NIPA(2) text form

Alan Freeman (
Tue, 13 Jan 1998 12:57:41 +0000

This is the text form of the document on measuring the UK. The formatted
version is available on Now hopefully I
can respond to some of the queries!


Alan Freeman, the University of Greenwich


Schumpeter (1994:7) offers an innocuous case for treating
economics as a science:
A science is any field of knowledge in which there are
people, so-called research workers or scientists or
scholars, who engage in the task of improving upon the
existing stock of facts and methods and who, in the
process of doing so, acquire a command of both that
differentiates them from the 'layman' and eventually
also from the mere 'practitioner'.
Since economics uses techniques that are not in use
among the general public, and since there are
economists to cultivate them, economics is obviously a
science within our meaning of the term

This is how the profession of economics today presents its
activities, and also the definition which the general
public accepts. It is not, however, without problems.
It is not self-evident that the use of 'techniques not in
use among the general public' by 'so-called research
workers' qualifies a study as a science, and it is indeed
not true. Any specialism has its own techniques, often
arcane - for example astrology and cabbalism. What
distinguishes science from the merely esoteric is not just
expertise, but a means of judging its results.

This is not reducible to the mere use of procedures or
checks, however rigorous, to reject or accept conclusions.
Economics carries out many internal checks, but. so did the
Spanish Inquisition. The problem common to the two is that
they qualify only their own specialists to administer them.
In a nutshell, economics judges its own results, and the
idea that the uninitiated might judge an economic matter
for themselves is simply not entertained, as testified by
the breathtaking presumption of an Economist leader
entitled, of all things, the Failure of Economics (23
August 1997:11):

Crucial ideas about the role of prices and markets,
the basic principles of microeconomics, are
uncontroversial among economists. These are the first
ideas that politicians and the public need to grasp if
they are to think intelligently about public policy.

Perhaps the first idea the economists need to grasp, if
they are to think intelligently about anything at all, is
that what they take to be uncontroversial might also be

The emperor's tailor cannot judge the emperor's suits. The
medical profession is highly specialised, but offers an
independent test of its methods, namely whether they cure
patients. Economics has never been known to cure anything.
In practice, when challenged, it seems to offer only three
means of demonstration: proof by expertise, proof by
authority, and under duress, proof by obscurity.

This chapter seeks to restore the scientific principles of
independent verification and transparency. Since both
computers and statistics are now easy to obtain, an average
person can with some hard work reconstruct the 'plain
facts' of the economy, and so study - and judge - policy
proposals and their outcomes.

But what is an economic fact? The question is daunting only
because every fact lies concealed in a maze of theory. This
is what I hope to redress: I hope to explain how the facts
are constructed, and so clear the way to an independent
judgment on what they really are. My purpose in this
article is to construct neither a definitive alternative
nor a new source of authority, but to restore to the public
the authority stolen from it, by providing it with the
means to construct its own alternatives.


I deal exclusively with macroeconomic statistics from the
National Income Accounts (NIAs). These principal measures
of the economic state of the nation are published in the
Blue Book1; its companion Sources and Methods, essential
for anyone wishing to go deeper, explains how these are
calculated from raw data. The accounts adhere to an
international standard, the UN's System of National
Accounts (SNA).2 I consider two NIA concepts: output, or
what the economy produces; and income, or who gets it.

The measurement of output

Output, like any statistic, is measured by transforming raw
data. The textbooks often present the accounts as if they
followed in some kind of obvious way from the properties of
the raw data itself. But the actual decisions made, about
what to include and what to omit, paint a different

Example 1: domestic work is not treated as part of output.
The national accounts do not treat it as productive because
it it isn't paid a wage; therefore, it adds no value.

Example 2: payment alone does not qualify an activity as a
contribution to output. Receiving a pension is not treated
as a contribution to output but a cost to be met from
elsewhere, because pensioners do not work.

Example 3: though landlords arguably work no harder than
pensioners, rent is accounted as a contribution to output.
The statistics treat them as supplying a factor of
production - a piece of land. The money they receive is
considered a measure of the value this land adds to the
social product.

