict: no clean overall era winner on safety and environment from this evidence. These unresolved and adverse outcomes reduce the strength of NDA's overall lead, while missing matched UPA baselines prevent a blanket alternative claim.

Registered cases are not all incidents. Road-death counts are not distance-adjusted risk. Suicide records do not prove the cause of an individual death.

Sources: 46, 47, 48, 49, 50.

Credit crosses the handover.

A national result usually has several contributors.

States run much of health, education, policing and local delivery. Kerala and Bihar do not start from the same conditions; their outcomes cannot be explained entirely by the government at the Centre. The state tables in Part 6 show why the national average leaves important differences hidden.

Inheritance works in both directions. NDA inherited institutions and projects that helped it deliver. It also inherited financial stress and unfinished work. UPA inherited an expanding economy and systems built by still earlier governments.

Verdict: NDA deserves credit for execution after the handover, without receiving ownership of all earlier preparation. UPA deserves credit for groundwork and its own results, without being excused from weaknesses left behind.

The report does not attempt to assign a numerical percentage of each project to a party. Funding, planning, construction and operation may have different timelines.

Sources: 18, 21, 23, 9, 42, 26.

The rankings give a mixed picture.

Scores, dates and methods matter more than a league-table headline.

Human development and logistics strengthen the case for progress. Gender, life evaluation, rights and environmental findings prevent an all-round success claim. Each index uses its own definition; no country rank supplies a direct party-performance score.

Verdict: the rankings support a mixed national record. They do not overturn the delivery evidence or justify treating every social concern as solved. Governments can influence these outcomes without being their only cause.

The full original-versus-recomputed comparisons and publisher-method caveats are in Part 6. Ranking position changes when other countries or coverage change too.

Sources: 51, 54, 52, 55, 56, 57, 58.

What each area can support.

These judgments are not added as equal points.

Transport, taps, accounts and digital scale are related delivery outcomes. Their appearance in several rows does not give NDA several automatic votes. Growth, rights and human outcomes can outweigh them if a reader assigns those areas greater importance.

The domain judgments narrow the claim. They distinguish strong evidence of improvement from weaker evidence about who caused it. They also show why a universal 'better on everything' verdict cannot survive the full report.

Sectoral evidence and qualifications appear on the preceding pages and in Parts 1-6. No equal-weight points total is calculated.

Sources: 3, 4, 16, 22, 25, 28, 32, 43.

My choice is NDA, with limits.

The deciding priority is useful delivery to more people.

If I must choose one government under the priorities I set for this report, I choose the BJP-led NDA after 2014. I give the greatest weight to expansion of useful public networks, basic-service reach and financial access. The assembled evidence shows enough substantial delivery in that broad group to give NDA a qualified overall advantage.

This choice also recognises bank cleanup, later stabilisation and meaningful production and capacity gains. Inheritance and favourable outside conditions reduce exclusive credit, but do not make the later execution disappear.

The strongest case against my choice is serious. UPA's observed ten-year real GDP growth was faster. Income per person grew at almost the same pace across the equal decades. Independent freedom assessments favour the earlier period. Jobs, learning, clean air and safety do not supply a decisive NDA victory.

COVID narrows the growth-based criticism of NDA, because it was an exceptional outside shock. The government also remains answerable for its response. The vaccination achievement and the documented hardship in emergency care and migrant support belong in the same judgment.

I cannot prove that my priority order is the only correct one. A growth-and-rights-first reader may choose UPA; someone asking only about networks and service reach has a stronger NDA answer. Those are differences about what matters most, as well as about how to interpret the evidence.

My final verdict is a qualified NDA lead for material delivery, not an unconditional finding that India was better governed in every respect after 2014. UPA's growth, poverty reduction, legislation and institutional foundations remain major parts of India's progress.

Overall verdict: the author's stated interpretation. The source data do not establish a statistically unique winner across all aspects of national life.

Sources: 3, 4, 16, 36, 43, 12, 11.

The verdict should stay open to evidence.

A useful conclusion can be clear without pretending to be final forever.

A fuller comparison of real household income, dependable paid work and service quality could strengthen or weaken the delivery-first judgment. Matched UPA and NDA baselines across more sectors would also make the pace comparison more reliable.

A consistent historical back series under the new GDP framework could change the numerical growth comparison. That is why this report names the old published snapshot and never treats its calculations as immune to revision.

Better evidence on pandemic outcomes and policy effects could change the assessment of the response. So could a more detailed examination of accountability, court access, minority rights and press conditions. Those issues deserve their own evidence, rather than a residual place after economic totals.

The Matrix charts were a starting point. Their unsupported pairs, mixed baselines and forecasts are not used to manufacture a winner here. A 2031 projection, a target, an approval or a cumulative registration is not an achievement already delivered.

For a reader who wants the shortest faithful statement: NDA delivered more widely in the areas this report gives greatest weight; UPA's observed decade grew faster, and its freedom assessment was stronger. Both contributed; neither has a complete success record. That is the conclusion I can defend from the evidence assembled.

I want this last volume to make disagreement easier to understand. You can disagree with my priority order and still use the same corrected numbers. India is better served when a claim remains open to a date, a definition and a source.

The next pages list clickable primary sources and original-publisher assessments. The six earlier volumes contain the extended numerical report card.

Sources: 3, 5, 59, 43.

Sources and reading notes