Almost every secondary-data economics dissertation in India can be built from four sources: the RBI’s Database on the Indian Economy, MoSPI’s national accounts and survey releases, the EPWRF India Time Series, and the Open Government Data platform. Everything else is either a subscription your department may or may not hold, or a specialised database you will use for a single chapter.
That single sentence saves most scholars a fortnight. The failure pattern in an M.A. or M.Phil. Economics dissertation is rarely the econometrics — it is discovering in the fourth week that the variable your model requires exists only as an annual figure when your specification needs quarterly observations, or that the state-wise break-up you assumed was public is actually inside a paid subscription. This guide sets out what each source contains, at what frequency, at what cost, and the checks to run before you let a dataset determine your chapter structure.
The comparison table: five sources, what each one is for
| Source | What it gives you | Frequency and span | Access and cost | Best suited to |
|---|---|---|---|---|
| RBI Database on the Indian Economy (DBIE) | Money and banking, interest rates, exchange rates, balance of payments, government finances, state-level indicators, banking outlets | Daily to annual depending on series; several macro series run back to the 1950s in the Handbook | Free, browser-based, downloadable to Excel or CSV | Monetary economics, banking, public finance, macro time series |
| MoSPI (Ministry of Statistics and Programme Implementation) and NSO | National Accounts Statistics, GDP and GVA series, Index of Industrial Production, Consumer Price Index, Annual Survey of Industries, Periodic Labour Force Survey, Household Consumption Expenditure Survey | Monthly for price and production indices; quarterly for GDP; annual for ASI and PLFS reports | Free reports; unit-level microdata through the ministry’s microdata portal, usually free after registration | Growth, industry, labour, inequality, consumption |
| EPWRF India Time Series | Curated, cleaned long time series across roughly two dozen modules — prices, agriculture, industry, banking, external sector, state domestic product | Long consistent runs, with series breaks documented | Institutional subscription; check your university library before paying anything yourself | Any study needing a long, consistent series without hand-stitching vintages |
| CMIE (Prowess, Economic Outlook, Consumer Pyramids) | Firm-level financial statements, high-frequency macro aggregates, household panel on income, spending and employment | Annual firm accounts; monthly and quarterly household waves | Commercial subscription, typically institutional; individual access is expensive | Corporate finance, industrial organisation, household-level labour and consumption |
| data.gov.in (Open Government Data Platform) | Departmental datasets published by individual ministries — agriculture, energy, transport, education, health | Highly uneven; some resources updated monthly, others abandoned mid-decade | Free, no subscription | Sector studies where the line ministry is the only publisher |
RBI DBIE: start here for anything monetary, fiscal or external
The Reserve Bank’s Database on the Indian Economy is the most usable free statistical portal in the Indian system. It is organised into subject folders, every series carries a code, and downloads come out in a shape you can read into R, Stata or EViews without cleaning half a day away.
Two RBI annual publications matter as much as the portal itself. The Handbook of Statistics on the Indian Economy gives long national series with consistent definitions and footnotes marking every break. The Handbook of Statistics on Indian States is the one most postgraduate scholars underuse — it carries state domestic product, per-capita income, fiscal indicators, power and agricultural data state by state, which is exactly what a panel-data dissertation needs.
The limitation to know before you build a model: RBI series are aggregates. If your research question needs a household, a firm or a district as the unit of observation, DBIE will not give it to you, and no amount of searching will change that. Decide your unit of observation first, then choose the source.
MoSPI and the NSO: national accounts, and the microdata most scholars never open
MoSPI publishes the national accounts, the price and production indices, and the large sample surveys. For an economics dissertation the surveys are where original work lives, because the published reports contain only the tabulations the ministry chose to print, while the unit-level microdata lets you construct your own.
- Periodic Labour Force Survey (PLFS) — the current national labour force survey, running since 2017-18, with annual reports plus more frequent urban bulletins. It is the standard source for unemployment rates, labour force participation, worker population ratio and informality.
- Annual Survey of Industries (ASI) — factory-sector data on output, inputs, employment and wages. Unit-level ASI is the backbone of most Indian productivity and industrial-organisation dissertations.