Example 4: two-thirds of all rent in the statistics is not
an actual payment at all.3 It is 'imputed': if you own a
house, you are deemed to pay yourself a rent equal to that
which would be paid, if the property was rented instead of

These four choices, taken together, cannot be squared with
any notion that the data itself informs us what to do with
it. For example, many argue that working in a house is more
productive than living in it. The accounts reject this
claim on theoretical, not practical grounds. It is not the
payment of money that qualifies an activity as a
contribution for output, or unpaid activities such as
imputed rent would be omitted. What's good enough for
houses is good enough for housewives, and it is simplicity
itself to impute a monetary value for domestic labour - for
example, by calculating the money which could be made by a
typical firm selling the same services. Moreover, even if
money does change hands, this is not in itself sufficient
to qualify an activity as a part of output, or pensions
would be included. The accounting decisions above are based
on a distinct, theoretical classification external to the

Productive and unproductive activities and their relation
to the accounts

Is there any 'correct' choice of categories for the
accounts? My own view, to place it in plain sight, is that
the choice of statistics varies with purpose. An audit of
our resources should probably include both nature and
unpaid labour. I would however start a study of our
capitalists - who command these resources with money - with
the way they use this money to reproduce these resources.
Since they can do so only by hiring labour, the labour
itself can reasonably be treated as their only universal
productive resource.

However nothing is sacred about any particular presentation
of the accounts, least of all the existing one. This is not
a relativist view: if science requires the clash of
theories, then this needs a level playing field; we cannot
accord one particular theory sole rights to present the
facts, though orthodox economics claims exactly this right.
Actually, to make a fully-informed judgement we must
compare the facts as presented by each theory with
independent evidence, from outside the accounts, which no
theory internal to the profession has had the chance to
tamper with.

Therefore, to every alternative theoretical perspective
there corresponds a presentation of the economic facts and
hence a transformed version of the accounts; indeed the
accountants themselves frequently adjust their views and
with them the accounts.4 Nevertheless the alternatives, and
the transformations, are not arbitrary. They are
subordinate to over-arching constraints dictated by the
logic of a money economy.

The most fundamental choice in presenting the macro-
economic facts, as shown by the examples above, is a
determination of what counts as output. This is the result
of a conceptual classification, either explicit or
implicit, of all economic activities into those which add
value, and those which do not; into productive and
unproductive activities or, in the language of neoclassical
theory, those which constitute factors of production, and
those which do not. This is the decisive theoretical
distinction in the national accounts, more primary than the
textbook categories such as expenditure and income, because
it determines what goes in and where it goes.

A further subdivision is evident. Some activities, like
domestic labour, are altogether absent from the accounts.
However there are unproductive activities like pensions,
where money changes hands and must be accounted for but
which do not contribute to output. These are treated as
transfers - they consume income produced elsewhere. These
choices are politically and ideologically sensitive, since
a transfer appears parasitical.

The measurement of income

The accounts treat value as the contribution of the owner
of a productive factor. Rent measures the value added by
the owner of land or a house, profits the value added by
the owner of money; wages the value added by the owner of a
human; and so on.5

However, this is not the same as the income actually
received by the owner of the factor. For example, everyone
pays tax. The money they actually spend - and what they
actually consume - is therefore reduced by transfers to
other consumers. The accounts call this disposable income,
introducing a clear distinction between the value added by
a factor and the income it receives.

Yet other transfers, such as interest payments, are not
treated as reducing income. This confusion has an
ideological origin, since economic theory holds that the
money paid to a factor ought to equal the value it adds, so
the accountants conveniently lose the difference. Whilst
the table giving factor incomes6 specifies the value added
by each class of person - wage-earners, property-owners,
and self-employed - the tables dealing with expenditure7
amalgamate these same classes, treating them alike as
'consumers' or households. They tell us, for example, what
wage-earners add to output, but not what they spend or

This can be overcome. Modern value-based reconstruction of
national accounts begins with the pioneer work of Shaikh,
Tonak and their co-workers on the net tax calculation,
which categorises all tax payments to show which income
they are deducted from, and all benefits to ascertain which
income they augment. This makes it possible to exhibit the
transfers which the state effects between different types
of income.8 A little ferreting extends the principle to
other transfers.

Example 5: Sources and Methods (p88) explains a dilemma
"which has always caused some difficulty in national
accounting statistics": measuring what the banks do.
Normally, a firm's contribution to GDP is

measured by its 'net output' - the excess of its
receipts from the sale of goods and services over its
operating expenditure on purchasing goods and services
from other enterprises ... However the application of
this definition to financial companies and
institutions produces a paradoxical result.

Profits are normally defined as the excess of sales over
costs. However, while bank costs are considerable (marble,
security, banquets, etc) their only sales are "bank charges
and commissions received from depositors." Hence net output
is small and profit is negative. The root problem is that

banks derive much of their income by lending money at
a higher rate of interest than they pay on money
deposited with them; and in the national accounts
interest receipts and payments are regarded as
transfers and not as receipts and payments for a
financial service.

