This dashboard is built for people who will be challenged on their
numbers. What follows is the complete ledger: the rules the data was assembled
under, every place two official sources disagree, every gap that could not be
closed, and every occasion on which a figure here is not simply lifted from a
published table. Nothing in this list is hidden elsewhere — it is repeated on
the individual charts too. It is gathered here so it can be read as a whole.
The rules
- National official sources only. PBS, SBP, the Economic
Survey, NEPRA, OGRA and the Power Division. Where an international source is the
only one available it is named as such on the chart — the gas circular debt
reading for December 2025 is an IMF country report, and is the only figure in the
dashboard that is not from a Pakistani body.
- Nothing estimated, interpolated or modelled. Where a year is
missing it is left missing. No trend has been fitted through a gap and no figure
has been carried forward to fill one.
- Sources are never spliced silently. Where two bodies measure
the same thing differently, both are shown and the disagreement is stated. The one
deliberate join in the dashboard is labelled as such (see below).
- Stocks and flows are treated differently. A flow —
monthly trade, remittances — is summed to a fiscal year, and only complete
twelve-month years are shown. A stock — debt outstanding, circular debt
— is sampled at period end, never summed and never averaged.
- Nine-month figures are never annualised. The Economic
Survey's main chapters report July–March because the Survey is published in
June. Those figures are charted as nine-month figures, separately from full-year
data, and scaling them to a year would be wrong by whatever the missing quarter
happens to be.
Where the official sources disagree with each other
- Nine-month electricity consumption. Four Survey editions
report it and one is defective: the 2023-24 edition printed both its
July–March columns about 21% below its neighbours, and the 2024-25 edition
restated July–March FY2023-24 from 68,559 GWh back up to 83,109. The
corrected figures are used. FY2022-23 takes the 2022-23 edition's 84,034 GWh
rather than the 2023-24 edition's 69,247 — taking the literally latest
figure would mean taking the withdrawn one. The test that settles it: a
July–March window is 71–80% of the matching full year everywhere
else, and the withdrawn columns come out at 61–62%.
- Petroleum consumption. OGRA's editions restate each other
heavily. FY2023-24 was cut from 16.79 to 15.69 million tonnes between consecutive
editions — 6.6%, unexplained in either. FY2022-23 moved from 16.03 to 17.49.
FY2018-19 moved from 19.56 to 20.03. Each year uses the latest edition that
prints it; the superseded readings are kept in the master workbook.
- Electricity generation. PBS, NEPRA and the Economic Survey
give three different figures for the same year, differing by up to 2,570 GWh.
NEPRA also restates itself: FY2018-19 reads 137,039 GWh in its 2019 report,
137,005 in the 2021 and 136,207 in the 2023.
- Installed capacity. Three official answers again, differing
by thousands of megawatts for the same date. There is no single correct capacity
figure for Pakistan.
- Circular debt, FY2019-20. NEPRA's 2021 report prints
Rs2,150bn, but its own line items sum to Rs2,198bn — the subtotal omits
Rs48bn of GENCO payables. The 2023 report prints the corrected figure, which is
what is used here.
- Revenue recovery, FY2018-19. NEPRA's 2019 report says 90.25%
in its narrative and 87.18% in its own table of billed and realised amounts. The
2023 report's table also says 87.18%, so the table figure is used.
- Savings. Economic Survey editions revise each other by up to
a full percentage point of GDP. National savings for FY2024-25 is 14.1% in the
2024-25 edition and 14.89% in the 2025-26. The most recent edition is used for
each year. Treat any single year as provisional until two editions agree.
- The SBP Handbook and the Economic Survey are not interchangeable.
This was tested rather than assumed: national savings implied by the Handbook runs
0.65 to 2.05 percentage points of GDP below the Survey's published figure across
every overlapping year, and the GDP denominators agree to 0.2%, so the gap is real
and sits in the savings numerator.
What could not be built, and why
- A poverty headcount. There is no national poverty rate in
the documents supplied. PBS's HIES 2024-25 provides the consumption data
for poverty estimation but publishes no headcount of its own; the
national poverty line is set by the Planning Commission. The scorecard for it
is shown empty and labelled rather than filled from a non-national source.
- A combined national peak electricity demand. CPPA-G and
K-Electric peaks are published separately and cannot be added: NEPRA's own country
figure is up to 10% below their sum — 24,022 MW against 26,475 for
FY2014-15 — because the two systems do not peak at the same instant. NEPRA's
country rows are charted where they exist, FY2014-15 to FY2017-18; the 2021 and
2023 editions dropped that row, so no national peak can be stated for later
years.
- Circular debt before FY2016-17. NEPRA's 2015 and 2019
editions discuss it only in narrative and print no year-by-year table, so
FY2013-14 to FY2015-16 cannot be recovered from NEPRA at all. Figures for those
years circulate from the Power Division and from lender documents, and they are
held in the master workbook, labelled — but every circular-debt
chart here is NEPRA and NEPRA alone, so those years are absent rather
than filled from a second body. The one place the Power Division figure appears
is as its own tile, beside NEPRA's and labelled with its source, because it is
the most current published stock and the one quoted in the press. Two tiles, two
named bases, never one line.