- Household Consumption Expenditure Survey (HCES) — the consumption survey used for poverty, inequality and demand estimation. Note that the survey design has changed across rounds, so comparability across decades has to be argued, not assumed.
- National Accounts Statistics — GDP, GVA, capital formation, savings, by sector and by state, with periodic base-year revisions.
Base-year revisions are the single most common source of a wrong result in an Indian macro dissertation. When the base year of a series changes, the pre-revision and post-revision numbers are not the same variable. Splicing them without saying so produces a break your external examiner will spot in your own plot. Say in your data section exactly which vintage you used, which base year, and how you handled the join.
EPWRF India Time Series: the shortcut, if your library subscribes
The EPW Research Foundation’s India Time Series exists precisely because assembling a long consistent Indian macro series by hand is a research project in itself. It takes official data, reconciles vintages, documents breaks and presents the result as one downloadable series with metadata attached.
It is a paid subscription, and the price is beyond a research scholar’s own budget. But a great many Indian university libraries, and almost all the older economics departments, already hold institutional access. Ask your librarian before you assume you cannot use it — a fifteen-minute email frequently replaces three weeks of manual series-stitching.
CMIE: firm-level and household-panel work, subscription permitting
CMIE’s Prowess database holds standardised financial statements for listed and large unlisted Indian companies, which makes it the default source for dissertations on capital structure, firm performance, ownership or industry concentration. Consumer Pyramids is a large household panel with waves on income, consumption and employment status.
Access is the whole question. Prowess is normally licensed to institutions, and the business schools and larger economics departments are the ones that hold it. If you do not have institutional access, do not design a firm-level dissertation and hope to solve access later — build the same question from company annual reports and stock-exchange filings, which are free but require you to standardise the accounts yourself, or change the unit of observation.
Specialised sources worth knowing by name
- DGCIS trade statistics, published through the commerce ministry’s export-import data bank, for commodity-level and country-level trade flows.
- India KLEMS, hosted by the RBI, for growth accounting — capital, labour, energy, materials and services inputs by industry.
- Census of India, for the demographic and district-level baseline; note that the 2011 Census remains the most recent completed full enumeration, so any post-2011 district population figure in your dissertation is a projection and must be labelled as one.
- National Family Health Survey (NFHS), for health, nutrition and household characteristics at state and district level.
- India Human Development Survey (IHDS), a nationally representative household panel available to researchers, useful when you need the same household observed twice.
- ICSSR Data Service, which archives Indian social science datasets for academic reuse.
- World Development Indicators, IMF databases and the Penn World Table, when the dissertation is cross-country rather than domestic.
For discipline-neutral guidance on official Indian repositories beyond economics, our overview of official data sources for an Indian thesis by discipline covers the sciences, education and management alongside this material.
Three checks to run before a dataset decides your chapter structure
1. Check frequency against your specification, not against your topic
A topic like “the effect of monetary policy on industrial output” sounds feasible until you notice that your policy variable is available daily, your output variable monthly, and your control variables annual. Aggregating everything to the lowest frequency may leave you with twenty-odd observations, which will not support the model your synopsis promised. Confirm frequency and span before the synopsis, not after. If the committee has already approved the proposal, our guide to repairing a synopsis after committee objections covers how to formally revise scope.
2. Check whether the break-up you need is actually published
National totals are almost always available. State-wise, district-wise, rural-urban and sector-wise break-ups are available only sometimes, and gender-disaggregated versions less often still. Download one year of the actual table before you write the objective that depends on it.
3. Check the revision history
Provisional, revised and final estimates coexist for the same year in Indian national accounts. Record the release date of the file you downloaded and cite the vintage. Examiners in Indian economics departments do ask this question at the viva, and “I downloaded it from the website” is not an answer.
How to write the data section your examiner is actually reading
The data section of an Indian economics dissertation is short and it is graded on precision. Six elements, in this order: the source with its publisher and the release you used; the unit of observation; the period covered and the number of observations; the definition and construction of each variable; the treatment of missing values, outliers and any deflation or splicing; and a summary statistics table.
Deflation deserves a sentence of its own. If you have converted nominal series to real, name the deflator, the base year and the reason you chose that deflator over the alternative. This is one of the most common corrections handed back after evaluation, and it takes two sentences to prevent.