In 1994 the banks recorded gross trading profits of
-#10,839m but the non-bank (industrial and commercial)
sector recorded #102,028m. Total value-added is the
difference, #91,189.9 That is, interest payments
transferred #10,839 from the industrial and commercial
companies to the banks.

However, this is not confined to corporations: applied
consistently, the same treatment would show transfers of
interest between consumers, each other, and the banks. We
can get a picture of the net effect from the household
accounts, which in 1994 show that consumers received #34bn
in rent, dividends and interest and paid #30,480.10 This
makes it look as if the consumers as a whole pay interest
to themselves, but there are evidently two classes of
consumer - those who pay interest and those who receive it
- just as there are two classes of company. The monetary
system transfers around #30bn (8% of wages) from wage-
earners to property-owners. This is the beginning of an
income account, shown in table 1, for the owners of the two
factors of production, wage-earners and property-owners.

(All figures in #million) Value Income
Value added by wage earners #362,758
Value added by self-employed #63,655
Interest transferred to (#30,480)
Income of wage-earners #395,933

Value added by property owners #157,048
Interest transferred from #30,480
Income of property-owners #187,528
Totals #583,461 #583,461

Table 3: income and value added for the personal
sector with interest transfers(11)

There is no principled objection to this presentation,
since even the existing accounts present interest as a
transfer in the corporate sector. Unfortunately, this
partial oubreak of common-sense is not applied
consistently. The accountants want to say that value
resides in things that get sold, but they also want to
classify incomes as rewards to factors. You can't have it
both ways: the moneylenders get an income without selling
anything. Pulling on this thread leads back to a
contradiction in the original theory: if interest is not a
source of value-added, then capital cannot be a factor of
production. The thread breaks at the turning marked 'Social

There are only two coherent solutions. The first is to say
that all property income adds value, and make interest a
component of value-added by imputing a sale to the banks
equal to their receipts from interest and called an
'adjustment for financial services', as is done for example
in the input-output accounts.

The equally coherent alternative is to say no property
income adds value. In that case, all value is added by wage-
earners and all profit income is a transfer out of this
value. The accounts then read as follows:

(All figures in #million) Value Income
Value added by wage earners and self- #583,461
Income transferred as interest (#30,480)
Income transferred through the (#157,048)

Income of wage-earners and self- #395,933
Value added by property owners #000,000
Interest transferred from wage- #30,480
earners and self-employed
Income received through the market #157,048

Combined income of the property- #187,528
Totals #583,461 #583,461

Table 3: income and value added for personal sector with
transfers from all property income

Output is now expressed as a total sum of the value-added
created by wage-earners, and then transferred through the
workings of the market (or, as it has been referred to
elsewhere 'exploitation') to a variety of property-owners.
We may then further break down the income of the property-
owners to show how the value created by the wage-earners is
disposed of. This is the source not only of all investment,
and hence all growth and change, but also military
activity, financial costs, and the profligate living of the
leisure classes; we can then see which of these has been
prioritised by the actual spending decisions of the
property-owners and the government. In my opinion this is
just as much a 'true fact of the economy' as the
conventional NIA categories.

Gross, net and constant capital: what is output, really?

Why not just eliminate property income altogether and
define output to be #362,758, the income of the wage-
earners, reducing output by an amount equal to profits and
self-employed earnings? Before playing such havoc with the
idea of output, we should ask whether it can be defined
independent of the whole notion of rewards to factors: that
is, we should connect up with the real world. It may be
argued that we should have done this in the first place.
However, we began not from the facts as they really are,
but as they are published. We must retrace the accountants'
steps to find a secure starting point for our own account
of reality.

The output of an enterprise consists of the things it
sells. If we add up all sales of new useful things, we get
a commonsense measure of the nation's output. The accounts
thus provide an independent definition of output, namely
sales. However this is greater than value-added because it
includes the materials and machinery used up by the
enterprises in making the useful things. It is normally
termed gross output: value-added corresponds to the concept
of net output.12 The difference is the value of the things
used up in production, which in the input-output accounts
are termed 'intermediate inputs.'13

But since all personal incomes arise from the difference
between sales and intermediate inputs, net output provides
an obvious independent definition of value added. The
simplest and most consistent definitions are that gross
output is the total annual sales of new useful things, and
net output - value added - is that component of this total
which constitutes the income of persons. This is why it is
inconsistent to omit profits from output.