- Energy consumption before FY2000-01. The Statistical
Supplement 2024-25 is the single source for all four fuels by sector, and its
Table 14.1 begins at FY2000-01. Earlier Supplement editions would extend it;
the Survey's own chapters would not, and are not used.
- Savings in rupees after FY2020-21. The Survey publishes
savings only as a share of GDP; the SBP Handbook gives rupees but stopped at
FY2020-21 and is no longer updated.
- Gas circular debt as a series. Two isolated point-in-time
readings exist. OGRA's own annual report mentions circular debt once, in a
narrative sentence, with no figure. No fiscal-year-end gas circular debt series
exists in any accessible source.
- Primary energy in MTOE. The HDIP energy yearbook is
paywalled.
Series that stop before the present
- The Labour Force Survey is not annual. Rounds exist for
2014-15, 2017-18, 2018-19, 2020-21 and 2024-25 and the years between them have
no survey at all. Those gaps are not missing data and nothing has been
interpolated across them. The 2024-25 round is also published on TWO
definitions — 6.9% unemployment on the 13th ICLS, 7.1% on the 19th. The
13th is what the dashboard shows, because it is the one comparable with the
earlier rounds; the 19th is held as a separate series and never joined.
- Fiscal operations start at FY2000-01 — that is where
the earliest edition supplied begins. The federal budget series starts at
FY2010-11: the FY2009-10 edition's own components do not add to its own total,
so that year is dropped rather than patched.
- REER and NEER end December 2020 — the SBP Handbook is no longer updated.
- SBP Handbook savings and investment end FY2020-21, for the same reason.
- External debt by creditor ends FY2016-17.
- Foreign private investment by source country ends FY2019-20.
- PBS electricity generation and capacity end FY2020-21.
- Most NEPRA series — circular debt and its two stock components,
generation and its fuel split, capacity and its fuel split, both system peaks
— now run to FY2024-25, from four tables of the 2025 State
of Industry Report (Tables 4, 7, 43 and 55). Each of those tables was checked
against its own printed sub-totals and grand total before being used.
- Revenue recovery still ends FY2023-24. The recovery table is
not among the four read from the 2025 edition, so that one series sits a year
behind the rest of the NEPRA block.
- The flow lines in the circular-debt breakdown — the cost of
excess losses and the amount billed but not recovered — still end FY2022-23,
for the same reason. The two stock components now reach FY2024-25.
- NEPRA's national peak-demand row ends FY2017-18; the 2021, 2023 and 2024
editions dropped it.
- Traction electricity ends FY2011-12, household coal is intermittent and gas
to the cement sector ends FY2023-24 — in each case the Statistical
Supplement itself prints “-”, its mark for not available. Nothing has
been carried forward or interpolated across those holes.
- Revenue recovery has a genuine hole at FY2016-17: no NEPRA edition held
prints an aggregate figure for that year, and nothing has been interpolated
across it.
The few places a figure here is not simply a published number
These are the only four, and each is stated on its own chart as well.
- Fiscal-year inflation, averaged from the monthly index.
PBS publishes the monthly CPI but prints a fiscal-year inflation row for one
year only, so the twelve monthly index values from July to June are averaged
and the change on the previous year's average is taken. That is the standard
method and the same arithmetic PBS uses; it reproduces the published headline
figures exactly — 29.2% for FY2022-23, 23.4% for FY2023-24, 4.5% for
FY2024-25. A year appears only if all twelve months exist.
- Monthly flows summed to fiscal years. Trade by country and
commodity, and remittances by country, are published monthly; complete
twelve-month years are summed here so the browser does not have to. Incomplete
years are not shown.
- Monthly debt stocks sampled at 30 June. The fiscal-year view
of central government debt selects the June reading from the monthly table. An
outstanding balance at 30 June is the year-end figure, so nothing is
averaged or summed.
- One generation series joins two sources. The default view of
electricity generation is PBS to FY2020-21 and NEPRA thereafter, with a single
join. Where the two overlap they differ by between −2.1% and +0.9%, so a
step of that order sits at the join. Both sources are also available alone.
- Two savings blocks are rescaled to one unit. The SBP Handbook
publishes the same indicator in Rs Million to FY2009-10 and Rs Billion after; both
are expressed in Rs Million so the row is continuous. That is a unit change, not a
revision.
Reading the dashboard
- Rupee levels do not chain across price-base changes
(1959-60, 1980-81, 1999-2000, 2005-06, 2015-16). Growth rates do. If a question
spans a base change, ask it in growth rates.
- Price indices on the 2015-16 and 2007-08 bases are different
indices, not two vintages of one. For July 2017 they read 107.1 and
216.2 for the same month.
- Year-on-year inflation carries a base effect. The fall to
0.3% in April 2025 is partly April 2024 having been extraordinarily high. Read
the index level alongside the rate.
- The SBP policy rate is event-dated. An observation exists
only in a month when the rate changed. A gap is not missing data, and these
points must not be averaged into an annual rate — how long each rate was in
force is what matters.
- Group rows contain their own sub-items in the trade,
sector and debt breakdowns. Summing every line double-counts; the charts say so
where it applies.
If a figure here disagrees with one you hold, the disagreement is
probably in this list. If it is not, it is worth reporting — the underlying
database carries the full source, note and vintage for all — series, and every one can be traced back to a
published table.