Once the data section is written, the analysis chapter follows the design you have chosen rather than the software you happen to know. If you are still deciding between specifications, our decision table for choosing a statistical test maps the question type to the appropriate method, and if your regression is behaving strangely with several correlated macro predictors, the guide to diagnosing multicollinearity with VIF covers the diagnosis and the four legitimate remedies.
A worked example of a defensible data paragraph
Here is the shape an examiner wants, written for a state-panel study:
“The study uses an unbalanced panel of 21 major Indian states for the period 2005-06 to 2022-23. Gross State Domestic Product at constant prices and per-capita net state domestic product are taken from the Reserve Bank of India’s Handbook of Statistics on Indian States (2024 release, 2011-12 base). State fiscal indicators are drawn from the same source. Labour force participation rates are computed from unit-level Periodic Labour Force Survey data for the survey years in which the state sample supports estimation; states with fewer than 300 sampled households in a given year are excluded, which is why the panel is unbalanced. All nominal magnitudes are deflated using the state-level Consumer Price Index (Combined) with 2012 as the base year.”
Note what that paragraph does: it names releases, it explains the unbalance rather than hiding it, it states an explicit exclusion rule, and it names the deflator and the base. Nothing there is difficult. It is simply written down before the examiner has to ask.
Build the data chapter without losing the fortnight
The data section, the variable definition table and the summary statistics commentary are the parts of an economics dissertation that take longest to write and carry the least intellectual reward. Tesify drafts them from your own dataset description and your own sources, keeps your citations formatted to the style your department requires, and leaves the interpretation — which is the part being examined — to you.
Start your dissertation chapter with Tesify and get the data section into shape this week.
Frequently asked questions
Can I complete an M.A. Economics dissertation entirely with secondary data?
Yes. Most Indian postgraduate economics dissertations are secondary-data studies, and departments accept them routinely. What is examined is not whether you collected the data yourself but whether the source is authoritative, the variables are correctly defined, the period is justified and the method matches the question. Primary collection is expected in field-based economics — agricultural economics, development studies with a survey component — where the department’s ordinance says so.
Is RBI DBIE data free to use in a dissertation?
The database is free to access and download, and the data may be used in academic work with proper attribution. Cite the Reserve Bank of India as the publisher, name the specific table or Handbook edition, and give the date on which you retrieved the file, because series are revised.
How do I get NSSO or PLFS unit-level data as a student?
Unit-level survey data is distributed through the ministry’s microdata portal, generally after a free registration. You download the data file along with its layout document and questionnaire, and you must use both — the layout tells you column positions and codes, and the multipliers or weights must be applied for any estimate to be nationally representative. Unweighted PLFS estimates are a recurring error in student dissertations.
My department does not subscribe to CMIE Prowess. What are the alternatives?
Company annual reports and stock-exchange filings are public and free, but you will have to standardise line items across firms yourself, which limits a realistic sample to perhaps thirty to sixty companies for a postgraduate timeline. Alternatively, ask whether a nearby institution allows library access, or reframe the question at industry level using ASI data, which is free.
How many years of data do I need for a time series dissertation?
There is no UGC rule; the requirement comes from your method. Unit root and cointegration procedures behave poorly on very short annual samples, which is why scholars working with annual Indian data commonly extend the period to thirty years or more, or move to quarterly frequency. Justify the span by the method and by structural events — a liberalisation break, a base-year change, a pandemic year — rather than by convenience.
Do I need ethics clearance for a secondary-data economics dissertation?
Usually not, because published aggregate data involves no human participants. Anonymised microdata is normally treated the same way, though some universities still require a declaration. If any part of your work involves interviewing respondents or collecting household information yourself, clearance is required and must be obtained before collection begins.
Which base year should I use when deflating Indian series?
Use the base year of the official series you are working with rather than converting to a base of your own choosing, and state it explicitly. If your period spans a base-year revision, either restrict the sample to one base or splice the series using the official linking factor and report that you have done so in the data section.
Can I cite a newspaper report of a statistic instead of the source?
No. Trace every figure to the publishing agency and cite that. Newspaper coverage frequently rounds, mislabels the reference period or reports a provisional estimate as final. An examiner who finds a press citation where an official one belongs will treat every other figure in the chapter as unverified.