However, this is not the end of the story. The suppression
of intermediate inputs by the national accounts leads to
paradoxical results because it categorises a substantial
part of the income transferred between different types of
property-owners as an intermediate input, and promptly
omits it. In consequence, value-added increases when we
draw up the full accounts for a transfer, because we must
include not just the transferred profits recorded by the
recipient, but their intermediate costs also.

Example 6: Company cars are often recorded as an expense of
production. Suppose, in deference to the tax authorities,
we designated them as a perk - a disguised part of wages.
This minor change in theoretical perspective increases
reported national output. Suppose a company previously
wrote its accounts (in condensed form) as in Table 4a:
The bottom part is value-added and the top part is cost.
But the change in perspective leads to Table 4b:

Car trips #20,000 Materials #100,000
Materials #100,000
===================== ===========================
Total inputs #120,000 Total inputs #100,000

Wages #30,000 Wages:
In money #30,000
In kind(cars) #20,000
Profits #10,000 Profits #10,000
===================== ===========================
Value-added #40,000 Value-added #60,00

Output #160,000 Output #160,000
===================== =============================
Table 4a: Company cars Table 4b: Company cars treated
as cost of production as perk

Although gross output is the same, value-added has risen by
#20,000 and national accounts show the same increase. This
is because the sums of money which figure in the accounts
pay for a whole thing, and not just the value-added in it.
If a company car is suddenly treated as unproductive, then
an expense that was previously suppressed from the accounts
suddenly appears in it like the bottom half of an iceberg.
Banks are like the company car. As the accountants observe,
they produce no commodity: they create no new use for
society, but merely circulate what is created elsewhere,14
adding no value to output as just defined. However, they
certainly get paid. The accounts make a partial concession
by recording their interest receipts as a transfer. But
this is just the tip of the financial iceberg, since
firstly even their receipts from sales are not a reward for
the creation of a new use, and since moreover these
receipts must cover not just their net output but their
total costs. If we determine that they add nothing to gross
output, then we cannot treat their costs as an intermediate
input. Profits properly stated thus include the full cost
of the financial sector.

Obtaining information about the full costs of the financial
sector is the only really difficult aspect of producing
reliable accounts on the principles suggested in this
chapter. From 1989 onwards input-output balances have been
produced15 showing some intermediate costs. However some
care is required as the costs assigned to 'financial
intermediation include costs that do not belong here at
all, such as telecommunications, postal communications, and
computer software, because all are allocated to the
ideologically confused accounting category of 'services'.

For this reason I adopted a slightly different procedure. I
took the wage costs of the financial sector directly from
Table 2.1 of the national accounts and applied to this a
ratio between the wages of the financial sector and its
intermediate costs calculated using the 123-sector i/o
tables at 3-year intervals and interpolated for intervening
years; this was then used to calculate the intermediate
costs of the financial sector for all of the years 1989-

Two further corrections are required. Strictly speaking,
transfers due to the tax system using the net-tax method
should be used to correct class incomes. As the difference
is not great and for reasons of simplicity, this was not
done. Fazeli 1996, Freeman 1991 offer a more detailed
analysis. Finally, until now we have not corrected the
output figure for depreciation. This is because the NIA
depreciation figures are imputed, being based on estimates
of asset lives, and do not necessarily represent the actual
decline in asset values corresponding, for example, to
moral depreciation. To preserve the principle of
transparency in the final figures I give NIA depreciation

Finally, we have broken down the income of the property-
owners to indicate how it is spent. Since, for example, the
proportion of this income spent on new investment is
decisive for growth, it is instructive to compare this with
the other uses to which the income has been put.
A radically different insight into the structure of the UK
economy emerges. Table 5 shows the distribution of national
output, including the use made of the transfers to property-
owners, in the years 1980 and 1994, and chart 1 shows the
distribution of the principal components of property income
between 1980 and 1994.
(chat omitted: too complex)

Over these core years of the conservative government,
employment income from this point of view fell by 15% from
53% of output to 44.7%. But this gain in income to the
property-owning classes was not spent on increasing the
productive capacity of the nation; over the same period the
proportion of investment in output fell by 22% - and it
should be remembered that 1980 was a recession year. The
unproductive employment of this income is what soaked up
the difference. Personal rentier income rose by 11% as a
proportion of output; and the cost of the banking system by

An independent check of our calculations which affords
further insight is given by the structure of employment.
The following figures show the number of people involved in
various branches of production over this period. A dramatic
structural shift has taken place; 4,361,000 people moved
out of production in the normal sense, and into financial
or commercial services.16

Employment in the UK 1980 1994
Agriculture 380 265
All mining and minerals 354 88
Manufacturing 7253 4330
Electricity, Gas and Water 368 223
Construction 1239 886
Trade 4257 4671
Transport and storage 1056 862
Post and Telecoms 423 366
Total Productive (being 15330 11691
generous to retail)
Finance 1647 2788
Other services 1584 2178
Total Unproductive 3231 4966
State services 4602 4800
Grand Total 23163 23451

Schumpeter, J.A. (1994) History of Economic Analysis
London: Routledge

Anderson, Victor (1989) Possible Reforms in National Income
Accounting, Alternative Economic Indicators Project.
London: New Economic Foundation (mimeo)

Gough, Iam (1979) The Political Economy of the Welfare
State, London:MacMillan

Fazeli, R(1996) The Economic Impact of The Welfare State
and the Social Wage: the British experience. Aldershot and
Brookfield, USA: Avebury

Freeman, A (1991) 'National Accounts in Value Terms: the
Social Wage and Profit Rate in Britain 1950-1986', in
Dunne(ed) Quantitative Marxism, Cambridge: Polity Press

Stone, Richard () A System of National Accounts


1 United Kingdom National Accounts, CSO. Unless otherwise
stated, all references to series and tables relate to the
1995 version.

2 Stone()

3 In 1994 imputed rent of owner-occupied buildings (series
CDDF, table 4.1) was #35,115m and total rent income (series
DIDF, Table 1.4) was #56,793m.

4 Thus in 1987 UK government spending was reclassified in
line with UN categories - a trap for the unwary, since
health and education spending show dramatic shifts
corresponding to nothing real at all.

5 The state occupies a contradictory position: economic
theory begrudges it the status of a independent factor of
production, which it could grant by treating taxes as the
price of this factor just as interest is treated as the
price of what banks do. State labour looks at first sight
like a transfer, being paid out of taxes. But a true
transfer would appear as someone else's income, which does
not happen: if the accounts were done this way then health
services would have to be treated as a benefit in kind.
Instead they enter value added in a disguised form, as the
labour of health workers. Taxes thus fall into two
categories, those which ultimately pay for services treated
as productive and appear in value-added mere transfers,
which appear as an income in the personal sector.

6 Table 1.4. Space does not let us to deal with self-
employed income, which amalgamates small business people or
professionals whose income is a combination of wages and
profits, and a new body of semi-lumpenised workers forced
to sell their services instead of their labour. In what
follows we record them with wage-earners, but separately so
that the reader can re-allocate them if desired. In Freeman
(1991) we suggest a method for allocating this between wage
and property income.

7 Tables 1.2, 1.3, and chapter 4.

8 See for example Fazeli (1996), Shaikh and Tonak (1994),

9 Series: AIAD (table 5.4), AIFB (Table 5.7), CIAC(tables
1.5, 5.1)

10 Source: DJAO(Table 1.4), GITP, GIUG (table 4.9). The UK
household accounts (Table 4.9), available only since 1984
though backdated to 1975, show personal sector interest
payments and receipts separately, instead of (as in the
normal personal sector accounts), the net figure. These
figures exclude private pensions and so represent true
rentier income.

11 Employment and self-employment income: DJAU, DJAO (Table
1.4), income from property calculated as sum of all other
categories. Interest payments: GIUG(Table 4.9). Transfers
to and from the state omitted: net-tax research indicates
this is small.

12 National accounts often use the word 'net' to mean output
net of depreciation. We mean the following: "Although
'value added' and 'net output' are often used synonymously,
net output in the census of production is calculated by
subtracting the value of materials purchased (allowing for
stock changes) from the value of each industry's sales."
(Bannock 1977:412).

13 The German and Spanish systems record this as 'commodity
flow' data. Older UK accounts do not follow this practice
but in 1992 the UK switched to input-output based
accounting, and from that date this information is

14 They are hence unproductive in the strict sense for which
Marx reserves the term: the activity as a whole, including
the labour of its wage-earners, adds no new value. There is
a strong case to treat commerce the same, insofar as it
merely changes the ownership of an existing thing, rather
than the nature of the thing. The input-output accounts
themselves distinguish retail price from producer price,
that part of price that does not depend commerce. For
reasons of space we have not made this adjustment.

15 Table 2.1. I am grateful to the ONS for providing input-
output balances backdated to 1989. For only #40., too.

16 The category 'other services' here refers almost
exclusively to financial services, although some caution is
needed in general with the term 'services' since it tends
to include a pot-pourri of productive services such as
communications and transport, mixed in with financial and
commercial activities.

17 NIA Table 12.3.